Datasets:

Jingmei

/

Pandemic_PMC

like 0

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4713654/

Genetic analysis of TNFST15 variants in ankylosing spondylitis

Aims: The purpose of this study was to explore the role of TNF-like ligand 1A (TL1A) gene ( TNFST15 ) polymorphisms (rs3810936, rs7848647, and rs6478109) in the generation of ankylosing spondylitis (AS). Methods: Polymerase chain reaction (PCR) and sequencing were used to conduct the genotyping of TNFSF15 polymorphisms in 113 AS patients and 120 healthy persons as the case and control groups. The frequencies comparison was performed by chi-square or t test between the two groups. Odds ratio (OR) and 95% confidence interval (95% CI) were calculated to represent the correlation between TNFSF15 polymorphism and AS. Besides, genotypes distribution of the former in controls was checked by Hardy-Weinberg equilibrium (HWE). Results: There was statistically significant difference in AS patients and controls based on family history. Among TNFSF15 polymorphisms, only TT genotype frequency of rs3810936 in cases was obviously high, compared with the controls ( P =0.04), the results indicated that TT was a high-risk genotype (OR=2.31, 95% CI=1.03-5.20). However, both of rs6478109, rs7848647 polymorphisms didn't show any association with AS. Conclusion: Rs3810936 of TNFSF15 were related to the risk of AS and we should pay more attention to the role of TNFSF15 polymorphisms in the pathogenesis of AS in the future.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7371112/

Microbial genomics amidst the Arctic crisis

The Arctic is warming – fast. Microbes in the Arctic play pivotal roles in feedbacks that magnify the impacts of Arctic change. Understanding the genome evolution, diversity and dynamics of Arctic microbes can provide insights relevant for both fundamental microbiology and interdisciplinary Arctic science. Within this synthesis, we highlight four key areas where genomic insights to the microbial dimensions of Arctic change are urgently required: the changing Arctic Ocean, greenhouse gas release from the thawing permafrost, 'biological darkening' of glacial surfaces, and human activities within the Arctic. Furthermore, we identify four principal challenges that provide opportunities for timely innovation in Arctic microbial genomics. These range from insufficient genomic data to develop unifying concepts or model organisms for Arctic microbiology to challenges in gaining authentic insights to the structure and function of low-biomass microbiota and integration of data on the causes and consequences of microbial feedbacks across scales. We contend that our insights to date on the genomics of Arctic microbes are limited in these key areas, and we identify priorities and new ways of working to help ensure microbial genomics is in the vanguard of the scientific response to the Arctic crisis. Introduction The accelerated warming of the Arctic is already resulting in the loss of sea ice, the recession of glaciers and the expansion of wildfires [ 1, 2 ], with the consequences of these impacts already reaching far beyond the Arctic region [ 3–5 ]. Within the 'business as usual' scenario presented by the Intergovernmental Panel on Climate Change (IPCC) (RCP8.5) [ 6 ], it is likely that regions of the Arctic will experience up to 10 °C warming by the end of the century [ 7 ]. For a region that can be defined by July monthly mean temperatures of 10 °C or less, it is clear that an additional warming of 10 °C will have extensive impacts [ 8 ]. As a result, the Arctic is one of the areas in greatest danger from the current climate crisis [ 9 ]. Since microbes inhabit many of the critical interfaces between the Arctic environment and its climate [ 10, 11 ], they will experience impacts and prompt feedbacks as a result of the Arctic crisis. However, the climate interactions of Arctic microbes are still somewhat overlooked within contemporary syntheses [ 12 ]. Herein, we contend that understanding microbial responses to Arctic warming, and indeed predicting whether Arctic microbes will fuel further feedbacks, requires exploration of Arctic microbial genomic potential and the fusion of genomic insights with those garnered from diverse academic disciplines. Arctic microbes as first responders Microbes are the first responders to the Arctic crisis. Small in size but large in number, microbes inhabit diverse niches across the Arctic. Their generation times are typically much shorter than plant or animal inhabitants of the Arctic and often well within seasonal or synoptic timescales [ 13, 14 ], allowing rapid changes in Arctic microbial populations in response to climate changes. Specifically, many microbial niches are interposed at the margins between frozen substrates (e.g. brine channels in sea ice, permafrost, glacial weathering crusts, englacial vein boundaries). Here, liquid water is limited, which hinders microbial activity [ 15–17 ]. Thus, changes in temperature that switch Arctic environments from frozen solids into melted liquids can radically alter the niches available to microbial populations. Through their growth and nutrient cycling, the collective regional-scale responses of Arctic microbes can influence biogeochemical cycles on regional and global scales [ 18–20 ]. Consequently, members of Arctic microbiota have been considered both sentinels and amplifiers of global climate change [ 10 ]. The key to how Arctic microbes both sense changes in their local environment and amplify the global impacts of these changes is to be found within their genomes. The microbiology of a crisis The purpose of this review article is to stimulate the response of microbial genomics as a field to the Arctic crisis. Some of the most innovative and significant conceptual and technical developments within the history of microbiology have been stimulated by crises: from germ theory [ 21 ] to real-time genomic epidemiology [ 22 ]. Confronting the Arctic crisis presents an imperative to address many gaps in our fundamental knowledge of cold-region microbiology, which potentially constrain climate models and the informing of policymakers. Furthermore, in crisis there is also opportunity, and bioprospecting of the Arctic is recognized as an emerging field [ 23 ]. However, beyond potential influence on climatology, policy or economy, the study of Arctic microbial genomics is merited in its own right, for many fundamental gaps remain in our knowledge of Arctic microbes. In any case, the prospect of rapid and radical change in the microbial ecosystems of the Arctic must prompt the systematic investigation of genomic diversity within these ecosystems before they are overridden by the effects of warming, the rate of which is unprecedented in human history [ 9 ]. Therefore, this review will offer a primer on key microbial habitats and processes in the Arctic, before considering important challenges and potential opportunities for microbial genomics in confronting the Arctic crisis. Arctic microbes as first responders Microbes are the first responders to the Arctic crisis. Small in size but large in number, microbes inhabit diverse niches across the Arctic. Their generation times are typically much shorter than plant or animal inhabitants of the Arctic and often well within seasonal or synoptic timescales [ 13, 14 ], allowing rapid changes in Arctic microbial populations in response to climate changes. Specifically, many microbial niches are interposed at the margins between frozen substrates (e.g. brine channels in sea ice, permafrost, glacial weathering crusts, englacial vein boundaries). Here, liquid water is limited, which hinders microbial activity [ 15–17 ]. Thus, changes in temperature that switch Arctic environments from frozen solids into melted liquids can radically alter the niches available to microbial populations. Through their growth and nutrient cycling, the collective regional-scale responses of Arctic microbes can influence biogeochemical cycles on regional and global scales [ 18–20 ]. Consequently, members of Arctic microbiota have been considered both sentinels and amplifiers of global climate change [ 10 ]. The key to how Arctic microbes both sense changes in their local environment and amplify the global impacts of these changes is to be found within their genomes. The microbiology of a crisis The purpose of this review article is to stimulate the response of microbial genomics as a field to the Arctic crisis. Some of the most innovative and significant conceptual and technical developments within the history of microbiology have been stimulated by crises: from germ theory [ 21 ] to real-time genomic epidemiology [ 22 ]. Confronting the Arctic crisis presents an imperative to address many gaps in our fundamental knowledge of cold-region microbiology, which potentially constrain climate models and the informing of policymakers. Furthermore, in crisis there is also opportunity, and bioprospecting of the Arctic is recognized as an emerging field [ 23 ]. However, beyond potential influence on climatology, policy or economy, the study of Arctic microbial genomics is merited in its own right, for many fundamental gaps remain in our knowledge of Arctic microbes. In any case, the prospect of rapid and radical change in the microbial ecosystems of the Arctic must prompt the systematic investigation of genomic diversity within these ecosystems before they are overridden by the effects of warming, the rate of which is unprecedented in human history [ 9 ]. Therefore, this review will offer a primer on key microbial habitats and processes in the Arctic, before considering important challenges and potential opportunities for microbial genomics in confronting the Arctic crisis. Key microbial processes in critical zones of Arctic change Within this section, we consider the changing microbiology of four ice-cold hot-spots of microbial diversity, activity and feedbacks in the Arctic climate system ( Fig. 1 ). These range from the changing Arctic Ocean and permafrost thaw to glacial ecosystems and human activities in the Arctic. For each of these critical zones, microbe-mediated processes interacting with climate change are highlighted and areas are identified where improved understanding of the genomic foundations of Arctic microbiomes is required. Fig. 1. Ice-cold hot-spots of microbial change within the warming Arctic. The images demonstrate: a controlled release of methane-saturated groundwater from High Arctic permafrost (a); first year sea ice in winter (b); a marine-terminating glacier meeting open water in a High Arctic fjord in winter (c); cruise ship visitors at Ny à lesund, a Svalbard settlement used for coal mining and scientific research for over a century (d); glacier algae growth and cryoconite accumulation in the Dark Zone of the Greenland Ice Sheet (e). All photographs are from the personal collection of A. Edwards. Sea ice habitat loss in the Arctic Ocean At its maximum, Arctic sea ice currently extends to around 15 million km 2 , blanketing almost all of the Arctic Ocean with a solid frozen cap. In the 40 years of satellite observations, the extent of sea ice in the Arctic Ocean has declined considerably, from a September minimum of 7.7 million km 2 in 1979 to the lowest extent of 3.6 million km 2 in 2012, with the last 13 years representing the thirteen lowest extents in the satellite era ( http://www.climate.gov/news-features/understanding-climate/climate-change-minimum-arctic-sea-ice-extent ). The loss of sea ice itself prompts changes in sea ice albedo feedback, where the high surface reflectance of sea ice to solar energy is decreased in contrast to the increased absorption of solar energy to the darker surface of open water [ 24 ]. These changes have the potential to influence the planetary energy budget [ 25 ]. Sea ice is a complex microbial habitat marked by profound gradients in temperature, chemistry and salinity across a vertical profile of a few metres [ 15 ]. Within the sea ice column, microbes inhabit highly saline waters within pore spaces and brine channels are created as ice formation excludes dissolved salts [ 15 ]. The interface between the base of the sea ice column and underlying seawater is marked by high densities of microbes [ 15, 26 ]. Sea ice microbes are perennially active [ 27, 28 ]. Long hours of sunlight during the polar day supports a net autotrophic ecosystem [ 15 ] driven by eukaryotic phototrophy, primarily from diatoms such as Fragilariopsis [ 29, 30 ]. The exudation of organic carbon from sea ice diatoms supports a diverse range of bacterial heterotrophs, Archaea, protists and meiofauna associated with the sea ice [ 15, 31–33 ]. The export of sea ice organic carbon nourishes the food web of the underlying water column and seabed [ 34 ], emphasizing the importance of sea ice habitats in the functioning of the Arctic Ocean ecosystem [ 35 ]. The declining extent of sea-ice coverage, therefore, has profound impacts on the broader Arctic Ocean ecosystem, and the loss of thicker, structurally more complex, multi-year sea ice diminishes the range of productive niches available to sea-ice microbes [ 36, 37 ]. The loss of multi-year sea ice in a region is associated with long-term taxonomic shifts in the microbial communities of underlying water, for example, communities in the Beaufort Sea [ 38, 39 ] showed a decline in the abundance of multiple microbial groups relevant to biogeochemical cycling within the region. These included Bacteroidetes , which are typically associated with processing complex organic carbon [ 40 ], likely from diatoms, and marked reduction in the abundance of ammonia-oxidizing Thaumarchaeota, which correlates with a decline in nitrate, the limiting nutrient for productivity in the region Concomitant with the loss of sea ice, the expansion of open water as a habitat in the Arctic Ocean is prompting the immigration of microbial groups once thought limited to lower latitudes, with potential effects on bacterial lineages that may be endemic in the Arctic Ocean [ 41, 42 ]. Most notably, coccoid picocyanobacteria ( Synechococcus sp.), major marine primary producers in lower-latitude oceans, were once considered essentially absent in the colder waters of the Arctic Ocean [ 43 ]; however, they are now found as far north as 82.5°N, with abundant populations in Atlantic water reaching north-west of Svalbard (79°N) [ 44 ]. Picocyanobacteria such as Synechococcus and Prochlorococcus can survive the dark season in the Arctic Ocean [ 45 ], through photoheterotrophic metabolism. Consequently, they may out-compete eukaryotic algae [ 46 ] within the autumn and winter seasons in the Arctic Ocean. Therefore, they may be well placed to respond to the continued warming and loss of sea ice [ 47 ]. Indeed, it has been assumed that solar radiation offers a bottom-up control on the overall structure and function of Arctic marine ecosystems, and that the lengthy dark period of the polar night represents a quiescent phase in the ecology of the Arctic Ocean. However, with the increased expanse of open water and the warming of the Arctic Ocean's surface, it has become clear that polar night represents a period of microbial activity that has hitherto been largely overlooked [ 48 ]. Historically considered synonymous, winter, as a period of sustained low temperatures resulting in extensive sea ice, is now discrete from polar night, as a period of sustained darkness [ 49 ]. The implications of this decoupling of polar night and winter beg for further research attention to its consequences for microbial dynamics in a warming Arctic [ 47, 48 ]. In summary, understanding the range shifts, genomic adaptations and population structures of important marine primary producers, such as these cyanobacteria [ 47 ], will inform our predictions of how the food webs of the Arctic Ocean will respond to a future in which the extent and longevity of sea ice is severely curtailed. Carbon release from thawing permafrost Nearly half of global soil organic carbon is found in Arctic soils (1330–1580 Pg C) [ 18 ]. Most of this carbon is stored within permanently frozen ground, including permafrost (i.e. soil that remains frozen for 2 or more consecutive years) [ 50 ]. However, with warming temperatures across Arctic lands, this accumulated stock of legacy carbon from past climates does not represent a permanent sink of carbon. Indeed, thawing permafrost reinvigorates microbial communities that consume these carbon stores and release greenhouse gases in the form of carbon dioxide (CO 2 ) and methane (CH 4 ) to the atmosphere, with potent climate consequences [ 18, 51 ]. If our current trajectory of global warming is continued (IPCC RCP8.5), by the end of this century 30–99 % of near surface permafrost in the Arctic will have been degraded [ 52 ], with an associated release of 37–174 Pg C due to microbial processes within the thawed soils [ 18 ]. Understanding the drivers, rates, mechanisms and extent of microbial carbon cycling in Arctic soils is, therefore, an urgent priority, considering the large uncertainties apparent in the estimated ranges of carbon release [ 53 ]. Changes in Arctic lands also affect Arctic rivers and the Arctic Ocean [ 54, 55 ]. Thawing and degradation of permafrost will liberate substantial terrestrially derived organic matter (tDOM) for riverine transport to the Arctic Ocean [ 56 ]. Approximately 44 Tg organic carbon is released annually to the ocean from coastal and terrestrial permafrost within the Siberian Arctic alone, the bulk of which is predicted to be respired to carbon dioxide [ 57 ]. Microbes within coastal Arctic fjords respond readily to the influx of tDOM. Experiments simulating the release of permafrost carbon to a Svalbard fjord showed Glaciecola populations expanded aggressively following the addition of terrestrial dissolved organic matter [ 58 ]. Since Glaciecola is a gammaproteobacterial lineage associated with rapid consumption of organic matter in cold waters [ 59 ] following spring phytoplankton blooms, it is possible this effect may have broader impacts on the bacterial processing of organic carbon and food webs in the coastal Arctic [ 58 ]. The influence of tDOM extends further than the Arctic shoreline. The first study to develop metagenome-assembled genomes (MAGs) from the Arctic Ocean recently revealed marine Chloroflexi populations with the capability for degrading highly aromatic tDOM [ 20 ]. It appears that this capability arose through lateral gene transfer from terrestrial lineages, since the aromatic metabolism genes detected within the marine Chloroflexi MAGs were closely related to homologues present in terrestrial actinobacteria, acidobacteria and proteobacteria, while the parent lineage of Chloroflexi SAR202 has been found in deeper and darker (see the previous section) Arctic waters [ 60 ]. While the lateral transfer of carbon between Arctic lands and the Arctic Ocean is well acknowledged [ 55 ], the interactions concerning lateral gene transfer between terrestrial and marine Arctic microbial genomes and climate-driven changes in the biogeochemical cycles of the Arctic clearly merit further exploration. Permafrost itself is considered an unusual microbial habitat [ 61 ], since it represents a structurally heterogeneous environmental matrix that combines a long-term deep-frozen store of microbial biomass and genomic diversity from past climates (even >1 million years old) [ 18 ], which may become reactivated upon the degradation of the permafrost [ 62 ]. Recent chronosequence surveys [ 63, 64 ] have examined the survival mechanisms of viable microbiota within permafrost dating from the Pleistocene period (samples that dated to 19 000–33 000 years before the present). These reveal that while endospore-forming taxa are prevalent, viable biomass from some of these taxa remain as vegetative cells [ 64 ], underlining the potential for long-term, low-growth-rate survival rather than sporulation as a persistence mechanism within permafrost. Cold laboratory incubations (reviewed by Nikrad et al . [ 65 ]) lend further support to the potential for activity and even growth of microbial populations in sub-zero conditions [ 66 ]. Analysis of pangenomes from isolate and MAGs from permafrost could, therefore, provide insights to survival mechanisms and microevolutionary processes across geological timescales. However, the urgent questions prompted by the prospect of extensive permafrost degradation relate to the rates and pathways of carbon metabolism in thawing permafrost. In particular, our ability to predict the relative magnitude of permafrost carbon released as either CO 2 or as CH 4 , which has an estimated global warming potential ~30 times greater than CO 2 [ 6 ], is essential for predicting the role of permafrost in climate warming feedbacks. In-field gas flux measurements or cold-lab incubations alone have not offered an integrative view of microbial contributions to greenhouse gas evasion from permafrost [ 16 ]; therefore, recent work has focused on the integration of multi-omics approaches with biogeochemical process measurements of permafrost thaw incubations [ 67, 68 ], as well as permafrost cores [ 69 ]. Most recently, large-scale genome-centred metagenomics conducted across a permafrost thaw gradient have underlined the importance of linking processes with pathways and taxa by revealing novel fungal pathways for plant polysaccharide degradation and syntrophic interactions resulting in CH 4 production [ 70 ]. Furthermore, the potential for viruses to modulate carbon cycling through methanogen infection or lateral gene transfer of carbon processing pathways has been revealed by metavirome sequencing [ 71 ]. In determining the magnitude of CH 4 release from thawing Arctic permafrost, a critical question is posed by the capability of CH 4 -oxidizing microbes (methanotrophs) to consume CH 4 formed by archaeal CH 4 producers (methanogens) before it can reach the atmosphere. Methanotrophs may modulate the release of between 20 and 60 % of the CH 4 formed in tundra wetlands [ 72–74 ]. Linking genome-centred metagenomics with metatranscriptomics is revealing the diversity of methane-processing genetic mechanisms present within thawing permafrost. One highlight has been the identification of a resilient, dynamic methanotrophic community that can utilize isozymes with differential affinities for CH 4 ; thus, aiding their ability to persist through fluctuating CH 4 availability [ 75 ]. Understanding the potential for methanotrophy to mitigate methane release from thawing permafrost will require pairing biogeochemical process measurements with the capability to resolve diverse pathways for methane oxidation. These pathways must also be attributed to multiple lineages present within dynamic microbial communities. Pairing biogeochemical process measurements, physical and chemical analyses with genomics seems to offer a promising strategy to resolve the extent to which methanotrophy can offset methanogenesis in thawing tundra over the coming decades. To support accurate predictions of methane release, such approaches must embrace the complexity offered by a net outcome that is the sum of interactions within dynamic microbial consortia in structurally heterogeneous habitats defined by fluctuating oxygen, methane, terminal electron acceptor and water levels. Microbial impacts on Arctic glaciers and the Greenland Ice Sheet Arctic glaciers, ice caps and the Greenland Ice Sheet have started to experience the crippling consequences of Arctic warming, yet the melt that has been experienced to date is a mere fraction of what can be expected as the Arctic continues to warm. The largest of these glacial ice masses is by far the Greenland Ice Sheet, which occupies 1.7 million km 2 and currently sequesters the water equivalent of 7.4 m of sea-level rise [ 76 ]. Glacial meltwater is currently the largest contributor to sea-level rise [ 8 ], making this relationship a key societal concern and research priority. Models of glacier mass balance presently used to constrain estimates of sea-level rise currently do not incorporate microbiological parameters [ 8 ]. Addressing the role of microbiota in the rate and magnitude of Arctic glacial ice loss and climate warming, topics that have historically been overlooked, therefore, are of considerable importance to ensure robust estimations of future sea-level rise and climate change. That glacial systems are home to abundant and diverse life forms has been long established [ 77, 78 ]. This century has seen a refreshed synthesis of evidence for biodiverse microbial ecosystems associated with glaciers and ice sheets, and an acknowledgement that these biomes contribute to global biogeochemical cycles [ 79 ]. Like permafrost, glacial ice is a vast, climate-sensitive repository of microbial biomass and genomic diversity [ 80 ]. Equally, microbial processes at the surfaces and beds of Arctic glaciers and the Greenland Ice Sheet have the potential to amplify the impacts of climate warming on glaciers. The subglacial zone beneath glaciers and ice sheets is perennially dark and cold. Microbial ecosystems are apparent here, subsisting on organic carbon washed from the surface [ 81 ], relict carbon overridden in the last ice age [ 82 ] or existing through chemolithotrophy [ 83 ]. Critically, the evolution of anoxic conditions in subglacial habitats can favour methanogenesis [ 84 ] through both hydrogenotrophic and acetoclastic pathways [ 82 ]. Methane oxidation at the oxygenated glacial margins may mitigate subglacial methane production [ 85, 86 ]; however, contrasting results from neighbouring outlets of the Greenland Ice Sheet prompt uncertainty on whether methane oxidation can adequately compensate against methane production [ 85, 87 ]. Hitherto, a genomic perspective on subglacial ecosystem structure and function is in its infancy, with very few publicly available metagenomic datasets [ 88 ]. In contrast, the surfaces of Arctic glaciers receive abundant solar radiation in summer. This prompts the seasonal development of a range of microbial community types predominantly supported by photoautotrophy [ 15, 89 ]. Importantly, the accumulation of microbial biomass replete with photosynthetic and photoprotective pigments and recalcitrant dark organic matter at the glacier surface has the potential to influence the melting rate of the glacier surface [ 8 ]. The reduced surface reflectance of glaciers consequent to microbial growth in surface habitats has been termed 'biological darkening' [ 90 ] or 'bioalbedo' [ 91 ] in recent years. Estimates of microbial contributions to glacier melting are emerging [ 92–94 ]; however, integration of microbial-associated parameters in estimates of sea-level rise remains an active research goal. Snowpacks on Arctic glacial surfaces are vast environments that both support distinctive microbial consortia and are highly sensitive to warming [ 95–97 ]. The potential for snowpack bacteria to cycle climate-relevant trace gases has recently been highlighted as an emerging area [ 98, 99 ]. However, the growth and pigmentation of green algae in the family Chlamydomonadaceae in discrete patches on snow is particularly apparent. The consequent formation of intracellular carotenoid pigments can modulate local solar energy balance [ 100 ] and colour snowpacks bright red. Snow algal productivity can support a diverse range of heterotrophic taxa [ 101 ]; thus, subsidizing the development of a snowpack carbon cycle. Similarly, on the bare ice surface, members of the Zygnematophyceae glacier algae [ 102 ] form ice-darkening biofilms [ 103 ] that are particularly prominent on the south-western margins of the Greenland Ice Sheet, exhibiting locally structured populations [ 104 ]. Glacier algae influence surface reflectivity through the accumulation of dark photoprotective purpurogallin pigments [ 102 ], and their expansive spatial coverage can promote both surface ice ablation and carbon cycling [ 94, 102, 105 ]. Finally, cryoconite ecosystems are among the most intensively studied habitats on the glacier surface [ 106 ]. These collections of granular microbe-mineral aggregates (cryoconite) darken patches of the ice surface, resulting in localized melting and the formation of quasi-circular cylindrical melt holes [ 106 ]. Cryoconite granules are maintained at thermodynamic equilibrium depths and generally as single-granule layers at the floor of the cryoconite hole [ 107 ]. These responses to solar energy balance and cryoconite debris loads ensures the major primary producers in cryoconite, cyanobacteria, are continually exposed to optimal levels of solar radiation for photosynthesis; thus, promoting high levels of carbon fixation in spite of low ambient temperatures [ 107–109 ] Across Arctic glacial surfaces, a single lineage of filamentous cyanobacteria, Phormidesmis sp., appears responsible for binding together each cryoconite granule [ 99, 110–115 ]. A recent global-scale survey identified a single 16S-23S ITS haplotype of Phormidesmis priestleyi predominant upon Arctic glaciers [ 112 ]. Since the population structure of Phormidesmis sp. and other lineages of glacier cyanobacteria shows increasing fragmentation following the decline in glacierized surface area since the peak of the last ice age, it is likely they have an enduring role as ecosystem engineers of glacier surfaces [ 112 ]. Consequently, it is likely that one lineage of glacial cyanobacteria is predominantly responsible for the formation and maintenance of productive island-like microbial ecosystems within the austere environs of Arctic glacier surfaces, making cryoconite holes attractive for studies requiring naturally occurring microcosms of community development (cf. the article by Rivett and Bell [ 116 ]). Human dimensions of the changing Arctic: bioprospecting and infectious disease risks The Arctic has been inhabited by humans for millennia. It is now home to four million people. Moreover, the rapid changes in Arctic climate expected this century are leading to renewed interest in the economic potential of the Arctic as mineral resources, maritime navigation and tourism all become more accessible. This poses microbial risks and opportunities [ 117 ]. Firstly, growing commercial and political interests coupled with increased logistical accessibility is likely to stimulate interest in Arctic bioprospecting [ 23 ]. Adaptations for life in the cold found within the reservoir of Arctic genomic diversity can be industrially useful [ 118 ]. Examples include enzymes with low temperature optima [ 119 ], low-alcohol yeast [ 120 ], antifreeze and ice-binding proteins [ 121 ], and potential antimicrobials [ 122 ] ( Table 1 , fully referenced and expanded in Table S1, available with the online version of this article). To date, the majority of antimicrobial compounds and cold-active enzymes have relied on cultured isolates, which are either screened directly for activity [ 122 ] or genome sequenced; followed by the cloning and expression of candidate genes or gene clusters in a heterologous host [ 123 ]. These strategies rely on the isolation and genome-sequencing of microbes, which is limiting because: (i) fully sequenced genomes of Arctic strains are limited in number (see below), and (ii) many microbes remain uncultivated. In addition, advances in sequence-based and functional metagenomic approaches [ 124 ] offer promising approaches to mine and exploit such potential. For example, specially engineered heterologous expression hosts, such as the ArcticExpress Escherichia coli competent cells (Agilent Technologies) and Pseudoalteromonas haloplanktis TAC125 strain [ 123 ] are noteworthy for being products of cryospheric bioprospecting, and tools by which further functional exploration of this environment can be accomplished. ArcticExpress cells co-express the chaperonin system Cpn60 and Cpn10 from Oleispira antarctica , which helps to ensure the proper folding of cold-active proteins and increase the growth rate of E. coli at low temperatures [ 125 ]. Pseudoalteromonas haloplanktis TAC125, however, is of cryospheric origin, and displays increased solubility and secretion of protein products over other Gram-negative expression hosts. Meanwhile, improvements in sequencing technologies have resulted in the assembly of longer contigs, with deeper coverage than ever before, unlocking types of analyses previously available only to whole-genome-sequenced cultured organisms. Tools such as antiSMASH [ 126 ], for example, can be used on metagenomic datasets (contigs or MAGs) to detect biosynthetic gene clusters responsible for the synthesis of industrially useful compounds, such as antibiotics, fatty acids, polysaccharides, antioxidants and UV-protective pigments. However, the greatest improvements in bioprospecting will likely come from the synergy of sequence-based and functional methods, because an understanding of the genomic background of the source organism (see below) is vital for the strategic genetic engineering of suitable hosts and the identification of optimal conditions for expression of the desired natural product. Table 1. Key cold adaptations relevant for biotechnology Target product Adaptation mechanism Application Cold-active enzymes Amino acid changes that increase enzyme flexibility Food industry, detergents, molecular biology tools Polyunsaturated fatty acids Polyunsaturated fatty acids increase membrane permeability at low temperatures Dietary supplements for humans, livestock and fish Ice nucleation proteins Seed small crystals instead of large damaging crystals Food industry, synthetic snow Antifreeze proteins and solutes Prevent water molecules from forming ice-crystal structure Cryoprotectants, food industry Antioxidants and UV pigments Protect micro-organisms from seasonally high UV irradiation in snow Biomedical, pharmaceutical, food technology and cosmetics Exopolysaccharide Trapping of liquid water, preventing freezing Biomedical, pharmaceutical, food technology and cosmetics Antimicrobial compounds Chemical defences and weapons against competing bacteria in low-resource environments Pharmaceuticals: antibiotics, antifungals, anti-tumour medications and pesticides Secondly, the Arctic is not pristine and has not been pristine for some time [ 127 ]. There is a long history of human activities that have contaminated the Arctic in many ways, from hydrocarbon exploitation [ 128, 129 ] to military activities, including the largest nuclear explosion to date [ 130 ]. These have resulted in locally derived contamination of the Arctic. Likewise, long-range atmospheric transport of pollutants and the global distillation effect has led to the deposition of pollutants in the Arctic from the mid-latitudes for at least 3 millennia [ 127 ]. The potential roles of microbes in modulating or exacerbating the threats posed by contaminants liberated by Arctic warming is a current focus for researchers addressing radionuclide [ 131 ], persistent organic pollutant [ 132 ], black carbon [ 133 ], mercury [ 134 ] and heavy metal contaminants [ 135 ]. Furthermore, both increased access to the Arctic and the potential release of long-frozen hazards is raising the prospect of the liberation of ancient infectious diseases [ 136, 137 ]. Such notions seem speculative, for they depend on the release of viable agents able to withstand severe freeze–thaw stresses as they migrate to the actively melting layers of glaciers or permafrost [ 138 ]. Hitherto, dedicated efforts to recover pathogens from the Arctic have failed, as human remains have degraded within the active layer of permafrost [ 139 ]. However, the release of Bacillus anthracis from frozen wildlife carcasses has been invoked as the cause of a recent Siberian anthrax outbreak [ 140 ]. A further hypothetical risk is presented by implementing synthetic biology approaches to resurrect poorly described viral genomes from Arctic ice [ 141 ], as these may generate highly concentrated infectious materials within laboratories [ 137 ]. The ethical debate and moratorium on resurrecting highly pathogenic influenza strains [ 142 ] offers a certain precedent for concern within this arena. Finally, these changes in the accessibility and ecology of the Arctic bring with them pressures for human healthcare [ 143 ]. These include increased demand on the limited healthcare services available or the immigration of emerging infectious diseases; for example, as vectors move polewards [ 144 ]. This may necessitate enhanced microbiological surveillance and diagnostic capability; and distributed or ubiquitous genomic sensing [ 145 ] may prove important in detecting and managing microbial threats as the Arctic experiences disruptive change. Sea ice habitat loss in the Arctic Ocean At its maximum, Arctic sea ice currently extends to around 15 million km 2 , blanketing almost all of the Arctic Ocean with a solid frozen cap. In the 40 years of satellite observations, the extent of sea ice in the Arctic Ocean has declined considerably, from a September minimum of 7.7 million km 2 in 1979 to the lowest extent of 3.6 million km 2 in 2012, with the last 13 years representing the thirteen lowest extents in the satellite era ( http://www.climate.gov/news-features/understanding-climate/climate-change-minimum-arctic-sea-ice-extent ). The loss of sea ice itself prompts changes in sea ice albedo feedback, where the high surface reflectance of sea ice to solar energy is decreased in contrast to the increased absorption of solar energy to the darker surface of open water [ 24 ]. These changes have the potential to influence the planetary energy budget [ 25 ]. Sea ice is a complex microbial habitat marked by profound gradients in temperature, chemistry and salinity across a vertical profile of a few metres [ 15 ]. Within the sea ice column, microbes inhabit highly saline waters within pore spaces and brine channels are created as ice formation excludes dissolved salts [ 15 ]. The interface between the base of the sea ice column and underlying seawater is marked by high densities of microbes [ 15, 26 ]. Sea ice microbes are perennially active [ 27, 28 ]. Long hours of sunlight during the polar day supports a net autotrophic ecosystem [ 15 ] driven by eukaryotic phototrophy, primarily from diatoms such as Fragilariopsis [ 29, 30 ]. The exudation of organic carbon from sea ice diatoms supports a diverse range of bacterial heterotrophs, Archaea, protists and meiofauna associated with the sea ice [ 15, 31–33 ]. The export of sea ice organic carbon nourishes the food web of the underlying water column and seabed [ 34 ], emphasizing the importance of sea ice habitats in the functioning of the Arctic Ocean ecosystem [ 35 ]. The declining extent of sea-ice coverage, therefore, has profound impacts on the broader Arctic Ocean ecosystem, and the loss of thicker, structurally more complex, multi-year sea ice diminishes the range of productive niches available to sea-ice microbes [ 36, 37 ]. The loss of multi-year sea ice in a region is associated with long-term taxonomic shifts in the microbial communities of underlying water, for example, communities in the Beaufort Sea [ 38, 39 ] showed a decline in the abundance of multiple microbial groups relevant to biogeochemical cycling within the region. These included Bacteroidetes , which are typically associated with processing complex organic carbon [ 40 ], likely from diatoms, and marked reduction in the abundance of ammonia-oxidizing Thaumarchaeota, which correlates with a decline in nitrate, the limiting nutrient for productivity in the region Concomitant with the loss of sea ice, the expansion of open water as a habitat in the Arctic Ocean is prompting the immigration of microbial groups once thought limited to lower latitudes, with potential effects on bacterial lineages that may be endemic in the Arctic Ocean [ 41, 42 ]. Most notably, coccoid picocyanobacteria ( Synechococcus sp.), major marine primary producers in lower-latitude oceans, were once considered essentially absent in the colder waters of the Arctic Ocean [ 43 ]; however, they are now found as far north as 82.5°N, with abundant populations in Atlantic water reaching north-west of Svalbard (79°N) [ 44 ]. Picocyanobacteria such as Synechococcus and Prochlorococcus can survive the dark season in the Arctic Ocean [ 45 ], through photoheterotrophic metabolism. Consequently, they may out-compete eukaryotic algae [ 46 ] within the autumn and winter seasons in the Arctic Ocean. Therefore, they may be well placed to respond to the continued warming and loss of sea ice [ 47 ]. Indeed, it has been assumed that solar radiation offers a bottom-up control on the overall structure and function of Arctic marine ecosystems, and that the lengthy dark period of the polar night represents a quiescent phase in the ecology of the Arctic Ocean. However, with the increased expanse of open water and the warming of the Arctic Ocean's surface, it has become clear that polar night represents a period of microbial activity that has hitherto been largely overlooked [ 48 ]. Historically considered synonymous, winter, as a period of sustained low temperatures resulting in extensive sea ice, is now discrete from polar night, as a period of sustained darkness [ 49 ]. The implications of this decoupling of polar night and winter beg for further research attention to its consequences for microbial dynamics in a warming Arctic [ 47, 48 ]. In summary, understanding the range shifts, genomic adaptations and population structures of important marine primary producers, such as these cyanobacteria [ 47 ], will inform our predictions of how the food webs of the Arctic Ocean will respond to a future in which the extent and longevity of sea ice is severely curtailed. Carbon release from thawing permafrost Nearly half of global soil organic carbon is found in Arctic soils (1330–1580 Pg C) [ 18 ]. Most of this carbon is stored within permanently frozen ground, including permafrost (i.e. soil that remains frozen for 2 or more consecutive years) [ 50 ]. However, with warming temperatures across Arctic lands, this accumulated stock of legacy carbon from past climates does not represent a permanent sink of carbon. Indeed, thawing permafrost reinvigorates microbial communities that consume these carbon stores and release greenhouse gases in the form of carbon dioxide (CO 2 ) and methane (CH 4 ) to the atmosphere, with potent climate consequences [ 18, 51 ]. If our current trajectory of global warming is continued (IPCC RCP8.5), by the end of this century 30–99 % of near surface permafrost in the Arctic will have been degraded [ 52 ], with an associated release of 37–174 Pg C due to microbial processes within the thawed soils [ 18 ]. Understanding the drivers, rates, mechanisms and extent of microbial carbon cycling in Arctic soils is, therefore, an urgent priority, considering the large uncertainties apparent in the estimated ranges of carbon release [ 53 ]. Changes in Arctic lands also affect Arctic rivers and the Arctic Ocean [ 54, 55 ]. Thawing and degradation of permafrost will liberate substantial terrestrially derived organic matter (tDOM) for riverine transport to the Arctic Ocean [ 56 ]. Approximately 44 Tg organic carbon is released annually to the ocean from coastal and terrestrial permafrost within the Siberian Arctic alone, the bulk of which is predicted to be respired to carbon dioxide [ 57 ]. Microbes within coastal Arctic fjords respond readily to the influx of tDOM. Experiments simulating the release of permafrost carbon to a Svalbard fjord showed Glaciecola populations expanded aggressively following the addition of terrestrial dissolved organic matter [ 58 ]. Since Glaciecola is a gammaproteobacterial lineage associated with rapid consumption of organic matter in cold waters [ 59 ] following spring phytoplankton blooms, it is possible this effect may have broader impacts on the bacterial processing of organic carbon and food webs in the coastal Arctic [ 58 ]. The influence of tDOM extends further than the Arctic shoreline. The first study to develop metagenome-assembled genomes (MAGs) from the Arctic Ocean recently revealed marine Chloroflexi populations with the capability for degrading highly aromatic tDOM [ 20 ]. It appears that this capability arose through lateral gene transfer from terrestrial lineages, since the aromatic metabolism genes detected within the marine Chloroflexi MAGs were closely related to homologues present in terrestrial actinobacteria, acidobacteria and proteobacteria, while the parent lineage of Chloroflexi SAR202 has been found in deeper and darker (see the previous section) Arctic waters [ 60 ]. While the lateral transfer of carbon between Arctic lands and the Arctic Ocean is well acknowledged [ 55 ], the interactions concerning lateral gene transfer between terrestrial and marine Arctic microbial genomes and climate-driven changes in the biogeochemical cycles of the Arctic clearly merit further exploration. Permafrost itself is considered an unusual microbial habitat [ 61 ], since it represents a structurally heterogeneous environmental matrix that combines a long-term deep-frozen store of microbial biomass and genomic diversity from past climates (even >1 million years old) [ 18 ], which may become reactivated upon the degradation of the permafrost [ 62 ]. Recent chronosequence surveys [ 63, 64 ] have examined the survival mechanisms of viable microbiota within permafrost dating from the Pleistocene period (samples that dated to 19 000–33 000 years before the present). These reveal that while endospore-forming taxa are prevalent, viable biomass from some of these taxa remain as vegetative cells [ 64 ], underlining the potential for long-term, low-growth-rate survival rather than sporulation as a persistence mechanism within permafrost. Cold laboratory incubations (reviewed by Nikrad et al . [ 65 ]) lend further support to the potential for activity and even growth of microbial populations in sub-zero conditions [ 66 ]. Analysis of pangenomes from isolate and MAGs from permafrost could, therefore, provide insights to survival mechanisms and microevolutionary processes across geological timescales. However, the urgent questions prompted by the prospect of extensive permafrost degradation relate to the rates and pathways of carbon metabolism in thawing permafrost. In particular, our ability to predict the relative magnitude of permafrost carbon released as either CO 2 or as CH 4 , which has an estimated global warming potential ~30 times greater than CO 2 [ 6 ], is essential for predicting the role of permafrost in climate warming feedbacks. In-field gas flux measurements or cold-lab incubations alone have not offered an integrative view of microbial contributions to greenhouse gas evasion from permafrost [ 16 ]; therefore, recent work has focused on the integration of multi-omics approaches with biogeochemical process measurements of permafrost thaw incubations [ 67, 68 ], as well as permafrost cores [ 69 ]. Most recently, large-scale genome-centred metagenomics conducted across a permafrost thaw gradient have underlined the importance of linking processes with pathways and taxa by revealing novel fungal pathways for plant polysaccharide degradation and syntrophic interactions resulting in CH 4 production [ 70 ]. Furthermore, the potential for viruses to modulate carbon cycling through methanogen infection or lateral gene transfer of carbon processing pathways has been revealed by metavirome sequencing [ 71 ]. In determining the magnitude of CH 4 release from thawing Arctic permafrost, a critical question is posed by the capability of CH 4 -oxidizing microbes (methanotrophs) to consume CH 4 formed by archaeal CH 4 producers (methanogens) before it can reach the atmosphere. Methanotrophs may modulate the release of between 20 and 60 % of the CH 4 formed in tundra wetlands [ 72–74 ]. Linking genome-centred metagenomics with metatranscriptomics is revealing the diversity of methane-processing genetic mechanisms present within thawing permafrost. One highlight has been the identification of a resilient, dynamic methanotrophic community that can utilize isozymes with differential affinities for CH 4 ; thus, aiding their ability to persist through fluctuating CH 4 availability [ 75 ]. Understanding the potential for methanotrophy to mitigate methane release from thawing permafrost will require pairing biogeochemical process measurements with the capability to resolve diverse pathways for methane oxidation. These pathways must also be attributed to multiple lineages present within dynamic microbial communities. Pairing biogeochemical process measurements, physical and chemical analyses with genomics seems to offer a promising strategy to resolve the extent to which methanotrophy can offset methanogenesis in thawing tundra over the coming decades. To support accurate predictions of methane release, such approaches must embrace the complexity offered by a net outcome that is the sum of interactions within dynamic microbial consortia in structurally heterogeneous habitats defined by fluctuating oxygen, methane, terminal electron acceptor and water levels. Microbial impacts on Arctic glaciers and the Greenland Ice Sheet Arctic glaciers, ice caps and the Greenland Ice Sheet have started to experience the crippling consequences of Arctic warming, yet the melt that has been experienced to date is a mere fraction of what can be expected as the Arctic continues to warm. The largest of these glacial ice masses is by far the Greenland Ice Sheet, which occupies 1.7 million km 2 and currently sequesters the water equivalent of 7.4 m of sea-level rise [ 76 ]. Glacial meltwater is currently the largest contributor to sea-level rise [ 8 ], making this relationship a key societal concern and research priority. Models of glacier mass balance presently used to constrain estimates of sea-level rise currently do not incorporate microbiological parameters [ 8 ]. Addressing the role of microbiota in the rate and magnitude of Arctic glacial ice loss and climate warming, topics that have historically been overlooked, therefore, are of considerable importance to ensure robust estimations of future sea-level rise and climate change. That glacial systems are home to abundant and diverse life forms has been long established [ 77, 78 ]. This century has seen a refreshed synthesis of evidence for biodiverse microbial ecosystems associated with glaciers and ice sheets, and an acknowledgement that these biomes contribute to global biogeochemical cycles [ 79 ]. Like permafrost, glacial ice is a vast, climate-sensitive repository of microbial biomass and genomic diversity [ 80 ]. Equally, microbial processes at the surfaces and beds of Arctic glaciers and the Greenland Ice Sheet have the potential to amplify the impacts of climate warming on glaciers. The subglacial zone beneath glaciers and ice sheets is perennially dark and cold. Microbial ecosystems are apparent here, subsisting on organic carbon washed from the surface [ 81 ], relict carbon overridden in the last ice age [ 82 ] or existing through chemolithotrophy [ 83 ]. Critically, the evolution of anoxic conditions in subglacial habitats can favour methanogenesis [ 84 ] through both hydrogenotrophic and acetoclastic pathways [ 82 ]. Methane oxidation at the oxygenated glacial margins may mitigate subglacial methane production [ 85, 86 ]; however, contrasting results from neighbouring outlets of the Greenland Ice Sheet prompt uncertainty on whether methane oxidation can adequately compensate against methane production [ 85, 87 ]. Hitherto, a genomic perspective on subglacial ecosystem structure and function is in its infancy, with very few publicly available metagenomic datasets [ 88 ]. In contrast, the surfaces of Arctic glaciers receive abundant solar radiation in summer. This prompts the seasonal development of a range of microbial community types predominantly supported by photoautotrophy [ 15, 89 ]. Importantly, the accumulation of microbial biomass replete with photosynthetic and photoprotective pigments and recalcitrant dark organic matter at the glacier surface has the potential to influence the melting rate of the glacier surface [ 8 ]. The reduced surface reflectance of glaciers consequent to microbial growth in surface habitats has been termed 'biological darkening' [ 90 ] or 'bioalbedo' [ 91 ] in recent years. Estimates of microbial contributions to glacier melting are emerging [ 92–94 ]; however, integration of microbial-associated parameters in estimates of sea-level rise remains an active research goal. Snowpacks on Arctic glacial surfaces are vast environments that both support distinctive microbial consortia and are highly sensitive to warming [ 95–97 ]. The potential for snowpack bacteria to cycle climate-relevant trace gases has recently been highlighted as an emerging area [ 98, 99 ]. However, the growth and pigmentation of green algae in the family Chlamydomonadaceae in discrete patches on snow is particularly apparent. The consequent formation of intracellular carotenoid pigments can modulate local solar energy balance [ 100 ] and colour snowpacks bright red. Snow algal productivity can support a diverse range of heterotrophic taxa [ 101 ]; thus, subsidizing the development of a snowpack carbon cycle. Similarly, on the bare ice surface, members of the Zygnematophyceae glacier algae [ 102 ] form ice-darkening biofilms [ 103 ] that are particularly prominent on the south-western margins of the Greenland Ice Sheet, exhibiting locally structured populations [ 104 ]. Glacier algae influence surface reflectivity through the accumulation of dark photoprotective purpurogallin pigments [ 102 ], and their expansive spatial coverage can promote both surface ice ablation and carbon cycling [ 94, 102, 105 ]. Finally, cryoconite ecosystems are among the most intensively studied habitats on the glacier surface [ 106 ]. These collections of granular microbe-mineral aggregates (cryoconite) darken patches of the ice surface, resulting in localized melting and the formation of quasi-circular cylindrical melt holes [ 106 ]. Cryoconite granules are maintained at thermodynamic equilibrium depths and generally as single-granule layers at the floor of the cryoconite hole [ 107 ]. These responses to solar energy balance and cryoconite debris loads ensures the major primary producers in cryoconite, cyanobacteria, are continually exposed to optimal levels of solar radiation for photosynthesis; thus, promoting high levels of carbon fixation in spite of low ambient temperatures [ 107–109 ] Across Arctic glacial surfaces, a single lineage of filamentous cyanobacteria, Phormidesmis sp., appears responsible for binding together each cryoconite granule [ 99, 110–115 ]. A recent global-scale survey identified a single 16S-23S ITS haplotype of Phormidesmis priestleyi predominant upon Arctic glaciers [ 112 ]. Since the population structure of Phormidesmis sp. and other lineages of glacier cyanobacteria shows increasing fragmentation following the decline in glacierized surface area since the peak of the last ice age, it is likely they have an enduring role as ecosystem engineers of glacier surfaces [ 112 ]. Consequently, it is likely that one lineage of glacial cyanobacteria is predominantly responsible for the formation and maintenance of productive island-like microbial ecosystems within the austere environs of Arctic glacier surfaces, making cryoconite holes attractive for studies requiring naturally occurring microcosms of community development (cf. the article by Rivett and Bell [ 116 ]). Human dimensions of the changing Arctic: bioprospecting and infectious disease risks The Arctic has been inhabited by humans for millennia. It is now home to four million people. Moreover, the rapid changes in Arctic climate expected this century are leading to renewed interest in the economic potential of the Arctic as mineral resources, maritime navigation and tourism all become more accessible. This poses microbial risks and opportunities [ 117 ]. Firstly, growing commercial and political interests coupled with increased logistical accessibility is likely to stimulate interest in Arctic bioprospecting [ 23 ]. Adaptations for life in the cold found within the reservoir of Arctic genomic diversity can be industrially useful [ 118 ]. Examples include enzymes with low temperature optima [ 119 ], low-alcohol yeast [ 120 ], antifreeze and ice-binding proteins [ 121 ], and potential antimicrobials [ 122 ] ( Table 1 , fully referenced and expanded in Table S1, available with the online version of this article). To date, the majority of antimicrobial compounds and cold-active enzymes have relied on cultured isolates, which are either screened directly for activity [ 122 ] or genome sequenced; followed by the cloning and expression of candidate genes or gene clusters in a heterologous host [ 123 ]. These strategies rely on the isolation and genome-sequencing of microbes, which is limiting because: (i) fully sequenced genomes of Arctic strains are limited in number (see below), and (ii) many microbes remain uncultivated. In addition, advances in sequence-based and functional metagenomic approaches [ 124 ] offer promising approaches to mine and exploit such potential. For example, specially engineered heterologous expression hosts, such as the ArcticExpress Escherichia coli competent cells (Agilent Technologies) and Pseudoalteromonas haloplanktis TAC125 strain [ 123 ] are noteworthy for being products of cryospheric bioprospecting, and tools by which further functional exploration of this environment can be accomplished. ArcticExpress cells co-express the chaperonin system Cpn60 and Cpn10 from Oleispira antarctica , which helps to ensure the proper folding of cold-active proteins and increase the growth rate of E. coli at low temperatures [ 125 ]. Pseudoalteromonas haloplanktis TAC125, however, is of cryospheric origin, and displays increased solubility and secretion of protein products over other Gram-negative expression hosts. Meanwhile, improvements in sequencing technologies have resulted in the assembly of longer contigs, with deeper coverage than ever before, unlocking types of analyses previously available only to whole-genome-sequenced cultured organisms. Tools such as antiSMASH [ 126 ], for example, can be used on metagenomic datasets (contigs or MAGs) to detect biosynthetic gene clusters responsible for the synthesis of industrially useful compounds, such as antibiotics, fatty acids, polysaccharides, antioxidants and UV-protective pigments. However, the greatest improvements in bioprospecting will likely come from the synergy of sequence-based and functional methods, because an understanding of the genomic background of the source organism (see below) is vital for the strategic genetic engineering of suitable hosts and the identification of optimal conditions for expression of the desired natural product. Table 1. Key cold adaptations relevant for biotechnology Target product Adaptation mechanism Application Cold-active enzymes Amino acid changes that increase enzyme flexibility Food industry, detergents, molecular biology tools Polyunsaturated fatty acids Polyunsaturated fatty acids increase membrane permeability at low temperatures Dietary supplements for humans, livestock and fish Ice nucleation proteins Seed small crystals instead of large damaging crystals Food industry, synthetic snow Antifreeze proteins and solutes Prevent water molecules from forming ice-crystal structure Cryoprotectants, food industry Antioxidants and UV pigments Protect micro-organisms from seasonally high UV irradiation in snow Biomedical, pharmaceutical, food technology and cosmetics Exopolysaccharide Trapping of liquid water, preventing freezing Biomedical, pharmaceutical, food technology and cosmetics Antimicrobial compounds Chemical defences and weapons against competing bacteria in low-resource environments Pharmaceuticals: antibiotics, antifungals, anti-tumour medications and pesticides Secondly, the Arctic is not pristine and has not been pristine for some time [ 127 ]. There is a long history of human activities that have contaminated the Arctic in many ways, from hydrocarbon exploitation [ 128, 129 ] to military activities, including the largest nuclear explosion to date [ 130 ]. These have resulted in locally derived contamination of the Arctic. Likewise, long-range atmospheric transport of pollutants and the global distillation effect has led to the deposition of pollutants in the Arctic from the mid-latitudes for at least 3 millennia [ 127 ]. The potential roles of microbes in modulating or exacerbating the threats posed by contaminants liberated by Arctic warming is a current focus for researchers addressing radionuclide [ 131 ], persistent organic pollutant [ 132 ], black carbon [ 133 ], mercury [ 134 ] and heavy metal contaminants [ 135 ]. Furthermore, both increased access to the Arctic and the potential release of long-frozen hazards is raising the prospect of the liberation of ancient infectious diseases [ 136, 137 ]. Such notions seem speculative, for they depend on the release of viable agents able to withstand severe freeze–thaw stresses as they migrate to the actively melting layers of glaciers or permafrost [ 138 ]. Hitherto, dedicated efforts to recover pathogens from the Arctic have failed, as human remains have degraded within the active layer of permafrost [ 139 ]. However, the release of Bacillus anthracis from frozen wildlife carcasses has been invoked as the cause of a recent Siberian anthrax outbreak [ 140 ]. A further hypothetical risk is presented by implementing synthetic biology approaches to resurrect poorly described viral genomes from Arctic ice [ 141 ], as these may generate highly concentrated infectious materials within laboratories [ 137 ]. The ethical debate and moratorium on resurrecting highly pathogenic influenza strains [ 142 ] offers a certain precedent for concern within this arena. Finally, these changes in the accessibility and ecology of the Arctic bring with them pressures for human healthcare [ 143 ]. These include increased demand on the limited healthcare services available or the immigration of emerging infectious diseases; for example, as vectors move polewards [ 144 ]. This may necessitate enhanced microbiological surveillance and diagnostic capability; and distributed or ubiquitous genomic sensing [ 145 ] may prove important in detecting and managing microbial threats as the Arctic experiences disruptive change. Challenges and opportunities for Arctic microbial genomics A recurring theme within the preceding sections is that there are significant lacunae in our understanding of Arctic microbial genomics. These constrain both our appreciation of the fundamental properties of Arctic microbial ecosystems, and our ability to predict their interactions with the aggressively changing climate of the Arctic. In a time where microbial genome sequencing in other study areas is all but routine [ 146, 147 ], and expensive, expansive efforts to catalogue microbial diversity across the planet yield transformative results [ 148, 149 ], why are the diverse and societally relevant genomes of Arctic microbes genomes left out in the cold? This part of the review will identify some of the salient challenges faced in Arctic microbial genomics and opportunities to address them. These challenges range from conceptual to technical and logistical considerations; thus, there is scope for innovation, which could prove both timely and transformative for Arctic microbial genomics. Challenge 1 – insufficient data to develop unifying concepts for life in the cold Most undergraduate microbiology textbooks may define psychrophiles (or cryophiles) in relation to organisms with relatively colder cardinal temperatures for in vitro growth [ 150 ]; therefore, this challenge may seem surprising. In many cases, it has been assumed that Arctic ecosystems are populated by such psychrophiles. Indeed, there are many organisms from cold regions that are isolated in culture and exhibit growth at low temperatures [ 151–153 ]. Furthermore, various well-described traits are linked to growth at low temperature in vitro , ranging from changes in enzyme structure to membrane fluidity to stress responses [ 154 ]. Table 2 summarizes key genes invoked in cold adaptation in laboratory studies of bacteria, which are then fully detailed in Table S2. These indicate a broad array of mechanisms for cold adaptations, including cold-shock responses to DNA topology modulation, protein synthesis and stabilization, and metabolic processes. Table 2. Summary of genes implicated in cold adaptation Gene Protein Assigned function dnaA DnaA DNA binding, replication initiation, global transcription regulator dnaG DnaG DNA primase gyrA GyrA DNA cleaving/binding/re-joining subunit of DNA gyrase hns H-NS Nucleoid protein, transcriptional repressor, DNA supercoiling hupB Hu-β Nucleoid protein, DNA supercoiling recA RecA General, homologous recombination, DNA repair, SOS response cspA CspA Cold-inducible RNA chaperone, RNA and DNA binding, anti-terminator, transcriptional enhancer cspB CspB Cold-shock-inducible, RNA binding cspE CspE Cold induced in lag phase RNA chaperone, RNA binding, transcriptional anti-terminator, inhibits PNPase and RNase E, regulation of and expression of stress response proteins RpoS and UspA deaD DeaD ATP-dependent RNA helicase, aids ribosome assembly, possibly involved in RNA degradation pnp PNPase Cold-shock protein required for growth at low temperatures, 3′→5′ exoribonuclease, component of RNA degradosome, purine phosphorylase nusA NusA Transcription termination/antitermination/elongation L factor infA IF-1 Protein chain (translation) initiation factor IF-1, RNA binding infB IF-2 Protein initiation factor, translation initiation, fMet-tRNA binding, chaperone infC IF-3 Protein initiation factor, translation initiation, initiation site selection, RNA binding, stimulates mRNA translation rbfA RbfA 30S ribosome-binding factor processing of 16S rRNA (3′→5′ exonucleases) rnr Ribonuclease R Cold-shock induced, ribosome assembly/maturation yfiA pY Protein Y, 30S ribosomal subunit linked, inhibits translation dnaJ DnaJ Chaperone dnaK DnaK Chaperone hscA Hsc66/HscA DnaK chaperone homologue (Hsp70-type protein chaperone) hscB HscB DnaJ co-chaperone homologue (for HscA) tig Trigger factor Multiple stress protein, chaperone, protein-folding, ribosome-binding aceE AceE Pyruvate dehydrogenase E1 component, decarboxylase aceF AceF Pyruvate dehydrogenase, dihydrolipoamide acetyltransferase lpxP Palmitoleoyltransferase Cold-inducible, lipid A biosynthesis otsA OtsA Cold-induced and essential, trehalose phosphate synthase otsB OtsB Cold-induced and essential, trehalose phosphate phosphatase cspC CspC Regulation of growth and the stress response proteins RpoS and UspA cspD CspD Stationary phase induced and nutrient starvation, DNA replication inhibition, biofilm development, persister cell development cspF CspF Very-low-level expression with no detected protein product cspG CspG Cold-inducible, cold-shock protein homologue cspH – Very-low-level expression with no detected protein product cspI CspI Cold-inducible, cold-shock protein homologue ves Ves Cold- and stress-inducible Nevertheless, psychrophily itself almost seems to be a conceptual afterthought defined by contrast to thermophily and mesophily [ 150 ]. Critically, whether psychrophily is ecologically relevant remains open to question, for colder optimal growth temperatures in the laboratory do not necessarily translate to increased fitness in low-temperature environments. Recently, Cavicchioli [ 155 ] provided a detailed critique of the relevance of psychrophily as a concept for life in the cold. It is clear that not only is in vitro psychrophily defined differently by different workers, but also there are striking examples of discordant patterns in the growth optima of organisms prominent within low-temperature environments. Moreover, defining psychrophily on the basis of colder cardinal growth temperatures shown by axenic cultures in vitro fails to embrace the diverse range of stresses likely experienced by organisms in Arctic ecosystems. These can include resource and nutrient limitations, energy constraints, UV radiation and reduced water activity [ 156 ]; all of which may also act in concert with biotic factors such as competition or predation [ 157 ]. Therefore, it could be argued that, by itself, the concept of psychrophily as defined by the growth rates of an axenic culture in vitro fails to offer an adequate framework for understanding the adaptations and functioning of Arctic microbes. Yet few ecologically meaningful alternatives have been advanced. Cavicchioli [ 155 ] offers the elegant suggestion that the term 'psychrophile' applies to any microbe that is indigenous to a cold environment. While this has a certain utilitarian advantage and is certainly inclusive, it is perhaps overly inclusive. Since our understanding of microbial biogeography remains patchy, with continued debate on the validity of an early 20th century concept on whether all microbes are indigenous everywhere [ 41, 158–160 ], and our techniques for detecting the presence of microbes extrapolative, classifying whether microbes (or their phylotypes) are indigenous or transient in a given environment remains problematic. It also fails to recognize the potential for transient and immigratory microbes to make important contributions to Arctic ecosystem functioning [ 44 ]. In spite of these limitations, if we define psychrophiles as microbes indigenous to cold environments ( sensu the paper by Cavicchioli [ 155 ]), it is clear we are profoundly limited in our ability to define the genomic basis of psychrophily. This is for the simple reason that we lack microbial genomes from the cold environments, including the Arctic. At the time of writing, fewer than a hundred microbial genomes from the Arctic are listed in public databases. Critically, to our knowledge, only one cyanobacterial genome from the terrestrial Arctic is publicly available [ 115 ], which frustrates our understanding of the comparative genomics of a major group of primary producers in the terrestrial Arctic. Representative genome sequences for other key microbial groups are also severely limited in their public availability. Importantly, this frustrates any effort to select representative model organism(s) for the genome-centred study of Arctic microbiology. Inspiration for solving this challenge can be found readily within neighbouring fields of microbial genomics. While high-throughput reconstruction of genomes from deeply sequenced environmental metagenomes offers culture-independent insights to landscape-scale processes in the Arctic (e.g. carbon release from permafrost or marine degradation of tDOM) [ 20 ], the potential for experimental validation of the MAGs or exploration of their corresponding phenotypes is curtailed. Furthermore, while the world is turning to the High Arctic to preserve genomic diversity in agricultural crops [ 161 ] and open source code ( https://archiveprogram.github.com/ ), there is no corresponding effort to conserve the microbial diversity endangered by Arctic change [ 162 ]. Therefore, there is value in the systematic isolation, cultivation, genome sequencing and experimental analysis of Arctic bacteria. This approach offers the advantage of high-quality reference genome sequences coupled with the curation of the source isolate for later experimental verification. The 'Hungate1000' collaboration for sequencing ruminant microbial genomes [ 163 ] offers one potential blueprint for community-led sequencing of Arctic microbial isolates. The establishment of a dedicated sequencing and strain curation facility provides an alternative, service-based model. Irrespective of the approach towards the generation of such a resource, systematic sequencing, curation and experimentation with Arctic isolates both conserves Arctic microbial biodiversity and creates an enabling platform. An 'Arctic1000' project would permit for selecting model organisms, testing discrete hypotheses, refining gene annotations and resolving the evolution of cold adaptation [ 155, 164 ]. In summary, bringing microbial genomes in from the cold is a necessary but tractable first step in gathering evidence for a unifying concept for psychrophily that will also likely be of relevance for understanding microbial life in other cold regions. Challenge 2 – phylotypes obscure genomic diversity in Arctic microbes The profound dearth of available Arctic microbial genomes means that most Arctic microbes known to science are outlined by their phylotypes. Most recently these are viewed through the lens of amplicon sequencing of specific loci [e.g. 16S rRNA genes or 16S rRNA (cDNA) or eukaryotic equivalents] but, historically, community fingerprinting commonly served similar purposes [ 14, 28, 165 ]. PCR-dependent amplification and targeted sequencing permits description and comparison of community composition from low-biomass density environments that are vast in scale (e.g. snow or ice across the Greenland Ice Sheet [ 97, 99 ] and Arctic air samples [ 166–168 ]). However, these applications amplify the challenges typical of amplicon sequencing approaches for cataloguing or comparing microbiomes. A particular challenge for Arctic microbiologists is that many of the taxa prevalent in amplicon sequencing studies of Arctic microbiomes are close relatives of frequently observed contaminants of the amplicon sequencing process. The 'kitome', or the contaminated reagent microbiome [ 169 ], typically comprises a range of organisms well adapted to oligotrophic conditions, stresses from low water activity (albeit in high-salt solutions) and cold storage. In short, molecular reagents are often facsimiles of the stresses common in polar environments. Moreover, low-biomass samples are typical of many habitat types across the Arctic, for example, snow, ice or freshwater habitats [ 15 ], and the impact of contamination is magnified in such samples [ 170, 171 ]. These trends are supported by the coincidence of many authentic members of Arctic microbial communities among blacklisted taxa from microbiome analyses [ 169 ] ( Fig. 2 ). Fig. 2. Neighbour-joining tree of partial 16S rRNA gene sequences from isolates in culture from Arctic habitats that are also reported [ 169 ] as frequently occurring contaminants in sequenced negative controls. The tree comprises 52 alignable sequences from the 56 available isolate sequences drawn from the 92 genera named in table 1 of the paper by Salter et al . [ 169 ]. All seven of the groups with named genera listed by Salter et al . [ 169 ] are represented in cultures from Arctic environments. Actinobacteria , red; Alphaproteobacteria , blue; Betaproteobacteria , purple; Gammaproteobacteria , brown; Firmicutes , pink; Deinococcus - Thermus , green; and Bacteroidetes, gold. Scale shows nucleotide substitutions per site. Disentangling authentic from contaminant taxa present in amplicon sequence data, therefore, poses particular challenges for Arctic microbiologists. As well as the type I error (inclusion of contaminants in microbiome profiles), the scope for type II error (the exclusion of authentic taxa) is enhanced. Therefore, rigorous experimental design and management must be emphasized. This can include the implementation of contamination-mitigation practices (cf. the article by Willerslev et al . [ 172 ]) during sample collection and processing. However, consistently achieving and verifying sterility within the laboratory, let alone during sampling activities in expeditionary conditions, is challenging if not impractical. Therefore, a suite of extraction, reagent-blank and mock community controls [ 170, 171 ], which are processed, sequenced and analysed alongside study samples, becomes critical. The use of automated contamination detection software requires careful manual curation, as their application on data from communities dominated by a small core of common taxa or keystone taxa (e.g. cryoconite [ 114 ]) may lead to the false negative rejection of those taxa [ 173 ]. The imperative for experimental good sense and good laboratory practices has recently been emphasized in microbiome analyses [ 174 ], and amplification-dependent studies of Arctic microbiota should be no exception. Furthermore, validation of key experimental trends should be normalized, which should include, as a minimum, identification of the closest environmental and cultured representatives of key phylotypes [ 175 ], and ideally orthogonal confirmation of their detection in culture or through phylogenetic staining of samples by fluorescent in situ hybridization (FISH). With the development of improved workflows (e.g Anvi'o [ 176 ]) for genome-resolved metagenomics, it may be that amplicon sequencing becomes an adjunct to the direct analysis of functional diversity represented within the genomes of Arctic microbes; for example, by selecting samples for more intensive study by genome-resolved metagenomics on the basis of community profiling. Challenge 3 – biomolecular stability as a pre-requisite for integrative multi-omics In spite of the critique of amplicon sequencing presented above, the capability to systematically compare the microbiomes of many samples has led to valuable collaborative efforts for the large-scale mapping of microbiomes [ 148 ]. While these efforts capture broad-scale trends in microbial biogeography, they are predominantly focused on the mid-latitudes [ 148, 149, 177 ]. Consequently, there is the risk of overlooking genomic diversity within Arctic microbial ecosystems and its interactions with climate change when conducting global-scale analyses [ 12 ]. It is likely that the limited availability of samples from high-latitude locations accounts for this bias, for the collection and recovery of microbial samples from the Arctic is non-trivial. Indeed, within the general field of ecology there is a profound station bias in the distribution of studies on Arctic climate change. A recent synthesis showed that 31 % of all study citations in 1840 publications on Arctic change relate to work performed in just two locations, Toolik Lake Station in Alaska and Abisko Station in Sweden [ 178 ]. Notable gaps in the literature include the microbiology of particularly rapidly changing regions of the Arctic, for example the Canadian Arctic Archipelago or the Russian Arctic coastline [ 178 ]. It is likely that the logistical challenges in accessing these vast and important areas are either too costly or practically prohibitive for many investigators. As an approximation, Mallory et al . [ 179 ] calculated the costs of animal ecological fieldwork are typically eight times greater in the Arctic than comparable work in the mid-latitudes, which increases the barrier for executing studies that then also require costly analyses in the home laboratory in the form of high-throughput sequencing. The high level of sample integrity required for most forms of microbial 'omics studies exacerbate this challenge. To transcend descriptive inventories of genomic diversity, the systematic capture of transient microbial gene products (transcripts, proteins, metabolites) can offer greater functional insight. Sampling forays straying beyond assured cold-chain archival of samples are, therefore, risky but essential for understanding microbial responses to the Arctic crisis in vast habitats far from the nearest freezer. Indeed, integrating different strands of 'omics methodologies can allow an investigator to reveal how a microbe is contributing to ecosystem function. Such contributions may be through many different routes, with important contributions by active, dormant, damaged or dead microbes in turn [ 180 ], which are difficult to disentangle both practically [ 180 ] and epistemologically [ 181 ]. These challenges are pronounced in harsh environments where microbes may be functioning under prolonged exposure to multiple stresses [ 156 ]. Representation within the RNA or protein pool of a habitat is consistent with contemporary contributions to ecosystem function, while the presence of functional genes within metagenomes indicates potential functional contributions, which may be in the past or future. Furthermore, differential extraction of vegetative or spore-associated DNA may indicate the potential for long-term storage of genes within an ecosystem [ 63, 64 ] and, finally, even the contribution of lysed microbes to the organic carbon and nutrient pool of an ecosystem can prove critical for foodwebs [ 182 ]. Within the realm of multi-omics studies, the paucity of in situ metatranscriptome studies of Arctic microbiota presents a notable lacuna in our understanding of ecosystem responses to climate stresses [ 67, 68, 75 ]. Multiple challenges must be addressed in implementing such experimental strategies. Firstly, microbial mRNA has a very brief half-life. For laboratory grown bacteria, this is typically in the order of minutes [ 183 ], but even assuming the potential for greater mRNA stability in slow-growing cells in low-temperature environments, the turnover of mRNA is likely within hours to a day [ 184 ]. Recovering representative high-quality mRNA (RNA Integrity Number >7; [ 185 ]) for transcriptome analyses in locations where liquid nitrogen flash-freezing may not be practical or permissible is, therefore, challenging. Such situations may include research performed at many Arctic stations, research performed over extended field campaigns and research performed in areas only accessible by small chartered aircraft. Furthermore, for aqueous or frozen sample matrices common in the Arctic (snow, ice, meltwater), samples require lengthy pre-processing that can include gentle melting and filtration over hours or days before chemical preservation or deep freezing becomes feasible. Secondly, mRNA is a minor species of RNA when compared to the abundance of rRNA within the cell [ 186 ]. Indeed Moran et al . [ 186 ] estimate that the typical marine bacterial cell may only contain ~200 transcripts at any one time. Using the same allometric assumptions [ 186 ] constrained by data on the median cell size of bacteria in glacial meltwater [ 90 ], it is possible that only ~40 transcripts are present in a bacterial cell eluted from glacial ice. When an investigator wishes to generate a snapshot inventory of transcripts in a sample, the low abundance of transcripts and the apparent stochasticity of transcription, therefore, pose practical and conceptual problems in quantifying transcripts and their relevance for ecological functions. Nevertheless, reducing the degradation of biomolecules within the parent environmental matrix appears vital for robust multi-omics studies. Two contrasting approaches are identified as a means of enhancing the fidelity of insights from multi-omics studies. Firstly, researchers have the option of relocating bulk environmental matrices to a controlled laboratory environment for incubation. This approach has the advantage of allowing experimental manipulation under precisely controlled conditions where confounding factors can be minimized, and treatments administered and measured precisely. For environmental matrices that can be sampled and transferred frozen in bulk, this strategy has proven fruitful. A key example is found within experimental studies of permafrost responses to controlled thawing [ 67, 68 ], which integrate process measurements of microbial activities with metatranscriptomics of samples incubated under controlled conditions. Such studies address the issue of biomolecular integrity by immediate extraction of nucleic acids (or liquid nitrogen flash-freezing) and have the conspicuous advantage of permitting experimentally replicated application of treatments under controlled conditions. For more labile environmental matrices (e.g. snow or water), habitats where low-activity states are typical (e.g. ice cores) or where replication of field conditions in the laboratory is more challenging [ 132 ], bringing the habitat to the laboratory is more problematic. Nevertheless, the establishment of faithful cold laboratory models of Arctic microbial communities or keystone taxa offers one potential route to enhanced functional insights. Secondly, the opposite strategy of relocating the microbial genomics laboratory to the sample is an increasingly viable option. The development of third-generation DNA sequencers that are field-deployable in austere conditions is prompting innovation in this area. In particular, nanopore sequencing using MinION from Oxford Nanopore Technologies permits a user to rapidly analyse crude extracts of nucleic acid through shotgun metagenomics or carry out amplicon-based analyses (e.g. 16S rRNA gene sequencing) on a USB-based device. Within the Arctic, this strategy has been implemented in a field station in the Canadian Arctic [ 187 ], field camps in Greenland [ 188 ] and as part of a student expedition on the Vatnajökull ice cap of Iceland [ 189 ]. The selection of miniaturized equipment for battery-powered DNA extraction, simplified library preparation using freeze-dried reagents and sequencing on solar-powered or battery-powered instruments [ 188, 189 ] permits on-site characterization of Arctic microbiota, which can focus sampling plans and reduce logistical risks from sample loss or degradation during transport. Moreover, sequencing on-site allows the real-time completion of the scientific method in situ , rather than incurring potential delays by returning samples to a remote laboratory ( Fig. 3 ). Fig. 3. Example of in-field DNA sequencing and analysis. Working in a remote field camp on the Greenland Ice Sheet during the Arctic winter (a), it was possible to extract nucleic acids and sequence them in ambient temperatures of circa −20 °C by using freeze-dried reagents and adapted protocols for nanopore sequencing (b). In situ data processing and analysis (c) permitted a refined experimental strategy (d) for genome-centred metagenomic comparison of glacial habitats (e). Images (a) and (b) by J. M. Cook; images (c) and (d) by A. Edwards. Image (e) is representative of unpublished data from A. Edwards, M. C. Hay and J. M. Cook. In summary, it is likely that a combination of these approaches will offer Arctic microbial genomics researchers the balance between capturing in situ processes and the option of precisely controlled laboratory experiments, both of which will be required to advance our knowledge of microbial responses to the warming Arctic. Challenge 4 – business as usual for Arctic microbial genomics? The study of Arctic change in the 21st century could be summarized as the measure of a biome-scale response to perturbation on a scale unprecedented in human history. Implicit within this is the knowledge that the resilience of both Arctic ecosystems and Arctic researchers is threatened. Indeed, the burden of ecological grief [ 190 ] upon environmental researchers has recently been acknowledged for its impact on their mental health [ 191 ]. Climate change impacts are also complicating the study of Arctic microbiology in other critical ways, from the rapid destruction or fragmentation of study habitats [ 192–194 ], to anomalous weather patterns [ 195, 196 ], to the disruption of infrastructure and logistics. Considering the inherent biases and undersampling of Arctic habitats to date (see the previous section), fieldwork will remain a core requirement for capturing the diversity and dynamics of Arctic microbial genomes across spatial and temporal scales. However, remote fieldwork in the Arctic brings with it additional logistical costs and safety risks [ 179 ], and exposure to the increased frequency of extreme weather conditions [ 8 ] and their impacts on avalanche, landslide or sea ice disintegration adds to these risks. Furthermore, it must be acknowledged that Arctic microbiology is a carbon-intensive research discipline, with considerable quantities of greenhouse gases released both in regular travel to remote field locations and in the maintenance of cold laboratories and ultra-freezers. Sustainable and innovative ways of exploring the microbiology of the Arctic, while minimizing the contribution of Arctic microbiologists to further environmental change, are required. The rapid technological developments within microbial genomics have meant that the wealth of data collected and archived in public repositories dates rapidly. Today's ultra-deep sequencing experiment becomes tomorrow's shallow metagenome 'skim' [ 197 ]. Welcome increases in capacities for sequencing and data processing [ 198 ], enhancements in standards for metadata reporting, and recognition of methodological biases, all contribute to the depreciation or obsolescence of unique datasets harvested with great costs and potential environmental impacts. The Antarctic microbiology research community has recognized the value of preserving community DNA samples in public archives in addition to high standards of data curation [ 199, 200 ]. This creates a resource that can be explored retrospectively and shared among researchers to reduce the requirement for additional fieldwork at additional cost and environmental impact. Furthermore, such sample archives provide both a baseline for contemporary genomic diversity and an insurance against habitat loss in the face of Arctic change [ 162, 199, 200 ]. Establishing microbial observatories with the capacity for longitudinal studies of changes in Arctic microbial ecosystems is one potential area of development. The Arctic is home to many scientific research stations ( https://eu-interact.org/ ) and research centres, most of which collect and archive a valuable range of environmental measurements, for example meteorological or atmospheric chemistry data [ 201 ]. These datasets underpin our quantitative view of the Arctic's present and past; thus, delivering the raw ingredients for forecasting changes in the Arctic and the entire planet [ 8 ]. To our knowledge, at present, there is no comparable effort aimed at monitoring Arctic microbiota, in spite of the critical interactions between Arctic microbes and climate change [ 202 ]. Creating deployable genomics resources can distribute the task of characterizing and monitoring Arctic microbiota to a network of such stations. By analogy, the inclusion of distributed, real-time pathogen genomics is considered a key data stream for a prospective global surveillance system, which integrates human, animal and ecosystem health [ 145 ]. Within the Arctic context, this 'sequencing singularity' [ 145 ] would entail the contemporaneous monitoring of Arctic microbiota, along with the physical and chemical environment of the warming Arctic; thus, capturing the complexity of microbe-mediated feedbacks in Arctic change and offering a rich new stream of data to enhance our predictive understanding of Arctic change. Hitherto, the study of Arctic microbial genomics has typically entailed either laboratory-based analyses, field surveys or plot scale experiments [ 67, 71, 98 ]. These are most pertinent for understanding the molecular ecology of the cryosphere at scales of microns to metres, with further insights typically extrapolative. However, it is the emergent macroscale effects of Arctic microbial activity that have relevance for contemporary climate change, requiring microscopic processes to be studied at the scale of entire landscapes. This scale mismatch between processes occurring at the 7; [ 185 ]) for transcriptome analyses in locations where liquid nitrogen flash-freezing may not be practical or permissible is, therefore, challenging. Such situations may include research performed at many Arctic stations, research performed over extended field campaigns and research performed in areas only accessible by small chartered aircraft. Furthermore, for aqueous or frozen sample matrices common in the Arctic (snow, ice, meltwater), samples require lengthy pre-processing that can include gentle melting and filtration over hours or days before chemical preservation or deep freezing becomes feasible. Secondly, mRNA is a minor species of RNA when compared to the abundance of rRNA within the cell [ 186 ]. Indeed Moran et al . [ 186 ] estimate that the typical marine bacterial cell may only contain ~200 transcripts at any one time. Using the same allometric assumptions [ 186 ] constrained by data on the median cell size of bacteria in glacial meltwater [ 90 ], it is possible that only ~40 transcripts are present in a bacterial cell eluted from glacial ice. When an investigator wishes to generate a snapshot inventory of transcripts in a sample, the low abundance of transcripts and the apparent stochasticity of transcription, therefore, pose practical and conceptual problems in quantifying transcripts and their relevance for ecological functions. Nevertheless, reducing the degradation of biomolecules within the parent environmental matrix appears vital for robust multi-omics studies. Two contrasting approaches are identified as a means of enhancing the fidelity of insights from multi-omics studies. Firstly, researchers have the option of relocating bulk environmental matrices to a controlled laboratory environment for incubation. This approach has the advantage of allowing experimental manipulation under precisely controlled conditions where confounding factors can be minimized, and treatments administered and measured precisely. For environmental matrices that can be sampled and transferred frozen in bulk, this strategy has proven fruitful. A key example is found within experimental studies of permafrost responses to controlled thawing [ 67, 68 ], which integrate process measurements of microbial activities with metatranscriptomics of samples incubated under controlled conditions. Such studies address the issue of biomolecular integrity by immediate extraction of nucleic acids (or liquid nitrogen flash-freezing) and have the conspicuous advantage of permitting experimentally replicated application of treatments under controlled conditions. For more labile environmental matrices (e.g. snow or water), habitats where low-activity states are typical (e.g. ice cores) or where replication of field conditions in the laboratory is more challenging [ 132 ], bringing the habitat to the laboratory is more problematic. Nevertheless, the establishment of faithful cold laboratory models of Arctic microbial communities or keystone taxa offers one potential route to enhanced functional insights. Secondly, the opposite strategy of relocating the microbial genomics laboratory to the sample is an increasingly viable option. The development of third-generation DNA sequencers that are field-deployable in austere conditions is prompting innovation in this area. In particular, nanopore sequencing using MinION from Oxford Nanopore Technologies permits a user to rapidly analyse crude extracts of nucleic acid through shotgun metagenomics or carry out amplicon-based analyses (e.g. 16S rRNA gene sequencing) on a USB-based device. Within the Arctic, this strategy has been implemented in a field station in the Canadian Arctic [ 187 ], field camps in Greenland [ 188 ] and as part of a student expedition on the Vatnajökull ice cap of Iceland [ 189 ]. The selection of miniaturized equipment for battery-powered DNA extraction, simplified library preparation using freeze-dried reagents and sequencing on solar-powered or battery-powered instruments [ 188, 189 ] permits on-site characterization of Arctic microbiota, which can focus sampling plans and reduce logistical risks from sample loss or degradation during transport. Moreover, sequencing on-site allows the real-time completion of the scientific method in situ , rather than incurring potential delays by returning samples to a remote laboratory ( Fig. 3 ). Fig. 3. Example of in-field DNA sequencing and analysis. Working in a remote field camp on the Greenland Ice Sheet during the Arctic winter (a), it was possible to extract nucleic acids and sequence them in ambient temperatures of circa −20 °C by using freeze-dried reagents and adapted protocols for nanopore sequencing (b). In situ data processing and analysis (c) permitted a refined experimental strategy (d) for genome-centred metagenomic comparison of glacial habitats (e). Images (a) and (b) by J. M. Cook; images (c) and (d) by A. Edwards. Image (e) is representative of unpublished data from A. Edwards, M. C. Hay and J. M. Cook. In summary, it is likely that a combination of these approaches will offer Arctic microbial genomics researchers the balance between capturing in situ processes and the option of precisely controlled laboratory experiments, both of which will be required to advance our knowledge of microbial responses to the warming Arctic. Challenge 4 – business as usual for Arctic microbial genomics? The study of Arctic change in the 21st century could be summarized as the measure of a biome-scale response to perturbation on a scale unprecedented in human history. Implicit within this is the knowledge that the resilience of both Arctic ecosystems and Arctic researchers is threatened. Indeed, the burden of ecological grief [ 190 ] upon environmental researchers has recently been acknowledged for its impact on their mental health [ 191 ]. Climate change impacts are also complicating the study of Arctic microbiology in other critical ways, from the rapid destruction or fragmentation of study habitats [ 192–194 ], to anomalous weather patterns [ 195, 196 ], to the disruption of infrastructure and logistics. Considering the inherent biases and undersampling of Arctic habitats to date (see the previous section), fieldwork will remain a core requirement for capturing the diversity and dynamics of Arctic microbial genomes across spatial and temporal scales. However, remote fieldwork in the Arctic brings with it additional logistical costs and safety risks [ 179 ], and exposure to the increased frequency of extreme weather conditions [ 8 ] and their impacts on avalanche, landslide or sea ice disintegration adds to these risks. Furthermore, it must be acknowledged that Arctic microbiology is a carbon-intensive research discipline, with considerable quantities of greenhouse gases released both in regular travel to remote field locations and in the maintenance of cold laboratories and ultra-freezers. Sustainable and innovative ways of exploring the microbiology of the Arctic, while minimizing the contribution of Arctic microbiologists to further environmental change, are required. The rapid technological developments within microbial genomics have meant that the wealth of data collected and archived in public repositories dates rapidly. Today's ultra-deep sequencing experiment becomes tomorrow's shallow metagenome 'skim' [ 197 ]. Welcome increases in capacities for sequencing and data processing [ 198 ], enhancements in standards for metadata reporting, and recognition of methodological biases, all contribute to the depreciation or obsolescence of unique datasets harvested with great costs and potential environmental impacts. The Antarctic microbiology research community has recognized the value of preserving community DNA samples in public archives in addition to high standards of data curation [ 199, 200 ]. This creates a resource that can be explored retrospectively and shared among researchers to reduce the requirement for additional fieldwork at additional cost and environmental impact. Furthermore, such sample archives provide both a baseline for contemporary genomic diversity and an insurance against habitat loss in the face of Arctic change [ 162, 199, 200 ]. Establishing microbial observatories with the capacity for longitudinal studies of changes in Arctic microbial ecosystems is one potential area of development. The Arctic is home to many scientific research stations ( https://eu-interact.org/ ) and research centres, most of which collect and archive a valuable range of environmental measurements, for example meteorological or atmospheric chemistry data [ 201 ]. These datasets underpin our quantitative view of the Arctic's present and past; thus, delivering the raw ingredients for forecasting changes in the Arctic and the entire planet [ 8 ]. To our knowledge, at present, there is no comparable effort aimed at monitoring Arctic microbiota, in spite of the critical interactions between Arctic microbes and climate change [ 202 ]. Creating deployable genomics resources can distribute the task of characterizing and monitoring Arctic microbiota to a network of such stations. By analogy, the inclusion of distributed, real-time pathogen genomics is considered a key data stream for a prospective global surveillance system, which integrates human, animal and ecosystem health [ 145 ]. Within the Arctic context, this 'sequencing singularity' [ 145 ] would entail the contemporaneous monitoring of Arctic microbiota, along with the physical and chemical environment of the warming Arctic; thus, capturing the complexity of microbe-mediated feedbacks in Arctic change and offering a rich new stream of data to enhance our predictive understanding of Arctic change. Hitherto, the study of Arctic microbial genomics has typically entailed either laboratory-based analyses, field surveys or plot scale experiments [ 67, 71, 98 ]. These are most pertinent for understanding the molecular ecology of the cryosphere at scales of microns to metres, with further insights typically extrapolative. However, it is the emergent macroscale effects of Arctic microbial activity that have relevance for contemporary climate change, requiring microscopic processes to be studied at the scale of entire landscapes. This scale mismatch between processes occurring at the <10 1 µm scale and their effects at the 10 1 –10 5 km 2 scale (a chasm spanning up to 14 orders of magnitude) makes emergent effects challenging to study. Scaling down and scaling up both present challenges ( Fig. 4 ). When downscaling, the analysis of biomass crammed into sample tubes or on to membrane filters blurs the fine-scale resolution of microbial interactions and physico-chemical heterogeneity typical of life in the interstices of soil grains or ice crystals [ 61, 203 ]. When upscaling, researchers can only cover a limited area and process a finite number of samples, obscuring critical spatiotemporal phenomena and relationships to synoptic scale meteorological, geomorphological, glaciological and hydrological processes. Quantitative predictions of community structure, function and stability at larger scales must integrate the sampling of genomic diversity with measurements of environmental processes and remote sensing to capture emergent macroscale effects. Fig. 4. Arctic genomes across scales. Many critical processes [e.g. greenhouse gas (GHG) cycling] mediated by Arctic microbiota occur within environments that are extremely heterogeneous at the microscopic scale (a). In particular, these include interstitial spaces in otherwise frozen substrates (e.g. sea ice, permafrost, glacial ice) where micro-scale gradients in biomass or physical and chemical conditions are apparent. These are disrupted at the sample scale (b) by the requirement to collect sufficient biomass for (meta-) genomic analysis and the distortion incurred by bulk chemical analyses of substrates. At the plot scale (c), undersampling of spatial heterogeneity at the meso-scale poses a further challenge. Finally, upscaling to the landscape or regional scale from plot-scale studies (d) is hampered by spatial and temporal biases in sampling. One example of such an emergent macroscale effect is the biological darkening of glacier surfaces, especially in a 'dark zone' along the western coast of the Greenland Ice Sheet as a result of glacier algae and cryoconite development, as detailed above [ 102 ]. These processes must be understood as part of an integrated glacial system, which in turn requires methods for inferring microbial processes at scale using aerial or orbital remote-sensing platforms. Remote detection and quantification of microbial communities, in particular for algae, is routine for lacustrine and oceanic systems. Earlier remote-sensing efforts for cryospheric algae have focused on snow algae [ 204, 205 ]. However, the complex and highly spatially and temporally variable optics of glacier ice, combined with the mixing of microbial cells, inorganic light absorbing particles and meltwater, make remote detection of glacier microbes more challenging. Nevertheless, there are several potential footholds that may enable remote biomass quantification and perhaps yield insights into microbial processes distributed over space and time. Glacier algae are discernible by their distinctive pigmentation, and it has been suggested that inversion of radiative transfer models could be used to reverse engineer algal pigmentation from spectral reflectance measurements [ 94, 206 ], offering a potential route to unpicking environmental stresses and responses in supraglacial microbial communities. These insights demonstrate that although the existing conceptual models require further empirical support, there are potential emergent phenomena that could be used to infer microbial processes on the ground. As field measurements become more abundant, aerial sensors continue to develop and high performance computing resources become increasingly accessible, our ability to measure, monitor and model the environmentally relevant emergent phenomena related to Arctic microbial processes and their feedbacks to microbial ecology, and to do so at scale, is enhanced. This fusion of genomics and geospatial techniques could enable better-informed climate mitigation strategies, and may well stimulate the next revolution in our understanding of Arctic microbial ecology and its feedbacks to the global climate. Concluding remarks Climate change is unfolding across the ocean and lands of the Arctic at a pace unprecedented in human history, and its consequences will profoundly affect the Earth and our society. We now appreciate microbes play pivotal roles in the response of the Arctic to anthropogenic warming. Arctic microbial genomics has the potential to inform us of this problem, but this information alone is problematic in that it does not present a solution [ 207 ]. Insight into the microbial dimensions of Arctic change can nevertheless support society in its search for solutions to the climate crisis, for example through improving our understanding of carbon sequestration in the Arctic Ocean, aiding models of greenhouse gas release from the permafrost, refining projections of sea-level rise or designing energy efficient catalysis. Within this review, we have identified conceptual and technical barriers, as well as potential routes to surmount these obstacles to progress in Arctic microbial genomics. As a priority ( Fig. 5 ), we must couple the pursuit of improved reference data for microbial genomes from the Arctic with capturing the actualité of physiological responses and population dynamics in Arctic microbial communities through field-based multi-omics and coupled measurements of processes and environmental parameters. To address this priority, we must gain the clearest insights on how microbes respond to environmental changes through transitioning from inventories of changes in phylotype distribution and relative abundance to studies that primarily focus on genes and their products which are encoded within the genomes of Arctic microbes. Moreover, microbial genomics needs to be better integrated within the interdisciplinary framework and infrastructure of Arctic change science. This will permit robust upscaling and numerical modelling of landscape-scale impacts driven by microbial genomes. Importantly, improving reference data (e.g. high-quality annotated genomes, comparative physiology, laboratory mesocosm studies) will provide crucial context for changes occurring within the field. Meanwhile, insights from real-time integrative studies of geospatial, meteorological, biogeochemical and genomic changes will also provide focus for laboratory-based studies. Fig. 5. Arctic microbial genomics will require the fusion of improved reference data and the real-time capture of microbial drivers and responses to changes in the 21st century Arctic. WGA: Whole Genome Amplification. Integrating both strands of Arctic microbial genomics research will require effective interaction with other disciplines and research infrastructures. Indeed, at the heart of our blueprint for Arctic microbial genomics is the blurring of traditional disciplinary boundaries, and the necessity of making the border between laboratory and field studies porous. For effective synthesis, accurate inventories of Arctic habitat types and metadata types must be collected, requiring genomics researchers to engage with geographical knowledge. All of these changes will require improved standards and architecture for sample, isolate, data and metadata collection, accessibility and analysis. By targeting these priorities, we anticipate Arctic microbial genomics would prove agile enough to respond to the contemporary Arctic crisis through providing quantitative predictions of microbial feedbacks in the changing Arctic. However, if nothing else, doing so will secure fundamental knowledge and genomic diversity for study by future generations, long after the global consequences of the Arctic crisis have become unequivocally clear for all humans. Supplementary Data Supplementary material 1 Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10217606/

Encapsulation of Bioactive Peptides by Spray-Drying and Electrospraying

Bioactive peptides derived from enzymatic hydrolysis are gaining attention for the production of supplements, pharmaceutical compounds, and functional foods. However, their inclusion in oral delivery systems is constrained by their high susceptibility to degradation during human gastrointestinal digestion. Encapsulating techniques can be used to stabilize functional ingredients, helping to maintain their activity after processing, storage, and digestion, thus improving their bioaccessibility. Monoaxial spray-drying and electrospraying are common and economical techniques used for the encapsulation of nutrients and bioactive compounds in both the pharmaceutical and food industries. Although less studied, the coaxial configuration of both techniques could potentially improve the stabilization of protein-based bioactives via the formation of shell–core structures. This article reviews the application of these techniques, both monoaxial and coaxial configurations, for the encapsulation of bioactive peptides and protein hydrolysates, focusing on the factors affecting the properties of the encapsulates, such as the formulation of the feed solution, selection of carrier and solvent, as well as the processing conditions used. Furthermore, this review covers the release, retention of bioactivity, and stability of peptide-loaded encapsulates after processing and digestion. 1. Introduction Bioactive peptides have received increasing interest in the last few decades due to the wide range of biological activities they can exert. Multiple studies have demonstrated their antioxidant, antihypertensive, antimicrobial or anti-inflammatory activities, among others, as well as their high potential for the treatment of various diseases, such as diabetes and different types of cancers [ 1 ]. This has boosted the research on the use of biopeptides as therapeutic agents, especially for the treatment of certain chronic conditions, through their incorporation in supplements, pharmaceutical compounds, or functional foods [ 2 ]. In addition to the bioactive properties previously mentioned, biopeptides have other advantages such as their low production cost, low allergenicity, high nutritional value and high digestibility ( Figure 1 ) [ 3 ]. Enzymatic hydrolysis is commonly used to produce bioactive peptides [ 4 ]. This technology releases the peptides encrypted in the original protein by means of breaking the peptide bonds with proteases, improving the technological and bioactive properties of the hydrolysates, enhancing their digestibility, and decreasing their antigenicity by degrading the allergenic epitopes [ 5 ]. Bioactive peptides are usually of 2–20 amino acids in length, and their activity is determined by the specific amino acid sequence and the relative abundance of certain residues (e.g., hydrophilic, hydrophobic, or aromatic) within the peptide [ 6 ]. Clinical application of bioactive peptides is severely limited by the difficulty to find an effective method of administration [ 7 ]. In the pharmacological field, most protein-based treatments are administered via parenteral injection. However, this administration approach presents several drawbacks, mainly (i) difficulty in self-administration, (ii) short half-life of the proteins, and (iii) proteins can be easily degraded in the bloodstream [ 8 ]. Although oral administration of bioactive peptides is a much easier and more practical approach to implement for consumers, there are multiple challenges in its use that must be overcome ( Figure 1 ): High hygroscopicity, which may result in physicochemical instability and loss of bioactivity [ 9 ]. Bitterness due to the exposure to taste receptors of hydrophobic amino acid residues generated from hydrolysis. It has a negative impact on consumer's acceptance [ 10 ]. Low water-solubility, limiting the introduction of hydrolysates or peptides into food matrices, which requires generating dispersed systems [ 11 ]. Physicochemical instability during storage, processing, and digestion, due to the exposure of peptides to environmental conditions (e.g., oxygen, heat) or their interaction with the digestive proteases and other compounds present in the food matrix [ 12 ]. Limited bioaccessibility. Once ingested, peptides must be able to remain intact until they are absorbed in the intestine in order to exert their bioactivity. This is challenging due to the harsh conditions found in the gastrointestinal tract, such as the strongly acidic pH in the stomach and the enzymatically active gastric and intestinal fluids [ 13 ]. To overcome these limitations, great efforts are focused on increasing the physicochemical stability and bioaccessibility of peptides. Indeed, stabilization of bioactive peptides is an essential process to ensure that their in vitro activity properly translates to in vivo animal and human models after processing, storage, and digestion [ 14 ]. Encapsulation of bioactive peptides, consisting in entrapping the peptides within a matrix or layer(s) of encapsulating agent(s), reduces hygroscopicity, masks bitterness and protects the biopeptides from degradation during processing, storage and digestion while maintaining their bioaccessibility [ 15 ]. The pharmaceutical industry has extensively used encapsulation techniques for the stabilization and controlled release of active compounds. Encapsulation processes have attracted much attention from the food industry, especially due to the growing interest during recent decades in the fortification of food with bioactive ingredients, including peptides, to produce functional food products [ 14 ]. Although significant research has been carried out on the encapsulation of lipids (e.g., omega-3 fatty acids), probiotics, vitamins and antioxidants (e.g., polyphenols), there are considerably fewer studies on the encapsulation of bioactive peptides [ 3 , 9 , 14 ]. Several techniques are available for the entrapment of food bioactives (i.e., peptides) within a biopolymer matrix (i.e., spray-drying, spray-cooling, fluid-bed coating, extrusion, electrospraying and complex coacervation followed by drying) [ 16 ]. Nevertheless, spray-drying is the most commonly used technique for the encapsulation of thermosensitive compounds, which permits obtaining dry microparticles at an industrial scale [ 17 ] without peptide degradation (e.g., changes in secondary structure) [ 18 ]. In recent years, electrospraying has emerged as a promising technique to encapsulate bioactive ingredients at room temperature, obtaining nano/microcapsules with narrow size distribution, low cost, and potential to be scaled up [ 19 ]. In any case, it should be noted that the potential degradation of peptides due to electrostatic stresses during electrospraying might need further investigation [ 18 ]. Furthermore, both spray-drying and electrospraying can work in coaxial configuration, resulting in capsules with a double layer of encapsulating wall(s), which might enhance the protection and delivery of the bioactives [ 20 ]. Contrarily to coaxial spray-drying and electrospraying, other encapsulating techniques such as fluidized-bed and spray-chilling processes require an additional production stage to provide a second coating of the encapsulates, which can result in double-layered capsules [ 21 ]. In the light of the above, this work presents a systematic review on the encapsulation of bioactive peptides and protein hydrolysates by spray-drying and electrospraying. This review focuses on the various factors affecting the properties of the encapsulates (e.g., morphology, size, encapsulation efficiency), such as the formulation of the feed solution including the type of carrier and solvent. The specific processing conditions used for both encapsulation techniques (i.e., inlet and outlet air temperature for spray-drying and voltage and flow rate for electrospraying, among others) were reviewed. The use of mono- or coaxial encapsulation methods for both spray-drying and electrospraying are discussed. Finally, this review focuses on release, bioactivity and stability after digestion of the encapsulated peptides. 2. Literature Search A literature search of the papers reporting the encapsulation of peptides, protein hydrolysates and proteins via spray-drying and electrospraying, both monoaxial and coaxial, published in the period between 2012 and November 2022 was carried out using Scopus ( https://www.scopus.com/ (accessed on 25 November 2022)). Research papers containing the keywords "peptide & encapsulation & spray drying," "hydrolysate & encapsulation & spray drying," "protein & encapsulation & spray drying," "peptide & encapsulation & electrospraying," "hydrolysate & encapsulation & electrospraying," "protein & encapsulation & electrospraying," "coaxial & encapsulation & electrospraying" and "coaxial & encapsulation & spray drying" were searched. Consequently, the literature search yielded 444 references that were manually screened. From all the works, 397 were excluded because: (a) the paper was published before 2012 ( n = 87), (b) the reference was not a research paper ( n = 71), (c) the reference was repeated in different searches ( n = 105), (d) peptides, protein hydrolysates or protein were not the active ingredient encapsulated ( n = 71), (e) spray-drying or electrospraying was not used to dry the formulations ( n = 46), and (f) no information about the spray-drying process or formulation was available ( n =  17). Additionally, three works not found in the Scopus literature research were deemed adequate to be added. A total of 50 experimental research papers were considered for the evaluation. 3. Encapsulation of Protein-Based Bioactives by Spray-Drying 3.1. Fundamentals of Spray-Drying Spray-drying is an encapsulation technique based on the atomization of a solution into droplets and their subsequent drying within a gas (e.g., air or nitrogen) at high temperature, producing dry particles [ 22 ]. More specifically, the encapsulation of bioactives by spray-drying consists in the dispersion/mixing of the bioactive together with a carrier (encapsulating agent) in the selected solvent. The solution is pumped and atomized at the entrance of the drying chamber using an atomizer. The atomizer type (i.e., rotary atomizer, pressure nozzle, or sonic nozzle) is selected depending on the characteristics and desired particle size of the final product. The pressure nozzle is the most used at laboratory scale ( Figure 2 A) [ 23 ]. A gas stream at high temperature is concurrently fed to the chamber, providing the driving force to the drying process (e.g., difference in temperature and relative humidity between the droplet and the inlet air). Most of the water is removed during the first drying stage, where the droplet surface is saturated with water. The evaporation of the solvent provides the cooling needed to maintain the surface temperature at a constant value (i.e., wet-bulb temperature). The second drying stage, known as the falling rate period, begins when the surface of the droplet is no longer saturated with water, resulting in the formation of a thin crust. This crust limits water diffusion to the surface, reducing the evaporation rate and causing an increase in the temperature of the dried particle. The dry particles are separated from the outlet drying gas in a cyclone ( Figure 2 A) [ 23 ]. Contrarily to other drying processes (e.g., freeze-drying), spray-drying is normally operated in continuous mode at industrial scale, resulting in high encapsulation efficiency (EE) and minimal degradation of thermolabile bioactive ingredients [ 24 , 25 ]. During spray-drying, multiple parameters must be optimized to achieve high encapsulation efficiency and reduced heat damage. The parameters to be taken into account are based on the formulation of the solution/dispersion/emulsions, i.e., the carrier type, the ratio between the mass of peptide and that of the carrier (core:wall), and the pretreatments needed, such as the formation of nanoliposomes or nanoemulsions [ 26 , 27 ]. Drying kinetics are governed by heat and mass transport. Processing variables of the drying process influencing these transport mechanisms are temperature and humidity of the inlet gas, feed flow rate or flow rate of the drying gas [ 23 , 28 ]. Inlet gas temperature is a key parameter that provides the driving force for the solvent evaporation. The temperature should be set at a level that is sufficiently high to promote water evaporation and ensure complete drying, without leading to agglomeration or deposition of wet particles on the chamber wall. Additionally, an increase in the temperature of the inlet air leads to a decrease in its relative humidity, which promotes water transport. However, excessively high inlet temperatures can lead to prompt crust formation, which limits water diffusion and subsequent evaporation. Therefore, careful control of the temperature of the inlet gas is necessary to ensure effective microencapsulation [ 29 ]. Feed flow rate determines the size of the atomized droplets as well as the amount of water to be evaporated, which influences the temperature of the outlet gas and of the resulting particles. Drying-gas flow rate determines the amount of water evaporated and the residence time of the particles in the drying chamber. Too low a flow rate results in higher water condensation, as well as agglomeration or deposition of particles in the drying chamber. On the other hand, too high a flow rate could lead to degradation of the particles by shear stress. The type of nozzle employed for atomization allows us to process: (1) only one liquid solution in a two-fluid nozzle, known as monoaxial spray-drying, or (2) two concentric liquid solutions in a three-fluid nozzle, known as coaxial spray-drying ( Figure 2 B). The type of process carried out (e.g., mono- or coaxial) affects the distribution of the bioactive compound within the matrix of encapsulating agent(s) ( Figure 2 B). For instance, in monoaxial spray-drying, the bioactive is dispersed within the carrier matrix. On the contrary, the encapsulation by coaxial spray-drying forms two layers of encapsulating agent(s), making it possible to disperse the bioactive within the core matrix of the carrier and forming an additional layer of encapsulating agent ( Figure 2 B). Nevertheless, it should be noted that both monoaxial and coaxial spray drying have some disadvantages, such as the wide size range (5–100 µm) and particle size distribution of the obtained powder. The latter might lead to capsules with different performance in terms of protection and delivery of the bioactive ingredient [ 23 , 30 ]. 3.2. Encapsulation by Monoaxial Spray-Drying 3.2.1. Formulation of the Feed Stream Different methods are currently available to formulate the feed stream containing both the bioactive and the encapsulating agent. Blending is commonly reported in the literature, with 20 of the 29 papers found (69%) using it ( Table 1 ). This method consists in mixing/dispersing the bioactive protein or peptide with the selected carrier, typically protein- or polysaccharide-based, in a selected solvent. The use of proteins as carriers is widely used for the encapsulation of other bioactive compounds, such as fatty acids or vitamins [ 62 , 63 , 64 , 65 ], due to their functional properties, such as emulsifying, water holding, gelling, and film-forming capacities [ 26 ]. However, their use for the encapsulation of protein-based bioactives is still very limited, being reported in only two of the articles found in the literature ( Table 1 ). This is mainly explained by the fact that the use of a carrier with a biochemical composition considerably similar to the bioactive compound to be encapsulated can lead to limitations in stability or expected bioactivity [ 66 ]. Among the few studies using proteins as carrier, Wang et al. [ 45 ] used soy protein isolate and maltodextrin (MD) (1:1) for the encapsulation of antioxidant soy protein hydrolysates at two different core:wall ratios (1.2:1 and 0.8:1). Similarly, Wang et al. used rapeseed protein isolates modified by acylation and high pressure for the encapsulation of rapeseed peptides [ 57 , 58 ] ( Table 1 ). These studies reported particle sizes and encapsulation efficiencies consistent with those achieved using different carriers ( Table 1 ). Polysaccharides are the most used carriers for encapsulation of protein-based bioactives by spray-drying using a blend feed ( Table 1 ). These biopolymers are abundant and inexpensive, as well as chemically stable [ 67 ]. Maltodextrin (MD), a derivative from starch with a dextrose equivalent ranging from 3 to 20, is commonly cited in the literature for the encapsulation of both protein hydrolysates [ 34 , 35 , 36 , 37 , 48 ] and peptides [ 53 , 59 ] ( Table 1 ). MD shows good water solubility, high glass transition temperature and no flavor or odor [ 68 ]. Salleh et al. [ 33 ] studied the use of MD for the encapsulation of edible bird's nest hydrolysates, as well as its combined use with other polysaccharide-based carriers such as carboxymethyl cellulose (CMC) and xanthan gum (XG), demonstrating that the encapsulates produced with MD-XG combination exhibited the best characteristics, with lower water activity, high solubility, and highest retention of the bioactivity. Another polysaccharide-based carrier found in the literature is chitosan (CS) ( Table 1 ). CS is characterized by being nontoxic, biocompatible, and biodegradable [ 67 ], which makes it very interesting as biomaterials for protein encapsulation. In addition, its mucoadhesive and intestinal epithelium-penetrating properties make it an ideal carrier for the oral delivery of proteins [ 67 ]. Despite these advantages, its use is notably limited. Aquino et al. [ 46 ] combined CS and mannitol for the oral delivery of spirulina bioactive peptide extracts. Gómez-Mascaraque et al. [ 49 ] compared the effect of encapsulating whey protein hydrolysate with both gelatin and CS, determining that CS was more effective at stabilizing the peptide while not affecting the profile of the peptides after digestion. An alternative approach for the generation of the feed solution involves the development of nanoliposomes. Nanoliposomes are lipid-based systems composed of a single or multiple concentric bilayers made of phospholipids constituting a vesicle, which allows the storage of the bioactive peptides both in the aqueous core or in the interior of the bilayer [ 69 ]. Sarabandi et al. [ 38 ] studied the encapsulation of flaxseed protein hydrolysates dissolved in a phosphate-buffered solution (PBS) via nanoliposome formation using cholesterol, as well as lecithin and Tween 80 as surfactants. A MD solution was added to the nanoliposomes in a 1:1 ( v / v ) ratio before spray-drying. The effect of CS coating of the flaxseed protein hydrolysate–MD nanoliposomes was further studied, showing improvement in the physical properties of the particles (e.g., smaller particles after reconstitution, improved solubility, higher encapsulation efficiency) [ 39 ]. Mohammadi et al. [ 47 ] studied the effect of adding a CS coating to encapsulated spirulina platensis nanoliposomes, finding that it effectively improved the physical stability of the vesicles during storage by reducing particle aggregation. Similarly to nanoliposomes, nanoemulsions also use lipidic carriers. Oil-in-water nanoemulsions are colloidal systems with a hydrophobic liquid core composed of the oily/organic phase dispersed in the aqueous phase [ 70 ]. De Figueiredo Furtado et al. [ 51 ] produced single emulsions dispersing whey protein isolate and a mix of MD and lactose in an aqueous phase, while the oily phase was composed of a blend of high oleic sunflower oil, coconut oil and caprylic/capric triglyceride. They concluded that using oil blends with medium-chain triacylglycerols favors the formation of smaller spray-dried particles. Thus, the use of chitosan coating and medium-chain triglycerides can improve the physical properties of the particles. Zhu et al. [ 54 ] obtained emulsified droplets loaded with exenatide, which is a synthetic incretin used in the treatment of diabetes mellitus type 2. In this case, the protein-based bioactive was dispersed in the aqueous phase and poly(lactic-co-glycolic acid) (PLGA) in the organic phase composed of dichloromethane (DCM) and dimethyl carbonate (DMC) (1:1), resulting in particles with higher release and degradation compared to an alternative encapsulation method, such as ultrafine particle processing, which is based on disk rotation principles. Double-emulsion systems have also been developed for the encapsulation of protein-based bioactives by monoaxial spray-drying. Ying et al. [ 59 ] produced a double emulsion (W1/O/W2) of soy peptides using different emulsifiers (i.e., polyglycerol polyricinoleate, lecithin and Span 60). To this end, a primary (W1/O) emulsion was obtained combining the soy peptides and different emulsifiers in the aqueous phase, whereas medium-chain triglycerides composed the oily phase. This primary emulsion was mixed with the W2 phase, which contained OSA modified starch and MD. The final double emulsion was spray-dried to produce peptide-loaded microcapsules; however, the authors reported that during drying, the double emulsion was ruptured, resulting in low encapsulation efficiency. Calderón-Oliver et al. [ 56 ] compared the use of blend and emulsion methods for the encapsulation of nisin, an antimicrobial peptide, using either pectin or alginate as carrier. They produced a double W1/O/W2 emulsion by combining the aqueous phase (nisin and avocado peel extract as an antioxidant) and the oily phase (soybean oil with monoglycerides as emulsifier) to produce the primary emulsion (W1/O). The final W1/O/W2 was obtained by mixing the primary emulsion with a collagen solution and the carrier (either pectin or alginate) solution. The emulsion method resulted in improved encapsulation efficiency. In summary, when it comes to selecting the optimal formulation method for the encapsulation of bioactive peptides by spray-drying, it is important to consider the advantages and disadvantages of each approach. While feeds obtained by blending the peptides and carriers are simply prepared, leading to high encapsulation efficiency ( Table 1 ), preparation of emulsion-based feeds can lead to degradation of the peptides due to their exposure to oil–water interfaces and the mechanical stress to which they are subjected during the emulsification process [ 46 ]. Moreover, the production of nanoliposomes has the disadvantage that the particles can suffer lipid oxidation during production and storage, limiting their shelf life [ 14 ]. Therefore, careful consideration should be given to the specific properties of the bioactive peptides and the intended application of the encapsulated product when selecting a formulation method. In addition to the method of preparation of the feed solution and the type of carrier used, there are other characteristics of the solution that affect the encapsulation process, as well as the morphology and release of the capsules produced. The peptide:carrier or core:wall ratios greatly affect the encapsulation efficiency, which decreases when increasing the load of peptide [ 14 ]. Akbarbaglu et al. [ 37 ] studied the effect that core:wall ratios of 1:1, 1:2 and 1:3 ( w / w ) had on the encapsulation of flaxseed protein hydrolysates using MD as carrier. They found that increasing the concentration of carrier resulted in a higher production yield, while the moisture content and water activity decreased. Likewise, Palamutoğlu and Sariçoban [ 34 ] encapsulated fish collagen hydrolysates with MD at 1:4 and 1:9 ( w / w ) core:wall ratios and found that moisture content and water activity decreased with increasing concentration of MD. Similar results were also obtained by encapsulation of rapeseed peptides using rapeseed protein isolates as wall materials at different core:wall ratios (1:1, 1:2 and 2:1 w / w ), and the production yield increased when more carrier was used [ 58 ]. Taken together, the core:wall ratio is an important factor that affects the encapsulation efficiency of bioactive peptides. Increasing the concentration of carrier can improve the production yield while reducing the moisture content and water activity of the encapsulates, leading to better entrapment of the bioactive peptides. 3.2.2. Processing Conditions Apart from the formulation of the feed stream, the processing conditions also influence the properties of the encapsulates obtained. The inlet temperature of the drying gas is a key variable determining the drying kinetics. The works reported in the literature on the encapsulation of protein-based bioactives use temperature values in the range of 60–200 °C ( Table 1 ). The lowest inlet temperature was employed by Zhu et al. [ 54 ], who carried out the encapsulation of exenatide by spray-drying using inlet air at 60 °C. This low temperature, out of the norm for spray-drying, was due to the use of organic solvents (DMC and DCM) instead of water, as these evaporate at lower temperatures. The highest inlet temperature was 200 °C, which was used for the encapsulation of whey protein hydrolysate [ 50 ] and glutathione [ 55 ] employing β-cyclodextrin (β-CD) or CS as encapsulating agents ( Table 1 ). Yang et al. [ 50 ] reported outlet temperatures of 90 ± 5 °C, while Webber et al. [ 55 ] reported 72 °C. This difference could be due to the joint effect of other parameters such as feed flow rate, drying air flow rate and solid concentration of the feed solution. Cao et al. [ 44 ] studied the effect of inlet temperature on the encapsulation of watermelon seed hydrolyzed protein with MD. They compared inlet temperatures of 150, 160, 170, and 180 °C, finding that increasing temperature led to lower moisture content and solubility of the capsules. Likewise, particle size slightly increased with increasing temperature, varying from an average particle diameter of 10 µm at 150–160 °C to 12.9 µm at 170 °C. These results agree with a previous study [ 71 ] on the encapsulation of whey protein concentrates, reporting larger particles with increasing inlet temperature. This was attributed to elevated inlet temperatures increasing the moisture removal rate, forming a crust more quickly and limiting the shrinking time of the particle. Thus, the inlet temperature must be selected based on the nature of the solvent and the bioactive to be encapsulated, as well as other parameters like feed flow rate, drying air flow rate, and solid concentration of the feed solution. While high inlet temperatures increase the moisture removal rate, which can lead rapid crust formation and then larger particles, it also increases the risk of thermal degradation of bioactives. Therefore, it seems reasonable to use inlet temperatures between 130 and 190 °C and to avoid outlet temperatures above 100 °C to obtain encapsulates with minimal thermal degradation. Feed flow rates reported in the literature range from 160 to 1380 mL/h ( Table 1 ). This variable affects the size of the atomized droplets and consequently the size of the dried particles [ 72 ]. The lowest feed flow rate (160 mL/h) value was applied to encapsulate buffalo whey protein hydrolysates, with gum arabic (GA) and MD as encapsulating agents, obtaining particles of average diameter between 2 and 20 μm when using a two-fluid nozzle with 700 μm inner diameter [ 31 ]. The highest feed flow rate was 1380 mL/h and was used for the encapsulation of P. lunatus hydrolysates with GA and MD, forming particles with diameter of 3.3–6.8 μm when using a two-fluid nozzle for atomization with diameter of 500 μm [ 42 ]. Although it is accepted that a higher feed flow rate results in larger particles due to collision and subsequent fusion of small drops [ 73 ], the combined effect of the other processing variables could justify that the particles obtained in the latter study being smaller. Although high feed flow rates are desired for higher production rates, they can also result in condensation on the walls or wet particles in the chamber or cyclone due to higher humidity increasing stickiness and agglomeration, leading to a decrease in particle recovery attributed to wall deposition [ 60 ]. Thus, ratios of 300–500 mL/h are the most preferred. Curiously, there was a wide difference between the pneumatic air pressures used (40 and 1 bar), which has been proven to affect the particle size during the atomization process [ 73 ]. Regarding the correlation between the nozzle diameter and the morphology of the particles obtained, Keogh et al. [ 74 ] studied the effect of the nozzle diameter in the processing of milk powders and observed that the particle diameter of the spray-dried powders increased when increasing the nozzle diameter. However, the literature in this regard is very limited. Only two nozzle sizes were used in the literature found (500 and 700 µm), making it not possible to establish a correlation between these values and the particle size results. It is noteworthy that both the solubility and shelf stability of the powdered product is affected by the physicochemical properties of the encapsulates, such as morphology, particle size distribution, moisture content, or encapsulation efficiency. Smallness (under 5 µm) and a tight distribution of particles is desired to improve shelf stability [ 28 ]. Regarding the morphology of the particles, it was found that the size of the encapsulates produced by monoaxial spray-drying ranged from 0.132 μm [ 39 ] to 183 μm [ 35 ], with 21 of 24 articles reporting diameter values from 0.25 to 50 μm. These results are consistent with previous studies reporting typical dry particle diameter in laboratory scale spray-dryers of 0.5 to 50 μm [ 75 ]. It is also noteworthy that four of the five studies reporting average size below 1 μm were obtained via formation of nanoliposomes and similar conditions were used for three of the works found, with inlet air temperature of 130–170 °C and feed flow rate of 300 mL/h [ 38 , 39 , 41 ]. Encapsulation efficiency (EE) is one of the main parameters determining shelf-life stability. EE can be defined as the percentage of bioactive compound, whether peptide, hydrolysate, or protein, that is trapped inside the carrier matrix with respect to the amount initially added. Lower EE values lead to more exposed bioactive on the surface of the capsules, which is more easily degraded, resulting in reduced bioactivity [ 14 ]. All studies reported in the literature on the encapsulation of protein-based bioactives by monoaxial spray-drying obtained EE values over 50%, except for Ying et al. [ 59 ], who reported EE values in the range 29.51–45.83% for the encapsulation of soy peptides via W1/O/W2 because the emulsion was not physically stable ( Table 1 ). The highest EE (~100%) was reached by Koker et al. [ 60 ], where ovalbumin was encapsulated using dextran sulfate and poly-l-arginine as carrier for the elaboration of vaccine antigens. EE of peptide-loaded encapsulates can be improved by the addition of surfactants. For instance, Tween 80 was used to reduce flaxseed peptide migration to the air–water interface in spray-drying when using MD as carrier [ 36 ]. The authors reported reduced degradation by shear stress and dehydration, leading as well to higher bioactivity (i.e., antioxidant activity), indicating that the addition of surfactants should be considered in the future to improve encapsulation of peptides by spray-drying. Altogether, processing conditions play a significant role in determining the properties of the encapsulates obtained. The inlet temperature of the drying gas is a crucial variable that affects drying kinetics, particle size, and morphology. Inlet temperatures between 130 and 190 °C are found to obtain desirable encapsulates with minimal thermal degradation. Feed flow rate also affects the size of the atomized droplets, and consequently the size of the dried particles. It was observed that particle size under 5 µm and a narrow distribution of particles improve shelf stability of the encapsulates. Finally, feed flow rate ratios of 300–500 mL/h are preferred to avoid the formation of wet particles or wall deposition in lab spray-dryers. 3.3. Encapsulation by Coaxial Spray-Drying Lately, several works have been reported on the coaxial encapsulation of food bioactives by spray-drying using a three-fluid nozzle. This configuration ( Figure 2 B) allows feeding two different solutions through two concentric channels in the nozzle while the pneumatic air responsible for the atomization flows in the external channel [ 76 ]. In this way, the bioactive compound dissolved in a solution of carrier (core solution) can be pumped through the inner channel, and a solution containing the same or a different carrier (shell solution) is pumped through the outer channel. Both the shell and core solutions meet concentrically at the tip of the nozzle being atomized and dried in the chamber to form microcapsules with a core–shell structure. Therefore, this method might offer greater protection and better control for the release of the bioactive compound when compared to the monoaxial process (two-fluid nozzle) [ 77 ]. Despite the advantages of this method and its widespread use in the encapsulation of bioactive compounds such as carotenoids, triglycerides or organosulfur compounds [ 78 , 79 , 80 ], the literature on the encapsulation of bioactive protein hydrolysates and peptides by coaxial spray-drying is very limited. To the authors' knowledge, only one paper has been published regarding the production of protein-loaded microparticles via spray-drying using a three-fluid nozzle [ 20 ]. In this work, the authors encapsulated lysozymes employing an aqueous solution of trehalose as core solution and a shell solution consisting in PLGA dissolved in a mixture of solvents (acetonitrile, DCM, and acetone). The effect of the core:shell flow rate ratio (4:1 or 10:1) was studied. The operating conditions selected were (i) inlet temperature = 60 °C, (ii) outlet temperature = 40–45 °C, and (iii) drying air flow rate = 37500 L/h. The study concluded that solvent selection did not affect particle size, while it did affect EE. On the other hand, the feed flow rate ratio did affect the particle size, which ranged from 1.07 to 1.60 μm. Hence, the coaxial encapsulation by spray-drying of bioactive protein hydrolysates and peptides deserves further research. 3.1. Fundamentals of Spray-Drying Spray-drying is an encapsulation technique based on the atomization of a solution into droplets and their subsequent drying within a gas (e.g., air or nitrogen) at high temperature, producing dry particles [ 22 ]. More specifically, the encapsulation of bioactives by spray-drying consists in the dispersion/mixing of the bioactive together with a carrier (encapsulating agent) in the selected solvent. The solution is pumped and atomized at the entrance of the drying chamber using an atomizer. The atomizer type (i.e., rotary atomizer, pressure nozzle, or sonic nozzle) is selected depending on the characteristics and desired particle size of the final product. The pressure nozzle is the most used at laboratory scale ( Figure 2 A) [ 23 ]. A gas stream at high temperature is concurrently fed to the chamber, providing the driving force to the drying process (e.g., difference in temperature and relative humidity between the droplet and the inlet air). Most of the water is removed during the first drying stage, where the droplet surface is saturated with water. The evaporation of the solvent provides the cooling needed to maintain the surface temperature at a constant value (i.e., wet-bulb temperature). The second drying stage, known as the falling rate period, begins when the surface of the droplet is no longer saturated with water, resulting in the formation of a thin crust. This crust limits water diffusion to the surface, reducing the evaporation rate and causing an increase in the temperature of the dried particle. The dry particles are separated from the outlet drying gas in a cyclone ( Figure 2 A) [ 23 ]. Contrarily to other drying processes (e.g., freeze-drying), spray-drying is normally operated in continuous mode at industrial scale, resulting in high encapsulation efficiency (EE) and minimal degradation of thermolabile bioactive ingredients [ 24 , 25 ]. During spray-drying, multiple parameters must be optimized to achieve high encapsulation efficiency and reduced heat damage. The parameters to be taken into account are based on the formulation of the solution/dispersion/emulsions, i.e., the carrier type, the ratio between the mass of peptide and that of the carrier (core:wall), and the pretreatments needed, such as the formation of nanoliposomes or nanoemulsions [ 26 , 27 ]. Drying kinetics are governed by heat and mass transport. Processing variables of the drying process influencing these transport mechanisms are temperature and humidity of the inlet gas, feed flow rate or flow rate of the drying gas [ 23 , 28 ]. Inlet gas temperature is a key parameter that provides the driving force for the solvent evaporation. The temperature should be set at a level that is sufficiently high to promote water evaporation and ensure complete drying, without leading to agglomeration or deposition of wet particles on the chamber wall. Additionally, an increase in the temperature of the inlet air leads to a decrease in its relative humidity, which promotes water transport. However, excessively high inlet temperatures can lead to prompt crust formation, which limits water diffusion and subsequent evaporation. Therefore, careful control of the temperature of the inlet gas is necessary to ensure effective microencapsulation [ 29 ]. Feed flow rate determines the size of the atomized droplets as well as the amount of water to be evaporated, which influences the temperature of the outlet gas and of the resulting particles. Drying-gas flow rate determines the amount of water evaporated and the residence time of the particles in the drying chamber. Too low a flow rate results in higher water condensation, as well as agglomeration or deposition of particles in the drying chamber. On the other hand, too high a flow rate could lead to degradation of the particles by shear stress. The type of nozzle employed for atomization allows us to process: (1) only one liquid solution in a two-fluid nozzle, known as monoaxial spray-drying, or (2) two concentric liquid solutions in a three-fluid nozzle, known as coaxial spray-drying ( Figure 2 B). The type of process carried out (e.g., mono- or coaxial) affects the distribution of the bioactive compound within the matrix of encapsulating agent(s) ( Figure 2 B). For instance, in monoaxial spray-drying, the bioactive is dispersed within the carrier matrix. On the contrary, the encapsulation by coaxial spray-drying forms two layers of encapsulating agent(s), making it possible to disperse the bioactive within the core matrix of the carrier and forming an additional layer of encapsulating agent ( Figure 2 B). Nevertheless, it should be noted that both monoaxial and coaxial spray drying have some disadvantages, such as the wide size range (5–100 µm) and particle size distribution of the obtained powder. The latter might lead to capsules with different performance in terms of protection and delivery of the bioactive ingredient [ 23 , 30 ]. 3.2. Encapsulation by Monoaxial Spray-Drying 3.2.1. Formulation of the Feed Stream Different methods are currently available to formulate the feed stream containing both the bioactive and the encapsulating agent. Blending is commonly reported in the literature, with 20 of the 29 papers found (69%) using it ( Table 1 ). This method consists in mixing/dispersing the bioactive protein or peptide with the selected carrier, typically protein- or polysaccharide-based, in a selected solvent. The use of proteins as carriers is widely used for the encapsulation of other bioactive compounds, such as fatty acids or vitamins [ 62 , 63 , 64 , 65 ], due to their functional properties, such as emulsifying, water holding, gelling, and film-forming capacities [ 26 ]. However, their use for the encapsulation of protein-based bioactives is still very limited, being reported in only two of the articles found in the literature ( Table 1 ). This is mainly explained by the fact that the use of a carrier with a biochemical composition considerably similar to the bioactive compound to be encapsulated can lead to limitations in stability or expected bioactivity [ 66 ]. Among the few studies using proteins as carrier, Wang et al. [ 45 ] used soy protein isolate and maltodextrin (MD) (1:1) for the encapsulation of antioxidant soy protein hydrolysates at two different core:wall ratios (1.2:1 and 0.8:1). Similarly, Wang et al. used rapeseed protein isolates modified by acylation and high pressure for the encapsulation of rapeseed peptides [ 57 , 58 ] ( Table 1 ). These studies reported particle sizes and encapsulation efficiencies consistent with those achieved using different carriers ( Table 1 ). Polysaccharides are the most used carriers for encapsulation of protein-based bioactives by spray-drying using a blend feed ( Table 1 ). These biopolymers are abundant and inexpensive, as well as chemically stable [ 67 ]. Maltodextrin (MD), a derivative from starch with a dextrose equivalent ranging from 3 to 20, is commonly cited in the literature for the encapsulation of both protein hydrolysates [ 34 , 35 , 36 , 37 , 48 ] and peptides [ 53 , 59 ] ( Table 1 ). MD shows good water solubility, high glass transition temperature and no flavor or odor [ 68 ]. Salleh et al. [ 33 ] studied the use of MD for the encapsulation of edible bird's nest hydrolysates, as well as its combined use with other polysaccharide-based carriers such as carboxymethyl cellulose (CMC) and xanthan gum (XG), demonstrating that the encapsulates produced with MD-XG combination exhibited the best characteristics, with lower water activity, high solubility, and highest retention of the bioactivity. Another polysaccharide-based carrier found in the literature is chitosan (CS) ( Table 1 ). CS is characterized by being nontoxic, biocompatible, and biodegradable [ 67 ], which makes it very interesting as biomaterials for protein encapsulation. In addition, its mucoadhesive and intestinal epithelium-penetrating properties make it an ideal carrier for the oral delivery of proteins [ 67 ]. Despite these advantages, its use is notably limited. Aquino et al. [ 46 ] combined CS and mannitol for the oral delivery of spirulina bioactive peptide extracts. Gómez-Mascaraque et al. [ 49 ] compared the effect of encapsulating whey protein hydrolysate with both gelatin and CS, determining that CS was more effective at stabilizing the peptide while not affecting the profile of the peptides after digestion. An alternative approach for the generation of the feed solution involves the development of nanoliposomes. Nanoliposomes are lipid-based systems composed of a single or multiple concentric bilayers made of phospholipids constituting a vesicle, which allows the storage of the bioactive peptides both in the aqueous core or in the interior of the bilayer [ 69 ]. Sarabandi et al. [ 38 ] studied the encapsulation of flaxseed protein hydrolysates dissolved in a phosphate-buffered solution (PBS) via nanoliposome formation using cholesterol, as well as lecithin and Tween 80 as surfactants. A MD solution was added to the nanoliposomes in a 1:1 ( v / v ) ratio before spray-drying. The effect of CS coating of the flaxseed protein hydrolysate–MD nanoliposomes was further studied, showing improvement in the physical properties of the particles (e.g., smaller particles after reconstitution, improved solubility, higher encapsulation efficiency) [ 39 ]. Mohammadi et al. [ 47 ] studied the effect of adding a CS coating to encapsulated spirulina platensis nanoliposomes, finding that it effectively improved the physical stability of the vesicles during storage by reducing particle aggregation. Similarly to nanoliposomes, nanoemulsions also use lipidic carriers. Oil-in-water nanoemulsions are colloidal systems with a hydrophobic liquid core composed of the oily/organic phase dispersed in the aqueous phase [ 70 ]. De Figueiredo Furtado et al. [ 51 ] produced single emulsions dispersing whey protein isolate and a mix of MD and lactose in an aqueous phase, while the oily phase was composed of a blend of high oleic sunflower oil, coconut oil and caprylic/capric triglyceride. They concluded that using oil blends with medium-chain triacylglycerols favors the formation of smaller spray-dried particles. Thus, the use of chitosan coating and medium-chain triglycerides can improve the physical properties of the particles. Zhu et al. [ 54 ] obtained emulsified droplets loaded with exenatide, which is a synthetic incretin used in the treatment of diabetes mellitus type 2. In this case, the protein-based bioactive was dispersed in the aqueous phase and poly(lactic-co-glycolic acid) (PLGA) in the organic phase composed of dichloromethane (DCM) and dimethyl carbonate (DMC) (1:1), resulting in particles with higher release and degradation compared to an alternative encapsulation method, such as ultrafine particle processing, which is based on disk rotation principles. Double-emulsion systems have also been developed for the encapsulation of protein-based bioactives by monoaxial spray-drying. Ying et al. [ 59 ] produced a double emulsion (W1/O/W2) of soy peptides using different emulsifiers (i.e., polyglycerol polyricinoleate, lecithin and Span 60). To this end, a primary (W1/O) emulsion was obtained combining the soy peptides and different emulsifiers in the aqueous phase, whereas medium-chain triglycerides composed the oily phase. This primary emulsion was mixed with the W2 phase, which contained OSA modified starch and MD. The final double emulsion was spray-dried to produce peptide-loaded microcapsules; however, the authors reported that during drying, the double emulsion was ruptured, resulting in low encapsulation efficiency. Calderón-Oliver et al. [ 56 ] compared the use of blend and emulsion methods for the encapsulation of nisin, an antimicrobial peptide, using either pectin or alginate as carrier. They produced a double W1/O/W2 emulsion by combining the aqueous phase (nisin and avocado peel extract as an antioxidant) and the oily phase (soybean oil with monoglycerides as emulsifier) to produce the primary emulsion (W1/O). The final W1/O/W2 was obtained by mixing the primary emulsion with a collagen solution and the carrier (either pectin or alginate) solution. The emulsion method resulted in improved encapsulation efficiency. In summary, when it comes to selecting the optimal formulation method for the encapsulation of bioactive peptides by spray-drying, it is important to consider the advantages and disadvantages of each approach. While feeds obtained by blending the peptides and carriers are simply prepared, leading to high encapsulation efficiency ( Table 1 ), preparation of emulsion-based feeds can lead to degradation of the peptides due to their exposure to oil–water interfaces and the mechanical stress to which they are subjected during the emulsification process [ 46 ]. Moreover, the production of nanoliposomes has the disadvantage that the particles can suffer lipid oxidation during production and storage, limiting their shelf life [ 14 ]. Therefore, careful consideration should be given to the specific properties of the bioactive peptides and the intended application of the encapsulated product when selecting a formulation method. In addition to the method of preparation of the feed solution and the type of carrier used, there are other characteristics of the solution that affect the encapsulation process, as well as the morphology and release of the capsules produced. The peptide:carrier or core:wall ratios greatly affect the encapsulation efficiency, which decreases when increasing the load of peptide [ 14 ]. Akbarbaglu et al. [ 37 ] studied the effect that core:wall ratios of 1:1, 1:2 and 1:3 ( w / w ) had on the encapsulation of flaxseed protein hydrolysates using MD as carrier. They found that increasing the concentration of carrier resulted in a higher production yield, while the moisture content and water activity decreased. Likewise, Palamutoğlu and Sariçoban [ 34 ] encapsulated fish collagen hydrolysates with MD at 1:4 and 1:9 ( w / w ) core:wall ratios and found that moisture content and water activity decreased with increasing concentration of MD. Similar results were also obtained by encapsulation of rapeseed peptides using rapeseed protein isolates as wall materials at different core:wall ratios (1:1, 1:2 and 2:1 w / w ), and the production yield increased when more carrier was used [ 58 ]. Taken together, the core:wall ratio is an important factor that affects the encapsulation efficiency of bioactive peptides. Increasing the concentration of carrier can improve the production yield while reducing the moisture content and water activity of the encapsulates, leading to better entrapment of the bioactive peptides. 3.2.2. Processing Conditions Apart from the formulation of the feed stream, the processing conditions also influence the properties of the encapsulates obtained. The inlet temperature of the drying gas is a key variable determining the drying kinetics. The works reported in the literature on the encapsulation of protein-based bioactives use temperature values in the range of 60–200 °C ( Table 1 ). The lowest inlet temperature was employed by Zhu et al. [ 54 ], who carried out the encapsulation of exenatide by spray-drying using inlet air at 60 °C. This low temperature, out of the norm for spray-drying, was due to the use of organic solvents (DMC and DCM) instead of water, as these evaporate at lower temperatures. The highest inlet temperature was 200 °C, which was used for the encapsulation of whey protein hydrolysate [ 50 ] and glutathione [ 55 ] employing β-cyclodextrin (β-CD) or CS as encapsulating agents ( Table 1 ). Yang et al. [ 50 ] reported outlet temperatures of 90 ± 5 °C, while Webber et al. [ 55 ] reported 72 °C. This difference could be due to the joint effect of other parameters such as feed flow rate, drying air flow rate and solid concentration of the feed solution. Cao et al. [ 44 ] studied the effect of inlet temperature on the encapsulation of watermelon seed hydrolyzed protein with MD. They compared inlet temperatures of 150, 160, 170, and 180 °C, finding that increasing temperature led to lower moisture content and solubility of the capsules. Likewise, particle size slightly increased with increasing temperature, varying from an average particle diameter of 10 µm at 150–160 °C to 12.9 µm at 170 °C. These results agree with a previous study [ 71 ] on the encapsulation of whey protein concentrates, reporting larger particles with increasing inlet temperature. This was attributed to elevated inlet temperatures increasing the moisture removal rate, forming a crust more quickly and limiting the shrinking time of the particle. Thus, the inlet temperature must be selected based on the nature of the solvent and the bioactive to be encapsulated, as well as other parameters like feed flow rate, drying air flow rate, and solid concentration of the feed solution. While high inlet temperatures increase the moisture removal rate, which can lead rapid crust formation and then larger particles, it also increases the risk of thermal degradation of bioactives. Therefore, it seems reasonable to use inlet temperatures between 130 and 190 °C and to avoid outlet temperatures above 100 °C to obtain encapsulates with minimal thermal degradation. Feed flow rates reported in the literature range from 160 to 1380 mL/h ( Table 1 ). This variable affects the size of the atomized droplets and consequently the size of the dried particles [ 72 ]. The lowest feed flow rate (160 mL/h) value was applied to encapsulate buffalo whey protein hydrolysates, with gum arabic (GA) and MD as encapsulating agents, obtaining particles of average diameter between 2 and 20 μm when using a two-fluid nozzle with 700 μm inner diameter [ 31 ]. The highest feed flow rate was 1380 mL/h and was used for the encapsulation of P. lunatus hydrolysates with GA and MD, forming particles with diameter of 3.3–6.8 μm when using a two-fluid nozzle for atomization with diameter of 500 μm [ 42 ]. Although it is accepted that a higher feed flow rate results in larger particles due to collision and subsequent fusion of small drops [ 73 ], the combined effect of the other processing variables could justify that the particles obtained in the latter study being smaller. Although high feed flow rates are desired for higher production rates, they can also result in condensation on the walls or wet particles in the chamber or cyclone due to higher humidity increasing stickiness and agglomeration, leading to a decrease in particle recovery attributed to wall deposition [ 60 ]. Thus, ratios of 300–500 mL/h are the most preferred. Curiously, there was a wide difference between the pneumatic air pressures used (40 and 1 bar), which has been proven to affect the particle size during the atomization process [ 73 ]. Regarding the correlation between the nozzle diameter and the morphology of the particles obtained, Keogh et al. [ 74 ] studied the effect of the nozzle diameter in the processing of milk powders and observed that the particle diameter of the spray-dried powders increased when increasing the nozzle diameter. However, the literature in this regard is very limited. Only two nozzle sizes were used in the literature found (500 and 700 µm), making it not possible to establish a correlation between these values and the particle size results. It is noteworthy that both the solubility and shelf stability of the powdered product is affected by the physicochemical properties of the encapsulates, such as morphology, particle size distribution, moisture content, or encapsulation efficiency. Smallness (under 5 µm) and a tight distribution of particles is desired to improve shelf stability [ 28 ]. Regarding the morphology of the particles, it was found that the size of the encapsulates produced by monoaxial spray-drying ranged from 0.132 μm [ 39 ] to 183 μm [ 35 ], with 21 of 24 articles reporting diameter values from 0.25 to 50 μm. These results are consistent with previous studies reporting typical dry particle diameter in laboratory scale spray-dryers of 0.5 to 50 μm [ 75 ]. It is also noteworthy that four of the five studies reporting average size below 1 μm were obtained via formation of nanoliposomes and similar conditions were used for three of the works found, with inlet air temperature of 130–170 °C and feed flow rate of 300 mL/h [ 38 , 39 , 41 ]. Encapsulation efficiency (EE) is one of the main parameters determining shelf-life stability. EE can be defined as the percentage of bioactive compound, whether peptide, hydrolysate, or protein, that is trapped inside the carrier matrix with respect to the amount initially added. Lower EE values lead to more exposed bioactive on the surface of the capsules, which is more easily degraded, resulting in reduced bioactivity [ 14 ]. All studies reported in the literature on the encapsulation of protein-based bioactives by monoaxial spray-drying obtained EE values over 50%, except for Ying et al. [ 59 ], who reported EE values in the range 29.51–45.83% for the encapsulation of soy peptides via W1/O/W2 because the emulsion was not physically stable ( Table 1 ). The highest EE (~100%) was reached by Koker et al. [ 60 ], where ovalbumin was encapsulated using dextran sulfate and poly-l-arginine as carrier for the elaboration of vaccine antigens. EE of peptide-loaded encapsulates can be improved by the addition of surfactants. For instance, Tween 80 was used to reduce flaxseed peptide migration to the air–water interface in spray-drying when using MD as carrier [ 36 ]. The authors reported reduced degradation by shear stress and dehydration, leading as well to higher bioactivity (i.e., antioxidant activity), indicating that the addition of surfactants should be considered in the future to improve encapsulation of peptides by spray-drying. Altogether, processing conditions play a significant role in determining the properties of the encapsulates obtained. The inlet temperature of the drying gas is a crucial variable that affects drying kinetics, particle size, and morphology. Inlet temperatures between 130 and 190 °C are found to obtain desirable encapsulates with minimal thermal degradation. Feed flow rate also affects the size of the atomized droplets, and consequently the size of the dried particles. It was observed that particle size under 5 µm and a narrow distribution of particles improve shelf stability of the encapsulates. Finally, feed flow rate ratios of 300–500 mL/h are preferred to avoid the formation of wet particles or wall deposition in lab spray-dryers. 3.2.1. Formulation of the Feed Stream Different methods are currently available to formulate the feed stream containing both the bioactive and the encapsulating agent. Blending is commonly reported in the literature, with 20 of the 29 papers found (69%) using it ( Table 1 ). This method consists in mixing/dispersing the bioactive protein or peptide with the selected carrier, typically protein- or polysaccharide-based, in a selected solvent. The use of proteins as carriers is widely used for the encapsulation of other bioactive compounds, such as fatty acids or vitamins [ 62 , 63 , 64 , 65 ], due to their functional properties, such as emulsifying, water holding, gelling, and film-forming capacities [ 26 ]. However, their use for the encapsulation of protein-based bioactives is still very limited, being reported in only two of the articles found in the literature ( Table 1 ). This is mainly explained by the fact that the use of a carrier with a biochemical composition considerably similar to the bioactive compound to be encapsulated can lead to limitations in stability or expected bioactivity [ 66 ]. Among the few studies using proteins as carrier, Wang et al. [ 45 ] used soy protein isolate and maltodextrin (MD) (1:1) for the encapsulation of antioxidant soy protein hydrolysates at two different core:wall ratios (1.2:1 and 0.8:1). Similarly, Wang et al. used rapeseed protein isolates modified by acylation and high pressure for the encapsulation of rapeseed peptides [ 57 , 58 ] ( Table 1 ). These studies reported particle sizes and encapsulation efficiencies consistent with those achieved using different carriers ( Table 1 ). Polysaccharides are the most used carriers for encapsulation of protein-based bioactives by spray-drying using a blend feed ( Table 1 ). These biopolymers are abundant and inexpensive, as well as chemically stable [ 67 ]. Maltodextrin (MD), a derivative from starch with a dextrose equivalent ranging from 3 to 20, is commonly cited in the literature for the encapsulation of both protein hydrolysates [ 34 , 35 , 36 , 37 , 48 ] and peptides [ 53 , 59 ] ( Table 1 ). MD shows good water solubility, high glass transition temperature and no flavor or odor [ 68 ]. Salleh et al. [ 33 ] studied the use of MD for the encapsulation of edible bird's nest hydrolysates, as well as its combined use with other polysaccharide-based carriers such as carboxymethyl cellulose (CMC) and xanthan gum (XG), demonstrating that the encapsulates produced with MD-XG combination exhibited the best characteristics, with lower water activity, high solubility, and highest retention of the bioactivity. Another polysaccharide-based carrier found in the literature is chitosan (CS) ( Table 1 ). CS is characterized by being nontoxic, biocompatible, and biodegradable [ 67 ], which makes it very interesting as biomaterials for protein encapsulation. In addition, its mucoadhesive and intestinal epithelium-penetrating properties make it an ideal carrier for the oral delivery of proteins [ 67 ]. Despite these advantages, its use is notably limited. Aquino et al. [ 46 ] combined CS and mannitol for the oral delivery of spirulina bioactive peptide extracts. Gómez-Mascaraque et al. [ 49 ] compared the effect of encapsulating whey protein hydrolysate with both gelatin and CS, determining that CS was more effective at stabilizing the peptide while not affecting the profile of the peptides after digestion. An alternative approach for the generation of the feed solution involves the development of nanoliposomes. Nanoliposomes are lipid-based systems composed of a single or multiple concentric bilayers made of phospholipids constituting a vesicle, which allows the storage of the bioactive peptides both in the aqueous core or in the interior of the bilayer [ 69 ]. Sarabandi et al. [ 38 ] studied the encapsulation of flaxseed protein hydrolysates dissolved in a phosphate-buffered solution (PBS) via nanoliposome formation using cholesterol, as well as lecithin and Tween 80 as surfactants. A MD solution was added to the nanoliposomes in a 1:1 ( v / v ) ratio before spray-drying. The effect of CS coating of the flaxseed protein hydrolysate–MD nanoliposomes was further studied, showing improvement in the physical properties of the particles (e.g., smaller particles after reconstitution, improved solubility, higher encapsulation efficiency) [ 39 ]. Mohammadi et al. [ 47 ] studied the effect of adding a CS coating to encapsulated spirulina platensis nanoliposomes, finding that it effectively improved the physical stability of the vesicles during storage by reducing particle aggregation. Similarly to nanoliposomes, nanoemulsions also use lipidic carriers. Oil-in-water nanoemulsions are colloidal systems with a hydrophobic liquid core composed of the oily/organic phase dispersed in the aqueous phase [ 70 ]. De Figueiredo Furtado et al. [ 51 ] produced single emulsions dispersing whey protein isolate and a mix of MD and lactose in an aqueous phase, while the oily phase was composed of a blend of high oleic sunflower oil, coconut oil and caprylic/capric triglyceride. They concluded that using oil blends with medium-chain triacylglycerols favors the formation of smaller spray-dried particles. Thus, the use of chitosan coating and medium-chain triglycerides can improve the physical properties of the particles. Zhu et al. [ 54 ] obtained emulsified droplets loaded with exenatide, which is a synthetic incretin used in the treatment of diabetes mellitus type 2. In this case, the protein-based bioactive was dispersed in the aqueous phase and poly(lactic-co-glycolic acid) (PLGA) in the organic phase composed of dichloromethane (DCM) and dimethyl carbonate (DMC) (1:1), resulting in particles with higher release and degradation compared to an alternative encapsulation method, such as ultrafine particle processing, which is based on disk rotation principles. Double-emulsion systems have also been developed for the encapsulation of protein-based bioactives by monoaxial spray-drying. Ying et al. [ 59 ] produced a double emulsion (W1/O/W2) of soy peptides using different emulsifiers (i.e., polyglycerol polyricinoleate, lecithin and Span 60). To this end, a primary (W1/O) emulsion was obtained combining the soy peptides and different emulsifiers in the aqueous phase, whereas medium-chain triglycerides composed the oily phase. This primary emulsion was mixed with the W2 phase, which contained OSA modified starch and MD. The final double emulsion was spray-dried to produce peptide-loaded microcapsules; however, the authors reported that during drying, the double emulsion was ruptured, resulting in low encapsulation efficiency. Calderón-Oliver et al. [ 56 ] compared the use of blend and emulsion methods for the encapsulation of nisin, an antimicrobial peptide, using either pectin or alginate as carrier. They produced a double W1/O/W2 emulsion by combining the aqueous phase (nisin and avocado peel extract as an antioxidant) and the oily phase (soybean oil with monoglycerides as emulsifier) to produce the primary emulsion (W1/O). The final W1/O/W2 was obtained by mixing the primary emulsion with a collagen solution and the carrier (either pectin or alginate) solution. The emulsion method resulted in improved encapsulation efficiency. In summary, when it comes to selecting the optimal formulation method for the encapsulation of bioactive peptides by spray-drying, it is important to consider the advantages and disadvantages of each approach. While feeds obtained by blending the peptides and carriers are simply prepared, leading to high encapsulation efficiency ( Table 1 ), preparation of emulsion-based feeds can lead to degradation of the peptides due to their exposure to oil–water interfaces and the mechanical stress to which they are subjected during the emulsification process [ 46 ]. Moreover, the production of nanoliposomes has the disadvantage that the particles can suffer lipid oxidation during production and storage, limiting their shelf life [ 14 ]. Therefore, careful consideration should be given to the specific properties of the bioactive peptides and the intended application of the encapsulated product when selecting a formulation method. In addition to the method of preparation of the feed solution and the type of carrier used, there are other characteristics of the solution that affect the encapsulation process, as well as the morphology and release of the capsules produced. The peptide:carrier or core:wall ratios greatly affect the encapsulation efficiency, which decreases when increasing the load of peptide [ 14 ]. Akbarbaglu et al. [ 37 ] studied the effect that core:wall ratios of 1:1, 1:2 and 1:3 ( w / w ) had on the encapsulation of flaxseed protein hydrolysates using MD as carrier. They found that increasing the concentration of carrier resulted in a higher production yield, while the moisture content and water activity decreased. Likewise, Palamutoğlu and Sariçoban [ 34 ] encapsulated fish collagen hydrolysates with MD at 1:4 and 1:9 ( w / w ) core:wall ratios and found that moisture content and water activity decreased with increasing concentration of MD. Similar results were also obtained by encapsulation of rapeseed peptides using rapeseed protein isolates as wall materials at different core:wall ratios (1:1, 1:2 and 2:1 w / w ), and the production yield increased when more carrier was used [ 58 ]. Taken together, the core:wall ratio is an important factor that affects the encapsulation efficiency of bioactive peptides. Increasing the concentration of carrier can improve the production yield while reducing the moisture content and water activity of the encapsulates, leading to better entrapment of the bioactive peptides. 3.2.2. Processing Conditions Apart from the formulation of the feed stream, the processing conditions also influence the properties of the encapsulates obtained. The inlet temperature of the drying gas is a key variable determining the drying kinetics. The works reported in the literature on the encapsulation of protein-based bioactives use temperature values in the range of 60–200 °C ( Table 1 ). The lowest inlet temperature was employed by Zhu et al. [ 54 ], who carried out the encapsulation of exenatide by spray-drying using inlet air at 60 °C. This low temperature, out of the norm for spray-drying, was due to the use of organic solvents (DMC and DCM) instead of water, as these evaporate at lower temperatures. The highest inlet temperature was 200 °C, which was used for the encapsulation of whey protein hydrolysate [ 50 ] and glutathione [ 55 ] employing β-cyclodextrin (β-CD) or CS as encapsulating agents ( Table 1 ). Yang et al. [ 50 ] reported outlet temperatures of 90 ± 5 °C, while Webber et al. [ 55 ] reported 72 °C. This difference could be due to the joint effect of other parameters such as feed flow rate, drying air flow rate and solid concentration of the feed solution. Cao et al. [ 44 ] studied the effect of inlet temperature on the encapsulation of watermelon seed hydrolyzed protein with MD. They compared inlet temperatures of 150, 160, 170, and 180 °C, finding that increasing temperature led to lower moisture content and solubility of the capsules. Likewise, particle size slightly increased with increasing temperature, varying from an average particle diameter of 10 µm at 150–160 °C to 12.9 µm at 170 °C. These results agree with a previous study [ 71 ] on the encapsulation of whey protein concentrates, reporting larger particles with increasing inlet temperature. This was attributed to elevated inlet temperatures increasing the moisture removal rate, forming a crust more quickly and limiting the shrinking time of the particle. Thus, the inlet temperature must be selected based on the nature of the solvent and the bioactive to be encapsulated, as well as other parameters like feed flow rate, drying air flow rate, and solid concentration of the feed solution. While high inlet temperatures increase the moisture removal rate, which can lead rapid crust formation and then larger particles, it also increases the risk of thermal degradation of bioactives. Therefore, it seems reasonable to use inlet temperatures between 130 and 190 °C and to avoid outlet temperatures above 100 °C to obtain encapsulates with minimal thermal degradation. Feed flow rates reported in the literature range from 160 to 1380 mL/h ( Table 1 ). This variable affects the size of the atomized droplets and consequently the size of the dried particles [ 72 ]. The lowest feed flow rate (160 mL/h) value was applied to encapsulate buffalo whey protein hydrolysates, with gum arabic (GA) and MD as encapsulating agents, obtaining particles of average diameter between 2 and 20 μm when using a two-fluid nozzle with 700 μm inner diameter [ 31 ]. The highest feed flow rate was 1380 mL/h and was used for the encapsulation of P. lunatus hydrolysates with GA and MD, forming particles with diameter of 3.3–6.8 μm when using a two-fluid nozzle for atomization with diameter of 500 μm [ 42 ]. Although it is accepted that a higher feed flow rate results in larger particles due to collision and subsequent fusion of small drops [ 73 ], the combined effect of the other processing variables could justify that the particles obtained in the latter study being smaller. Although high feed flow rates are desired for higher production rates, they can also result in condensation on the walls or wet particles in the chamber or cyclone due to higher humidity increasing stickiness and agglomeration, leading to a decrease in particle recovery attributed to wall deposition [ 60 ]. Thus, ratios of 300–500 mL/h are the most preferred. Curiously, there was a wide difference between the pneumatic air pressures used (40 and 1 bar), which has been proven to affect the particle size during the atomization process [ 73 ]. Regarding the correlation between the nozzle diameter and the morphology of the particles obtained, Keogh et al. [ 74 ] studied the effect of the nozzle diameter in the processing of milk powders and observed that the particle diameter of the spray-dried powders increased when increasing the nozzle diameter. However, the literature in this regard is very limited. Only two nozzle sizes were used in the literature found (500 and 700 µm), making it not possible to establish a correlation between these values and the particle size results. It is noteworthy that both the solubility and shelf stability of the powdered product is affected by the physicochemical properties of the encapsulates, such as morphology, particle size distribution, moisture content, or encapsulation efficiency. Smallness (under 5 µm) and a tight distribution of particles is desired to improve shelf stability [ 28 ]. Regarding the morphology of the particles, it was found that the size of the encapsulates produced by monoaxial spray-drying ranged from 0.132 μm [ 39 ] to 183 μm [ 35 ], with 21 of 24 articles reporting diameter values from 0.25 to 50 μm. These results are consistent with previous studies reporting typical dry particle diameter in laboratory scale spray-dryers of 0.5 to 50 μm [ 75 ]. It is also noteworthy that four of the five studies reporting average size below 1 μm were obtained via formation of nanoliposomes and similar conditions were used for three of the works found, with inlet air temperature of 130–170 °C and feed flow rate of 300 mL/h [ 38 , 39 , 41 ]. Encapsulation efficiency (EE) is one of the main parameters determining shelf-life stability. EE can be defined as the percentage of bioactive compound, whether peptide, hydrolysate, or protein, that is trapped inside the carrier matrix with respect to the amount initially added. Lower EE values lead to more exposed bioactive on the surface of the capsules, which is more easily degraded, resulting in reduced bioactivity [ 14 ]. All studies reported in the literature on the encapsulation of protein-based bioactives by monoaxial spray-drying obtained EE values over 50%, except for Ying et al. [ 59 ], who reported EE values in the range 29.51–45.83% for the encapsulation of soy peptides via W1/O/W2 because the emulsion was not physically stable ( Table 1 ). The highest EE (~100%) was reached by Koker et al. [ 60 ], where ovalbumin was encapsulated using dextran sulfate and poly-l-arginine as carrier for the elaboration of vaccine antigens. EE of peptide-loaded encapsulates can be improved by the addition of surfactants. For instance, Tween 80 was used to reduce flaxseed peptide migration to the air–water interface in spray-drying when using MD as carrier [ 36 ]. The authors reported reduced degradation by shear stress and dehydration, leading as well to higher bioactivity (i.e., antioxidant activity), indicating that the addition of surfactants should be considered in the future to improve encapsulation of peptides by spray-drying. Altogether, processing conditions play a significant role in determining the properties of the encapsulates obtained. The inlet temperature of the drying gas is a crucial variable that affects drying kinetics, particle size, and morphology. Inlet temperatures between 130 and 190 °C are found to obtain desirable encapsulates with minimal thermal degradation. Feed flow rate also affects the size of the atomized droplets, and consequently the size of the dried particles. It was observed that particle size under 5 µm and a narrow distribution of particles improve shelf stability of the encapsulates. Finally, feed flow rate ratios of 300–500 mL/h are preferred to avoid the formation of wet particles or wall deposition in lab spray-dryers. 3.3. Encapsulation by Coaxial Spray-Drying Lately, several works have been reported on the coaxial encapsulation of food bioactives by spray-drying using a three-fluid nozzle. This configuration ( Figure 2 B) allows feeding two different solutions through two concentric channels in the nozzle while the pneumatic air responsible for the atomization flows in the external channel [ 76 ]. In this way, the bioactive compound dissolved in a solution of carrier (core solution) can be pumped through the inner channel, and a solution containing the same or a different carrier (shell solution) is pumped through the outer channel. Both the shell and core solutions meet concentrically at the tip of the nozzle being atomized and dried in the chamber to form microcapsules with a core–shell structure. Therefore, this method might offer greater protection and better control for the release of the bioactive compound when compared to the monoaxial process (two-fluid nozzle) [ 77 ]. Despite the advantages of this method and its widespread use in the encapsulation of bioactive compounds such as carotenoids, triglycerides or organosulfur compounds [ 78 , 79 , 80 ], the literature on the encapsulation of bioactive protein hydrolysates and peptides by coaxial spray-drying is very limited. To the authors' knowledge, only one paper has been published regarding the production of protein-loaded microparticles via spray-drying using a three-fluid nozzle [ 20 ]. In this work, the authors encapsulated lysozymes employing an aqueous solution of trehalose as core solution and a shell solution consisting in PLGA dissolved in a mixture of solvents (acetonitrile, DCM, and acetone). The effect of the core:shell flow rate ratio (4:1 or 10:1) was studied. The operating conditions selected were (i) inlet temperature = 60 °C, (ii) outlet temperature = 40–45 °C, and (iii) drying air flow rate = 37500 L/h. The study concluded that solvent selection did not affect particle size, while it did affect EE. On the other hand, the feed flow rate ratio did affect the particle size, which ranged from 1.07 to 1.60 μm. Hence, the coaxial encapsulation by spray-drying of bioactive protein hydrolysates and peptides deserves further research. 4. Encapsulation of Protein-Based Bioactives by Electrospraying 4.1. Fundamentals of Electrospraying Electrospraying or electrohydrodynamic atomization is a drying and encapsulation technique based on the application of an electric field to a solution to obtain dried nano/microstructures at room temperature [ 81 ]. Electrospraying consists of pumping a solution, dispersion, or emulsion that contains the protein-based bioactive through a capillary injector/needle of a conductive material [ 82 ]. A grounded collector is placed opposite the needle at a given distance, and an electric field is applied between the injector and the collector. The solution is pumped through the needle at a regulated flow rate, and if no voltage is applied, as the drop of solution emerges from the needle a meniscus is formed. When the electrostatic field is sufficiently high, the air–liquid interface of the meniscus is polarized, causing it to deform into a conical shape, known as a Taylor cone [ 81 ]. As the voltage continues to increase, it reaches a point at which the surface tension is no longer able to hold the liquid in the droplet, resulting in the emission of a jet from the tip of the cone directed towards the collector. The jet breaks into a spray of charged particles due to the low viscoelasticity of the solution and the electrostatic repulsion forces that take place. In the travel of the droplets towards the collector, the solvent(s) used is evaporated and dry nano/microparticles are obtained ( Figure 3 ) [ 83 ]. Depending on the properties of the solution and the processing parameters used, two different methodologies can be applied, mainly (i) electrospraying, where the intermolecular cohesion of the fluid is low enough that the electrostatic forces break the jet emitted from the solution into small droplets that result in the formation of nano/microparticles after solvent evaporation, or (ii) electrospinning, where the high molecular cohesion avoids jet fragmentation, and after the evaporation of the solvent, it gives rise to the formation of ultrafine fibers [ 82 ]. Electrospraying is the preferred process for obtaining food ingredients, since electrospun fibers, contrarily to nano/microcapsules, result in continuous mats that are difficult to disperse in any food matrix without prior breakage [ 19 ]. Electrospraying, as spray-drying, also allows working in both mono- and coaxial configurations. Monoaxial electrospraying typically results in the formation of amorphous solid dispersions containing the protein-based bioactives dispersed within the carrier matrix ( Figure 3 ). Alternatively, coaxial electrospraying is a customized version of electrospraying, in which two different liquids are separately delivered through individual coaxial capillary needles ( Figure 4 ). The solution containing the bioactive compound and potentially the encapsulating agent is pumped through the inner needle (core), while a second solution containing the same or a different encapsulating agent is delivered through an outer concentrical needle (shell) [ 84 ]. Therefore, a concentric Taylor cone of both solutions is formed at the tip of the needles, and when the solution and processing parameters are appropriately selected, it results in the formation of nano/microcapsules with a core–shell structure [ 85 , 86 ] ( Figure 4 ). Coaxial electrospraying combines the advantages of monoaxial electrospraying, adding the ability to precisely control the core–shell shape, as well as better protecting the bioactive peptides from process-induced denaturation and aggregation [ 87 ]. Electrospraying parameters such as solution properties, processing variables and environmental conditions can affect the morphology, particle size and EE [ 88 ]. A high concentration of encapsulating agent leading to high viscosity and density of the solution can derive in the formation of larger particles, while increasing electrical conductivity of the solution results in particles with smaller diameter. Regarding the processing variables, a high electric potential between injector and collector results in smaller particle diameter, whereas increasing solution flow rate increases particle size. Long injector–collector distances allow for better evaporation of the solvent, while short distances may result in wet and collapsed particles. Environmental conditions such as temperature and humidity also affect the drying kinetics, since they determine the driving forces for the drying process. In addition, other factors should be considered when encapsulating protein-based bioactives by electrospraying: (1) proteins lead to highly conductive solutions that prevent charge formation, which reduces the stability of the Taylor cone, and (2) the use of food-grade solvents such as water leads to solutions with high surface tension that hinder jetting [ 86 ]. Electrospraying encapsulation has been widely used in the pharmacological field due to its low cost, easy operation and improved bioaccessibility of the nano/microcapsules obtained [ 8 ]. However, its use in food applications is still limited due to its low production capacity. To solve this limitation, several modifications have been reported, including: (i) multineedle electrospraying systems [ 89 ], (ii) free surface electrospraying systems [ 90 ], or (iii) pressurized-gas-assisted electrospraying [ 91 ]. However, this absence of application to the food industry is reflected in a lack of literature on the subject. The available information is mainly oriented to oral pharmacological supplementation, and no data were found on bioactive protein hydrolysates. Thus, further research should be carried out on the application of electrospraying encapsulation in foods. 4.2. Encapsulation by Monoaxial Electrospraying 4.2.1. Formulation of the Feed Stream The encapsulation of protein-based bioactives by monoaxial electrospraying requires the drying of only one solution containing the bioactive. The most common method to produce this feed stream is blending (e.g., dissolving the bioactive in a solution containing the carrier). Nine of the eleven works found in the literature used this approach ( Table 2 ). Bock et al. [ 92 ] encapsulated bovine serum albumin (BSA) by electrospraying a blend feed stream where BSA was dissolved in chloroform or DCM using poly(ethylene glycol) (PEG) and poly(ε-caprolactone)(PCL)/PLGA as carrier. Similarly, Musaei et al. [ 93 ] prepared a blend feed stream using an ethanol–acetic acid mixture to encapsulate BSA using PLGA as encapsulating agent. Blend electrospraying has also been used to encapsulate larger molecules, such as the hormone angiotensin II using N-octyl-O-sulfate chitosan (NOSC) as a carrier [ 94 ], or the enzymes alkaline phosphatase with poly(ethylene oxide) (PEO) [ 95 ] and streptokinase with PLGA [ 96 ]. Although electrospraying is carried out at room temperature, which avoids thermal degradation of thermosensitive ingredients, the use of specific solvents may induce protein denaturation and loss of activity when exposure is prolonged [ 8 ]. Hence, an alternative approach to a blend for producing the feed stream is to obtain emulsions that prevent contact between specific solvents and the bioactives [ 8 ]. According to previous studies, encapsulation by emulsion electrospraying allows the formation of particles with core–shell structures similar to those that could be obtained by coaxial electrospraying [ 103 ]. This process is often used to mix two immiscible fluids, typically through a single W/O or double W1/O/W2 emulsion [ 8 ]. The two articles found in the literature using emulsion electrospraying were based on the drying of water-in-water (W/W) emulsions. Yao et al. [ 99 ] used this approach to encapsulate BSA in PLGA. For that, two immiscible solutions were prepared: the organic phase was composed of PLGA in chloroform and the aqueous phase was composed of the BSA dissolved in water. Similarly, Y. Song et al. [ 102 ] produced a W/W emulsion by dissolving β-amylase in the aqueous phase composed of dextran and sodium alginate, which was electrosprayed into a water solution containing CaCl 2 and PEG, forming a calcium alginate shell containing the amylase core. As mentioned in Section 3.2.1 , it should be noted that emulsion feed preparation is less used, as it presents difficulties in producing stable emulsions and the shear stress of mechanical mixing required for emulsion preparation could modify the protein-based bioactives [ 8 ]. The type of carrier and solvent used determine the main properties of the feed stream influencing the electrospraying process, such as viscoelasticity, conductivity, and surface tension [ 86 ]. A wide variety of natural and synthetic polymers are used as encapsulating agents in electrospraying, including biocompatible and biodegradable polymers such as gelatin, MD, pullulan, glucose syrup, dextran, hyaluronan, CS, PCL, poly(lactic acid) (PLA), PEG, PLGA, alginate, PEO, and NOSC, among many others. Carriers commonly used were alginate, PEO and NOSC, all of which are particularly used for the formulation of oral delivery drugs since they are all safe and present high biocompatibility. Alginate was used to encapsulate α-calcitonin gene-related peptide (α-CGRP) [ 97 ] and BSA/porcine interleukin-1β (pIL-1β) [ 100 ], both by blend electrospraying and resulting in particles with widely differing sizes, ranging from 194.23 ± 10.08 to 20 μm, respectively. PEO is a synthetic semicrystalline polymer mostly used for electrospinning due to its rheological characteristics, and thus only one work used it for electrospraying the enzyme alkaline phosphatase [ 95 , 104 ]. Likewise, NOSC was only used for the encapsulation of the hormone angiotensin II [ 94 ]. The most reported polymer carrier in the literature for the encapsulation of protein-based bioactives was PLGA ( Table 2 ), a US Food and Drug Administration (FDA)-approved biocompatible copolymer that has been extensively used in biomedical devices with excellent application records in vivo [ 99 ]. Interestingly, Musaei et al. [ 93 ] found that increasing PLGA concentration did indeed affect the particle size of the capsules, increasing the size of the nanocapsules from 120 nm to 225 nm, which is related to increasing EE. Although all these biopolymers have shown good encapsulating capacity, studies have focused on drug release formulation and research on food application is very limited. Only PLGA has been studied for application in food fortification, with good results [ 105 ]. Polysaccharide- and protein-based carriers are commonly employed as encapsulating agents for encapsulation of protein-based bioactives by spray-drying; however their use for the encapsulation of these bioactives by electrospraying was not reported in the literature. These kinds of carriers are especially suitable for the food industry since they are food-grade and soluble in water, which avoids the use of non-food-grade solvents [ 105 ]. Therefore, further research on the use of food-grade, low-cost biopolymers for the encapsulation of protein-based bioactives by electrospraying is required. 4.2.2. Processing Conditions Processing variables (voltage, injector-to-collector distance and feed flow rate), together with feed solution properties, affect the characteristics of the nano/microcapsules obtained (e.g., morphology, size). The applied voltage for all the studies reported in the literature was kept between 2 and 20 kV ( Table 2 ). The effect of voltage was assessed for the encapsulation of BSA by electrospraying using PLGA as carrier [ 93 ]. The authors compared three different voltages (10, 15 and 20 kV) and found that increasing voltage from 10 kV to 20 kV resulted in decreasing average diameter from 0.185 µm to 0.085 µm, which is desired to increase surface area and thus improve solubility and permeability. This effect was also found in electrosprayed amylase particles with PEG and dextran as carrier, where the applied voltages were adjusted from 2.6 kV to 2.85 kV to produce particles of different diameters [ 102 ]. However, applied voltages over 20 kV were found to alter protein-based bioactives. For instance, a study on the encapsulation of angiotensin II by electrospraying using NOSC as encapsulating agent showed that its stability was significantly reduced at 20 kV. Since the electric field strength is determined by both applied voltage and distance between nozzle and collector (N-C), changes in both parameters affect the stability of the bioactive compounds during processing. Increasing feed flow rate is desired to increase productivity; however, it is linked to higher particle diameter. Low feed flow rate results in better encapsulation as well, and thus a compromise between productivity and quality of the capsules must be reached. Onyekuru et al. [ 95 ] studied the effect of feed flow rates ranging from 0.3 to 1.5 mL/h on the encapsulation of alkaline phosphatase with PEO, determining that although low flow rates produced better encapsulation, the optimum flow rate was 0.6 mL/h. Different feed flow rates were also compared for the encapsulation of serum albumin (SA) by electrospraying using PEG and PCL/PLGA as encapsulating agents [ 92 ]. It was reported that increasing the feed flow rate from 0.5 mL/h to 1 mL/h resulted in average diameters increasing from 5.6 ± 0.8 μm to 7.1 ± 1.7 μm. However, higher flow rates also resulted in uneven spread of the solution at the nozzle and an uncontrolled electrospraying of large droplets. This work also utilized different nozzle diameters (450–800 μm), but no effect on particle size was reported. On the other hand, Y. Song et al. [ 102 ] determined that the size of the particles could be reduced by using nozzles with a smaller diameter after comparing three different diameters (40, 170 and 320 μm). The consensus seems to be working at a low feed flow rate, but process productivity must not be compromised to be cost-effective. The studies shown in Table 2 reported EE values for the nano/microcapsules loaded with protein-based bioactives ranging from 20% to 92%. The lowest EE was obtained for encapsulation of SA using PEG and PCL/PLGA as carriers [ 92 ]. Since higher EE has been linked to larger particles and lower protein loading [ 106 ], these parameters were studied. Indeed, the authors confirmed that increased particle diameter corresponded with higher EE values, with a critical size allowing optimum encapsulation. In the same study, lower protein loading also resulted in improved EE, but the extraction method used to measure EE presented limitations due to protein aggregation and the lack of use of surfactants. The highest EE was achieved by encapsulating BSA by emulsion electrospraying using PLGA as encapsulating agent [ 99 ]. This work showed that increasing the aqueous phase volume ratio (e.g., increasing bioactive load) resulted in decreased EE, varying from 92% at 5 µL/mL to 80% at 100 µL/mL. It was explained as being due to increased density of emulsion droplets in the feed solution, and thus increasing migration of the aqueous phase containing the BSA to the surface of the particle. 4.3. Encapsulation by Coaxial Electrospraying The literature found regarding the encapsulation by coaxial electrospraying of protein-based bioactives was focused exclusively on the pharmacological/medical field. No works on the encapsulation by coaxial electrospraying of bioactive protein hydrolysates or peptides have been reported in the literature. Only four works studying the coaxial electrospraying of proteins were found ( Table 3 ). None of them used carrier in the formulation of the inner solution (core). For the outer solution (shell) PLGA was the most used encapsulating agent, appearing in two studies. This follows the trend established in monoaxial electrospraying, since, as previously mentioned, all the literature found was mainly focused on oral drug delivery, where PLGA was the most frequently used biopolymer. One study focused on the encapsulation of a water solution of BSA using an outer solution of PLGA dissolved in either DCM or a combination of DCM and DMF [ 107 ]. The other work encapsulated ranibizumab, a protein drug used for the treatment of age-related macular degeneration, using PLGA dissolved in a combination of DCM and acetonitrile as the outer solution [ 108 ]. Regarding the use of solvents, six of eight works used organic solvents, mainly for the outer feed. This is because the use of two immiscible solutions provides better core–shell separation by minimizing interdiffusion between layers [ 109 ]. A solution of ethyl acetate and n-butanol, along with acetylated dextran as carrier, was used as the outer feed for the encapsulation of anthrax protective antigens dissolved in the inner water solution [ 110 ]. Rasekh et al. [ 94 ] coaxially electrosprayed angiotensin II using NOSC as carrier for the inner solution and tristearin dissolved in DCM as outer solution. Since the literature found was focused on the production of oral delivered drugs, it would be necessary to take into consideration the need to apply two completely immiscible food-grade solvents to produce encapsulates oriented for food fortification. Voltages applied ranged from 5 to 22.5 kV, similar to the values used for monoaxial electrospraying (2.67–20 kV). The effect of voltage was studied for the encapsulation of angiotensin II using tristearin and NOSC as carriers inner and outer carriers, respectively [ 94 ]. The applied voltage values were 20 and 30 kV, and the authors compared the stability of the enzyme using an ELISA, finding that at 30 kV the concentration of angiotensin II in the microparticles was reduced by approximately 20%. For the encapsulation of alkaline phosphatase with PEO as outer carrier [ 95 ], the voltage was optimized to 22.5 kV. Similarly, these authors found that this high voltage resulted in a loss of activity of the enzyme up to 40% compared to the activity obtained by monoaxial electrospraying at 15.5 kV. Other parameters affecting particle characteristics are feed flow rates (inner and outer) and nozzle diameters. For the inner solutions (core), feed flow rates of 0.02–3.6 mL/h were used, while for the outer solutions (shell), 0.1–18 mL/h was used. Regarding the nozzle diameters, they ranged from 184 to 1000 μm for inner capillary and 603 to 2000 μm for the outer capillary. As previously mentioned in the previous section, increasing feed flow rate and nozzle diameters typically results in larger particles. This agreed with the data obtained by Zhao et al. [ 112 ], where alkaline phosphatase was encapsulated using CMC as inner carrier and alginate and PEGDA as outer carriers. They reported the highest feed flow rates (1.8 mL/h for the core and 3.96 mL/h for the shell) in the literature and obtained the largest particles at 440 μm. However, the opposite conclusion was obtained after comparing the encapsulation of angiotensin II (using NOSC as inner carrier and tristearin outer carriers) [ 94 ] and the encapsulation of alkaline phosphatase with PEO as carrier [ 95 ]. Both studies used similar nozzle diameters (1000 μm (inner)–2000 μm (outer), and 900 μm (inner)–1900 μm (outer), respectively), but the first study used feed flow rates 10 times higher. Even though larger particles would be expected for the angiotensin II encapsulation, due to the higher flow rates, their size was up to 86% smaller. In fact, they obtained the smallest particles, which could be due to the nozzle–collector distance, the highest reported in the literature at 20 cm, and the slightly higher voltage used. Coaxial electrospraying of bovine hemoglobin also resulted in small particles of 0.37 μm, as it was particularly important to obtain nano/microcapsules in the range of 0.1 to 3 μm to effectively avoid extravasation through the blood vessel wall and act as oxygen carriers [ 111 ]. High EE values were obtained for all the studies reported in the literature ( Table 3 ), ranging from 70% to 99%. These values are higher than the ones obtained for monoaxial electrospraying, where four of the nine reported EE values were under 50%. Zamani et al. [ 107 ] reported ranges of EE from 46.7 ± 4.3% to 74.6 ± 2.9%, which were linked to incomplete encapsulation due to inner feed flow rates being too high as well as high concentrations of BSA in the core. The highest EE found was obtained for the encapsulation of alkaline phosphatase with PEO as outer carrier [ 95 ]. They also compared the effect of monoaxial and coaxial electrospraying, confirming that the EE was increased in core–shell structures. Although coaxial electrospraying has exhibited promising outcomes, the encapsulation of bioactive protein hydrolysates or peptides has only been minimally investigated. Thus, further studies are required to fully evaluate the feasibility of this technology for the encapsulation of bioactive peptides. Particularly, there is a need to investigate the use of food-grade solvents and to optimize processing conditions that lead to encapsulates with potential use in food fortification. 4.1. Fundamentals of Electrospraying Electrospraying or electrohydrodynamic atomization is a drying and encapsulation technique based on the application of an electric field to a solution to obtain dried nano/microstructures at room temperature [ 81 ]. Electrospraying consists of pumping a solution, dispersion, or emulsion that contains the protein-based bioactive through a capillary injector/needle of a conductive material [ 82 ]. A grounded collector is placed opposite the needle at a given distance, and an electric field is applied between the injector and the collector. The solution is pumped through the needle at a regulated flow rate, and if no voltage is applied, as the drop of solution emerges from the needle a meniscus is formed. When the electrostatic field is sufficiently high, the air–liquid interface of the meniscus is polarized, causing it to deform into a conical shape, known as a Taylor cone [ 81 ]. As the voltage continues to increase, it reaches a point at which the surface tension is no longer able to hold the liquid in the droplet, resulting in the emission of a jet from the tip of the cone directed towards the collector. The jet breaks into a spray of charged particles due to the low viscoelasticity of the solution and the electrostatic repulsion forces that take place. In the travel of the droplets towards the collector, the solvent(s) used is evaporated and dry nano/microparticles are obtained ( Figure 3 ) [ 83 ]. Depending on the properties of the solution and the processing parameters used, two different methodologies can be applied, mainly (i) electrospraying, where the intermolecular cohesion of the fluid is low enough that the electrostatic forces break the jet emitted from the solution into small droplets that result in the formation of nano/microparticles after solvent evaporation, or (ii) electrospinning, where the high molecular cohesion avoids jet fragmentation, and after the evaporation of the solvent, it gives rise to the formation of ultrafine fibers [ 82 ]. Electrospraying is the preferred process for obtaining food ingredients, since electrospun fibers, contrarily to nano/microcapsules, result in continuous mats that are difficult to disperse in any food matrix without prior breakage [ 19 ]. Electrospraying, as spray-drying, also allows working in both mono- and coaxial configurations. Monoaxial electrospraying typically results in the formation of amorphous solid dispersions containing the protein-based bioactives dispersed within the carrier matrix ( Figure 3 ). Alternatively, coaxial electrospraying is a customized version of electrospraying, in which two different liquids are separately delivered through individual coaxial capillary needles ( Figure 4 ). The solution containing the bioactive compound and potentially the encapsulating agent is pumped through the inner needle (core), while a second solution containing the same or a different encapsulating agent is delivered through an outer concentrical needle (shell) [ 84 ]. Therefore, a concentric Taylor cone of both solutions is formed at the tip of the needles, and when the solution and processing parameters are appropriately selected, it results in the formation of nano/microcapsules with a core–shell structure [ 85 , 86 ] ( Figure 4 ). Coaxial electrospraying combines the advantages of monoaxial electrospraying, adding the ability to precisely control the core–shell shape, as well as better protecting the bioactive peptides from process-induced denaturation and aggregation [ 87 ]. Electrospraying parameters such as solution properties, processing variables and environmental conditions can affect the morphology, particle size and EE [ 88 ]. A high concentration of encapsulating agent leading to high viscosity and density of the solution can derive in the formation of larger particles, while increasing electrical conductivity of the solution results in particles with smaller diameter. Regarding the processing variables, a high electric potential between injector and collector results in smaller particle diameter, whereas increasing solution flow rate increases particle size. Long injector–collector distances allow for better evaporation of the solvent, while short distances may result in wet and collapsed particles. Environmental conditions such as temperature and humidity also affect the drying kinetics, since they determine the driving forces for the drying process. In addition, other factors should be considered when encapsulating protein-based bioactives by electrospraying: (1) proteins lead to highly conductive solutions that prevent charge formation, which reduces the stability of the Taylor cone, and (2) the use of food-grade solvents such as water leads to solutions with high surface tension that hinder jetting [ 86 ]. Electrospraying encapsulation has been widely used in the pharmacological field due to its low cost, easy operation and improved bioaccessibility of the nano/microcapsules obtained [ 8 ]. However, its use in food applications is still limited due to its low production capacity. To solve this limitation, several modifications have been reported, including: (i) multineedle electrospraying systems [ 89 ], (ii) free surface electrospraying systems [ 90 ], or (iii) pressurized-gas-assisted electrospraying [ 91 ]. However, this absence of application to the food industry is reflected in a lack of literature on the subject. The available information is mainly oriented to oral pharmacological supplementation, and no data were found on bioactive protein hydrolysates. Thus, further research should be carried out on the application of electrospraying encapsulation in foods. 4.2. Encapsulation by Monoaxial Electrospraying 4.2.1. Formulation of the Feed Stream The encapsulation of protein-based bioactives by monoaxial electrospraying requires the drying of only one solution containing the bioactive. The most common method to produce this feed stream is blending (e.g., dissolving the bioactive in a solution containing the carrier). Nine of the eleven works found in the literature used this approach ( Table 2 ). Bock et al. [ 92 ] encapsulated bovine serum albumin (BSA) by electrospraying a blend feed stream where BSA was dissolved in chloroform or DCM using poly(ethylene glycol) (PEG) and poly(ε-caprolactone)(PCL)/PLGA as carrier. Similarly, Musaei et al. [ 93 ] prepared a blend feed stream using an ethanol–acetic acid mixture to encapsulate BSA using PLGA as encapsulating agent. Blend electrospraying has also been used to encapsulate larger molecules, such as the hormone angiotensin II using N-octyl-O-sulfate chitosan (NOSC) as a carrier [ 94 ], or the enzymes alkaline phosphatase with poly(ethylene oxide) (PEO) [ 95 ] and streptokinase with PLGA [ 96 ]. Although electrospraying is carried out at room temperature, which avoids thermal degradation of thermosensitive ingredients, the use of specific solvents may induce protein denaturation and loss of activity when exposure is prolonged [ 8 ]. Hence, an alternative approach to a blend for producing the feed stream is to obtain emulsions that prevent contact between specific solvents and the bioactives [ 8 ]. According to previous studies, encapsulation by emulsion electrospraying allows the formation of particles with core–shell structures similar to those that could be obtained by coaxial electrospraying [ 103 ]. This process is often used to mix two immiscible fluids, typically through a single W/O or double W1/O/W2 emulsion [ 8 ]. The two articles found in the literature using emulsion electrospraying were based on the drying of water-in-water (W/W) emulsions. Yao et al. [ 99 ] used this approach to encapsulate BSA in PLGA. For that, two immiscible solutions were prepared: the organic phase was composed of PLGA in chloroform and the aqueous phase was composed of the BSA dissolved in water. Similarly, Y. Song et al. [ 102 ] produced a W/W emulsion by dissolving β-amylase in the aqueous phase composed of dextran and sodium alginate, which was electrosprayed into a water solution containing CaCl 2 and PEG, forming a calcium alginate shell containing the amylase core. As mentioned in Section 3.2.1 , it should be noted that emulsion feed preparation is less used, as it presents difficulties in producing stable emulsions and the shear stress of mechanical mixing required for emulsion preparation could modify the protein-based bioactives [ 8 ]. The type of carrier and solvent used determine the main properties of the feed stream influencing the electrospraying process, such as viscoelasticity, conductivity, and surface tension [ 86 ]. A wide variety of natural and synthetic polymers are used as encapsulating agents in electrospraying, including biocompatible and biodegradable polymers such as gelatin, MD, pullulan, glucose syrup, dextran, hyaluronan, CS, PCL, poly(lactic acid) (PLA), PEG, PLGA, alginate, PEO, and NOSC, among many others. Carriers commonly used were alginate, PEO and NOSC, all of which are particularly used for the formulation of oral delivery drugs since they are all safe and present high biocompatibility. Alginate was used to encapsulate α-calcitonin gene-related peptide (α-CGRP) [ 97 ] and BSA/porcine interleukin-1β (pIL-1β) [ 100 ], both by blend electrospraying and resulting in particles with widely differing sizes, ranging from 194.23 ± 10.08 to 20 μm, respectively. PEO is a synthetic semicrystalline polymer mostly used for electrospinning due to its rheological characteristics, and thus only one work used it for electrospraying the enzyme alkaline phosphatase [ 95 , 104 ]. Likewise, NOSC was only used for the encapsulation of the hormone angiotensin II [ 94 ]. The most reported polymer carrier in the literature for the encapsulation of protein-based bioactives was PLGA ( Table 2 ), a US Food and Drug Administration (FDA)-approved biocompatible copolymer that has been extensively used in biomedical devices with excellent application records in vivo [ 99 ]. Interestingly, Musaei et al. [ 93 ] found that increasing PLGA concentration did indeed affect the particle size of the capsules, increasing the size of the nanocapsules from 120 nm to 225 nm, which is related to increasing EE. Although all these biopolymers have shown good encapsulating capacity, studies have focused on drug release formulation and research on food application is very limited. Only PLGA has been studied for application in food fortification, with good results [ 105 ]. Polysaccharide- and protein-based carriers are commonly employed as encapsulating agents for encapsulation of protein-based bioactives by spray-drying; however their use for the encapsulation of these bioactives by electrospraying was not reported in the literature. These kinds of carriers are especially suitable for the food industry since they are food-grade and soluble in water, which avoids the use of non-food-grade solvents [ 105 ]. Therefore, further research on the use of food-grade, low-cost biopolymers for the encapsulation of protein-based bioactives by electrospraying is required. 4.2.2. Processing Conditions Processing variables (voltage, injector-to-collector distance and feed flow rate), together with feed solution properties, affect the characteristics of the nano/microcapsules obtained (e.g., morphology, size). The applied voltage for all the studies reported in the literature was kept between 2 and 20 kV ( Table 2 ). The effect of voltage was assessed for the encapsulation of BSA by electrospraying using PLGA as carrier [ 93 ]. The authors compared three different voltages (10, 15 and 20 kV) and found that increasing voltage from 10 kV to 20 kV resulted in decreasing average diameter from 0.185 µm to 0.085 µm, which is desired to increase surface area and thus improve solubility and permeability. This effect was also found in electrosprayed amylase particles with PEG and dextran as carrier, where the applied voltages were adjusted from 2.6 kV to 2.85 kV to produce particles of different diameters [ 102 ]. However, applied voltages over 20 kV were found to alter protein-based bioactives. For instance, a study on the encapsulation of angiotensin II by electrospraying using NOSC as encapsulating agent showed that its stability was significantly reduced at 20 kV. Since the electric field strength is determined by both applied voltage and distance between nozzle and collector (N-C), changes in both parameters affect the stability of the bioactive compounds during processing. Increasing feed flow rate is desired to increase productivity; however, it is linked to higher particle diameter. Low feed flow rate results in better encapsulation as well, and thus a compromise between productivity and quality of the capsules must be reached. Onyekuru et al. [ 95 ] studied the effect of feed flow rates ranging from 0.3 to 1.5 mL/h on the encapsulation of alkaline phosphatase with PEO, determining that although low flow rates produced better encapsulation, the optimum flow rate was 0.6 mL/h. Different feed flow rates were also compared for the encapsulation of serum albumin (SA) by electrospraying using PEG and PCL/PLGA as encapsulating agents [ 92 ]. It was reported that increasing the feed flow rate from 0.5 mL/h to 1 mL/h resulted in average diameters increasing from 5.6 ± 0.8 μm to 7.1 ± 1.7 μm. However, higher flow rates also resulted in uneven spread of the solution at the nozzle and an uncontrolled electrospraying of large droplets. This work also utilized different nozzle diameters (450–800 μm), but no effect on particle size was reported. On the other hand, Y. Song et al. [ 102 ] determined that the size of the particles could be reduced by using nozzles with a smaller diameter after comparing three different diameters (40, 170 and 320 μm). The consensus seems to be working at a low feed flow rate, but process productivity must not be compromised to be cost-effective. The studies shown in Table 2 reported EE values for the nano/microcapsules loaded with protein-based bioactives ranging from 20% to 92%. The lowest EE was obtained for encapsulation of SA using PEG and PCL/PLGA as carriers [ 92 ]. Since higher EE has been linked to larger particles and lower protein loading [ 106 ], these parameters were studied. Indeed, the authors confirmed that increased particle diameter corresponded with higher EE values, with a critical size allowing optimum encapsulation. In the same study, lower protein loading also resulted in improved EE, but the extraction method used to measure EE presented limitations due to protein aggregation and the lack of use of surfactants. The highest EE was achieved by encapsulating BSA by emulsion electrospraying using PLGA as encapsulating agent [ 99 ]. This work showed that increasing the aqueous phase volume ratio (e.g., increasing bioactive load) resulted in decreased EE, varying from 92% at 5 µL/mL to 80% at 100 µL/mL. It was explained as being due to increased density of emulsion droplets in the feed solution, and thus increasing migration of the aqueous phase containing the BSA to the surface of the particle. 4.2.1. Formulation of the Feed Stream The encapsulation of protein-based bioactives by monoaxial electrospraying requires the drying of only one solution containing the bioactive. The most common method to produce this feed stream is blending (e.g., dissolving the bioactive in a solution containing the carrier). Nine of the eleven works found in the literature used this approach ( Table 2 ). Bock et al. [ 92 ] encapsulated bovine serum albumin (BSA) by electrospraying a blend feed stream where BSA was dissolved in chloroform or DCM using poly(ethylene glycol) (PEG) and poly(ε-caprolactone)(PCL)/PLGA as carrier. Similarly, Musaei et al. [ 93 ] prepared a blend feed stream using an ethanol–acetic acid mixture to encapsulate BSA using PLGA as encapsulating agent. Blend electrospraying has also been used to encapsulate larger molecules, such as the hormone angiotensin II using N-octyl-O-sulfate chitosan (NOSC) as a carrier [ 94 ], or the enzymes alkaline phosphatase with poly(ethylene oxide) (PEO) [ 95 ] and streptokinase with PLGA [ 96 ]. Although electrospraying is carried out at room temperature, which avoids thermal degradation of thermosensitive ingredients, the use of specific solvents may induce protein denaturation and loss of activity when exposure is prolonged [ 8 ]. Hence, an alternative approach to a blend for producing the feed stream is to obtain emulsions that prevent contact between specific solvents and the bioactives [ 8 ]. According to previous studies, encapsulation by emulsion electrospraying allows the formation of particles with core–shell structures similar to those that could be obtained by coaxial electrospraying [ 103 ]. This process is often used to mix two immiscible fluids, typically through a single W/O or double W1/O/W2 emulsion [ 8 ]. The two articles found in the literature using emulsion electrospraying were based on the drying of water-in-water (W/W) emulsions. Yao et al. [ 99 ] used this approach to encapsulate BSA in PLGA. For that, two immiscible solutions were prepared: the organic phase was composed of PLGA in chloroform and the aqueous phase was composed of the BSA dissolved in water. Similarly, Y. Song et al. [ 102 ] produced a W/W emulsion by dissolving β-amylase in the aqueous phase composed of dextran and sodium alginate, which was electrosprayed into a water solution containing CaCl 2 and PEG, forming a calcium alginate shell containing the amylase core. As mentioned in Section 3.2.1 , it should be noted that emulsion feed preparation is less used, as it presents difficulties in producing stable emulsions and the shear stress of mechanical mixing required for emulsion preparation could modify the protein-based bioactives [ 8 ]. The type of carrier and solvent used determine the main properties of the feed stream influencing the electrospraying process, such as viscoelasticity, conductivity, and surface tension [ 86 ]. A wide variety of natural and synthetic polymers are used as encapsulating agents in electrospraying, including biocompatible and biodegradable polymers such as gelatin, MD, pullulan, glucose syrup, dextran, hyaluronan, CS, PCL, poly(lactic acid) (PLA), PEG, PLGA, alginate, PEO, and NOSC, among many others. Carriers commonly used were alginate, PEO and NOSC, all of which are particularly used for the formulation of oral delivery drugs since they are all safe and present high biocompatibility. Alginate was used to encapsulate α-calcitonin gene-related peptide (α-CGRP) [ 97 ] and BSA/porcine interleukin-1β (pIL-1β) [ 100 ], both by blend electrospraying and resulting in particles with widely differing sizes, ranging from 194.23 ± 10.08 to 20 μm, respectively. PEO is a synthetic semicrystalline polymer mostly used for electrospinning due to its rheological characteristics, and thus only one work used it for electrospraying the enzyme alkaline phosphatase [ 95 , 104 ]. Likewise, NOSC was only used for the encapsulation of the hormone angiotensin II [ 94 ]. The most reported polymer carrier in the literature for the encapsulation of protein-based bioactives was PLGA ( Table 2 ), a US Food and Drug Administration (FDA)-approved biocompatible copolymer that has been extensively used in biomedical devices with excellent application records in vivo [ 99 ]. Interestingly, Musaei et al. [ 93 ] found that increasing PLGA concentration did indeed affect the particle size of the capsules, increasing the size of the nanocapsules from 120 nm to 225 nm, which is related to increasing EE. Although all these biopolymers have shown good encapsulating capacity, studies have focused on drug release formulation and research on food application is very limited. Only PLGA has been studied for application in food fortification, with good results [ 105 ]. Polysaccharide- and protein-based carriers are commonly employed as encapsulating agents for encapsulation of protein-based bioactives by spray-drying; however their use for the encapsulation of these bioactives by electrospraying was not reported in the literature. These kinds of carriers are especially suitable for the food industry since they are food-grade and soluble in water, which avoids the use of non-food-grade solvents [ 105 ]. Therefore, further research on the use of food-grade, low-cost biopolymers for the encapsulation of protein-based bioactives by electrospraying is required. 4.2.2. Processing Conditions Processing variables (voltage, injector-to-collector distance and feed flow rate), together with feed solution properties, affect the characteristics of the nano/microcapsules obtained (e.g., morphology, size). The applied voltage for all the studies reported in the literature was kept between 2 and 20 kV ( Table 2 ). The effect of voltage was assessed for the encapsulation of BSA by electrospraying using PLGA as carrier [ 93 ]. The authors compared three different voltages (10, 15 and 20 kV) and found that increasing voltage from 10 kV to 20 kV resulted in decreasing average diameter from 0.185 µm to 0.085 µm, which is desired to increase surface area and thus improve solubility and permeability. This effect was also found in electrosprayed amylase particles with PEG and dextran as carrier, where the applied voltages were adjusted from 2.6 kV to 2.85 kV to produce particles of different diameters [ 102 ]. However, applied voltages over 20 kV were found to alter protein-based bioactives. For instance, a study on the encapsulation of angiotensin II by electrospraying using NOSC as encapsulating agent showed that its stability was significantly reduced at 20 kV. Since the electric field strength is determined by both applied voltage and distance between nozzle and collector (N-C), changes in both parameters affect the stability of the bioactive compounds during processing. Increasing feed flow rate is desired to increase productivity; however, it is linked to higher particle diameter. Low feed flow rate results in better encapsulation as well, and thus a compromise between productivity and quality of the capsules must be reached. Onyekuru et al. [ 95 ] studied the effect of feed flow rates ranging from 0.3 to 1.5 mL/h on the encapsulation of alkaline phosphatase with PEO, determining that although low flow rates produced better encapsulation, the optimum flow rate was 0.6 mL/h. Different feed flow rates were also compared for the encapsulation of serum albumin (SA) by electrospraying using PEG and PCL/PLGA as encapsulating agents [ 92 ]. It was reported that increasing the feed flow rate from 0.5 mL/h to 1 mL/h resulted in average diameters increasing from 5.6 ± 0.8 μm to 7.1 ± 1.7 μm. However, higher flow rates also resulted in uneven spread of the solution at the nozzle and an uncontrolled electrospraying of large droplets. This work also utilized different nozzle diameters (450–800 μm), but no effect on particle size was reported. On the other hand, Y. Song et al. [ 102 ] determined that the size of the particles could be reduced by using nozzles with a smaller diameter after comparing three different diameters (40, 170 and 320 μm). The consensus seems to be working at a low feed flow rate, but process productivity must not be compromised to be cost-effective. The studies shown in Table 2 reported EE values for the nano/microcapsules loaded with protein-based bioactives ranging from 20% to 92%. The lowest EE was obtained for encapsulation of SA using PEG and PCL/PLGA as carriers [ 92 ]. Since higher EE has been linked to larger particles and lower protein loading [ 106 ], these parameters were studied. Indeed, the authors confirmed that increased particle diameter corresponded with higher EE values, with a critical size allowing optimum encapsulation. In the same study, lower protein loading also resulted in improved EE, but the extraction method used to measure EE presented limitations due to protein aggregation and the lack of use of surfactants. The highest EE was achieved by encapsulating BSA by emulsion electrospraying using PLGA as encapsulating agent [ 99 ]. This work showed that increasing the aqueous phase volume ratio (e.g., increasing bioactive load) resulted in decreased EE, varying from 92% at 5 µL/mL to 80% at 100 µL/mL. It was explained as being due to increased density of emulsion droplets in the feed solution, and thus increasing migration of the aqueous phase containing the BSA to the surface of the particle. 4.3. Encapsulation by Coaxial Electrospraying The literature found regarding the encapsulation by coaxial electrospraying of protein-based bioactives was focused exclusively on the pharmacological/medical field. No works on the encapsulation by coaxial electrospraying of bioactive protein hydrolysates or peptides have been reported in the literature. Only four works studying the coaxial electrospraying of proteins were found ( Table 3 ). None of them used carrier in the formulation of the inner solution (core). For the outer solution (shell) PLGA was the most used encapsulating agent, appearing in two studies. This follows the trend established in monoaxial electrospraying, since, as previously mentioned, all the literature found was mainly focused on oral drug delivery, where PLGA was the most frequently used biopolymer. One study focused on the encapsulation of a water solution of BSA using an outer solution of PLGA dissolved in either DCM or a combination of DCM and DMF [ 107 ]. The other work encapsulated ranibizumab, a protein drug used for the treatment of age-related macular degeneration, using PLGA dissolved in a combination of DCM and acetonitrile as the outer solution [ 108 ]. Regarding the use of solvents, six of eight works used organic solvents, mainly for the outer feed. This is because the use of two immiscible solutions provides better core–shell separation by minimizing interdiffusion between layers [ 109 ]. A solution of ethyl acetate and n-butanol, along with acetylated dextran as carrier, was used as the outer feed for the encapsulation of anthrax protective antigens dissolved in the inner water solution [ 110 ]. Rasekh et al. [ 94 ] coaxially electrosprayed angiotensin II using NOSC as carrier for the inner solution and tristearin dissolved in DCM as outer solution. Since the literature found was focused on the production of oral delivered drugs, it would be necessary to take into consideration the need to apply two completely immiscible food-grade solvents to produce encapsulates oriented for food fortification. Voltages applied ranged from 5 to 22.5 kV, similar to the values used for monoaxial electrospraying (2.67–20 kV). The effect of voltage was studied for the encapsulation of angiotensin II using tristearin and NOSC as carriers inner and outer carriers, respectively [ 94 ]. The applied voltage values were 20 and 30 kV, and the authors compared the stability of the enzyme using an ELISA, finding that at 30 kV the concentration of angiotensin II in the microparticles was reduced by approximately 20%. For the encapsulation of alkaline phosphatase with PEO as outer carrier [ 95 ], the voltage was optimized to 22.5 kV. Similarly, these authors found that this high voltage resulted in a loss of activity of the enzyme up to 40% compared to the activity obtained by monoaxial electrospraying at 15.5 kV. Other parameters affecting particle characteristics are feed flow rates (inner and outer) and nozzle diameters. For the inner solutions (core), feed flow rates of 0.02–3.6 mL/h were used, while for the outer solutions (shell), 0.1–18 mL/h was used. Regarding the nozzle diameters, they ranged from 184 to 1000 μm for inner capillary and 603 to 2000 μm for the outer capillary. As previously mentioned in the previous section, increasing feed flow rate and nozzle diameters typically results in larger particles. This agreed with the data obtained by Zhao et al. [ 112 ], where alkaline phosphatase was encapsulated using CMC as inner carrier and alginate and PEGDA as outer carriers. They reported the highest feed flow rates (1.8 mL/h for the core and 3.96 mL/h for the shell) in the literature and obtained the largest particles at 440 μm. However, the opposite conclusion was obtained after comparing the encapsulation of angiotensin II (using NOSC as inner carrier and tristearin outer carriers) [ 94 ] and the encapsulation of alkaline phosphatase with PEO as carrier [ 95 ]. Both studies used similar nozzle diameters (1000 μm (inner)–2000 μm (outer), and 900 μm (inner)–1900 μm (outer), respectively), but the first study used feed flow rates 10 times higher. Even though larger particles would be expected for the angiotensin II encapsulation, due to the higher flow rates, their size was up to 86% smaller. In fact, they obtained the smallest particles, which could be due to the nozzle–collector distance, the highest reported in the literature at 20 cm, and the slightly higher voltage used. Coaxial electrospraying of bovine hemoglobin also resulted in small particles of 0.37 μm, as it was particularly important to obtain nano/microcapsules in the range of 0.1 to 3 μm to effectively avoid extravasation through the blood vessel wall and act as oxygen carriers [ 111 ]. High EE values were obtained for all the studies reported in the literature ( Table 3 ), ranging from 70% to 99%. These values are higher than the ones obtained for monoaxial electrospraying, where four of the nine reported EE values were under 50%. Zamani et al. [ 107 ] reported ranges of EE from 46.7 ± 4.3% to 74.6 ± 2.9%, which were linked to incomplete encapsulation due to inner feed flow rates being too high as well as high concentrations of BSA in the core. The highest EE found was obtained for the encapsulation of alkaline phosphatase with PEO as outer carrier [ 95 ]. They also compared the effect of monoaxial and coaxial electrospraying, confirming that the EE was increased in core–shell structures. Although coaxial electrospraying has exhibited promising outcomes, the encapsulation of bioactive protein hydrolysates or peptides has only been minimally investigated. Thus, further studies are required to fully evaluate the feasibility of this technology for the encapsulation of bioactive peptides. Particularly, there is a need to investigate the use of food-grade solvents and to optimize processing conditions that lead to encapsulates with potential use in food fortification. 5. Activity Retention and Release of the Encapsulated Protein-Based Bioactives Spray-drying and electrospraying techniques are viable encapsulation methods of protein-based bioactives. After encapsulation, it is fundamental that these bioactive compounds maintain their original activity. Moreover, they should remain active until reaching the target organ, where they will exert their activity [ 114 ]. However, research studying the factors that affect the preservation of activity and release of encapsulated protein-based bioactives is scarce. Maintaining activity after encapsulation is one of the most important challenges facing protein delivery, as this can be limited by protein aggregation or denaturation in the particles [ 115 ]. Regarding the works in the literature studying the retention of activity, it is striking that 7 out of 12 deal with antioxidant activity (i.e., DPPH free radical scavenging activity, ferric reducing antioxidant power, metal-chelating activity). This trend might be due to the increasing focus on the use of natural antioxidants as both bioactives in humans and functional ingredients in food products to avoid lipid oxidation [ 116 , 117 , 118 ]. Fish collagen hydrolysates exhibiting 2,2-diphenyl-1-picrylhydrazyl (DPPH)-inhibitory activity were spray-dried using MD as carrier at different ratios. Antioxidant activity was lost when MD was used as carrier, with a decrease in activity from 78.36 ± 12.29% when free collagen was spray-dried to 33.59 ± 6.47% when 80% of MD with 10–12 dextrose equivalent was used [ 34 ]. Other studies spray-dried fish hydrolysates with MD and GA, reporting that the presence of carbohydrates from the carriers decreased antioxidant activity [ 52 ]. High retention of activity (>60%) was found by two studies analyzing the antioxidant activity of spray-dried flaxseed protein hydrolysates with MD by different methods (i.e., DPPH free radical scavenging, ABTS free radical scavenging, hydroxyl radical scavenging, reducing power, nitric oxide scavenging) [ 36 , 37 ]. On the other hand, spray-drying of oyster protein hydrolysate included in emulsions, with MD as carrier and cholesterol for the oily phase, resulted in reduced free radical scavenging activity [ 41 ]. This was related to the high inlet temperature (170 °C) altering the lipid membrane and partially degrading the bioactive compound. The activity retention capacity of different encapsulated enzymes by electrospraying was also measured. Alkaline phosphatase with PEO was encapsulated both by monoaxial and coaxial electrospraying, retaining ~100% activity when monoaxially electrosprayed. However, when coaxial electrospraying was applied, only 60% of the activity was retained, which was attributed to the high voltage (22.5 kV) applied compared with monoaxial spraying (15.5 kV) [ 95 ]. Amylase included in emulsions using dextran and sodium alginate as carriers was encapsulated by electrospraying; however, the high activity observed for amylase (96%) was related to a low EE, indicating that the substrate had better access to the enzyme due to the location of the enzyme at the surface of the particle [ 102 ]. In addition to the efficient encapsulation of protein-based bioactives, it is essential for the application of these methods that these compounds can be released from the carrier matrix while maintaining their activity. Despite the importance of studying these parameters, literature on release kinetics or tailored release approaches for the encapsulation of protein-based bioactives is limited. In general, the release of nano/microencapsulated bioactive compounds occurs in three stages: i) surface release, which may be caused by inadequate entrapment (low EE) in the carrier matrix, ii) diffusion via swelling of the carrier matrix, and iii) erosion of the carrier matrix [ 119 ]. It is usually desired to avoid burst release as much as possible. For that, it is necessary to take into consideration the characteristics of the carrier, which should not totally solubilize in the release medium, not interact with protein-based bioactives, and be protective against external factors, such as acid pH during gastric digestion [ 120 ]. The most common method to measure protein release from the delivery systems found in the literature is by agitation in PBS release buffer. Bock et al. [ 92 ] studied the release during 81 days of SA electrosprayed using PEG as carrier. The initial burst release was dependent on the protein load, PEG load and PEG molecular weight (MW). Higher SA load resulted in a strong burst release, up to 60% in the first 24 h. A burst-free release of the SA with sustained release up to 84 days was achieved combining low protein loading (1%) and low MW PEG (6 kDA). These results agreed with the data obtained for the electrospraying of BSA emulsions using PLGA as the carrier [ 99 ]. In order to compare the release profile obtained by coaxial electrospraying versus emulsion monoaxial electrospraying, Zamani et al. encapsulated BSA using PLGA as carrier [ 107 ], finding that although coaxial electrospraying enhanced the EE, it also resulted in a stronger burst release (24–27%) than emulsion electrospraying (8–12%). This was due to the centralized distribution of the bioactive in the coaxially produced encapsulates allowing fast release once the solvent reached the bioactive, contrary to the compartmentalized distribution of the bioactive in the encapsulates produced by emulsion electrospraying ( Figure 4 ). Although this method does not take into consideration the gastrointestinal conditions to which the capsules would be subjected during digestion, it allows us to estimate the stability of the capsules. 6. Bioaccessibility of Encapsulated Protein-Based Bioactives and Enrichment of Food Matrices Bioactive peptides and proteins must be bioaccessible for their potential use in oral delivery systems for the development of supplements and functional foods. Bioaccessibility refers to the fraction of the biocomponent that, after digestion, becomes accessible for absorption through the epithelial layer of the gastrointestinal tract [ 121 ]. Bioaccessibility can be determined with good results using in vitro methods that simulate the biochemical and mechanical conditions of the gastrointestinal digestion. The effect of gastrointestinal digestion of spray-dried nanoliposomes of red tilapia viscera hydrolysates over antioxidant and ACE-inhibitory activities were studied [ 43 ]. Gastric digestion with pepsin induced degradation of the peptides, resulting in loss of antioxidant activity, while increasing ACE-inhibitory activity. For both bioactivities, the intestinal digestion resulted in enhanced inhibition, up to 10% compared to the original hydrolysates. This was attributed to the release of new oligopeptides after digestion with pancreatin, as bile salts in the intestinal phase promoted swelling and disruption of the vesicles, thus leading to leakage of bioactive peptides. Spray-dried peptides derived from spent brewer's grain were encapsulated with locust bean gum, P. columbina phycocolloids, or DM as carriers, and their bioaccessibility was measured by analysis of the size distribution of the peptides and the retention of their ACE-inhibitory activity [ 48 ]. They found that carrier selection affected greatly to the protection of the peptides during digestion, achieving highest protection when P. columbine phycocolloids were used as wall material. Regarding ACE-inhibitory activity, encapsulated peptides showed higher activity than free peptides after digestion, demonstrating that partial protection of bioactive peptides against digestive enzymes was possible. Similar results were obtained by spray-drying P. lunatus hydrolysates encapsulated by spray-drying with MD and GA as carriers [ 42 ], where the ACE-inhibitory activity, as well the DPP-IV and α-amylase activity, of free hydrolysates was severely affected by simulated gastrointestinal digestion (IC50 ≈ 300 µg/mL) compared to the retained activity of encapsulated hydrolysates (IC50 < 200 µg/mL). Coaxial electrospraying of alkaline phosphatase, with CMC as core carrier and a mixture of alginate and poly(ethylene glycol) diacrylate (PEGDA) as a carrier in the shell, was able to protect the enzyme from potential degradation during simulated gastrointestinal digestion [ 112 ]. Gómez-Mascaraque et al. [ 49 ] evaluated the changes in the profile of peptides obtained from whey protein hydrolysate, which was spray-dried using gelatin and CS as encapsulating agents, after digestion. After comparing the chromatograms of the original hydrolysate and the digested capsules, they found that simulated digestion of the free hydrolysate resulted in a "remarkable change" in the identified peptides, compromising bioaccessibility, whereas digestion of CS microcapsules retained the highest number of identified peptides. The higher digestion of the hydrolysate-loaded gelatin capsules was linked to the proteinaceous origin of the carrier, possibly also digested during the assay. Another work studied the release kinetics of bioactive compounds during and after in vitro digestion by measuring changes in absorbance of spray-dried rapeseed peptides [ 57 ]. This study found that microparticles did not produce an initial burst during gastric digestion, but rather a slow release of encapsulated peptides during the intestinal stage, which would increase bioaccessibility. Paz-Samaniego et al. [ 113 ] performed a simulation of gastrointestinal digestion with a complex Simulator of the Human Gastrointestinal Tract (Simgi). Insulin-loaded microcapsules obtained by coaxial electrospraying with maize bran arabinoxylans (core) and maize wastewater arabinoxylans (shell) as encapsulating agents were passed through five different reactors, simulating the stomach, small intestine, and the three regions of the colon: ascending, transverse, and descending. In this way, they found that 76% of the encapsulated insulin reached the colon without being degraded in the stomach. In addition to potential degradation during gastrointestinal digestion, protein-based biocompounds can easily react with complex food matrices, leading to alterations in their bioactivity. These matrices, such as soups or baked goods, can undergo thermal and high-pressure conditions that would make it difficult to maintain peptide stability [ 122 ]. Thus, encapsulation of bioactive peptides and proteins could potentially benefit their stability when incorporated into food matrices. Only three studies regarding the inclusion in food matrices of protein-based bioactives encapsulated by spray-drying or electrospraying were found in the literature. Yogurt was fortified with spray-dried weakfish hydrolysates exerting antioxidant and ACE-inhibitory activities using MD as carrier [ 48 ]. After a week of storage, not only were antioxidant and ACE-inhibitory activities maintained, but greater rheological stability was provided by the encapsulated hydrolysates. Whey protein hydrolysate-loaded gelatin or CS capsules were used to enrich yogurt produced by lactic acid fermentation [ 49 ]. During the fermentation process, a large part of the peptides present in the hydrolysate was lost due to susceptibility to the living starter cultures. After fermentation, the peptide profile of the yogurt enriched with free hydrolysate, encapsulation with CS or encapsulation with gelatin was analyzed. Enrichment with free hydrolysates resulted in the protection of 30 of the 58 peptides initially identified. The same amount was protected in hydrolysate-loaded CS capsules, but five different peptides were found. When hydrolysates were encapsulated with gelatin, only 21 peptides were protected; however, it is difficult to determine with certainty the protective effect of gelatin, since its proteinaceous nature resulted in very complex chromatograms that did not allow conclusive conclusions. Spray-dried hydrolysates from pink perch meat were used to fortify a sweet-corn vegetable soup [ 52 ]. Both particles with and without encapsulating agents (MD and GA) were used, and although higher activity was retained when no carrier was used, their use improved sensory acceptability. The activity loss on MD/GA hydrolysate particles was linked to interaction between the encapsulating agents and the hydrolysates. 7. Conclusions and Future Perspectives Bioactive peptides and protein hydrolysates are interesting ingredients for the production of functional foods and nutraceuticals due to their high bioactive potential and nutritional value. However, their physicochemical properties (e.g., bitter taste) and potential degradation during digestion have been shown to hinder their use. This work addresses the application of encapsulation technologies such as spray-drying and electrospraying for encapsulation, protection, and release of bioactive peptides and protein hydrolysates. For each encapsulation technology, both monoaxial and coaxial configurations were considered. Various parameters that may affect particle morphology and encapsulation efficiency (e.g., formulation processing method or carrier) were investigated, as well as the specific process parameters for both technologies (e.g., inlet and outlet temperature of the drying air, electrical potential, feed flow, injector–collector distance). It is worth noting that while there has been a significant effort to produce new bioactive peptide sequences, there has been a lack of attention towards their stabilization. Additionally, research on encapsulating protein-based bioactives using monoaxial spray-drying is scarce, and practically nonexistent for coaxial spray-drying. Although some research has been conducted on the encapsulation of bioactive peptides by both monoaxial and coaxial electrospraying, it has mainly been focused on pharmaceutical applications and parenteral supplementation. As a result, there is a significant gap in research regarding food application and oral supplementation. Altogether, encapsulation has an important role in maintaining the efficacy of functional foods containing bioactive peptides and proteins. However, the current widely used spray-drying method has only been studied in monoaxial configuration and the potential of the coaxial mode remains to be investigated. Although advanced methods such as electrospraying encapsulation have shown promise in pharmaceutical development, there is a need to explore their applications in the food industry. As research in this field continues, we can expect to see advances in the nano/microencapsulation of bioactive peptides for their application in functional foods. Finally, works focusing on release studies, retention of activity, and bioaccessibility are limited, especially regarding the study of monoaxial electrospraying. Overall, there is a lack of research concerning the bioaccessibility of protein-based bioactives encapsulated by spray-drying or electrospraying, as well as their stability in food matrices. To the authors' knowledge, no work has been directed towards the bioaccessibility analysis of monoaxial electrospraying of protein-based bioactives, nor has any study yet compared the effect of monoaxial versus coaxial electrospraying of these biocompounds in their bioaccessibility or use for food matrix enrichment. Hence, future research should prioritize two aspects for activity retention of bioactive peptides: (i) exploring the possibilities of coaxial encapsulation techniques that can improve the entrapment of bioactive peptides, and (ii) investigating the effects of nano/microencapsulation on the stability and release of peptides in the gastrointestinal environment to enhance bioavailability.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6582207/

Programmed necrotic cell death of macrophages: Focus on pyroptosis, necroptosis, and parthanatos

Macrophages are highly plastic cells of the innate immune system. Macrophages play central roles in immunity against microbes and contribute to a wide array of pathologies. The processes of macrophage activation and their functions have attracted considerable attention from life scientists. Although macrophages are highly resistant to many toxic stimuli, including oxidative stress, macrophage death has been reported in certain diseases, such as viral infections, tuberculosis, atherosclerotic plaque development, inflammation, and sepsis. While most studies on macrophage death focused on apoptosis, a significant body of data indicates that programmed necrotic cell death forms may be equally important modes of macrophage death. Three such regulated necrotic cell death modalities in macrophages contribute to different pathologies, including necroptosis, pyroptosis, and parthanatos. Various reactive oxygen and nitrogen species, such as superoxide, hydrogen peroxide, and peroxynitrite have been shown to act as triggers, mediators, or modulators in regulated necrotic cell death pathways. Here we discuss recent advances in necroptosis, pyroptosis, and parthanatos, with a strong focus on the role of redox homeostasis in the regulation of these events. 1 Introduction Macrophages (MΦs) are phagocytic cells of the innate immune system that play central roles in tissue homeostasis and response to pathogenic stimuli. Circulating monocytes can enter the tissues and differentiate into MΦs in response to CSF1 and GM-CSF. On the other hand, tissue resident MΦs (e.g. Kupffer cells in the liver and brain microglia) have been shown to arise from yolk sac-derived (and not bone marrow-derived) primitive MΦs and settle in organs prenatally [ 1 , 2 ] or may even transdifferentiate from smooth muscle cells, as demonstrated in the wall of atherosclerotic arteries [ 3 ]. In tissues, MΦs sample the environment and respond to the pathogen- or damage-associated molecular patterns (PAMPs and DAMPs, respectively), toxins, cytokines, and chemokines. Depending on the composition of the sampled environment, MΦs differentiate into a diverse set of phenotypes, which are usually characterized by two extremes of the MΦ differentiation spectrum, M1 (inflammatory) and M2 (resolutory). However, MΦs are characterized by high plasticity and their phenotype can easily change in response to alterations of the microenvironment. Classically activated (M1-like) MΦs produce large amounts of reactive oxygen and nitrogen species (ROS and RNS, respectively), mostly via activation of NADPH Oxidase 2 (NOX2) and inducible nitric oxide synthase (iNOS). These reactive species are effective tools to combat pathogens but also contribute to organ damage in inflamed tissues. MΦs also succumb to this ROS and RNS-filled environment and undergo cell death. While research on programmed cell death initially focused on apoptosis, in recent decades several novel forms of programmed necrotic cell death (necroptosis, pyroptosis, parthanatos, oxytosis, ferroptosis, NETosis) have attracted increased attention [ 4 ]. Many of these modalities are relevant to MΦs. Here, we will discuss the mechanisms and pathological consequences of programmed necrotic cell death of MΦs focusing on pyroptosis, necroptosis, and parthanatos. 2 Pyroptosis The term pyroptosis (from Greek "pyro,"- fire or fever, and "ptosis" – falling) was initially proposed by Cookson and Brennan [ 5 ] for a novel form of inflammatory programmed cell death. The first observations of pyroptosis were made with the invasive pathogenic bacteria Shigella and Salmonella , which triggered lytic cell death by activating caspase-1 in MΦs via the secreted effector proteins, SipB and IpaB, respectively [ 6 , 7 ]. Initially, this process was incorrectly classified as apoptosis, but was subsequently recognized as a form of programmed cell death different from apoptosis. Pyroptosis is now recognized as its own entity of programmed cell death. Initially, pyroptosis was described as being dependent on caspase-1 and gasdermin D (GSDMD), discriminating it on the mechanistic level from other forms of cell death, such as apoptosis, which depends on caspase-8, and necroptosis, which depends on RIP3 and mixed lineage kinase domain-like (MLKL) [ 8 ]. The prototypical form of pyroptosis is triggered by activation of proinflammatory caspases (caspase-1,-4,-5 in man and caspase-1 and -11 in mice). The terminal cell lysis is then mediated by cleavage of GSDMD by one of these caspases [ 9 , 10 ]. The active form of GSDMD, which consists of an N-terminal domain, can assemble to form pores in the cell membrane [ 11 , 12 ]. These pores lead to collapse of the membrane, potentially initiating death of the cell, but at the same time leading to the release of inflammatory mediators, including IL-1β and IL-18 [ 9 , 10 , 13 ]. Compared to apoptosis, where the cell content of the dying cell is sealed in apoptotic bodies, this form of cell death is inflammatory and the highly inflammatory cytokines of the IL-1 family and cellular danger signals (damage-associated molecular pattern - DAMP) are released. Besides GSDMD, other members of the gasdermin family participate in pyroptotic cell death. GSDME (DFNA5) cleavage by caspase-3 can induce pyroptosis in certain cancer cells upon chemotherapy [ 14 ], a process termed non-canonical pyroptosis. GSDMB is highly expressed in septic shock and might contribute to GSDMD processing [ 15 ]. Recently, it was shown that caspase-8 also cleaves GSDMD, leading to caspase-8-mediated GSDMD-dependent cell death in response to extrinsic triggers of apoptosis [ 16 ]. This form of pyroptosis is induced by inhibition of pro-survival signals, mainly by pharmaceutical or bacterial targeting of TAK1 kinase [ [17] , [18] , [19] ]. These examples illustrate that different forms of pyroptotic cell death exist and that these are interconnected with apoptotic and necroptotic pathways. The main function of the classical pyroptosis pathway is thought to be defense against infection [ 12 ]. Pathogens and pathogen-derived toxins are sensed by cytosolic pattern-recognition receptors (PRRs). Pyrin domain (PYD)-containing members of the Nod-like receptor (NLR) family and AIM2 are the most thoroughly studied members of the cytosolic PRRs. Upon activation, PYD-containing NLR proteins and AIM2 form a multimeric high-molecular weight complex in the cell, containing the adaptor protein apoptosis associated speck-like protein containing a CARD (ASC). These complexes, named "inflammasomes" by Jürg Tschopp, are scaffolds for the activation of pro-caspase-1, which is induced by proximity induced auto-processing [ 20 ]. ASC consists of a PYD and a CARD domain. The PYD domain can form homo-oligomeric interactions with the PYD from NLR or AIM2 and the CARD forms homo-oligomeric interactions with the CARD of caspase-1. In addition to the recruitment of effector cells by the released cytokines, one effector mechanism induced by the pyroptotic program is the expulsion of invaded bacterial pathogens from the affected cell in small membrane-enclosed structures. These pore-induced intracellular traps (PIT) can induce efferocytosis of the engulfed material by macrophages and neutrophils [ 21 ]. Currently, five PRRs are described as sensor proteins that induce pyroptosis (see Table 1 ). PRRs, including AIM2, NAIP/NLRC4 oligomer, NLRP3, Pyrin (TRIM20), and NLRP1, sense a variety of structurally different PAMPs. AIM2 reacts to double-stranded DNA in the cytosol [ 22 ], NAIP/NLRC4 senses bacterial type III secretion apparatus proteins and flagellin [ 23 ], NLRP3 is activated by different kinds of membrane damage [ 24 ], Pyrin is activated by bacterial modification of host proteins [ 25 ], and NLRP1 senses anthrax lethal toxin and Toxoplasma [ 26 ]. All of these pathways induce the activation of caspase-1, which ultimately processes GSDMD ( Fig. 1 ). An alternative pathway to activate pyroptosis is triggered by activation of caspase-11/4/5. This non-canonical inflammasome is activated by cytosolic lipopolysaccharide (LPS) [ 27 , 28 ]. Caspase-11 in mice and caspase-4 and -5 in man bind to LPS, leading to activation of these caspases [ 29 ]. Although caspase-11/4/5 can cleave and activate GSDM in human myeloid cells to induce pyroptosis, IL-1β production upon cytosolic LPS sensing by this pathway depends on NLRP3 activation [ 30 ]. GSDMD cleavage is needed to induce activation of NLRP3 [ 9 ]. The mechanism of NLRP3 activation is poorly understood; however, GSDMD and GSDME both target mitochondrial membranes and stimulate the release of ROS, which can trigger NLRP3 activation [ 31 , 32 ]. Table 1 Pyroptosis inducers (see text for details and references). Table 1 Inducer PRR Pathway cytosolic dsDNA AIM2 Canonical pyroptosis bacterial type III SA, flagellin NAIP/NLRC4 Canonical pyroptosis membrane damage NLRP3 Canonical pyroptosis Bacterial modification of GTPases Pyrin Canonical pyroptosis Anthrax lethal toxin, Toxoplasma ssp. NLRP1 Canonical pyroptosis Cytosolic LPS Caspase4,5/11 Canonical pyroptosis Chemotherapy drugs GSDME Non-canonical, caspase-3, GSDME Targeting of TAK1 Caspase-8-mediated GSDMD cleavage Fig. 1 Schematic representation of the pyroptosis pathways in a mammalian cell. See text for details. Fig. 1 Release of the potent inflammatory cytokine, IL-1β, is important in controlling infection. However, IL-1β has detrimental effects in sepsis, a life-threatening organ dysfunction caused by an overwhelming cytokine response towards bacterial pathogens. Neutrophils are emerging as important players in this condition and targeting the modulation of pyroptosis in neutrophils might be a viable treatment option [ 33 ]. 3 Redox control of pyroptosis As eluded to above, the pyroptosome triggering inflammasomes can be activated by microbial substances. In the case of the NLRP3 inflammasome, there is good evidence that its activity is also controlled by the redox state of the cell (for an excellent overview the reader is referred to Ref. [ 34 ]). The presence of ROS, produced by the MΦs upon microbial insult, was shown to contribute to NLRP3 activation by the redox sensor thioredoxin-interacting protein (TXNIP) (reviewed in Ref. [ 35 ]). However, the contribution of TXNIP to NLRP3 activation is controversial, as TXNIP knockout mice were reported to have no defects in IL-1β production [ 36 ]. NADPH oxidase, the enzyme that produces ROS for an oxidative burst in MΦs, is also not needed for this activation, as both NADPH oxidase knock out mice and chronic granulomatosis patients show normal IL-1β production [ [37] , [38] , [39] ]. In contrast, several reports suggest a role for mitochondrial-derived ROS (mitoROS) [ 40 ]. In a model of Shiga toxin and LPS-induced cell activation, mitoROS plays a critical role in IL-1β release and pyroptosis, mediated by both NLRP3 and GSDMD [ 31 ]. The master transcriptional regulator of redox homeostasis nuclear factor E2-related factor 2 (Nrf2) further contributes to NLRP3 activation and IL-1β secretion is inhibited by Nrf2 silencing [ 41 , 42 ]. The functional mechanism remains elusive, but is likely to be indirect, as no contribution of Nrf2 to inflammasome complex formation has yet been identified. The redox status can directly affect the activity of initiator caspases. Caspase-1 can be regulated by superoxide via reversible oxidation and glutathionylation of redox-sensitive cysteine residues. Accordingly, depletion of superoxide dismutase 1 (SOD1) leads to an oxygen-dependent reduction of caspase-1 activation [ 43 ]. For caspase-11, extracellular ROS can induce its expression and activation, which involves JNK activation [ 44 ]. Our knowledge about the fine-tuning of the final steps of pyroptosis is still very fragmentary, but redox status seems to contribute here as well. ROS have recently been shown to oxidize GSDM, which enhances GSDM cleavage by caspase-1 [ 45 ]. 4 Necroptosis Caspases are the executioner proteins of both apoptosis and pyroptosis. In contrast, necroptosis is a caspase-independent necrotic cell death program regulated by receptor-interacting protein (RIP) kinases. Necroptosis was initially discovered when cells, stimulated with FasL, tumor necrosis factor (TNF), or a TNF ligand, were additionally treated with the pan-caspase inhibitor, Z-VAD-FMK. Necrosis is an unprogrammed cell death and occurs due to an irreversible injury to the cell. In contrast, necroptosis is programmed and regulated by receptor-interacting protein kinase 1 (RIPK1) and receptor-interacting protein kinase 3 (RIPK3) [ 46 ]. Most of the current knowledge about necroptosis is primarily derived from investigating tumor necrosis factor (TNF) signaling. Engagement of TNF with its cognizant receptor results in the formation of complex I at the cell membrane. Complex I is composed of tumor necrosis factor receptor (TNFR)-associated death domain (TRADD), Fas-associated death domain (FADD), RIPK1, TNFR-associated factors (TRAF), and cellular inhibitor of apoptosis protein 1 (cIAP1) and cIAP2. TRAF proteins ubiquinate and stabilize RIP1 at the plasma membrane, leading to the activation of nuclear factor kappa B (NFκB) and cell survival [ 47 ]. Activation of necroptosis is initiated by the deubiquitination of RIPK1 by cylindromatosis protein (CYLD), which dislodges RIPK1 from complex I and forms a complex II with FADD, TRADD, and caspase-8. Active caspase-8 can cleave RIPK1 and RIPK3; however, inhibition of caspase-8 facilitates the interaction of RIPK1 and RIPK3 through their RIP homotypic interaction motives (RHIM) [ 48 ]. The RIP complex induces the phosphorylation of the pseudokinase mixed-lineage kinase domain-like protein (MLKL) [ 49 , 50 ]. Phosphorylation of MLKL exposes the amino acid-terminal 4-helical bundle domain, which forms a pore in the cell membrane by interacting with negatively charged phospholipids. Pore formation in the cell membrane ultimately leads to cell death [ 51 ]. The pathway leading to necroptosis has been largely deciphered through the discovery of small molecule inhibitors that block RIPK1 (necrostatin-1) [ 52 ], RIPK3 (GSK′872) [ 53 ], and MLKL (necrosulfonamide) [ 49 ]. 4.1 Necroptosis in MΦs Necroptosis in MΦs was first reported in the second mitochondria-derived activator of caspase (SMAC) mimetics mediated inhibition of cellular inhibitor of apoptosis proteins (cIAPs) and the expression of the caspase inhibitor XIAP in TNF-α stimulated bone marrow-derived MΦs (BMDMs) [ 54 ]. Consistently, inhibition of proteasomes in MΦs, using PS-341, suppresses the degradation of cIAPs, and, thus, attenuates necroptosis [ 55 ]. More recent reports suggest that the expression of pro-inflammatory cytokines is elevated, while RIPK1-dependent cell death is reduced, during the differentiation of MΦs and RIPK3-caspase-8 is important in the differentiation of MΦs. Resistance to cell death in differentiated MΦs is mediated by the p38/MK2 pathway [ 56 ]. Necroptosis plays a crucial role in pathophysiology. Unlike apoptosis, which is immunologically quiescent, necroptosis results in the release of cytoplasmic contents, which can activate MΦs during infection and other sterile inflammatory conditions. Atherogenic ligands stimulate the expression of the necroptotic genes, RIPK3 and MLKL. Thus, necroptosis occurs in MΦs associated with human atherosclerotic plaques, which becomes the driver of necrotic core formation in atherosclerosis [ 57 ]. Plant sterols, such as sitosterol, promote atherosclerosis by inducing necroptosis in MΦs [ 58 ]. Heme released during hemolysis is also known to induce necroptosis through ROS and TNF production due to the activation of TLR [ 59 ]. On the other hand, bacterial pathogens target MΦs and mitigate host defense mechanisms by eliminating MΦs via necroptosis. Excess TNF induces RIPK1-RIPK3-dependent mitochondrial ROS in Mycobacterium tuberculosis infected MΦs. The authors propose that induction of necroptosis is through the modulation of cyclophilin D that regulates mitochondrial membrane permeability pore formation and ceramide synthesis [ 60 ]. A similar phenomenon has also been observed in ischemia-associated oxidative damage wherein p53 associates with cyclophilin D and opens the mitochondrial permeability transition pore resulting in necrotic cell death [ 61 ]. Mycobacterium tuberculosis also secretes tuberculosis necrotizing toxin (TNT), a nicotinamide adenine dinucleotide (NAD + ) glycohydrolase that induces necroptosis in infected MΦs. Interestingly, depletion of NAD + is sufficient to induce necroptosis in MΦs [ 62 ]. Loss of NAD + may also result from PARP activation, which suggests that necroptosis and parthanatos, which is reviewed in the next section of the paper, may be linked. Interestingly in other cellular models, administration of NAD + decreases oxidative stress induced by H 2 O 2 and protects cells from necrosis [ 63 ]. Other bacterial pathogens are also known to produce toxins that induce necroptosis in MΦs. For instance, pathogens such as Serratia marcescens, Listeria monocytogenes, Staphylococcus aureus, Streptococcus pneumoniae, and uropathogenic Escherichia coli (UPEC) produce pore-forming toxins that trigger necroptosis and disrupt the cell membrane, damage mitochondria, decrease ATP, and increase ROS-generation [ 64 , 65 ]. These findings demonstrate that the necroptotic death of MΦs is a major cause of lung pathology in pneumonia. A strain of Streptococcus pneumoniae (TIGR4) is able to invade the heart and cause cardiac damage. Authors have demonstrated that the damage is critically associated with the necroptotic death of MΦs induced by pneumolysin, the toxin produced by the bacteria [ 66 ]. Similarly, Yersinia outer protein J (YopJ) of Yersinia pestis induces necroptotic death in MΦs, thus allowing the lymphatic spread of the pathogen [ 67 ]. Although most studies describe TNFα as the primary factor regulating necroptosis, TNFα-independent necroptosis upon the activation of TLRs in MΦs has also been reported [ 68 ]. Our findings demonstrate that Salmonella enterica ssp. enterica ser. Typhimurium ( S. Typhimurium) promotion of necroptotic death in MΦs is dependent on type I interferon (IFN–I) signaling. IFN-induced inflammatory pathology is predominantly due to necroptosis. Mice lacking the cognate receptor for IFN-I (IFNAR) showed reduced bacterial burden and pathology associated with the infection [ 69 ]. IFN–I drives necroptosis through IFN-stimulated gene factor 3 (ISGF3) signaling, which leads to persistent expression of STAT1, STAT2, and IRF9. Strikingly, MLKL is one of the interferon-stimulated genes (ISGs) [ 70 ]. Constitutive interferon signaling, such as in autoimmune diseases, primes MΦs to undergo necroptosis by maintaining adequate levels of MLKL [ 71 ]. Oxidative stress is a common feature in macrophage necroptosis induced by pathogens or under sterile inflammatory conditions. Bacterial toxin-stimulated necroptosis is prevented using Coenzyme Q10 in combination with a RIPK1 inhibitor. Antioxidants have also been shown to ameliorate heme-induced necroptosis. Moreover, heme oxygenase 1 (HO-1) reduces oxidative stress and, thus, provides cytoprotection [ 59 ]. The transcription factor, Nrf2, transcriptionally regulates HO-1 and other cytoprotective genes by binding to cis-acting antioxidant responsive elements (ARE). Thus, the Nrf2-regulated antioxidative response inhibits heme-induced cell death [ 72 ]. More recently, we have provided evidence that IFN-I-regulated RIPK3 activation sequesters Nrf2 in the cytoplasm by activating PGAM5 during S. Typhimurium infection. Importantly, pharmacological activation of Nrf2, using the synthetic triterpenoid compound, CDDO (2-cyano-3,12-dioxooleana-1,9-dien-28-oic-acid), was able to prevent necroptosis [ 73 ]. Upon TNFα-induced necroptosis, PGAM5 is recruited to the RIPK1/RIPK3 complex on the outer mitochondrial membrane, where it triggers Drp1-mediated mitochondrial fragmentation, which is considered an obligatory step in necroptosis [ 74 ]. Intriguingly, RIPK3 has been shown to regulate mitochondrial metabolism by targeting the pyruvate dehydrogenase complex [ 75 ]. Mitochondrial oxidative stress associated cell death also coincides with another form of cell death known as ferroptosis, which is caused by the accumulation of lipid-based ROS. Expression of glutathione peroxidase, which repairs oxidized lipid species, is also transcriptionally driven by Nrf2. Hence, Nrf2 is known to play a critical role in mitigating ferroptosis [ 76 ]. Inhibition of PARP has also been shown to reduce ROS generation and protect mitochondria [ 77 , 82 ]. Consistently, we and others have reported that during Mycobacterium tuberculosis [ 62 ] and S. Typhimurium induced necroptosis in MΦs, NAD + is depleted [ 78 ] and PARP-1 is activated [ 69 ]. Taken together, this evidence suggests that oxidative stress plays a significant part in the execution of necroptosis ( Fig. 2 ). Fig. 2 Schematic diagram representing the convergence of necroptotic signaling and oxidative stress. Fig. 2 4.1 Necroptosis in MΦs Necroptosis in MΦs was first reported in the second mitochondria-derived activator of caspase (SMAC) mimetics mediated inhibition of cellular inhibitor of apoptosis proteins (cIAPs) and the expression of the caspase inhibitor XIAP in TNF-α stimulated bone marrow-derived MΦs (BMDMs) [ 54 ]. Consistently, inhibition of proteasomes in MΦs, using PS-341, suppresses the degradation of cIAPs, and, thus, attenuates necroptosis [ 55 ]. More recent reports suggest that the expression of pro-inflammatory cytokines is elevated, while RIPK1-dependent cell death is reduced, during the differentiation of MΦs and RIPK3-caspase-8 is important in the differentiation of MΦs. Resistance to cell death in differentiated MΦs is mediated by the p38/MK2 pathway [ 56 ]. Necroptosis plays a crucial role in pathophysiology. Unlike apoptosis, which is immunologically quiescent, necroptosis results in the release of cytoplasmic contents, which can activate MΦs during infection and other sterile inflammatory conditions. Atherogenic ligands stimulate the expression of the necroptotic genes, RIPK3 and MLKL. Thus, necroptosis occurs in MΦs associated with human atherosclerotic plaques, which becomes the driver of necrotic core formation in atherosclerosis [ 57 ]. Plant sterols, such as sitosterol, promote atherosclerosis by inducing necroptosis in MΦs [ 58 ]. Heme released during hemolysis is also known to induce necroptosis through ROS and TNF production due to the activation of TLR [ 59 ]. On the other hand, bacterial pathogens target MΦs and mitigate host defense mechanisms by eliminating MΦs via necroptosis. Excess TNF induces RIPK1-RIPK3-dependent mitochondrial ROS in Mycobacterium tuberculosis infected MΦs. The authors propose that induction of necroptosis is through the modulation of cyclophilin D that regulates mitochondrial membrane permeability pore formation and ceramide synthesis [ 60 ]. A similar phenomenon has also been observed in ischemia-associated oxidative damage wherein p53 associates with cyclophilin D and opens the mitochondrial permeability transition pore resulting in necrotic cell death [ 61 ]. Mycobacterium tuberculosis also secretes tuberculosis necrotizing toxin (TNT), a nicotinamide adenine dinucleotide (NAD + ) glycohydrolase that induces necroptosis in infected MΦs. Interestingly, depletion of NAD + is sufficient to induce necroptosis in MΦs [ 62 ]. Loss of NAD + may also result from PARP activation, which suggests that necroptosis and parthanatos, which is reviewed in the next section of the paper, may be linked. Interestingly in other cellular models, administration of NAD + decreases oxidative stress induced by H 2 O 2 and protects cells from necrosis [ 63 ]. Other bacterial pathogens are also known to produce toxins that induce necroptosis in MΦs. For instance, pathogens such as Serratia marcescens, Listeria monocytogenes, Staphylococcus aureus, Streptococcus pneumoniae, and uropathogenic Escherichia coli (UPEC) produce pore-forming toxins that trigger necroptosis and disrupt the cell membrane, damage mitochondria, decrease ATP, and increase ROS-generation [ 64 , 65 ]. These findings demonstrate that the necroptotic death of MΦs is a major cause of lung pathology in pneumonia. A strain of Streptococcus pneumoniae (TIGR4) is able to invade the heart and cause cardiac damage. Authors have demonstrated that the damage is critically associated with the necroptotic death of MΦs induced by pneumolysin, the toxin produced by the bacteria [ 66 ]. Similarly, Yersinia outer protein J (YopJ) of Yersinia pestis induces necroptotic death in MΦs, thus allowing the lymphatic spread of the pathogen [ 67 ]. Although most studies describe TNFα as the primary factor regulating necroptosis, TNFα-independent necroptosis upon the activation of TLRs in MΦs has also been reported [ 68 ]. Our findings demonstrate that Salmonella enterica ssp. enterica ser. Typhimurium ( S. Typhimurium) promotion of necroptotic death in MΦs is dependent on type I interferon (IFN–I) signaling. IFN-induced inflammatory pathology is predominantly due to necroptosis. Mice lacking the cognate receptor for IFN-I (IFNAR) showed reduced bacterial burden and pathology associated with the infection [ 69 ]. IFN–I drives necroptosis through IFN-stimulated gene factor 3 (ISGF3) signaling, which leads to persistent expression of STAT1, STAT2, and IRF9. Strikingly, MLKL is one of the interferon-stimulated genes (ISGs) [ 70 ]. Constitutive interferon signaling, such as in autoimmune diseases, primes MΦs to undergo necroptosis by maintaining adequate levels of MLKL [ 71 ]. Oxidative stress is a common feature in macrophage necroptosis induced by pathogens or under sterile inflammatory conditions. Bacterial toxin-stimulated necroptosis is prevented using Coenzyme Q10 in combination with a RIPK1 inhibitor. Antioxidants have also been shown to ameliorate heme-induced necroptosis. Moreover, heme oxygenase 1 (HO-1) reduces oxidative stress and, thus, provides cytoprotection [ 59 ]. The transcription factor, Nrf2, transcriptionally regulates HO-1 and other cytoprotective genes by binding to cis-acting antioxidant responsive elements (ARE). Thus, the Nrf2-regulated antioxidative response inhibits heme-induced cell death [ 72 ]. More recently, we have provided evidence that IFN-I-regulated RIPK3 activation sequesters Nrf2 in the cytoplasm by activating PGAM5 during S. Typhimurium infection. Importantly, pharmacological activation of Nrf2, using the synthetic triterpenoid compound, CDDO (2-cyano-3,12-dioxooleana-1,9-dien-28-oic-acid), was able to prevent necroptosis [ 73 ]. Upon TNFα-induced necroptosis, PGAM5 is recruited to the RIPK1/RIPK3 complex on the outer mitochondrial membrane, where it triggers Drp1-mediated mitochondrial fragmentation, which is considered an obligatory step in necroptosis [ 74 ]. Intriguingly, RIPK3 has been shown to regulate mitochondrial metabolism by targeting the pyruvate dehydrogenase complex [ 75 ]. Mitochondrial oxidative stress associated cell death also coincides with another form of cell death known as ferroptosis, which is caused by the accumulation of lipid-based ROS. Expression of glutathione peroxidase, which repairs oxidized lipid species, is also transcriptionally driven by Nrf2. Hence, Nrf2 is known to play a critical role in mitigating ferroptosis [ 76 ]. Inhibition of PARP has also been shown to reduce ROS generation and protect mitochondria [ 77 , 82 ]. Consistently, we and others have reported that during Mycobacterium tuberculosis [ 62 ] and S. Typhimurium induced necroptosis in MΦs, NAD + is depleted [ 78 ] and PARP-1 is activated [ 69 ]. Taken together, this evidence suggests that oxidative stress plays a significant part in the execution of necroptosis ( Fig. 2 ). Fig. 2 Schematic diagram representing the convergence of necroptotic signaling and oxidative stress. Fig. 2 5 Parthanatos Parthanatos is a relatively new addition to the growing list of established cell death forms. The term parthanatos was coined to reflect the dependence of this cell death pathway on the formation of poly (ADP-ribose) (PAR), while the second part of the name refers to Thanatos, the personification of death in Greek mythology [ 79 ]. The PAR polymer is synthesized by some poly (ADP-ribose) polymerase (PARP) enzyme family members (PARP1, PARP2 and tankyrases) and parthanatos is triggered by the DNA breakage-induced activation of the founding member of this enzyme family, PARP1. Poly (ADP-ribosyl)ation (PARylation) of proteins, including PARP1 itself (auto-PARylation), near the DNA damage site facilitates the recruitment of DNA repair effector proteins and, thus, contributes to DNA repair. While PARylation is primarily a survival mechanism, in cells experiencing excessive DNA damage, high PARylation activity can cause regulated necrotic cell death. PARP1-mediated cell suicide was first described by Nathan Berger [ 80 ] and was thought to result from the depletion of the enzyme's substrate, NAD + , and, consequently, ATP in the cells. The signaling pathway for PARP1-mediated cell death, however, proved to be more complex than a metabolic collapse. Virág et al. demonstrated that PARP1-mediated cell death displays the features of necrosis[ [81] , [82] ]. Inhibition of PARP1 (e.g. by PARP inhibitors or inactivation of PARP1 gene) diverts cells to the "default" apoptotic route [ 81 , 82 ]. Key features of parthanatos include: a) its independence from caspases [ 81 ]; b) mitochondrial membrane depolarization and secondary ROS production [ 82 ]; c) dependence on calcium signaling [ 83 ]; d) independence from the cytoprotective effect of Bcl-2 [ 84 ]; e) synergism between PARG and PARP1 in cell death regulation [ 85 ]. The central mediator of this cell death pathway is the DNA damage response protein PARP1. PARP1 activation is considered a hallmark of oxidative stress. As a DNA nick sensor enzyme, PARP1 binds to broken DNA resulting in its activation. Activated PARP1 cleaves NAD + into ADP-ribose and nicotinamide and attaches ADP-ribose to acceptor proteins near the damage site. The enzyme can add further ADP-ribose units to the protein proximal moiety to generate an (ADP-ribose) n polymer known as poly (ADP-ribose) (PAR). PAR polymers are degraded by poly (ADP-ribose) glycohydrolase (PARG) and ADP-ribosylhydrolase-3 (ARH3) enzymes [ 86 ]. 5.1 The canonical route of parthanatos The fact that PARP1-mediated cell death is now recognized as a stand-alone cell death entity is mainly due to discoveries made in the lab of Valina and Ted Dawson at Johns Hopkins University. Their lab showed that cerebral ischemic injury [ 87 ], N-methyl- d -aspartate (NMDA) excitotoxicity [ 87 ], 1-metil-4-fenil-1,2,3,6-tetrahidropiridin (MPTP)-induced Parkinsonism, and neurodegenerative diseases [ 88 , 89 ] are mediated by PARP1 activation. While characterizing the molecular events leading to and following PARP1 activation in excitotoxicity, they identified apoptosis-inducing factor (AIF) as the downstream mediator of cell death [ 90 ]. Their proposed model relies on the following key events of parthanatos: stimulation of the NMDA receptor in neurons, glutamate triggering of calcium signaling, and calcium-dependent activation of the neuronal isoform of nitric oxide synthase (nNOS). Neuronal NOS produces nitric oxide, which combines with superoxide anion radical to form peroxynitrite (ONOO − ). Peroxynitrite is highly reactive and causes, among other macromolecular damages, DNA strand breaks [ 91 ]. Once DNA breaks are formed, the pathway may show an overlap with other cell death models in which DNA damage is triggered by direct DNA damaging stimuli (e.g. treatment with exogenous peroxynitrite, hydrogen peroxide, or DNA alkylating or crosslinking agents). DNA breaks are recognized by PARP1 and the enzyme becomes activated. Active PARP1 synthesizes the PAR polymer from NAD + to mark the site of DNA damage. PARG and ARH3 enzymes cleave the polymer off of PARylated proteins and the polymers (whether protein-bound or "naked" polymers is not known) leave the nucleus and translocate to the mitochondria ( Fig. 3 ). PAR triggers the release of AIF from the mitochondrial intermembrane space and AIF begins its journey towards the nucleus [ 79 ]. The mechanism of AIF release is not fully understood. PARP1-mediated NAD + depletion may trigger mitochondrial membrane depolarization, facilitating AIF release. Alternatively, the polymer may directly interact with the C-terminus of membrane-bound AIF. As an alternative to the consumption of NAD + , suppression of glycolysis may result from PAR binding to and inhibition of hexokinase, a key regulatory control point of glycolysis. Either way, AIF leaves the mitochondria and interacts with macrophage migration inhibitory factor (MIF) in the cytoplasm. Nuclear translocation of MIF is followed by MIF-mediated DNA fragmentation, as a result of the newly discovered nuclease activity of this multifunctional protein [ 92 ]. Fig. 3 Parthanatos. Nuclear mitochondrial crosstalk in parthanatos is triggered by DNA damaging stimuli activating PARP1. PAR synthesized in response to DNA breaks travels to the mitochondria and induces liberation of AIF. In turn, AIF interacts with MIF and the latter degrades DNA. Fig. 3 5.2 Parthanatos in MΦs Zingarelli et al. [ 93 ] was the first to demonstrate that PARP1-mediated cell death is an important cell death modality in activated MΦs. [Of note, at the time of their investigation PARP1 was the only known PARylating enzyme and was referred to as poly (ADP-ribose) synthase; PARS in this study.] They showed that exposure of murine peritoneal MΦs or J774 macrophage cells to high concentrations (10 μg/ml) of LPS triggered a rapid burst of superoxide production and a slower upregulation of iNOS. As a result, the ideal conditions were created for ONOO − formation, which triggered the following sequence of events: DNA breakage - PARP1 activation – NAD + /ATP depletion – cell death pathway. Follow up in vivo studies showed that inhibition of this pathway led to reduced organ damage and improved survival in endotoxemic or septic animals [ 91 , 94 , 95 ]. TLR ligands, such as LPS, also induce production of inflammatory cytokines, with TNFα considered to be the master cytokine regulator of inflammation [ 96 ]. MΦs are not only the primary source of TNFα, but also important targets of TNFα. MΦs express both TNFα receptors (TNFR1 and TNFR2) and respond to TNFα stimulation [ 96 ]. TNFα –induced cell death has been shown to be mediated by PARP1 in Actinomycin D-pretreated L929 and ME-180 human cervical carcinoma cells, but whether or not MΦs react similarly to the toxic effect of this cytokine remains to be seen [ 97 ]. Cytotoxicity of exogenous DNA damaging agents, such as H 2 O 2 , has also been investigated in MΦs. MΦs are often exposed to reactive oxygen species and are quite resistant to H 2 O 2 , mainly due to the constitutively active PI3K-Akt pathway [ 98 ]. Nonetheless, high concentrations of H 2 O 2 cause PARP1-mediated necrosis-like cell death characterized by plasma membrane permeabilization and lack of caspase activation [ 99 ]. Interestingly, even though this cell death is PARP1 dependent, no sign of AIF translocation could be observed. Thus, this type of cell death showed some key signs of parthanatos (PARP1-dependence, necrotic phenotype, energetic collapse), while lacking other key features (e.g. AIF translocation). Similar non-canonical parthanatos could also be observed in other models of cell death [ [100] , [101] , [102] ] raising the question as to which cell death events should be considered essential for the definition of parthanatos. Based on the above controversy, we propose a less restrictive definition that defines parthanatos simply as PARP1-dependent regulated necrosis. 5.1 The canonical route of parthanatos The fact that PARP1-mediated cell death is now recognized as a stand-alone cell death entity is mainly due to discoveries made in the lab of Valina and Ted Dawson at Johns Hopkins University. Their lab showed that cerebral ischemic injury [ 87 ], N-methyl- d -aspartate (NMDA) excitotoxicity [ 87 ], 1-metil-4-fenil-1,2,3,6-tetrahidropiridin (MPTP)-induced Parkinsonism, and neurodegenerative diseases [ 88 , 89 ] are mediated by PARP1 activation. While characterizing the molecular events leading to and following PARP1 activation in excitotoxicity, they identified apoptosis-inducing factor (AIF) as the downstream mediator of cell death [ 90 ]. Their proposed model relies on the following key events of parthanatos: stimulation of the NMDA receptor in neurons, glutamate triggering of calcium signaling, and calcium-dependent activation of the neuronal isoform of nitric oxide synthase (nNOS). Neuronal NOS produces nitric oxide, which combines with superoxide anion radical to form peroxynitrite (ONOO − ). Peroxynitrite is highly reactive and causes, among other macromolecular damages, DNA strand breaks [ 91 ]. Once DNA breaks are formed, the pathway may show an overlap with other cell death models in which DNA damage is triggered by direct DNA damaging stimuli (e.g. treatment with exogenous peroxynitrite, hydrogen peroxide, or DNA alkylating or crosslinking agents). DNA breaks are recognized by PARP1 and the enzyme becomes activated. Active PARP1 synthesizes the PAR polymer from NAD + to mark the site of DNA damage. PARG and ARH3 enzymes cleave the polymer off of PARylated proteins and the polymers (whether protein-bound or "naked" polymers is not known) leave the nucleus and translocate to the mitochondria ( Fig. 3 ). PAR triggers the release of AIF from the mitochondrial intermembrane space and AIF begins its journey towards the nucleus [ 79 ]. The mechanism of AIF release is not fully understood. PARP1-mediated NAD + depletion may trigger mitochondrial membrane depolarization, facilitating AIF release. Alternatively, the polymer may directly interact with the C-terminus of membrane-bound AIF. As an alternative to the consumption of NAD + , suppression of glycolysis may result from PAR binding to and inhibition of hexokinase, a key regulatory control point of glycolysis. Either way, AIF leaves the mitochondria and interacts with macrophage migration inhibitory factor (MIF) in the cytoplasm. Nuclear translocation of MIF is followed by MIF-mediated DNA fragmentation, as a result of the newly discovered nuclease activity of this multifunctional protein [ 92 ]. Fig. 3 Parthanatos. Nuclear mitochondrial crosstalk in parthanatos is triggered by DNA damaging stimuli activating PARP1. PAR synthesized in response to DNA breaks travels to the mitochondria and induces liberation of AIF. In turn, AIF interacts with MIF and the latter degrades DNA. Fig. 3 5.2 Parthanatos in MΦs Zingarelli et al. [ 93 ] was the first to demonstrate that PARP1-mediated cell death is an important cell death modality in activated MΦs. [Of note, at the time of their investigation PARP1 was the only known PARylating enzyme and was referred to as poly (ADP-ribose) synthase; PARS in this study.] They showed that exposure of murine peritoneal MΦs or J774 macrophage cells to high concentrations (10 μg/ml) of LPS triggered a rapid burst of superoxide production and a slower upregulation of iNOS. As a result, the ideal conditions were created for ONOO − formation, which triggered the following sequence of events: DNA breakage - PARP1 activation – NAD + /ATP depletion – cell death pathway. Follow up in vivo studies showed that inhibition of this pathway led to reduced organ damage and improved survival in endotoxemic or septic animals [ 91 , 94 , 95 ]. TLR ligands, such as LPS, also induce production of inflammatory cytokines, with TNFα considered to be the master cytokine regulator of inflammation [ 96 ]. MΦs are not only the primary source of TNFα, but also important targets of TNFα. MΦs express both TNFα receptors (TNFR1 and TNFR2) and respond to TNFα stimulation [ 96 ]. TNFα –induced cell death has been shown to be mediated by PARP1 in Actinomycin D-pretreated L929 and ME-180 human cervical carcinoma cells, but whether or not MΦs react similarly to the toxic effect of this cytokine remains to be seen [ 97 ]. Cytotoxicity of exogenous DNA damaging agents, such as H 2 O 2 , has also been investigated in MΦs. MΦs are often exposed to reactive oxygen species and are quite resistant to H 2 O 2 , mainly due to the constitutively active PI3K-Akt pathway [ 98 ]. Nonetheless, high concentrations of H 2 O 2 cause PARP1-mediated necrosis-like cell death characterized by plasma membrane permeabilization and lack of caspase activation [ 99 ]. Interestingly, even though this cell death is PARP1 dependent, no sign of AIF translocation could be observed. Thus, this type of cell death showed some key signs of parthanatos (PARP1-dependence, necrotic phenotype, energetic collapse), while lacking other key features (e.g. AIF translocation). Similar non-canonical parthanatos could also be observed in other models of cell death [ [100] , [101] , [102] ] raising the question as to which cell death events should be considered essential for the definition of parthanatos. Based on the above controversy, we propose a less restrictive definition that defines parthanatos simply as PARP1-dependent regulated necrosis. 6 Outlook Several open questions remain regarding the execution of necrotic MΦ cell death. The pathways leading to various forms of cell death are thought to be distinct. However, some evidence suggests that the necrotic cell death modalities discussed in this review could be interlinked. Studies have demonstrated that IAPs prevent RIPK3-dependent necroptosis and IL-1 activation [ 103 , 104 ]. Consistently, loss of XIAP triggers RIPK3- and caspase-8-driven IL-1β activation and cell death [ 105 ]. Activated MLKL also triggers the NLRP3 inflammasome in a cell intrinsic manner [ 106 ]. Staphylococcus toxin-induced necroptosis also promotes MLKL-NLRP3-mediated inflammation [ 65 ]. On the other hand, RIPK3 has also been shown to activate the inflammasome independent of MLKL [ 107 ]. These studies provide vital evidence of cross-talk between necroptosis and pyroptosis. Interestingly, a recent report demonstrated that the MLKL inhibitor, necrosulfonamide (NSA), which has been frequently used to prevent necroptosis in cells of human origin, also inhibits Gasdermin D, the pyroptosis-executioner [ 108 ]. Prevention of cell death using NSA has been interpreted as necroptosis, but it is possible that the drug also inhibited pyroptosis. Therefore, the existing data using NSA to inhibit necroptosis need to be revisited to understand if necroptosis and pyroptosis coexist. The simultaneous existence of multiple forms of necrotic cell death has not been scrutinized. For instance, S. Typhimurium infection in macrophages induces caspase-1/11 activation and IL-1β secretion. Additionally, we have shown that the pathogen induces necroptosis and also a differential PARP-1 cleavage and activation [ 69 ]. Similarly, Rhodococcus equi induced necrosis is also associated with PARP-1 activation [ 109 ]. PARP-1 activation leads to depletion of NAD + , which is sufficient to induce necroptosis during M. tuberculosis infection in MΦs. These findings provoke us to ask if pyroptosis, necroptosis, and parthanatos occur in different subsets of MΦs and how does the mode of cell death switch from one form to the other, depending on the functional state of the cell? Moreover, it is increasingly evident that mitochondrial metabolism, as well as glycolytic regulation, is a major contributor to regulated necrotic cell death modalities in MΦs. Compelling factors that might link various forms of cell death are reactive oxygen and nitrogen species, which sensitize macrophages to necrosis. However, this aspect of cell death has not yet been investigated in detail. For necroptosis, the modus operandi of RIP signaling in causing oxidative stress in addition to MLKL dependent membrane damage requires further investigation. For pyroptosis, a key question is if cell death related to IL-1 release can be uncoupled from IL-1 release in the cells. The notion that the initiation of events of programmed cell death pathways can be revoked and cells rescued from different forms of death is emerging [ 110 ]. The endosomal sorting complexes required for transport (ESCRT) machinery can repair damaged cell membranes and inhibit cell death [ 111 ], suggesting that this might allow for non-cell death mediated release of IL-1 cytokines. As for parthanatos, a detailed molecular characterization of this death pathway in MΦs, as well as proof for its in vivo relevance, is missing. Parthanatotic cell death is likely to contribute to inflammation via the release of DAMPs, due to its programmed necrotic phenotype. The release of the prototypical DAMP protein, HMGB1, is induced by PARP1 activation [ 112 ]. In LPS-treated MΦs, HMGB1 is PARylated, leading to acetylation, which is essential for HMGB1 release [ 113 ]. In MΦs, the interplay between PARP1 and DAMPs is a two-way communication. For example, in brain injury, microglial activation is mediated in part by the release of alarmins from damaged cells. Microglial signaling induced by the alarmin S100B is mediated by PARP1 [ 114 ]. Thus, the hypothesis that parthanatos is an inflammatory cell death pathway is plausible. However, the role of parthanatos in various forms of inflammation requires further investigation. Understanding the regulation of MΦ cell death modalities may open new avenues for therapeutic interventions in a wide range of inflammations ranging from microbial infections and atherosclerosis to toxic liver injuries. Exploiting our expanding knowledge of MΦ cell death pathways, the molecular switches diverting cells from one route to another, and the molecular determinants of MΦ sensitivity to cytotoxic stimuli may provide new opportunities for potential clinical treatments of various inflammatory diseases, atherosclerosis, microbial infections, toxic organ damage, and even cancer. Appendix A Supplementary data The following is the Supplementary data to this article: Multimedia component 1 Multimedia component 1

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5915198/

Conducting Science in Disasters: Recommendations from the NIEHS Working Group for Special IRB Considerations in the Review of Disaster Related Research

Summary: Research involving human subjects after public health emergencies and disasters may pose ethical challenges. These challenges may include concerns about the vulnerability of prospective disaster research participants, increased research burden among disaster survivors approached by multiple research teams, and potentially reduced standards in the ethical review of research by institutional review boards (IRBs) due to the rush to enter the disaster field. The NIEHS Best Practices Working Group for Special IRB Considerations in the Review of Disaster Related Research was formed to identify and address ethical and regulatory challenges associated with the review of disaster research. The working group consists of a diverse collection of disaster research stakeholders across a broad spectrum of disciplines. The working group convened in July 2016 to identify recommendations that are instrumental in preparing IRBs to review protocols related to public health emergencies and disasters. The meeting included formative didactic presentations and facilitated breakout discussions using disaster-related case studies. Major thematic elements from these discussions were collected and documented into 15 working group recommendations, summarized in this article, that address topics such as IRB disaster preparedness activities, informed consent, vulnerable populations, confidentiality, participant burden, disaster research response integration and training, IRB roles/responsibilities, community engagement, and dissemination of disaster research results. https://doi.org/10.1289/EHP2378 Introduction Public health emergencies and disaster events have challenged the world's preparedness and response capabilities for decades ( Figure 1 ). Since the attack on the World Trade Center in 2001, U.S. government agencies have redoubled their efforts to strengthen national preparedness, response, and recovery. With this national multiagency effort and continued exposure to new public health emergencies and disasters, the field of disaster research has evolved and has become commonplace in the post-disaster setting. Figure 1. Timeline of major global public health emergencies and disasters, 2001–2016. Figure 1 is adapted with permission from Lurie et al. ( 2013 ). We modified it significantly by extending the timeline, formatting color, and stratifying disasters by international and domestic. Timeline showing international disasters from 2001 through 2016. These include World Trade Center attacks (2001), Anthrax mailings (2001), SARS outbreak (2002), re-emergence of Avian influenza (H5N1) (2002), Bam earthquake in Iran (2003), Madrid train bombings (2004), Indian Ocean tsunami (2004), London bombings (2005), Hurricane Katrina (2005), Hurricane Rita (2005), Hurricane Wilma (2005), Prudhoe oil spill (2006), E. coli outbreak in spinach in the U.S. (2006), Mumbai attacks (2008), Sichuan earthquake (2008), Hurricane Gustav (2008), Hurricane Ike (2008), H1N1 influenza pandemic (2009), Haiti earthquake (2010), Pakistan floods (2010), Deepwater Horizon oil spill (2010), Japan earthquake, tsunami, and nuclear event (2011), series of tornadoes in U.S. (2011), Hurricane Irene (2011), North American drought (2012), Hurricane Sandy (2012), Typhoon Haiyan in the Philippines (2013), West African Ebola virus epidemic (2014), Flint water crisis (2014), Nepal earthquake (2015), Zika virus (2015), Paris terrorism attacks (2015) and U.S. Gulf Coast flooding (LA and TX) (2016), and Hurricane Matthew (2016). From a government agency perspective, disaster research is the study of individual, community and organizational preparedness, response, and recovery from a broad range of disaster types. Disaster research is essential to understanding how to prepare for and respond to catastrophic events such as hurricanes, earthquakes, disease outbreaks and pandemics, hazardous material spills, and large-scale acts of terrorism, as well as understanding their impact on human health. A unique feature of most disaster studies is the urgency of initiating data collection soon after the event to capture ephemeral baseline data that may be lost or subject to recall bias if collected later. Although the value of well-designed research studies in the immediate aftermath of disasters is recognized, there remain significant challenges that must be addressed to facilitate their administration. Some key challenges include time pressures related to the development of protocols and study materials, acquisition of rapid funding to support research work, concerns of the study team interfering with life-saving disaster response activities, and compromising a frail community ( Lurie et al. 2013 ; Miller et al. 2016 ). Despite its recognized value, research involving human subjects after disasters may pose ethical concerns ( Ferreira et al. 2015 ; O'Mathúna 2009 ). For example, the lack of coordination across investigators conducting research after a disaster can result in survivors being approached to join research by multiple research teams asking similar questions and requesting duplicative sample collections. In addition to the burden this may place on survivors, it can lead to unnecessary confusion when representatives from aid organizations offering direct assistance are in the field at the same time ( Taylor 2016 ). Most importantly, concerns about the vulnerability of prospective disaster research participants have been raised and evaluated ( Macklin 2014 ; Levine 2004 ). Although the Code of Federal Regulations (45 CFR Part 46—Protection of Human Subjects; DHHS 2009 ) does establish research protections for certain groups such as children, prisoners, women, and fetuses, there is no explicit protection for potentially vulnerable disaster survivor research participants. These human subject concerns led the National Institute of Environmental Health Sciences (NIEHS) to create an initiative to consider how institutional review boards (IRBs) can play a role in the preservation of ethical standards in the conduct of disaster research. Objective and Approach To address ethical and regulatory challenges in the oversight of post-disaster research, the Office of Human Research Compliance at NIEHS formed the new Best Practices Working Group for Special IRB Considerations in the Review of Disaster Related Research, as part of a larger effort at the National Institutes of Health (NIH) to enhance research oversight capacity after disasters. This effort, called the Disaster Research Response (DR2) program ( Lurie et al. 2013 ; Miller et al. 2016 ), began as a trans-NIH initiative in 2013 with the aim of developing a national framework to guide and facilitate research on the medical and public health aspects of disasters and public health emergencies. The working group was officially formed in September 2015 with the goals of exploring factors relevant to potential research participation in disasters and preparing IRBs for the review of disaster research protocols. The multidisciplinary working group consisted of 60 members from 23 U.S. states who were recruited through a nomination process that sought to assemble a diverse group of stakeholders, including academic researchers; bioethicists; disaster responders; local, state, and federal officials; disaster survivors; community advocates; and IRB/regulatory experts, and officials. The diverse nature of the group reflects the fact that disaster research is often a collaborative venture and the recognition that complex disaster issues cannot be adequately addressed through the lens of a single discipline but requires multidisciplinary expertise ( NRC 2006 ). The multi-sector working group was formed with the objective of developing IRB disaster-related research recommendations for the human research protection, IRB/regulatory, and disaster research, and response communities. At a meeting in July 2016, the working group was charged with addressing four overarching specific aims: 1. Preparing IRBs for the review of disaster research protocols 2. Exploring unique factors or heightened concerns as it relates to potential research participants and communities affected by disasters 3. Identifying participant burden for populations after disasters 4. Outlining duties and considerations of the IRB in the review of research involving disaster-affected communities. Breakout groups of participants were given a different disaster scenario and disaster research case study, and all groups were asked to react to the same set of discussion questions. The disaster scenarios included earthquake, terrorist attack with detonation of a radiological dispersion device (i.e., a "dirty bomb"), hurricane, pandemic influenza outbreak, and a toxic industrial chemical spill. The disaster scenarios were based on the Planning Scenarios developed by the Homeland Security Council in partnership with the Department of Homeland Security ( Homeland Security Council 2006 ). The case studies that followed the scenarios were hypothetical, condensed disaster research protocols that were designed to be implemented during the immediate response stage of a particular disaster. Major thematic elements from these discussions were collected and documented as 15 recommendations of the working group. Here we provide the recommendations as a framework and guidance for IRBs engaged in the review of disaster research protocols. Working Group Recommendations Recommendation 1: Prior to Consent, Prospective Participants Should Be Asked, to the Extent Feasible, about Unmet Needs and Provided Assistance Including Referrals and Resources to Reduce Risk and Maximize Benefit In the immediate aftermath of a disaster, survivors are often left behind with acute physical and mental health needs. Additionally, disasters can cause chronic impacts that impair social and economic stability including loss of employment and the dissolution of social networks. It is imperative that the life-sustaining and essential needs of potential research subjects are met for them to have adequate capacity to make a voluntary decision about enrollment in research. Researchers may be the first outsiders to face a disaster survivor, and they therefore should be trained in this regard and should identify unmet needs created by the disaster—for example, asthmatics and diabetics who no longer have access to their medication, or renal patients who are cut off from their dialysis center. Researchers who encounter urgent concerns among survivors have a responsibility to immediately notify the appropriate response officials. Researchers also should be prepared to provide participants with information on official disaster relief resources that are available (e.g., location of Red Cross tent, FEMA assistance centers) as well as referrals to local medical and/or mental health providers. Although referrals and resources could provide a benefit to potential participants, research should not interfere with potential research participants' efforts to meet their survival-related needs. Critical unmet needs must be the priority over enrollment in research. Recommendation 2: Close Monitoring of the Consent Process Is Key to Address Any Misconceptions about the Research IRBs should ensure close monitoring of the consenting process during recruitment in disaster studies, especially in the immediate aftermath of a disaster. Research teams must establish a standard plan (e.g., which may include a capacity or competence assessment screening questionnaire) for determining the decision-making ability of disaster-affected research participants to provide informed consent. As a precaution to eliminate confusion concerning the exchange of disaster aid for participating in research ( Ahmad and Mahmud 2010 ), consent forms may include a section requiring the participants to initial for indication they understand that they are participating in research and that their participation in the study is independent of disaster aid administered by local, state, or federal agencies or other entities. Additionally, research teams should distinguish themselves from responders by wearing vests, shirts, hats, and the like with clear labeling to establish their independence from the official responder community and clearly articulate to potential subjects that they are researchers asking them to engage in an optional research activity. As with all clinical studies, participants should be reassured throughout the consent process that they may opt out of the research at any time, and the process of opting out should be discussed with them. Consistent with good clinical practice, researchers may consider re-consenting participants weeks to months after enrollment as an additional tool to ensure ongoing maintenance of a robust informed consent process and remind participants of the voluntary nature of study participation, especially for those who enrolled during the initial response phase to the disaster. Recommendation 3: IRBs Should Guard against Any Reclassification of Minimal Risk Studies Due to the Establishment of New Post-Disaster Norms, and Should Ensure Transparency on Risks and Benefits of Research When the probability and magnitude of harm anticipated in the research is not greater than those ordinarily encountered in daily life, the research is properly classified as minimal risk. Because disasters can establish new daily norms, one might assume that an IRB could adopt a relative standard for minimal risk studies established in their wake. However, it is inappropriate to tolerate increased research risks even in post-disaster settings where a "new normal" has been established. There is a strong need for transparency in the research enterprise and clear identification and delineation of all potential risks and benefits of participating in a disaster-related study. Additionally, investigators should make it clear to potential participants, in the consent form and during the consent process, when the research offers no direct benefit. Researchers may want to consider a suitable level of remuneration commensurate with research participant time and effort and pay special attention to avoiding undue inducement under extreme post-disaster circumstances. Recommendation 4: Research Teams Should Ensure Private Areas to Conduct Study Procedures to Minimize Risk of Confidentiality Breaches Research procedures conducted in the disaster field may be out in the open because of damage to buildings and the set-up of temporary shelters. The loss of confidentiality may be particularly damaging in disaster studies when the release of personally identifiable information can create a long-lasting stigma of victimhood and potential discrimination experienced by survivors ( Harada et al. 2015 ). Research participants may also be concerned about the disclosure of sensitive medical information to their employers and/or insurance companies (e.g., disaster workers who participate in longitudinal research related to onsite exposures may potentially be banned from current or future work sites because their employer deems them unfit for deployment). To address privacy and confidentiality issues, research teams should plan in advance how they would assemble private areas to conduct interviews, examinations, or other study procedures. Additionally, researchers may consider applying for a Certificate of Confidentiality issued by the National Institutes of Health, which may serve to protect identifiable research information from forced disclosure and provide additional reassurance to research participants that their research data will be kept confidential. Although there have been rare legal challenges to a Certificate of Confidentiality that have resulted in the loss of confidentiality ( Beskow et al. 2008 ), there is substantial evidence that these certificates fulfill their intended purpose ( Wolf et al. 2015 ). Recommendation 5: Encourage Research on Groups (as Defined in 45 CFR 46 Such as Pregnant Women and Children) That Require Special Protections per Human Subjects Protection Regulations. Disaster Research Should Also Be Encouraged for Members of Vulnerable Groups That Are Underrepresented in the Disaster Research Literature Such as Women, Racial/Ethnic Minorities, and Elderly and Disabled Populations Researchers should develop new strategies to overcome the perceived barriers to the conduct of disaster research with groups that require special protections or who may have unique vulnerabilities. Valuable, informative research data may be lost if studies do not include these populations in their disaster studies. This is especially true when conducting research to assess behavioral and mental health outcomes. Indeed, there is mounting evidence that members of vulnerable groups may experience significant long-term mental and physical consequences following disaster events ( Lai et al. 2014 ; King et al. 2012 ). Justice demands that research be carried out for the benefit of the population as a whole; therefore, systematic exclusion of protected or vulnerable groups from disaster research studies should be avoided ( Mastroianni et al. 1994 ). Failure to include these groups leaves a knowledge gap in our understanding of the impact of disasters across the entire population. If the inclusion of one or more protected groups introduces unacceptable risks, researchers must justify why they are appropriately excluded from the research. IRBs must be aware of this knowledge gap and question whether such groups are unfairly excluded (e.g., due to perceived regulatory burdens rather than actual increased risks of participation in research procedures) from disaster research proposals. In situations when there is no clear rationale to exclude, IRBs must require research teams to outline a plan for conducting outreach and recruitment of such underrepresented groups into the study. Recommendation 6: Minimize Participant Burden Associated with Multiple Duplicative Studies in the Field through the Development of a Registry for Disaster Research Projects Survivors of disasters are often approached by many investigators, all seeking the same or similar information ( IOM 2014 ). This can result in survey and specimen collection fatigue and an overall increase in participant burden ( IOM 2014 ). A coordinated effort among researchers and funders could reduce duplication. One potential solution is the creation of a registry of disaster research projects to centralize and make more transparent the overall disaster research enterprise. Although development of such a registry is not an IRB function, it is consistent with the mission of the IRB to identify potential risks that may act to increase participant burden. Federal agencies and funders must play a leadership role in organizing such efforts by linking funding decisions to unique disaster research needs. An open and transparent database of disaster research studies, similar to ClinicalTrials.gov, would allow a central point for funders and government agencies to list disaster-related projects and requests for funding opportunities, reducing overall duplication. Recommendation 7: IRBs That Are Likely to Receive Disaster Research Protocols for Review Should Engage the Disaster Researcher and Responder Community Prior to Disaster Events Proactive engagement between IRBs, principal investigators (PIs), and the responder community may overcome some barriers to the timely review of disaster research protocols. Examples of engagement provided included inviting first responders and PIs to IRB trainings and meetings, securing responders with disaster expertise as ad hoc consultants to the IRB as a resource in the review of disaster research protocols, and setting up use agreements between IRBs and response agencies to ensure collaborative engagement during a disaster. Additionally, any perception of an antagonistic relationship between PIs and IRBs could be improved by proactive pre-disaster collaborative engagement. Recommendation 8: Disaster Researchers Should Consider the Development of Pre-event Generic Protocols for Provisional Approval by Their Local IRB. IRBs May Consider the Use of "Contingent Approval" Status for Time-Sensitive Disaster Studies Development of modular template protocols prior to disasters would facilitate protocol coordination and submission for approval after a disaster. A modular protocol would be one that is sufficiently flexible to fit a range of potential disaster scenarios. Activation of specific modular components that match the type and magnitude of the disaster and research interests could allow researchers to enter into the disaster field faster for time-sensitive disaster studies. The NIH DR2 program has developed such a protocol (i.e., Rapid Acquisition of Pre- and Post-Incident Disaster Data—RAPIDD) for the study of disaster workers, and the NIEHS IRB provisionally approved it in May 2015 ( Miller et al. 2016 ). The IRB preapproval of RAPIDD as an advancement in disaster research can be emulated in other jurisdictions. Indeed, RAPIDD has already been used as a model to develop such protocols at the University of Iowa and the University of Texas Medical Branch. Due to the variability that exists with different types and magnitudes of disasters, and depending on when the researcher wants to enter the disaster field, monitoring disaster research implementation in near real-time may help ensure the protection of research participants. IRBs are recommended to contingently approve disaster research protocols with the provision that the research team would report back to the IRB early in the implementation process and follow a fixed time schedule outlined by the IRB regarding any field related concerns or unanticipated issues. Additionally, an IRB may ask for the team to submit a continuing review report more frequently than the once a year required by federal regulations. Recommendation 9: Outsource Disaster Research Protocols to Specialized IRBs or Designate a Specialized IRB for Review of Disaster-Related Research IRBs should determine whether they have the appropriate expertise, review experience, training, and resources to properly review time-sensitive disaster-related research protocols. If an IRB determines that it lacks any of these elements, an alternate IRB with more disaster-related review experience should be made available when needed. An expansion of that idea could be the establishment of local or regional IRBs to act as specialized bodies for the review of disaster research protocols; inexperienced IRBs could then set up prepackaged reliance agreements with such entities. An example of such an entity is the Public Health Emergency Research Review Board (PHERRB), which has been put in place by the U.S. Department of Health and Human Services (DHHS) and NIH to serve as a single IRB exclusively for public health emergency research ( Lurie et al. 2013 ). Generally, the PHERRB may only be used for protocols that are conducted, supported, or regulated by HHS; that are subject to 45 CFR 46; and that require multiple IRB review. Recommendation 10: IRBs Should Develop Disaster and Community Profile Templates to Be Used by Research Teams to Gather Contextual Information to Guide IRB Review and Decision Making Disaster and community context is essential for IRBs to make informed decisions on disaster research protocols. IRBs should develop templates that would be populated by disaster researchers to provide the board with essential information about the disaster context. This template should include information on affected neighborhoods, morbidity and mortality associated with the event, post-disaster hazards and risks, and evacuation patterns among other variables. The template could also include detailed information on the community targeted for research (e.g., demographics, influential community groups, functional public health or medical infrastructure). Recommendation 11: Researchers Must Be Aware of a Disaster's Contextual Factors to Determine How They Impact Their Studies and to Optimize Timing of the Research Activities to Minimize Any Additional Stressors on Potential Research Participants while Maximizing Data Acquisition Optimal timing of research in the post-disaster setting is of paramount importance. IRBs need to have access to near real-time data on the nature and impact of the disaster, as it unfolds, on the affected community targeted for research. Depending upon the type, timing, and magnitude of a disaster, there may be certain time periods after a disaster when prospective research participants may have multiple unmet needs and lack specific survival-related resources. During this time, research would be inappropriate, especially when it does not offer goods or services needed to meet survivors' needs. Disaster events that result in mass casualties and/or cause long-term disruptions in critical infrastructure (e.g., utilities, health care systems) are more likely to lead to periods of acute stress and uncertainty among survivors. When post-disaster settings become normalized, a window of opportunity for research may present itself. Conversely, periods of stress and uncertainty may increase over time, especially when social and economic systems continue to erode after a disaster or when the disaster evolves slowly (e.g., the Flint water crisis). Recommendation 12: Encourage Mechanisms to Provide Pre-Disaster Local Community Knowledge to IRBs to Provide Context Specific to a Local Community IRBs based in localities at risk for disaster should, in the pre-disaster phase, identify community advisory groups and stakeholders that represent the broader community and who can serve as ad hoc consultants. The engagement of existing community advisory groups is an effective avenue to understanding community concerns and pre-disaster context so that post-disaster context can be accurately assessed. IRBs should be sure to give adequate attention to disadvantaged socioeconomic populations that may be at risk for undue inducement or exploitation. Although it is recognized that community knowledge on IRBs has value for the review of all types of research, it is especially true in disaster studies when affected communities may be particularly challenged. Because disasters are unpredictable in the communities they impact, preparedness efforts may only go so far. In the post-disaster setting, IRBs should make a concerted effort to contact community advisory groups in close proximity to the disaster to provide assistance in the review of a disaster research protocol. National organizations such as the Community-Campus Partnerships for Health provide access to community groups and academic institutions that can assist IRBs in their efforts. Disaster researchers can provide additional context by ensuring that their protocols include current information on the community to be studied and define strategies for gathering input from and ensuring participation by members of the community. Recommendation 13: IRBs That Wish to Establish Competency in the Review of Disaster Research Protocols Should Create and Adopt a Disaster Research Training Program and Resource Guide; Disaster Research Teams Would Also Benefit from Emergency Response Training Few IRBs have significant experience reviewing disaster protocols. IRB members should receive training on the basics of disaster management and specific human subject protection issues that can arise during the phases of disaster response and recovery as well as critical elements of IRB review for disaster-related research. PIs and their research teams could, in turn, be targeted for training on the regulatory aspects of the IRB review of disaster-related research. IRBs also should strongly encourage PIs and research teams to receive emergency response training (e.g., Incident Command System, National Incident Management System) before entering the disaster field, particularly during the immediate aftermath and especially when the research requires formal integration with the emergency response structure, in part to avoid impeding disaster response operations. In the pre-disaster phase, collaboratively training PIs, IRBs, and disaster responders together would be beneficial for the entire disaster research enterprise. An excellent example of such preparedness training could be the development of tabletop and field exercises that simulate the planning and implementation of disaster studies in the midst of a disaster response. Recommendation 14: IRBs Can Play an Important Role in Assessing the Feasibility of Disaster Research and Identifying Research That Might Not Lead to Generalizable Knowledge Due to the Disaster Context The IRB review process includes an assessment of the feasibility of the research. If the research is unlikely to be successful in testing its hypotheses due to logistical constraints in the disaster field (e.g., lack of stable utilities, difficulty of ingress and egress to disaster sites), IRBs should require research teams to establish contingency plans and modify their research protocols. IRBs can also play a role in identifying and rejecting disaster research that may pose unacceptable risks to the study participants or the research team itself, or that clearly interfere with the life and property-saving work of disaster responders. Recommendation 15: IRBs Should Assure That All Approved Disaster Research Specify and Confirm a Plan for the Timely Dissemination of Actionable Research Results Back to Key Stakeholders One of the principles of ethical research is to provide results and feedback to stakeholders, and disaster research is no different in that regard ( Emanuel et al. 2008 ). IRBs should require researchers to develop a dissemination plan for the results that clearly describes how the data will be reported back to participants and the community throughout the life cycle of the study. The plan must ensure a timely report back and should consider specific entities such as community groups and health educators that can help translate scientific findings into lay language. Methods of dissemination should be carefully considered to optimize information exchange with the community and may include town hall forums, newsletters, and use of social media. Recommendation 1: Prior to Consent, Prospective Participants Should Be Asked, to the Extent Feasible, about Unmet Needs and Provided Assistance Including Referrals and Resources to Reduce Risk and Maximize Benefit In the immediate aftermath of a disaster, survivors are often left behind with acute physical and mental health needs. Additionally, disasters can cause chronic impacts that impair social and economic stability including loss of employment and the dissolution of social networks. It is imperative that the life-sustaining and essential needs of potential research subjects are met for them to have adequate capacity to make a voluntary decision about enrollment in research. Researchers may be the first outsiders to face a disaster survivor, and they therefore should be trained in this regard and should identify unmet needs created by the disaster—for example, asthmatics and diabetics who no longer have access to their medication, or renal patients who are cut off from their dialysis center. Researchers who encounter urgent concerns among survivors have a responsibility to immediately notify the appropriate response officials. Researchers also should be prepared to provide participants with information on official disaster relief resources that are available (e.g., location of Red Cross tent, FEMA assistance centers) as well as referrals to local medical and/or mental health providers. Although referrals and resources could provide a benefit to potential participants, research should not interfere with potential research participants' efforts to meet their survival-related needs. Critical unmet needs must be the priority over enrollment in research. Recommendation 2: Close Monitoring of the Consent Process Is Key to Address Any Misconceptions about the Research IRBs should ensure close monitoring of the consenting process during recruitment in disaster studies, especially in the immediate aftermath of a disaster. Research teams must establish a standard plan (e.g., which may include a capacity or competence assessment screening questionnaire) for determining the decision-making ability of disaster-affected research participants to provide informed consent. As a precaution to eliminate confusion concerning the exchange of disaster aid for participating in research ( Ahmad and Mahmud 2010 ), consent forms may include a section requiring the participants to initial for indication they understand that they are participating in research and that their participation in the study is independent of disaster aid administered by local, state, or federal agencies or other entities. Additionally, research teams should distinguish themselves from responders by wearing vests, shirts, hats, and the like with clear labeling to establish their independence from the official responder community and clearly articulate to potential subjects that they are researchers asking them to engage in an optional research activity. As with all clinical studies, participants should be reassured throughout the consent process that they may opt out of the research at any time, and the process of opting out should be discussed with them. Consistent with good clinical practice, researchers may consider re-consenting participants weeks to months after enrollment as an additional tool to ensure ongoing maintenance of a robust informed consent process and remind participants of the voluntary nature of study participation, especially for those who enrolled during the initial response phase to the disaster. Recommendation 3: IRBs Should Guard against Any Reclassification of Minimal Risk Studies Due to the Establishment of New Post-Disaster Norms, and Should Ensure Transparency on Risks and Benefits of Research When the probability and magnitude of harm anticipated in the research is not greater than those ordinarily encountered in daily life, the research is properly classified as minimal risk. Because disasters can establish new daily norms, one might assume that an IRB could adopt a relative standard for minimal risk studies established in their wake. However, it is inappropriate to tolerate increased research risks even in post-disaster settings where a "new normal" has been established. There is a strong need for transparency in the research enterprise and clear identification and delineation of all potential risks and benefits of participating in a disaster-related study. Additionally, investigators should make it clear to potential participants, in the consent form and during the consent process, when the research offers no direct benefit. Researchers may want to consider a suitable level of remuneration commensurate with research participant time and effort and pay special attention to avoiding undue inducement under extreme post-disaster circumstances. Recommendation 4: Research Teams Should Ensure Private Areas to Conduct Study Procedures to Minimize Risk of Confidentiality Breaches Research procedures conducted in the disaster field may be out in the open because of damage to buildings and the set-up of temporary shelters. The loss of confidentiality may be particularly damaging in disaster studies when the release of personally identifiable information can create a long-lasting stigma of victimhood and potential discrimination experienced by survivors ( Harada et al. 2015 ). Research participants may also be concerned about the disclosure of sensitive medical information to their employers and/or insurance companies (e.g., disaster workers who participate in longitudinal research related to onsite exposures may potentially be banned from current or future work sites because their employer deems them unfit for deployment). To address privacy and confidentiality issues, research teams should plan in advance how they would assemble private areas to conduct interviews, examinations, or other study procedures. Additionally, researchers may consider applying for a Certificate of Confidentiality issued by the National Institutes of Health, which may serve to protect identifiable research information from forced disclosure and provide additional reassurance to research participants that their research data will be kept confidential. Although there have been rare legal challenges to a Certificate of Confidentiality that have resulted in the loss of confidentiality ( Beskow et al. 2008 ), there is substantial evidence that these certificates fulfill their intended purpose ( Wolf et al. 2015 ). Recommendation 5: Encourage Research on Groups (as Defined in 45 CFR 46 Such as Pregnant Women and Children) That Require Special Protections per Human Subjects Protection Regulations. Disaster Research Should Also Be Encouraged for Members of Vulnerable Groups That Are Underrepresented in the Disaster Research Literature Such as Women, Racial/Ethnic Minorities, and Elderly and Disabled Populations Researchers should develop new strategies to overcome the perceived barriers to the conduct of disaster research with groups that require special protections or who may have unique vulnerabilities. Valuable, informative research data may be lost if studies do not include these populations in their disaster studies. This is especially true when conducting research to assess behavioral and mental health outcomes. Indeed, there is mounting evidence that members of vulnerable groups may experience significant long-term mental and physical consequences following disaster events ( Lai et al. 2014 ; King et al. 2012 ). Justice demands that research be carried out for the benefit of the population as a whole; therefore, systematic exclusion of protected or vulnerable groups from disaster research studies should be avoided ( Mastroianni et al. 1994 ). Failure to include these groups leaves a knowledge gap in our understanding of the impact of disasters across the entire population. If the inclusion of one or more protected groups introduces unacceptable risks, researchers must justify why they are appropriately excluded from the research. IRBs must be aware of this knowledge gap and question whether such groups are unfairly excluded (e.g., due to perceived regulatory burdens rather than actual increased risks of participation in research procedures) from disaster research proposals. In situations when there is no clear rationale to exclude, IRBs must require research teams to outline a plan for conducting outreach and recruitment of such underrepresented groups into the study. Recommendation 6: Minimize Participant Burden Associated with Multiple Duplicative Studies in the Field through the Development of a Registry for Disaster Research Projects Survivors of disasters are often approached by many investigators, all seeking the same or similar information ( IOM 2014 ). This can result in survey and specimen collection fatigue and an overall increase in participant burden ( IOM 2014 ). A coordinated effort among researchers and funders could reduce duplication. One potential solution is the creation of a registry of disaster research projects to centralize and make more transparent the overall disaster research enterprise. Although development of such a registry is not an IRB function, it is consistent with the mission of the IRB to identify potential risks that may act to increase participant burden. Federal agencies and funders must play a leadership role in organizing such efforts by linking funding decisions to unique disaster research needs. An open and transparent database of disaster research studies, similar to ClinicalTrials.gov, would allow a central point for funders and government agencies to list disaster-related projects and requests for funding opportunities, reducing overall duplication. Recommendation 7: IRBs That Are Likely to Receive Disaster Research Protocols for Review Should Engage the Disaster Researcher and Responder Community Prior to Disaster Events Proactive engagement between IRBs, principal investigators (PIs), and the responder community may overcome some barriers to the timely review of disaster research protocols. Examples of engagement provided included inviting first responders and PIs to IRB trainings and meetings, securing responders with disaster expertise as ad hoc consultants to the IRB as a resource in the review of disaster research protocols, and setting up use agreements between IRBs and response agencies to ensure collaborative engagement during a disaster. Additionally, any perception of an antagonistic relationship between PIs and IRBs could be improved by proactive pre-disaster collaborative engagement. Recommendation 8: Disaster Researchers Should Consider the Development of Pre-event Generic Protocols for Provisional Approval by Their Local IRB. IRBs May Consider the Use of "Contingent Approval" Status for Time-Sensitive Disaster Studies Development of modular template protocols prior to disasters would facilitate protocol coordination and submission for approval after a disaster. A modular protocol would be one that is sufficiently flexible to fit a range of potential disaster scenarios. Activation of specific modular components that match the type and magnitude of the disaster and research interests could allow researchers to enter into the disaster field faster for time-sensitive disaster studies. The NIH DR2 program has developed such a protocol (i.e., Rapid Acquisition of Pre- and Post-Incident Disaster Data—RAPIDD) for the study of disaster workers, and the NIEHS IRB provisionally approved it in May 2015 ( Miller et al. 2016 ). The IRB preapproval of RAPIDD as an advancement in disaster research can be emulated in other jurisdictions. Indeed, RAPIDD has already been used as a model to develop such protocols at the University of Iowa and the University of Texas Medical Branch. Due to the variability that exists with different types and magnitudes of disasters, and depending on when the researcher wants to enter the disaster field, monitoring disaster research implementation in near real-time may help ensure the protection of research participants. IRBs are recommended to contingently approve disaster research protocols with the provision that the research team would report back to the IRB early in the implementation process and follow a fixed time schedule outlined by the IRB regarding any field related concerns or unanticipated issues. Additionally, an IRB may ask for the team to submit a continuing review report more frequently than the once a year required by federal regulations. Recommendation 9: Outsource Disaster Research Protocols to Specialized IRBs or Designate a Specialized IRB for Review of Disaster-Related Research IRBs should determine whether they have the appropriate expertise, review experience, training, and resources to properly review time-sensitive disaster-related research protocols. If an IRB determines that it lacks any of these elements, an alternate IRB with more disaster-related review experience should be made available when needed. An expansion of that idea could be the establishment of local or regional IRBs to act as specialized bodies for the review of disaster research protocols; inexperienced IRBs could then set up prepackaged reliance agreements with such entities. An example of such an entity is the Public Health Emergency Research Review Board (PHERRB), which has been put in place by the U.S. Department of Health and Human Services (DHHS) and NIH to serve as a single IRB exclusively for public health emergency research ( Lurie et al. 2013 ). Generally, the PHERRB may only be used for protocols that are conducted, supported, or regulated by HHS; that are subject to 45 CFR 46; and that require multiple IRB review. Recommendation 10: IRBs Should Develop Disaster and Community Profile Templates to Be Used by Research Teams to Gather Contextual Information to Guide IRB Review and Decision Making Disaster and community context is essential for IRBs to make informed decisions on disaster research protocols. IRBs should develop templates that would be populated by disaster researchers to provide the board with essential information about the disaster context. This template should include information on affected neighborhoods, morbidity and mortality associated with the event, post-disaster hazards and risks, and evacuation patterns among other variables. The template could also include detailed information on the community targeted for research (e.g., demographics, influential community groups, functional public health or medical infrastructure). Recommendation 11: Researchers Must Be Aware of a Disaster's Contextual Factors to Determine How They Impact Their Studies and to Optimize Timing of the Research Activities to Minimize Any Additional Stressors on Potential Research Participants while Maximizing Data Acquisition Optimal timing of research in the post-disaster setting is of paramount importance. IRBs need to have access to near real-time data on the nature and impact of the disaster, as it unfolds, on the affected community targeted for research. Depending upon the type, timing, and magnitude of a disaster, there may be certain time periods after a disaster when prospective research participants may have multiple unmet needs and lack specific survival-related resources. During this time, research would be inappropriate, especially when it does not offer goods or services needed to meet survivors' needs. Disaster events that result in mass casualties and/or cause long-term disruptions in critical infrastructure (e.g., utilities, health care systems) are more likely to lead to periods of acute stress and uncertainty among survivors. When post-disaster settings become normalized, a window of opportunity for research may present itself. Conversely, periods of stress and uncertainty may increase over time, especially when social and economic systems continue to erode after a disaster or when the disaster evolves slowly (e.g., the Flint water crisis). Recommendation 12: Encourage Mechanisms to Provide Pre-Disaster Local Community Knowledge to IRBs to Provide Context Specific to a Local Community IRBs based in localities at risk for disaster should, in the pre-disaster phase, identify community advisory groups and stakeholders that represent the broader community and who can serve as ad hoc consultants. The engagement of existing community advisory groups is an effective avenue to understanding community concerns and pre-disaster context so that post-disaster context can be accurately assessed. IRBs should be sure to give adequate attention to disadvantaged socioeconomic populations that may be at risk for undue inducement or exploitation. Although it is recognized that community knowledge on IRBs has value for the review of all types of research, it is especially true in disaster studies when affected communities may be particularly challenged. Because disasters are unpredictable in the communities they impact, preparedness efforts may only go so far. In the post-disaster setting, IRBs should make a concerted effort to contact community advisory groups in close proximity to the disaster to provide assistance in the review of a disaster research protocol. National organizations such as the Community-Campus Partnerships for Health provide access to community groups and academic institutions that can assist IRBs in their efforts. Disaster researchers can provide additional context by ensuring that their protocols include current information on the community to be studied and define strategies for gathering input from and ensuring participation by members of the community. Recommendation 13: IRBs That Wish to Establish Competency in the Review of Disaster Research Protocols Should Create and Adopt a Disaster Research Training Program and Resource Guide; Disaster Research Teams Would Also Benefit from Emergency Response Training Few IRBs have significant experience reviewing disaster protocols. IRB members should receive training on the basics of disaster management and specific human subject protection issues that can arise during the phases of disaster response and recovery as well as critical elements of IRB review for disaster-related research. PIs and their research teams could, in turn, be targeted for training on the regulatory aspects of the IRB review of disaster-related research. IRBs also should strongly encourage PIs and research teams to receive emergency response training (e.g., Incident Command System, National Incident Management System) before entering the disaster field, particularly during the immediate aftermath and especially when the research requires formal integration with the emergency response structure, in part to avoid impeding disaster response operations. In the pre-disaster phase, collaboratively training PIs, IRBs, and disaster responders together would be beneficial for the entire disaster research enterprise. An excellent example of such preparedness training could be the development of tabletop and field exercises that simulate the planning and implementation of disaster studies in the midst of a disaster response. Recommendation 14: IRBs Can Play an Important Role in Assessing the Feasibility of Disaster Research and Identifying Research That Might Not Lead to Generalizable Knowledge Due to the Disaster Context The IRB review process includes an assessment of the feasibility of the research. If the research is unlikely to be successful in testing its hypotheses due to logistical constraints in the disaster field (e.g., lack of stable utilities, difficulty of ingress and egress to disaster sites), IRBs should require research teams to establish contingency plans and modify their research protocols. IRBs can also play a role in identifying and rejecting disaster research that may pose unacceptable risks to the study participants or the research team itself, or that clearly interfere with the life and property-saving work of disaster responders. Recommendation 15: IRBs Should Assure That All Approved Disaster Research Specify and Confirm a Plan for the Timely Dissemination of Actionable Research Results Back to Key Stakeholders One of the principles of ethical research is to provide results and feedback to stakeholders, and disaster research is no different in that regard ( Emanuel et al. 2008 ). IRBs should require researchers to develop a dissemination plan for the results that clearly describes how the data will be reported back to participants and the community throughout the life cycle of the study. The plan must ensure a timely report back and should consider specific entities such as community groups and health educators that can help translate scientific findings into lay language. Methods of dissemination should be carefully considered to optimize information exchange with the community and may include town hall forums, newsletters, and use of social media. Conclusion The burgeoning field of disaster research has placed greater demands on IRBs to ensure that the welfare and rights of human research subjects are protected during disaster studies. The review of disaster research protocols requires new tools and training for IRBs to assure the protection of disaster survivors from research-related harms. These recommendations are currently being evaluated and prioritized by NIH officials to determine the process for moving forward with implementation. Although disaster research conducted during response may be challenging, IRBs can play useful roles in achieving careful, balanced, thoughtful procedures that both consider the value of the research to advance science and reduce suffering—and that also consider the potential for harm based on the unique vulnerabilities of disaster survivors in a disaster aftermath.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8923189/

Analysis of the Clinical Pipeline of Treatments for Drug-Resistant Bacterial Infections: Despite Progress, More Action Is Needed

ABSTRACT There is an urgent global need for new strategies and drugs to control and treat multidrug-resistant bacterial infections. In 2017, the World Health Organization (WHO) released a list of 12 antibiotic-resistant priority pathogens and began to critically analyze the antibacterial clinical pipeline. This review analyzes "traditional" and "nontraditional" antibacterial agents and modulators in clinical development current on 30 June 2021 with activity against the WHO priority pathogens mycobacteria and Clostridioides difficile . Since 2017, 12 new antibacterial drugs have been approved globally, but only vaborbactam belongs to a new antibacterial class. Also innovative is the cephalosporin derivative cefiderocol, which incorporates an iron-chelating siderophore that facilitates Gram-negative bacteria cell entry. Overall, there were 76 antibacterial agents in clinical development (45 traditional and 31 nontraditional), with 28 in phase 1, 32 in phase 2, 12 in phase 3, and 4 under regulatory evaluation. Forty-one out of 76 (54%) targeted WHO priority pathogens, 16 (21%) were against mycobacteria, 15 (20%) were against C. difficile , and 4 (5%) were nontraditional agents with broad-spectrum effects. Nineteen of the 76 antibacterial agents have new pharmacophores, and 4 of these have new modes of actions not previously exploited by marketed antibacterial drugs. Despite there being 76 antibacterial clinical candidates, this analysis indicated that there were still relatively few clinically differentiated antibacterial agents in late-stage clinical development, especially against critical-priority pathogens. We believe that future antibacterial research and development (R&D) should focus on the development of innovative and clinically differentiated candidates that have clear and feasible progression pathways to the market. INTRODUCTION The need for new antibacterial drugs to treat multidrug-resistant (MDR) bacterial infections is a critical global health issue, which has been recognized by many governmental, nongovernmental, and intergovernmental organizations ( 1 , 2 ), including the World Health Organization (WHO). In February 2017, the WHO released a list of 12 antibiotic-resistant priority pathogens ( Fig. 1 ), which are still among the most important bacterial infectious threats to human health ( 3 – 5 ). The WHO has also been critically analyzing the antibacterial pipeline since 2017, along with The Pew Charitable Trusts ( 6 ), which has resulted in the publication of four reports in 2017 ( 7 ), 2018 ( 8 ), 2019 ( 9 ), and 2021 ( 10 ). These pipeline reports and the WHO bacterial priority pathogen list have been used by policy makers, funders/sponsors, researchers, and developers to help guide the discovery and development of new antibacterial treatments. The U.S. Centers for Disease Control and Prevention (CDC) also released important pathogen (bacteria and fungi) threat lists in 2019 ( 11 ) and 2013 ( 12 ). While the WHO and CDC lists mostly overlap, there are some differences: the WHO list has ampicillin-resistant Haemophilus influenzae as a medium-priority pathogen, while the CDC list has Clostridioides difficile as an urgent threat, erythromycin-resistant group A Streptococcus and clindamycin-resistant group B Streptococcus as concerning threats, and drug-resistant Mycoplasma genitalium and Bordetella pertussis on a watch list. In March 2021, India released its own priority pathogen list ( 13 ), which included two pathogens, coagulase-negative staphylococci (CoNS) and Neisseria meningitidis (meningococcal disease), that are not in the WHO and CDC lists. The WHO is planning to update their priority pathogen list in 2022. FIG 1 List of the WHO's critical-, high-, and medium-priority pathogens ( 3 , 4 ) and mycobacteria. *, Enterobacteriaceae ( Escherichia coli , Enterobacter spp., and Klebsiella pneumoniae ) and Enterobacterales ( Morganella spp., Proteus spp., Providencia spp., and Serratia spp.). The discovery of new antibacterial drugs with activity against MDR bacteria is very challenging due to difficulties in designing products with suitable physicochemical properties (leading to desirable pharmacokinetics/pharmacodynamics properties) and acceptable toxicity profiles. Another major challenge is the lack of a suitable economic model that can provide long-term support for biotech and small companies developing new antibacterial agents ( 14 – 17 ). Factors underlying the lack of support include (i) the fact that antibacterial treatments are available for most bacterial infections, with most available as inexpensive generics, (ii) the typical short treatment duration of acute bacterial infections ( 18 ), (iii) the time and cost associated with traditional research and development (R&D) models, (iv) stewardship measures that—aiming at preserving new antibiotics efficacy—appropriately encourage prescribers to reserve new antibiotics and place them in the bottom of clinical guidelines as last-resort treatments, and (v) a lack of funding for phase 2 and 3 trials ( 19 ). All these elements have led to a market environment that is only marginally, if at all, profitable for most antibacterial drug developers. For example, the highest revenue for a patent protected antimicrobial in the United States in 2018 was US$138 million for the cephalosporin ceftaroline ( 17 ). The top 10 antimicrobials by sales in the United States in 2018, which included nine antibacterial drugs and the antifungal isavuconazole, had a total revenue of US$644 million. This drops down even further for the antibacterial drugs ranked 6 to 10 in sales (total sales revenues US$136 million, average $27.2 million). This is in stark contrast to the top-selling 2018 drug, adalimumab (therapeutic area: rheumatology), which had a total U.S. revenue of US$13.680 billion; even the revenue from the 10th highest selling drug in 2018, the anticoagulant apixaban, had US$3.76 billion revenue. This significant discrepancy in revenue helps to explain why most of the large pharmaceutical companies have either stopped or reduced their antibacterial R&D programs ( 19 ). The WHO has published an economic model that demonstrates these financial challenges ( 20 ). To address this issue, several "push" and "pull" development incentives are being proposed and implemented in several countries ( 21 – 25 ). Push-funding policies aim to reduce early development costs of developers by providing funding (e.g., grant support, contract funding, tax incentives, and private/public partnerships), while pull-funding policies aim to optimize the late stage of drug development and create viable market demand for sponsors (e.g., market entry rewards, extended exclusivity period, tradable market voucher, and higher reimbursement) ( 26 ). For example, the United Kingdom's antibiotic subscription pilot is the first ever fully delinked antibiotic pull incentive ( 27 , 28 ). In the United States, the PASTEUR Act is a bipartisan bill that, if passed into law, would similarly create a delinked reward model for novel and clinically needed new antimicrobials ( 29 , 30 ). In the last few years, there has been an increase in so-called "nontraditional" approaches to antibacterial therapy, developing drugs that have different modes of action compared to the "traditional" direct-acting antibacterial agents ( 31 , 32 ). These nontraditional antibacterial agents can prevent or treat bacterial infections through several modes of action, including directly or indirectly inhibiting bacterial growth, inhibiting virulence, ameliorating resistance, restoring the gut microbiome, or boosting the immune system to clear infections ( Table 1 ). However, most of these candidates are being clinically evaluated ( 33 ) as adjuvant therapies in combination with "standard of care" antibiotics. To date, there have been only three nontraditional antibacterial agents approved, all of which are monoclonal antibodies (MAb). Bezlotoxumab (approved by the FDA in 2016) binds and neutralizes Clostridioides difficile toxin B and was approved after the completion of two phase 3 clinical trials (NCT01241552 and NCT01513239) ( 34 ). Raxibacumab was authorized for the treatment of inhalational anthrax in adults and children (approved 2012 by the FDA) ( 35 ). Obiltoxaximab (US FDA, 2016; EMA, 2020) ( 36 ), like raxibacumab, has been approved to control the symptoms of inhaled anthrax toxins; while the safety profiles of raxibacumab and obiltoxaximab have been investigated in healthy volunteers, fortunately they have not yet been used clinically ( 37 ). TABLE 1 The five classification categories of nontraditional antibacterial agents Nontraditional classification Definition Antibodies A protein component of the immune system (or synthetic equivalent) that circulates in the blood and recognizes foreign substances like bacteria and viruses Bacteriophages and phage-derived enzymes Substances that directly cause pathogen lysis that are phage-derived recombinant enzymes or phages (including those engineered as nano-delivery vehicles) Microbiome-modulating agents Approaches that seek to modify the microbiome to eliminate or prevent carriage of resistant or pathogenic bacteria manipulating the metabolism of microbiota Immunomodulating agents Compounds that augment, stimulate, or suppress host immune responses that modify the course of infection Miscellaneous agents Group of strategies that seek to (i) inhibit the production or the activity of virulence factors such as toxins, (ii) impede bacterial adhesion to host cells and biofilm formation, (iii) interrupt or inhibit bacterial communication, and (iv) inhibit resistance mechanisms In this review, we discuss traditional and nontraditional antibacterial agents that were being evaluated in clinical trials on 30 June 2021 for the treatment of infections caused by the WHO priority pathogens Mycobacterium tuberculosis ( 38 ) and nontuberculosis mycobacteria (NTM) ( 39 ) and C. difficile , which is not a WHO priority pathogen but is considered by the CDC to be an urgent threat ( 11 ). A brief overview of the drug development process and drug regulatory agencies are provided in the supplemental information. Data for this review were based on the WHO's antibacterial agents in clinical development reports published in 2021 ( 10 ), 2018 ( 8 ), 2019 ( 9 ), and 2017 ( 7 ), as well as WHO's preclinical pipeline analyses ( 10 , 40 ). First, we examined antibacterial drugs that had been approved anywhere in the world between 1 July 2017 and 30 June 2021 ( Table 2 , Fig. 2 ). Next, we analyzed the traditional and nontraditional antibacterial agents being evaluated in phase 1 to 3 clinical trials or those having a new drug application (NDA)/market authorization application (MAA) submitted to a regulatory body with a cutoff date of 30 June 2021 that had not previously been granted market authorization for human use anywhere in the world ( Tables 3 to 8 ). The traditional and nontraditional antibacterial drug candidates were then analyzed by development phase ( Fig. 3 ), target organism type ( Fig. 4 ), and new pharmacophore types ( Fig. 5 ). TABLE 2 Antibacterial drugs that gained market authorization between July 2017 and June 2021 a Name (trade name) Market authorization holder(s) Agency/agencies granting approval (date) Antibacterial class Route of administration Indication(s) WHO EML & AWaRe Expected activity against priority pathogens b Innovation c CRAB CRPA CRE OPP NCR CC T MoA Delafloxacin (Baxdela) Melinta (Menarini, EU) FDA (6/2017 ABSSSI, 10/2019 CAP), EMA (12/2019 ABSSSI) Fluoroquinolone i.v. & oral ABSSSI, CAP AWaRe: Watch ○ ○ ○ ● — — — — Vaborbactam + meropenem (Vabomere) Melinta (Menarini, EU) FDA (8/2017), EMA (11/2018) Boronate BLI + β-lactam (carbapenem) i.v. cUTI WHO EML & AWaRe: Reserve ○ ○ ● d NA ? e ✓ — — Plazomicin (Zemdri) Achaogen (Cipla USA / QiLu Antibiotics, China) FDA (8/2018) Aminoglycoside i.v. cUTI WHO EML & AWaRe: Reserve ○ ○ ● NA — — — — Eravacycline (Xerava) Tetraphase (La Jolla, Everest Medicines) FDA (8/2018), EMA (9/2018) Tetracycline i.v. cIAI AWaRe: Reserve ? ○ ● NA — — — — Omadacycline (Nuzyra) Paratek FDA (10/2018) Tetracycline i.v. & oral CAP (iv), ABSSSI (iv, oral) AWaRe: Reserve ○ ○ ○ ● — — — — Relebactam + imipenem/cilastatin (Recarbrio) MSD FDA (7/2019 cUTI/cIAI, 7/2020 HAP/VAP), EMA (2/2020 Gram -ve) DBO-BLI + β-lactam (carbapenem)/ degradation inhibitor i.v. cUTI, cIAI, HAP/VAP AWaRe: Reserve ○ ? ● d NA — — — — Lefamulin (Xenleta) Nabriva (Sunovion Pharmaceuticals Canada) FDA (8/2019), EMA (7/2020) Pleuromutilin i.v. & oral CAP AWaRe: Reserve NA NA NA ● ? ✓ f — — Pretomanid (Dovprela) Viatris (TB Alliance) g FDA (8/2019), EMA (8/2020), CDSCO (7/2020) Nitroimidazole Oral XDR-TB NA NA NA ● h — — — — Lascufloxacin (Lasvic) Kyorin Pharmaceutical PDMA (8/2019) Fluoroquinolone i.v. & oral CAP, otorhinolaryngological AWaRe: Watch ○ ○ ○ ● — — — — Cefiderocol (Fetroja) Shionogi FDA (11/2019, cUTI; 9/21 HAP/VAP), EMA (4/2020) Siderophore β-lactam (cephalosporin) i.v. FDA: cUTI, HAP/VAP, EMA: aerobic G-ve WHO EML & AwaRe: Reserve ● ● ● NA ? — — — Levonadifloxacin (Emrok), Alalevonadifloxacin (Emrok-O) Wockhardt CDSCO (1/2020) Fluoroquinolone i.v. & oral ABSSSI AwaRe: Watch ○ ○ ○ ● — — — — Contezolid (Youxitai), Contezolid acefosamil MicuRx NMPA (6/2021) Oxazolidinone i.v. & oral cSSTI NA NA NA ● — — — — a Abbreviations: ABSSSI, acute bacterial skin and skin structure infections; AwaRe, A ccess Wa tch Re serve; CAP, community-acquired pneumonia; CC, new chemical class; cIAI, complicated interabdominal infection; CRAB, carbapenem-resistant Acinetobacter baumannii ; CRE, carbapenem-resistant Enterobacterales ; CRPA, carbapenem-resistant P. aeruginosa ; CDSCO, Central Drugs Standard Control Organization of the Government of India; cSSTI, complicated skin and soft tissue infections; cUTI, complicated urinary tract infection; EMA, European Medicines Agency; EML, essential medicines list; FDA, Food and Drug Administration (USA); HAP, hospital-acquired pneumonia; i.v., intravenous; KPC, K. pneumoniae carbapenemase; MBL, metallo-β-lactamase; OPP, other priority pathogens; MoA, new mode of action; NCR, no cross-resistance to other antibiotic classes; NMPA, China National Medical Products Administration; PDMA, Pharmaceuticals and Medical Devices Agency (Japan); T, new target; VAP, ventilator-acquired pneumonia; XDR, extensively drug-resistant. b Pathogen activity: ●, active; ?, possibly active; ○, not or insufficiently active; NA, activity not assessed, as the antibiotic is focused and developed for only either Gram-positive cocci or Gram-negative rods. Agents not active against critical-priority pathogens were assessed for activity against other priority pathogens (OPP), which includes the high and medium WHO priority pathogens. c Innovation assessment: ✓, criterion fulfilled; ?, inconclusive data; —, criterion not fulfilled. d Active against KPC- but not MBL-producing Enterobacteriaceae . e Cross-resistance can be obtained when the levels of the porin OmpK36 are varied. f First systemic formulation of this class, which was previously used in animals and topically in humans. g The approvals were obtained by the TB Alliance and then transferred to Viatris. h Approved for the treatment of XDR-TB or treatment-intolerant/nonresponsive MDR-TB, in combination with bedaquiline and linezolid. TABLE 3 Traditional antibacterial agents and combinations in NDA and phase 3 clinical development against WHO priority pathogens Name (synonym) Phase Antibacterial class Route of administration Developer Expected activity against priority pathogens a Innovation b CRAB CRPA CRE OPP NCR CC T MoA Solithromycin (T-4288) NDA c Macrolide i.v. & oral iFUJIFILM Toyama Chemical NA NA NA ● d — — — — Sulopenem, Sulopenem etzadroxil/probenecid NDA e β-Lactam (penem) i.v. & oral Iterum ○ ○ ○ f NA — — — — Durlobactam (ETX-2514) + sulbactam 3 DBO-BLI/PBP2 binder + β-lactam-BLI/PBP1,3 binder i.v. Entasis ● ○ ○ NA — — — — Taniborbactam (VNRX-5133) + cefepime 3 Boronate BLI + β-lactam (cephalosporin) i.v. VenatoRx/GARDP ○ ● ● NA ? ✓ — — Enmetazobactam (AAI-101) + cefepime 3 BLI + β-lactam (cephalosporin) i.v. Allecra ○ ○ ○ g NA — — — — Zoliflodacin 3 Spiropyrimidenetrione (topoisomerase inhibitor) Oral Entasis/GARDP NA NA NA ● d ✓ ✓ — ✓ Gepotidacin 3 Triazaacenaphthylene (topoisomerase inhibitor) i.v. & oral GSK NA NA NA ● d ? ✓/? h — ✓ Nafithromycin (WCK-4873) 3 Macrolide Oral Wockhardt NA NA NA ● d — — — — Benapenem 2/3 β-Lactam (carbapenem) i.v. Sichuan Pharmaceutical ○ ○ ○ NA — — — — a Pathogen activity: ●, active; ?, possibly active; ○, not or insufficiently active; NA, activity not assessed, as the antibiotic is focused and developed for only either Gram-positive cocci or Gram-negative rods. Agents not active against critical-priority pathogens were assessed for activity against OPP, which includes the high and medium WHO priority pathogens. b Innovation assessment: ✓, criterion fulfilled; ?, inconclusive data; —, criterion not fulfilled. CC, chemical class; MOA, new mode of action; NCR, no cross-resistance; T, new target. c Solithromycin NDA for otorhinolaryngological infections submitted in Japan in April 2019. d OPP target pathogens: solithromycin, S. pneumoniae ; nafithromycin, S. aureus and S. pneumoniae ; gepotidacin, N. gonorrhoeae and E. coli ; zoliflodacin, N. gonorrhoeae . e Sulopenem etzadroxil NDA submitted in USA for uncomplicated UTI (uUTI) in November 2020. f Active against ESBL-producing cephalosporin-resistant but not carbapenem-resistant Enterobacterales. g Active against ESBL-producing cephalosporin-resistant and some KPC-producing CRE. h Gepotidacin is being tested in two distinct phase 3 programs: gonorrhea (NCR ✓) and uUTI (NCR ?). TABLE 4 Traditional antibacterial agents and combinations in phase 1 and 2 clinical development against WHO priority pathogens Name (synonym) Phase Antibacterial class Route of administration Developer Expected activity against priority pathogens a Innovation b CRAB CRPA CRE OPP NCR CC T MoA Afabicin (Debio-1450) 2 Pyrido-enamide (FabI inhibitor) i.v. & oral Debiopharm NA NA NA ● c ✓ ✓ ✓ ✓ TNP-2092 2 Rifamycin-quinolizinone hybrid i.v. & oral TenNor Therapeutics NA NA NA ● c — — — — TNP-2198 1b/2a Rifamycin-nitroimidazole conjugate Oral TenNor Therapeutics NA NA NA ● c — — — — Zidebactam + cefepime 1 d DBO-BLI/PBP2 binder e + cephalosporin i.v. Wockhardt ● ● ● NA — — — — Nacubactam (OP0595) + meropenem 1 DBO-BLI/PBP2 binder e + β-lactam (carbapenem) i.v. Meiji Seika ○ ○ f ● NA — — — — ETX0282 + cefpodoxime 1 DBO-BLI/PBP2 binder e + β-lactam (cephalosporin) Oral Entasis Therapeutics ○ ○ ● NA — — — — XNW-4107+ imipenem + cilastatin 1 BLI + β-lactam (carbapenem) / degradation inhibitor i.v. Sinovent ? ? ? ? ? ? ? ? VNRX-7145 + ceftibuten 1 Boronate BLI + β-lactam (cephalosporin) Oral VenatoRx Pharmaceuticals ○ ○ ● NA ? ✓ — — SPR-206 1 Polymyxin i.v. Spero Therapeutics ● ● ● NA — — — — MRX-8 1 Polymyxin i.v. MicuRx ● ● ● NA — — — — QPX-9003 1 Polymyxin i.v. Qpex Biopharma ? ? ? ? ? ? ? ? KBP-7072 1 Tetracycline Oral KBP BioSciences ● ○ ○ ● c — — — — EBL-1003 (apramycin) 1 f Aminoglycoside i.v. Juvabis ● ○ ● NA — — — — TXA-709 1 "Difluorobenzamide" (FtsZ inhibitor) i.v. & oral TAXIS Pharmaceutical ○ ○ ○ ● c ✓ ✓ ✓ ✓ ARX-1796 (oral avibactam prodrug) 1 DBO-BLI + β-lactam (undisclosed) Oral Arixa/Pfizer g ○ ○ ● h NA — — — — PLG0206 (WLBU2) 1 Cationic peptide i.v. h Peptilogics ? i ? i ? i ● c , j ? ✓ ? ? QPX7728 k + QPX2014 / QPX7728 + QPX2015 1 Boronate-BLI + β-lactam (undisclosed) / boronate-BLI + β-lactam (undisclosed) i.v. / i.v. & oral Qpex Biopharma ● ● ● NA ? — — — a Pathogen activity: ●, active; ?, possibly active; ○, not or insufficiently active; NA, activity not assessed, as the antibiotic is focused and developed for only either Gram-positive cocci or Gram-negative rods. Agents not active against critical-priority pathogens were assessed for activity against OPP, which includes the high and medium WHO priority pathogens. b Innovation assessment: ✓, criterion fulfilled; ?, inconclusive data; —, criterion not fulfilled. CC, chemical class; MOA, new mode of action; NCR, no cross-resistance; T, new target. c OPP target pathogens: TNP-2198, H. pylori ; afabicin, TNP-2092, KBP-7072, TXA-109, and PLG0206, S. aureus . d A phase 3 trial for zidebactam + cefepime was registered in July 2021 for cUTI or acute pyelonephritis (NCT04979806). e The DBO-BLIs zidebactam, nacubactam, and ETX0282 also have some antibacterial activity and have been classified as β-lactam enhancers (BLE) ( 97 – 99 ). f Previously used as an antibacterial treatment in animals. g Activity against AmpC-producing and KPC-producing CRPA. Active against KPC- but not MBL-producing Enterobacteriaceae . h The original developer, Arixa Pharmaceuticals, was acquired by Pfizer in October 2020. i PLG0206 was evaluated in phase 1 using i.v. administration, but development is currently focused on use as an irrigation solution for prosthetic joint infections. j Peptilogics recently reported that coagulase-negative staphylococci, E. coli , Enterobacter cloacae , Citrobacter freundii , P. aeruginosa , and A. baumannii ( 100 ). k QPX7728 is being evaluated with two separate β-lactams, QPX-2014 and QPX2015. TABLE 5 Traditional antibacterial agents in clinical development for the treatment of TB and nontuberculous mycobacteria (NTM) Name (synonym) Phase Antibiotic class Route of administration Developer Innovation a NCR CC T MoA GSK-3036656 (GSK070) 2 Oxaborole (Leu-Rs inhibitor) Oral GSK ✓ ✓ ✓ ✓ Delpazolid (LCB01-0371) 2b Oxazolidinone Oral LegoChem Biosciences/HaiHe Biopharma — — — — Sutezolid 2 Oxazolidinone Oral TB Alliance/Sequella — — — — Telacebec (Q-203) 2 Imidazopyridine amide Oral Qurient ✓ ✓ ✓ ✓ TBA-7371 2 Azaindole (DprE1 inhibitor) Oral TB Alliance/Bill & Melinda Gates Foundation/Foundation for Neglected Disease Research ✓ ✓ ✓ ✓ SPR720 2a b Benzimidazole ethyl urea (GyrB inhibitor c ) Oral Spero/Bill & Melinda Gates Foundation — ✓ — — TBI-166 (pyrifazimine) d 2 Riminophenazine (clofazimine-analogue) Oral Institute of Materia Medica/TB Alliance/Chinese Academy of Medical Sciences/Peking Union Medical College — — — — OPC-167832 1/2 3,4-Dihydrocarbostyril (DprE1 inhibitor) Oral Otsuka ✓ ✓ ✓ ✓ BTZ-043 1/2 Benzothiazinone (DprE1 inhibitor) Oral University of Munich/Hans Knöll Institute, Jena/German Center for Infection Research ✓ ✓ ✓ ✓ Macozinone (PBTZ-169) 1 Benzothiazinone (DprE1 inhibitor) Oral Innovative Medicines for Tuberculosis Foundation/Nearmedic Plus ✓ ✓ ✓ ✓ TBI-223 1 Oxazolidinone Oral TB Alliance/Institute of Materia Medica — — — — TBAJ-876 1 Diarylquinoline Oral TB Alliance — — — — TBAJ-587 1 Diarylquinoline Oral TB Alliance — — — — GSK 2556286 (GSK-286) 1 Undisclosed Oral GSK/TB Drug Accelerator/Bill & Melinda Gates Foundation ? ✓ ✓ ? a Innovation assessment: ✓, criterion fulfilled; ?, inconclusive data; —, criterion not fulfilled. CC, chemical class; MOA, new mode of action; NCR, no cross-resistance; T, new target. b This phase 2a trial (NCT04553406) was on FDA clinical hold, but this was lifted in January 2022. c This is not considered to be a new mode of action, as the GyrB/ParE inhibitor novobiocin was once marketed but is no longer in clinical use. d The lead drug clofazimine is approved to treat leprosy and has been used off-label for TB treatment. TABLE 6 Traditional antibacterial agents in clinical development for the treatment of C. difficile infections Name (synonym) Phase Antibiotic class Route of administration Developer Innovation a NCR CC T MoA Ridinilazole 3 Bis -benzimidazole Oral Summit Therapeutics ✓ ✓ ✓ ✓ DNV-3837 (MCB-3837) 2 Oxazolidinone-quinolone hybrid i.v. Deinove ? — — — MGB-BP-3 2 Distamycin (DNA minor groove binder) Oral MGB Biopharma ? ✓ ✓ ✓ Ibezapolstat (ACX-362E) 2 "Substituted guanine" (DNA polymerase IIIC inhibitor) Oral Acurx Pharmaceuticals ? ✓ ✓ ✓ CRS3123 2 "Diaryldiamine" (methionyl-tRNA synthetase inhibitor; MetRS) Oral Crestone/NIAID ✓ ✓ ✓ ✓ a Innovation assessment: ✓, criterion fulfilled; ?, inconclusive data; —, criterion not fulfilled. CC, chemical class; MOA, new mode of action; NCR, no cross-resistance; T, new target. These agents are being developed for C. difficile infections, and their activity against WHO priority pathogens was not assessed. TABLE 7 Nontraditional antibacterial agents in phase 3 clinical development a Submitted to the Australian Therapeutic Goods Association (TGA) as a potential treatment for recurrent C. difficile infections in June 2021. b Submitted to the U.S. FDA as a potential treatment for necrotizing soft tissue infections (NSTI) in December 2020. TABLE 8 Nontraditional antibacterial agents in phase 1 and 2 clinical development for WHO priority pathogens, mycobacteria, and C. difficile FIG 2 Structures of antibacterial drugs approved worldwide since 2017 and their approved indications and targeted priority pathogens with country and year of first approval. FIG 3 Number of traditional and nontraditional antibacterials by (A) development phase and (B) development against WHO priority pathogens, TB and NTM, C. difficile , and G+ve/G−ve. FIG 4 Traditional and nontraditional antibacterials categorized by development phase and activity against WHO critical pathogens, WHO high- and medium-priority pathogens TB and NTM, C. difficile , and nontraditional nonspecific G+ve/G−ve activity. FIG 5 Antibacterials with new pharmacophores not previously found in human antibacterial drugs by target class, target, antibacterial name (current development phase), and antibacterial class. Abbreviations: TB, tuberculosis; Sa, S. aureus ; Cd, C. difficile , Ec, E. coli ; Ng, N. gonorrhoeae ; NTM, nontuberculosis mycobacteria; G−, Gram-negative bacteria. METHODOLOGY Scope and inclusion/exclusion criteria. This review details the antibacterial drugs that have been approved for the treatment of WHO priority pathogens anywhere in the world between 1 July 2017 and 30 June 2021. Also included in this analysis are traditional and nontraditional antibacterial agents administered by intravenous (i.v.), intramuscular (i.m.), oral, inhalation, enema, and colonoscopy administration routes that are currently being evaluated in phase 1 to 3 clinical trials or have NDA/MAA applications under consideration that have not previously been granted market authorization for human use anywhere in the world. Antibacterial agents were restricted to those being developed or that have the potential to treat bacterial infections caused by the WHO priority pathogens ( Fig. 1 ), mycobacteria, or C. difficile and are included only if they are new chemical entities (NCEs) (traditional or nontraditional) or new biological entities (NBEs) (nontraditional) not already accorded market authorization for human use anywhere in the world. Antibacterial agents whose development programs have been terminated, are no longer listed on a company's development pipeline, or have not had any development update for three or more years have been excluded in this analysis and are listed in the supplemental information (Table S1). This review does not include new formulations of approved antibacterial drugs, vaccines, topical decolonizing agents, nonspecific inorganic substances, and antibacterial agents developed only for topical applications such as creams, ointments, or eye drops. Fixed-dose combinations of potentiators and antibacterial agents are included if they contain an NCE or an NBE. Search strategy. Data from the 2020 WHO antibacterial pipeline report ( 10 ) were used as a starting point for this updated analysis. Recent antibacterial pipeline reviews ( 41 – 43 ), previous WHO reports ( 7 , 9 ), and The Pew Charitable Trusts' antibiotic development pipeline reviews ( 6 ) were also consulted. Additional references were identified using searches of antibacterial compound names and their synonyms from PubMed ( https://pubmed.ncbi.nlm.nih.gov/ ), Google Scholar ( https://scholar.google.com.au/ ), and conference abstracts and posters. The U.S. NIH ( https://clinicaltrials.gov/ ) and WHO International Clinical Trials Registry Platform (ICTRP) clinical trial databases ( https://www.who.int/clinical-trials-registry-platform ), the commercial database AdisInsight ( https://adisinsight.springer.com/ ), and the Access to Medicine Foundation's Antimicrobial Resistance Benchmark 2020 Antibacterials data ( 44 ) were searched. The websites of pharmaceutical companies active in antibacterial R&D and antibacterial development funders and foundations were also searched. Scope and inclusion/exclusion criteria. This review details the antibacterial drugs that have been approved for the treatment of WHO priority pathogens anywhere in the world between 1 July 2017 and 30 June 2021. Also included in this analysis are traditional and nontraditional antibacterial agents administered by intravenous (i.v.), intramuscular (i.m.), oral, inhalation, enema, and colonoscopy administration routes that are currently being evaluated in phase 1 to 3 clinical trials or have NDA/MAA applications under consideration that have not previously been granted market authorization for human use anywhere in the world. Antibacterial agents were restricted to those being developed or that have the potential to treat bacterial infections caused by the WHO priority pathogens ( Fig. 1 ), mycobacteria, or C. difficile and are included only if they are new chemical entities (NCEs) (traditional or nontraditional) or new biological entities (NBEs) (nontraditional) not already accorded market authorization for human use anywhere in the world. Antibacterial agents whose development programs have been terminated, are no longer listed on a company's development pipeline, or have not had any development update for three or more years have been excluded in this analysis and are listed in the supplemental information (Table S1). This review does not include new formulations of approved antibacterial drugs, vaccines, topical decolonizing agents, nonspecific inorganic substances, and antibacterial agents developed only for topical applications such as creams, ointments, or eye drops. Fixed-dose combinations of potentiators and antibacterial agents are included if they contain an NCE or an NBE. Search strategy. Data from the 2020 WHO antibacterial pipeline report ( 10 ) were used as a starting point for this updated analysis. Recent antibacterial pipeline reviews ( 41 – 43 ), previous WHO reports ( 7 , 9 ), and The Pew Charitable Trusts' antibiotic development pipeline reviews ( 6 ) were also consulted. Additional references were identified using searches of antibacterial compound names and their synonyms from PubMed ( https://pubmed.ncbi.nlm.nih.gov/ ), Google Scholar ( https://scholar.google.com.au/ ), and conference abstracts and posters. The U.S. NIH ( https://clinicaltrials.gov/ ) and WHO International Clinical Trials Registry Platform (ICTRP) clinical trial databases ( https://www.who.int/clinical-trials-registry-platform ), the commercial database AdisInsight ( https://adisinsight.springer.com/ ), and the Access to Medicine Foundation's Antimicrobial Resistance Benchmark 2020 Antibacterials data ( 44 ) were searched. The websites of pharmaceutical companies active in antibacterial R&D and antibacterial development funders and foundations were also searched. ANALYSIS AND DISCUSSION Antibacterial drugs approved since 2017. Twelve new antibacterial drugs ( Table 2 , Fig. 2 ) have been approved since the WHO's first analysis of the clinical antibacterial pipeline in 2017 ( 7 ). The most recent approval was in China in June 2021 for the oxazolidinone contezolid as a treatment for complicated skin and soft tissue infections (cSSTI) caused by Gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) ( 45 ). Only 1 of the 12 approved antibacterial drugs, the boronate β-lactamase inhibitor (BLI) vaborbactam, which is used in combination with meropenem, has a new antibacterial drug-related pharmacophore. Also, the cephalosporin cefiderocol ( 46 , 47 ) is noteworthy, as it is the first marketed β-lactam (cephalosporin) that has an iron-chelating siderophore incorporated into the structure, which facilitates Gram-negative bacteria outer membrane entry. The other 10 are members of previously approved antibacterial classes: three fluoroquinolones (delafloxacin, lascufloxacin, and levonadifloxacin), two tetracyclines (eravacycline and omadacycline), one aminoglycoside (plazomicin), one pleuromutilin (lefamulin), one nitroimidazole (pretomanid), one diazabicyclooctane (DBO) BLI (relebactam), and one oxazolidinone (contezolid). Of the 12 new antibacterial drugs, 6 target carbapenem-resistant Enterobacterales (CRE), 5 target other WHO priority pathogens (high and medium priority), and 1 was approved to treat MDR/extensively drug-resistant tuberculosis (XDR-TB) in combination with two other drugs, bedaquiline and linezolid. However, it is noted that lefamulin ( 48 ) is the first systemically administered pleuromutilin approved for human use (retapamulin was approved for human topical use and valnemulin and tiamulin for veterinary medicine). Traditional antibacterial agents in phase 3 clinical trials or with NDAs submitted against WHO priority pathogens. We identified two compounds, solithromycin and sulopenem, that have had NDAs submitted to the Japanese PMDA and U.S. FDA, respectively. Although after this review's cutoff date, in late July 2021, the FDA indicated that sulopenem will require further clinical trials to be undertaken ( 49 ). Three BLI/β-lactam combinations (durlobactam/sulbactam, taniborbactam/cefepime, and enmetazobactam/cefepime) and two compounds being developed to treat Neisseria gonorrhoeae infections (zoliflodacin and gepotidacin) are currently in phase 3 trials ( Table 3 ). Gepotidacin is also being evaluated in a phase 3 trial to treat urinary tract infections (UTIs). Updating the 2020 WHO report ( 10 ), nafithromycin is now being evaluated in a phase 3 trial (CTRI/2019/11/021964) for treatment of community-associated pneumonia (CAP) in India, and the phase 2/3 trial (NCT04505683) of benapenem in China has been completed for complicated urinary tract infections (cUTI), including acute pyelonephritis. Other phase 3 agents of interest: ATM-AVI and tebipenem pivoxil. There are two antibacterial agents in phase 3 development that did not meet this review's inclusion criteria (see "Scope and inclusion/exclusion criteria") that are noteworthy. The first is the aztreonam (monobactam-type β-lactam) and avibactam (BLI) combination (ATM-AVI), which was not included, as both components are previously approved drugs. ATM-AVI is being studied by Pfizer in a phase 3 trial (NCT03580044) to treat serious infection due to metallo-β-lactamase (MBL)-producing Gram-negative bacteria ( 50 , 51 ) with support from the Biomedical Advanced Research and Development Authority (BARDA), Innovative Medicines Initiative (IMI), and AbbVie, through their 2020 acquisition of Allergan. The second is the carbapenem prodrug tebipenem pivoxil ( 52 , 53 ), which was first approved for pediatric use in Japan in 2009 but has not been used elsewhere. Spero Therapeutics recently completed a phase 3 trial (NCT03788967) for tebipenem pivoxil (SPR994) as an oral treatment for Gram-negative cUTI and acute pyelonephritis infections. Clinically, the oral administration of tebipenem pivoxil could provide an alternative to i.v. administered carbapenems. Traditional antibacterial agents in phase 1 and 2 clinical trials being developed against WHO priority pathogens. We identified 2 compounds in phase 2, 1 in phase 1/2a, and 14 in phase 1 development ( Table 4 ). Since the 1 September 2020 cutoff date of the 2020 WHO pipeline report ( 10 ), two new polymyxin derivatives, MRX-8 ( 54 , 55 ) (NCT04649541) and QPX-9003 ( 56 , 57 ) (NCT04808414), started phase 1 trials in November 2020 and June 2021, respectively, as potential treatments for MDR Gram-negative pathogens. XNW-4107 is a BLI being that is developed in combination with imipenem and cilastatin that started a phase 1 in June 2021 (NCT04801043). The developer, Sinovent, has indicated that the XNW-4107 combination will be developed to treat patients with CRE, CRAB, and drug-resistant Pseudomonas aeruginosa ( 58 ). The structures of MRX-8, QPX-9003, and XNW-4107 have not been publicly disclosed. Finally, the rifamycin-quinolizinone hybrid TNP-2198 has moved into a Helicobacter pylori phase 1/2a trial ( 59 ). Traditional antibacterial agents against M. tuberculosis and nontuberculosis mycobacteria. There are currently 14 traditional antibacterial agents being evaluated in clinical trials against mycobacteria: 7 in phase 2, 2 in phase 1/2a, and 6 in phase 1. There also are two nontraditional antibacterial agents, CYT107 (NTM) and BVL-GSK098 ( M. tuberculosis ), under clinical investigation. In addition to these trials, there are also approximately 20 ongoing phase 3 trials ( 60 ) investigating new combinations and dosing regimens of previously approved TB drugs, which were not included in this review due to the selection criteria. Thirteen of the traditional antibacterial candidates target M. tuberculosis , and only one, SPR720, is in development for lung infections caused by the NTMs, Mycobacterium avium complex, and Mycobacterium abscessus ( Table 5 ). Eight of the 14 traditional antibacterial agents belong to new classes, and 9 have new antibacterial pharmacophores (see below). Since the 2020 WHO report ( 10 ) was released, two new compounds, TBAJ-587 (NCT04890535) and GSK 2556286 (NCT04472897), have entered phase 1 trials. TBAJ-587 is a bedaquiline analog with enhanced in vitro potency and a projected reduced cardiovascular liability ( 61 , 62 ). GSK 2556286 (GSK-286) acts directly on M. tuberculosis , and its mode of action has not been disclosed but has been proposed to involve cholesterol catabolism ( 63 , 64 ). In addition, TBI-166 (pyrifazimine) has moved to phase 2 (NCT04670120) and BTZ-043 has started a new phase 1/2 trial (NCT04044001). Traditional antibacterial agents being developed against C. difficile . There are currently five traditional antibacterial agents ( Table 6 ), one in phase 3 (ridinilazole) and four in phase 2 (DNV-3837, MGB-BP-3, ibezapolstat, and CRS3123), being developed to treat C. difficile . Antibacterial drugs used to treat C. difficile infections (CDI) are usually administered orally and absorbed poorly, as the infection is localized in the colon. This has encouraged the development of four new antibacterial classes and modes of action (ridinilazole, MGB-BP-3, ibezapolstat, and CRS3123; Fig. 5 ), while DNV-3837 is differentiated through its administration via i.v. infusion and is being targeted for use in patients who are unable to receive oral administration. Since the 2020 WHO report ( 10 ) was published, CRS3123 has started a phase 2 trial (NCT04781387). Nontraditional antibacterial agents in phase 3 clinical trials. We found two nontraditional antibacterial agents in the NDA/MAA phase and four in phase 3 development ( Table 7 ). Three are being developed to treat S. aureus infections: tosatoxumab is an MAb ( 65 ), exebacase is a phage-derived recombinant protein ( 66 ), and reltecimod is an immune modulator (CD28 T-lymphocyte receptor mimetic) ( 67 ). Three are being developed to treat C. difficile infections: SER-109 consists of purified Firmicutes spores ( 68 ), RBX2660 is a liquid suspension of screened donor fecal microbiota ( 69 , 70 ), and BB128 is a lyophilized donor fecal microbiota product. SER-109 has already successfully completed one phase 3 trial (NCT03183128) ( 71 ). Since the publication of the 2020 WHO pipeline report ( 10 ), a phase 3 trial (NCT03931941) has started to evaluate RBX2660 as a treatment of C. difficile , while BiomeBank has submitted an MAA to the Australian Therapeutic Goods Association (TGA) for BB128 as a potential treatment of recurrent C. difficile and ulcerative colitis ( 72 ). BiomeBank already has provisional approval for its use in Australia as a class 2 biologic. In addition, Atox Bio applied for an NDA in December 2020 for reltecimod as a potential supportive treatment for necrotizing soft tissue infections (NSTI) ( 73 ). Nontraditional antibacterial agents in phases 1 and 2. There are 10 nontraditional antibacterial agents in phase 2, 6 in phase 1/2a, 8 in phase 1 clinical trials, and 1 not disclosed by clinical phase ( Table 8 ). Combined with the 4 in phase 3 and 2 at the NDA/MAA stage, there are 31 nontraditional agents overall in clinical development. There are seven nontraditional antibacterial agents not detailed in the 2020 WHO report ( 10 ): CYT107, TRL1068, BVL-GSK098, and four bacteriophage products. CYT107 is a glycosylated recombinant human interleukin (IL-7) that is being tested in a phase 2 trial (NCT04154826) to evaluate its immunotherapeutic response in patients with NTM lung disease. CYT107 has been evaluated in other clinical trials, including a phase 2b trial (NCT02640807) that reported a 3- to 4-fold increase in the absolute lymphocyte count and in circulating CD4 + and CD8 + T cells with CYT107 in sepsis patients (predominantly secondary to pneumonia and abdominal infections) ( 74 ). TRL1068 is an MAb that binds to a DNABII epitope conserved across both Gram-positive and Gram-negative bacteria, which leads to bacterial biofilm disintegration, and is being evaluated in a phase 1 trial (NCT04763759) for prosthetic joint infections ( 75 , 76 ). BVL-GSK098 recently entered phase 1 (NCT04654143) and works through inactivation of a TetR-like repressor, EthR2, thereby enhancing ethionamide activation ( 77 ). BVL-GSK098 is intended to be used clinically in combination with ethionamide or prothionamide ( 78 ). Adaptive Phage Therapeutics are undertaking a phase 1/2 trial (APT.UTI.001, NCT04287478) to evaluate its PhageBank therapy in patients with UTI. There are also three other phage products, AP-PA02 (NCT04596319), YPT-01 (NCT04684641), and BX004-A (NCT05010577), being evaluated in phase 1/2 trials with cystic fibrosis patients with P. aeruginosa infections. For the phage-derived endolysin tonabacase (N-Rephasin SAL200), a new phase 1 trial has been initiated and it has been renamed LSVT-1701 ( 79 , 80 ). Antibacterial candidates in clinical trials with new pharmacophores. Although there have been significant efforts to identify antibacterial agents with new modes of action, most marketed antibacterial drugs still fall into four overarching mechanistic classes: inhibition of cell envelope biogenesis, DNA homeostasis, RNA homeostasis, and protein synthesis ( 81 ). A pharmacophore describes the part of a molecular structure that is responsible for a particular biological or pharmacological activity and is a key component, along with antibacterial activity differences, to decide whether an antibacterial agent belongs to a new class or subclass of antibiotics. It is possible to have antibacterial drugs and clinical candidates with the same mode of action but with different pharmacophores, which can have significant effects on biological activity, metabolism, and pharmacokinetics. For example, there are four compounds in clinical development that inhibit the M. tuberculosis cell wall synthesis enzyme decaprenylphosphoryl-β- d -ribose 2′-epimerase (DprE1) that have three distinct pharmacophores: benzothiazinone (BTZ), azaindole, and 3,4-dicarbostyril ( Fig. 5 , Fig. S2). There are 19 antibacterial agents with 18 new pharmacophores (macozinone and BTZ-043 are both BTZs) with seven inhibiting cell envelope synthesis, two acting at the protein synthesis level and five affecting DNA synthesis ( Fig. 5 , structures in supplemental information Fig. S1 to S3). Telacebec inhibits the mycobacterial respiratory system ( 82 , 83 ), which was first targeted by bedaquiline, via inhibition of the respiratory complex bc 1 ( 84 ). Half of the 18 new pharmacophores target mycobacteria, and 4 target C. difficile . Four of 19 antibacterial agents have new overarching modes of actions not previously exploited by marketed antibacterial drugs. Two target virulence: fluorothyazinon (phase 2, NCT03638830) and GSK 3882347 (phase 1, NCT04488770). Ftortiazinon inhibits the Gram-negative type III secretion system ( 85 ) and is being evaluated in a trial in combination with cefepime, and GSK 3882347 functions as an antagonist of the Gram-negative type 1 pilus adhesin (FimH) ( 86 ). BVL-GSK098 (phase 1, NCT04654143) ( 77 ) directly inhibits ethionamide-acquired resistance, while GSK 2556286 (phase 1 trial, NCT04472897) is proposed to involve M. tuberculosis cholesterol catabolism ( 63 , 64 ). Antibacterial agents that have halted or stopped clinical development. Drug development is inherently risky, and it is not uncommon for clinical development programs to be terminated or halted. The most common reasons for stopping development include lack of clinical efficacy, off-target toxicity, and commercial considerations ( 87 ). While antibacterial agents can drop out of the pipeline due to efficacy and resistance issues, it is more common to be due to toxicity and commercial concerns ( 16 , 41 , 88 ). A list of antibacterial compounds and nontraditional moieties whose development has been terminated or halted are listed in the supplemental information (Table S1). Current pipeline analysis. There are 76 antibacterial agents in clinical development using this review's inclusion criteria, which are divided into 45 traditional and 31 nontraditional antibacterial agents ( Fig. 3 ). Seventy-nine percent (60/76) of the antibacterial drug candidates are in phase 1 ( 28 ) and phase 2 ( 32 ), but as expected, this number falls away for late-stage development agents (12 in phase 3 and 4 NDA/MAA). This relatively low number of candidates in the later stages of drug development generally reflects the usual level of attrition in the pipeline, which is caused by several factors, including lack of efficacy, unacceptable toxicity, and market factors ( 87 , 88 ). The number of early development candidates is encouraging and reinforces research efforts and recent funding that have been invested into discovery and preclinical development. For example, CARB-X has funded 92 early-stage R&D drug and diagnostics projects since its inception 5 years ago ( 89 ), while GARDP ( 90 ) has signed license and codevelopment agreements with the companies that are developing two innovative products, zoliflodacin (target: gonorrhea) and cefepime-taniborbactam (target: cUTI). In addition, the AMR Action Fund plans to help support late-stage development of 3 to 4 new antibacterial candidates by 2030, which could help increase the number of new approvals ( 17 , 91 ). The fact that there are 76 antibacterial candidates currently being evaluated in clinical trials is promising, but it needs to be asked whether these agents will address future clinical needs. To evaluate this, pipeline agents are analyzed here for activity versus each of the major pathogen categories. The potential impact of nontraditional antibacterial agents. Nontraditional antibacterial agents have the potential to improve the clinical outcomes using alternative mechanisms to traditional antibacterial drugs. Although the number of nontraditional antibacterial agents entering clinical trials continues to increase, only one has been approved that successfully completed phase 3 trials: bezlotoxumab ( 34 ), which is a C. difficile toxin B-binding MAb. One of the main issues facing nontraditional agent developers has been clinical trial design ( 32 , 33 ), except for adjunctive agents that are being developed in combination with standard of care drugs. WHO priority pathogens. A total of 26/76 (34%) and 16/76 (21%) antibacterial agents are being developed to target the critical- and high/medium-priority WHO priority pathogens, respectively. This represents 55% of the total pipeline, and it is encouraging to observe product development being directed against the key pathogens. For the traditional antibacterial agents, only two compounds, gepotidacin and zoliflodacin, which are both new chemical classes, target priority pathogens ( E. coli critical and N. gonorrhoeae high). β-Lactams, with and without BLI inhibitors, account for a majority of the other antibacterial agents in development against critical-priority pathogens. Although there are six nontraditional agents in late-stage clinical development ( Table 7 ), only three of these target the high-priority pathogen S. aureus ; although there is need for innovative drugs to treat S. aureus infection, there are already several treatment options currently available to clinicians. There are nine agents in phase 1 and 2 trials being developed against critical WHO priority pathogens: four bacteriophages, one CRISP-Cas3 enhanced phage (LBP-EC01), two antivirulence (ftortiazinon and GSK-3882347), MAb-like recombinant protein (LMN-01), and an alginate oligosaccharide fragment (OligoG). Antibacterial developers and funders need to continue to develop pathways that allow the most promising of these antibacterial agents to rapidly progress through to late-stage clinical trials and beyond. TB and NTM. There are 16/76 (21%) candidates being developed to treat mycobacterial infections (14 TB and 2 NTM), which includes nine small molecules with new pharmacophores ( Fig. 5 , Fig. S2). Despite the considerable challenges associated with TB drug development ( 92 ), progress has been accelerated from sustained funding and guidance by organizations such as the TB Alliance and the Gates Foundation ( 93 ). The next challenge will be to move the most promising candidates through the pipeline and select and clinically evaluate the optimal drug regimens. C. difficile . There are also 15/76 (20%) agents in development to treat C. difficile infections, with 10 of these being nontraditional and 5 traditional antibacterial agents. Four of the five traditional antibacterial agents would be new classes if approved (Fig. S3). Of the 10 nontraditional agents, 7 are biotherapeutic products, 1 is an MAb (IM-01), and 2 are antibiotic inactivators (ribaxamase and DAV132). There are already several C. difficile drugs on the market, and it will be interesting to monitor the impact of any new approvals of small molecular antibacterial drugs and nontraditional biotherapeutic products and the effect that these will have on clinical practice and the market. Broad-spectrum agents active against Gram-negative and Gram-positive bacteria. Four of the 76 (5%) are nontraditional antibacterial agents with broad-spectrum antibacterial effects, which was achieved through a variety of mechanisms: a recombinant gelsolin protein Rhu-pGSN boosts the immune system, the MAb TRL1068 disrupts biofilms, the synthetic glycan KB109 modulates the gut microbiome composition and metabolic output, and the liposomal agent CAL02 captures and neutralizes bacterial toxins. Antibacterial drugs approved since 2017. Twelve new antibacterial drugs ( Table 2 , Fig. 2 ) have been approved since the WHO's first analysis of the clinical antibacterial pipeline in 2017 ( 7 ). The most recent approval was in China in June 2021 for the oxazolidinone contezolid as a treatment for complicated skin and soft tissue infections (cSSTI) caused by Gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) ( 45 ). Only 1 of the 12 approved antibacterial drugs, the boronate β-lactamase inhibitor (BLI) vaborbactam, which is used in combination with meropenem, has a new antibacterial drug-related pharmacophore. Also, the cephalosporin cefiderocol ( 46 , 47 ) is noteworthy, as it is the first marketed β-lactam (cephalosporin) that has an iron-chelating siderophore incorporated into the structure, which facilitates Gram-negative bacteria outer membrane entry. The other 10 are members of previously approved antibacterial classes: three fluoroquinolones (delafloxacin, lascufloxacin, and levonadifloxacin), two tetracyclines (eravacycline and omadacycline), one aminoglycoside (plazomicin), one pleuromutilin (lefamulin), one nitroimidazole (pretomanid), one diazabicyclooctane (DBO) BLI (relebactam), and one oxazolidinone (contezolid). Of the 12 new antibacterial drugs, 6 target carbapenem-resistant Enterobacterales (CRE), 5 target other WHO priority pathogens (high and medium priority), and 1 was approved to treat MDR/extensively drug-resistant tuberculosis (XDR-TB) in combination with two other drugs, bedaquiline and linezolid. However, it is noted that lefamulin ( 48 ) is the first systemically administered pleuromutilin approved for human use (retapamulin was approved for human topical use and valnemulin and tiamulin for veterinary medicine). Traditional antibacterial agents in phase 3 clinical trials or with NDAs submitted against WHO priority pathogens. We identified two compounds, solithromycin and sulopenem, that have had NDAs submitted to the Japanese PMDA and U.S. FDA, respectively. Although after this review's cutoff date, in late July 2021, the FDA indicated that sulopenem will require further clinical trials to be undertaken ( 49 ). Three BLI/β-lactam combinations (durlobactam/sulbactam, taniborbactam/cefepime, and enmetazobactam/cefepime) and two compounds being developed to treat Neisseria gonorrhoeae infections (zoliflodacin and gepotidacin) are currently in phase 3 trials ( Table 3 ). Gepotidacin is also being evaluated in a phase 3 trial to treat urinary tract infections (UTIs). Updating the 2020 WHO report ( 10 ), nafithromycin is now being evaluated in a phase 3 trial (CTRI/2019/11/021964) for treatment of community-associated pneumonia (CAP) in India, and the phase 2/3 trial (NCT04505683) of benapenem in China has been completed for complicated urinary tract infections (cUTI), including acute pyelonephritis. Other phase 3 agents of interest: ATM-AVI and tebipenem pivoxil. There are two antibacterial agents in phase 3 development that did not meet this review's inclusion criteria (see "Scope and inclusion/exclusion criteria") that are noteworthy. The first is the aztreonam (monobactam-type β-lactam) and avibactam (BLI) combination (ATM-AVI), which was not included, as both components are previously approved drugs. ATM-AVI is being studied by Pfizer in a phase 3 trial (NCT03580044) to treat serious infection due to metallo-β-lactamase (MBL)-producing Gram-negative bacteria ( 50 , 51 ) with support from the Biomedical Advanced Research and Development Authority (BARDA), Innovative Medicines Initiative (IMI), and AbbVie, through their 2020 acquisition of Allergan. The second is the carbapenem prodrug tebipenem pivoxil ( 52 , 53 ), which was first approved for pediatric use in Japan in 2009 but has not been used elsewhere. Spero Therapeutics recently completed a phase 3 trial (NCT03788967) for tebipenem pivoxil (SPR994) as an oral treatment for Gram-negative cUTI and acute pyelonephritis infections. Clinically, the oral administration of tebipenem pivoxil could provide an alternative to i.v. administered carbapenems. Traditional antibacterial agents in phase 1 and 2 clinical trials being developed against WHO priority pathogens. We identified 2 compounds in phase 2, 1 in phase 1/2a, and 14 in phase 1 development ( Table 4 ). Since the 1 September 2020 cutoff date of the 2020 WHO pipeline report ( 10 ), two new polymyxin derivatives, MRX-8 ( 54 , 55 ) (NCT04649541) and QPX-9003 ( 56 , 57 ) (NCT04808414), started phase 1 trials in November 2020 and June 2021, respectively, as potential treatments for MDR Gram-negative pathogens. XNW-4107 is a BLI being that is developed in combination with imipenem and cilastatin that started a phase 1 in June 2021 (NCT04801043). The developer, Sinovent, has indicated that the XNW-4107 combination will be developed to treat patients with CRE, CRAB, and drug-resistant Pseudomonas aeruginosa ( 58 ). The structures of MRX-8, QPX-9003, and XNW-4107 have not been publicly disclosed. Finally, the rifamycin-quinolizinone hybrid TNP-2198 has moved into a Helicobacter pylori phase 1/2a trial ( 59 ). Traditional antibacterial agents against M. tuberculosis and nontuberculosis mycobacteria. There are currently 14 traditional antibacterial agents being evaluated in clinical trials against mycobacteria: 7 in phase 2, 2 in phase 1/2a, and 6 in phase 1. There also are two nontraditional antibacterial agents, CYT107 (NTM) and BVL-GSK098 ( M. tuberculosis ), under clinical investigation. In addition to these trials, there are also approximately 20 ongoing phase 3 trials ( 60 ) investigating new combinations and dosing regimens of previously approved TB drugs, which were not included in this review due to the selection criteria. Thirteen of the traditional antibacterial candidates target M. tuberculosis , and only one, SPR720, is in development for lung infections caused by the NTMs, Mycobacterium avium complex, and Mycobacterium abscessus ( Table 5 ). Eight of the 14 traditional antibacterial agents belong to new classes, and 9 have new antibacterial pharmacophores (see below). Since the 2020 WHO report ( 10 ) was released, two new compounds, TBAJ-587 (NCT04890535) and GSK 2556286 (NCT04472897), have entered phase 1 trials. TBAJ-587 is a bedaquiline analog with enhanced in vitro potency and a projected reduced cardiovascular liability ( 61 , 62 ). GSK 2556286 (GSK-286) acts directly on M. tuberculosis , and its mode of action has not been disclosed but has been proposed to involve cholesterol catabolism ( 63 , 64 ). In addition, TBI-166 (pyrifazimine) has moved to phase 2 (NCT04670120) and BTZ-043 has started a new phase 1/2 trial (NCT04044001). Traditional antibacterial agents being developed against C. difficile . There are currently five traditional antibacterial agents ( Table 6 ), one in phase 3 (ridinilazole) and four in phase 2 (DNV-3837, MGB-BP-3, ibezapolstat, and CRS3123), being developed to treat C. difficile . Antibacterial drugs used to treat C. difficile infections (CDI) are usually administered orally and absorbed poorly, as the infection is localized in the colon. This has encouraged the development of four new antibacterial classes and modes of action (ridinilazole, MGB-BP-3, ibezapolstat, and CRS3123; Fig. 5 ), while DNV-3837 is differentiated through its administration via i.v. infusion and is being targeted for use in patients who are unable to receive oral administration. Since the 2020 WHO report ( 10 ) was published, CRS3123 has started a phase 2 trial (NCT04781387). Nontraditional antibacterial agents in phase 3 clinical trials. We found two nontraditional antibacterial agents in the NDA/MAA phase and four in phase 3 development ( Table 7 ). Three are being developed to treat S. aureus infections: tosatoxumab is an MAb ( 65 ), exebacase is a phage-derived recombinant protein ( 66 ), and reltecimod is an immune modulator (CD28 T-lymphocyte receptor mimetic) ( 67 ). Three are being developed to treat C. difficile infections: SER-109 consists of purified Firmicutes spores ( 68 ), RBX2660 is a liquid suspension of screened donor fecal microbiota ( 69 , 70 ), and BB128 is a lyophilized donor fecal microbiota product. SER-109 has already successfully completed one phase 3 trial (NCT03183128) ( 71 ). Since the publication of the 2020 WHO pipeline report ( 10 ), a phase 3 trial (NCT03931941) has started to evaluate RBX2660 as a treatment of C. difficile , while BiomeBank has submitted an MAA to the Australian Therapeutic Goods Association (TGA) for BB128 as a potential treatment of recurrent C. difficile and ulcerative colitis ( 72 ). BiomeBank already has provisional approval for its use in Australia as a class 2 biologic. In addition, Atox Bio applied for an NDA in December 2020 for reltecimod as a potential supportive treatment for necrotizing soft tissue infections (NSTI) ( 73 ). Nontraditional antibacterial agents in phases 1 and 2. There are 10 nontraditional antibacterial agents in phase 2, 6 in phase 1/2a, 8 in phase 1 clinical trials, and 1 not disclosed by clinical phase ( Table 8 ). Combined with the 4 in phase 3 and 2 at the NDA/MAA stage, there are 31 nontraditional agents overall in clinical development. There are seven nontraditional antibacterial agents not detailed in the 2020 WHO report ( 10 ): CYT107, TRL1068, BVL-GSK098, and four bacteriophage products. CYT107 is a glycosylated recombinant human interleukin (IL-7) that is being tested in a phase 2 trial (NCT04154826) to evaluate its immunotherapeutic response in patients with NTM lung disease. CYT107 has been evaluated in other clinical trials, including a phase 2b trial (NCT02640807) that reported a 3- to 4-fold increase in the absolute lymphocyte count and in circulating CD4 + and CD8 + T cells with CYT107 in sepsis patients (predominantly secondary to pneumonia and abdominal infections) ( 74 ). TRL1068 is an MAb that binds to a DNABII epitope conserved across both Gram-positive and Gram-negative bacteria, which leads to bacterial biofilm disintegration, and is being evaluated in a phase 1 trial (NCT04763759) for prosthetic joint infections ( 75 , 76 ). BVL-GSK098 recently entered phase 1 (NCT04654143) and works through inactivation of a TetR-like repressor, EthR2, thereby enhancing ethionamide activation ( 77 ). BVL-GSK098 is intended to be used clinically in combination with ethionamide or prothionamide ( 78 ). Adaptive Phage Therapeutics are undertaking a phase 1/2 trial (APT.UTI.001, NCT04287478) to evaluate its PhageBank therapy in patients with UTI. There are also three other phage products, AP-PA02 (NCT04596319), YPT-01 (NCT04684641), and BX004-A (NCT05010577), being evaluated in phase 1/2 trials with cystic fibrosis patients with P. aeruginosa infections. For the phage-derived endolysin tonabacase (N-Rephasin SAL200), a new phase 1 trial has been initiated and it has been renamed LSVT-1701 ( 79 , 80 ). Antibacterial candidates in clinical trials with new pharmacophores. Although there have been significant efforts to identify antibacterial agents with new modes of action, most marketed antibacterial drugs still fall into four overarching mechanistic classes: inhibition of cell envelope biogenesis, DNA homeostasis, RNA homeostasis, and protein synthesis ( 81 ). A pharmacophore describes the part of a molecular structure that is responsible for a particular biological or pharmacological activity and is a key component, along with antibacterial activity differences, to decide whether an antibacterial agent belongs to a new class or subclass of antibiotics. It is possible to have antibacterial drugs and clinical candidates with the same mode of action but with different pharmacophores, which can have significant effects on biological activity, metabolism, and pharmacokinetics. For example, there are four compounds in clinical development that inhibit the M. tuberculosis cell wall synthesis enzyme decaprenylphosphoryl-β- d -ribose 2′-epimerase (DprE1) that have three distinct pharmacophores: benzothiazinone (BTZ), azaindole, and 3,4-dicarbostyril ( Fig. 5 , Fig. S2). There are 19 antibacterial agents with 18 new pharmacophores (macozinone and BTZ-043 are both BTZs) with seven inhibiting cell envelope synthesis, two acting at the protein synthesis level and five affecting DNA synthesis ( Fig. 5 , structures in supplemental information Fig. S1 to S3). Telacebec inhibits the mycobacterial respiratory system ( 82 , 83 ), which was first targeted by bedaquiline, via inhibition of the respiratory complex bc 1 ( 84 ). Half of the 18 new pharmacophores target mycobacteria, and 4 target C. difficile . Four of 19 antibacterial agents have new overarching modes of actions not previously exploited by marketed antibacterial drugs. Two target virulence: fluorothyazinon (phase 2, NCT03638830) and GSK 3882347 (phase 1, NCT04488770). Ftortiazinon inhibits the Gram-negative type III secretion system ( 85 ) and is being evaluated in a trial in combination with cefepime, and GSK 3882347 functions as an antagonist of the Gram-negative type 1 pilus adhesin (FimH) ( 86 ). BVL-GSK098 (phase 1, NCT04654143) ( 77 ) directly inhibits ethionamide-acquired resistance, while GSK 2556286 (phase 1 trial, NCT04472897) is proposed to involve M. tuberculosis cholesterol catabolism ( 63 , 64 ). Antibacterial agents that have halted or stopped clinical development. Drug development is inherently risky, and it is not uncommon for clinical development programs to be terminated or halted. The most common reasons for stopping development include lack of clinical efficacy, off-target toxicity, and commercial considerations ( 87 ). While antibacterial agents can drop out of the pipeline due to efficacy and resistance issues, it is more common to be due to toxicity and commercial concerns ( 16 , 41 , 88 ). A list of antibacterial compounds and nontraditional moieties whose development has been terminated or halted are listed in the supplemental information (Table S1). Current pipeline analysis. There are 76 antibacterial agents in clinical development using this review's inclusion criteria, which are divided into 45 traditional and 31 nontraditional antibacterial agents ( Fig. 3 ). Seventy-nine percent (60/76) of the antibacterial drug candidates are in phase 1 ( 28 ) and phase 2 ( 32 ), but as expected, this number falls away for late-stage development agents (12 in phase 3 and 4 NDA/MAA). This relatively low number of candidates in the later stages of drug development generally reflects the usual level of attrition in the pipeline, which is caused by several factors, including lack of efficacy, unacceptable toxicity, and market factors ( 87 , 88 ). The number of early development candidates is encouraging and reinforces research efforts and recent funding that have been invested into discovery and preclinical development. For example, CARB-X has funded 92 early-stage R&D drug and diagnostics projects since its inception 5 years ago ( 89 ), while GARDP ( 90 ) has signed license and codevelopment agreements with the companies that are developing two innovative products, zoliflodacin (target: gonorrhea) and cefepime-taniborbactam (target: cUTI). In addition, the AMR Action Fund plans to help support late-stage development of 3 to 4 new antibacterial candidates by 2030, which could help increase the number of new approvals ( 17 , 91 ). The fact that there are 76 antibacterial candidates currently being evaluated in clinical trials is promising, but it needs to be asked whether these agents will address future clinical needs. To evaluate this, pipeline agents are analyzed here for activity versus each of the major pathogen categories. The potential impact of nontraditional antibacterial agents. Nontraditional antibacterial agents have the potential to improve the clinical outcomes using alternative mechanisms to traditional antibacterial drugs. Although the number of nontraditional antibacterial agents entering clinical trials continues to increase, only one has been approved that successfully completed phase 3 trials: bezlotoxumab ( 34 ), which is a C. difficile toxin B-binding MAb. One of the main issues facing nontraditional agent developers has been clinical trial design ( 32 , 33 ), except for adjunctive agents that are being developed in combination with standard of care drugs. WHO priority pathogens. A total of 26/76 (34%) and 16/76 (21%) antibacterial agents are being developed to target the critical- and high/medium-priority WHO priority pathogens, respectively. This represents 55% of the total pipeline, and it is encouraging to observe product development being directed against the key pathogens. For the traditional antibacterial agents, only two compounds, gepotidacin and zoliflodacin, which are both new chemical classes, target priority pathogens ( E. coli critical and N. gonorrhoeae high). β-Lactams, with and without BLI inhibitors, account for a majority of the other antibacterial agents in development against critical-priority pathogens. Although there are six nontraditional agents in late-stage clinical development ( Table 7 ), only three of these target the high-priority pathogen S. aureus ; although there is need for innovative drugs to treat S. aureus infection, there are already several treatment options currently available to clinicians. There are nine agents in phase 1 and 2 trials being developed against critical WHO priority pathogens: four bacteriophages, one CRISP-Cas3 enhanced phage (LBP-EC01), two antivirulence (ftortiazinon and GSK-3882347), MAb-like recombinant protein (LMN-01), and an alginate oligosaccharide fragment (OligoG). Antibacterial developers and funders need to continue to develop pathways that allow the most promising of these antibacterial agents to rapidly progress through to late-stage clinical trials and beyond. TB and NTM. There are 16/76 (21%) candidates being developed to treat mycobacterial infections (14 TB and 2 NTM), which includes nine small molecules with new pharmacophores ( Fig. 5 , Fig. S2). Despite the considerable challenges associated with TB drug development ( 92 ), progress has been accelerated from sustained funding and guidance by organizations such as the TB Alliance and the Gates Foundation ( 93 ). The next challenge will be to move the most promising candidates through the pipeline and select and clinically evaluate the optimal drug regimens. C. difficile . There are also 15/76 (20%) agents in development to treat C. difficile infections, with 10 of these being nontraditional and 5 traditional antibacterial agents. Four of the five traditional antibacterial agents would be new classes if approved (Fig. S3). Of the 10 nontraditional agents, 7 are biotherapeutic products, 1 is an MAb (IM-01), and 2 are antibiotic inactivators (ribaxamase and DAV132). There are already several C. difficile drugs on the market, and it will be interesting to monitor the impact of any new approvals of small molecular antibacterial drugs and nontraditional biotherapeutic products and the effect that these will have on clinical practice and the market. Broad-spectrum agents active against Gram-negative and Gram-positive bacteria. Four of the 76 (5%) are nontraditional antibacterial agents with broad-spectrum antibacterial effects, which was achieved through a variety of mechanisms: a recombinant gelsolin protein Rhu-pGSN boosts the immune system, the MAb TRL1068 disrupts biofilms, the synthetic glycan KB109 modulates the gut microbiome composition and metabolic output, and the liposomal agent CAL02 captures and neutralizes bacterial toxins. CONCLUSION Since its release in 2017, the WHO's priority pathogen list ( Fig. 1 ) has become a focus for antibacterial R&D and stewardship initiatives. The WHO also started analyzing the antibacterial pipeline in 2017, and since then, only vaborbactam (boronate BLI) of the 12 approved antibacterial drugs ( Table 2 , Fig. 2 ) is not a derivative of a previously approved class. Importantly, vaborbactam is used in combination with meropenem to treat Enterobacterales infections, which are critical-priority pathogens. The cephalosporin derivative cefiderocol is also noteworthy, as it displays activity against all three critical-priority pathogens, CRAB, CRPA, and CRE, regardless of the carbapenemase mechanism, and is the first marketed antibacterial drug that incorporates an iron-chelating siderophore. Renewed focus to identify new antibacterial drugs against MDR bacteria, combined with several recent financing mechanisms, has helped to increase the number of traditional and nontraditional antibacterial agents moving through the preclinical ( 94 ) and clinical development pipelines ( 41 , 90 , 95 , 96 ). Despite a total of 76 antibacterial candidates (45 traditional and 31 nontraditional) being evaluated in clinical trials on 30 June 2021, our analysis indicated that there were still relatively few clinically differentiated antibacterial agents in late-stage clinical development, especially against critical-priority pathogens. In addition, we identified 18 new antibacterial pharmacophores, but only 2 had activity against priority pathogens with most targeted mycobacteria and C. difficile . It is important to try to keep on identifying and developing antibacterial agents with new modes of action to try to slow down antibacterial drug resistance. Furthermore, we believe that future antibacterial R&D should focus on the development of innovative and clinically differentiated candidates that have clear and feasible progression pathways right through development and onto the market. There needs to be a development focus on quality over quantity, especially with limited development resources, ever-increasing numbers of MDR infections, and potential return-on-investment issues associated with development, manufacture, regulatory compliance, and distribution costs. Formidable challenges that we believe remain that still need further attention are as follows: • Difficulty in discovering novel antibacterial leads with selective activity against MDR bacteria that are nontoxic and have suitable pharmacokinetic and pharmacodynamic properties, especially with new modes of action • Current unmet medical need for new drugs to treat drug-resistant A. baumannii (e.g. CRAB) and P. aeruginosa (e.g. CRPA) infections • Development of antibacterial agents for use in neonates and children • Development of efficient progression pathways for nontraditional antibacterial candidates through the manufacturing, clinical trials, and approval processes • Difficulties in optimal trial design and selection of relevant intended target population • Sustained advocacy for strong and sustainable political support and governmental commitments to promote R&D and help developers overcome economic, scientific, and technical barriers • Implementation of business models that improve the current market dynamics with a focus on developing and securing approval of truly innovative and clinically differentiated antibacterial treatments

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5825070/

Neither Lys- and DAP-type peptidoglycans stimulate mouse or human innate immune cells via Toll-like receptor 2

Peptidoglycan (PGN), a major component of bacterial cell walls, is a pathogen-associated molecular pattern (PAMP) that causes innate immune cells to produce inflammatory cytokines that escalate the host response during infection. In order to better understand the role of PGN in infection, we wanted to gain insight into the cellular receptor for PGN. Although the receptor was initially identified as Toll-like receptor 2 (TLR2), this receptor has remained controversial and other PGN receptors have been reported. We produced PGN from live cultures of Bacillus anthracis and Staphylococcus aureus and tested samples of PGN isolated during the purification process to determine at what point TLR2 activity was removed, if at all. Our results indicate that although live B . anthracis and S . aureus express abundant TLR2 ligands, highly-purified PGN from either bacterial source is not recognized by TLR2. Introduction During bacterial sepsis, the bacterial-produced pathogen-associated molecular patterns (PAMPs) are responsible for activation of the complement and coagulation systems, contributing to organ failure and death of the host [ 1 , 2 ]. Peptidoglycans (PGN), a disaccharide polymer with peptide cross-linkers and present in bacterial cell walls, have been established as important bacterial PAMPs: PGN has been shown to induce inflammatory mediators from human innate immune cells, to stimulate platelet aggregation and prothrombinase activity, and to stimulate complement consumption [ 3 – 5 ]. Animals challenged with PGN show features of sepsis pathology [ 6 ], suggesting that PGN is an important PAMP in this condition. PGN is a large glycan polymer composed of alternating units of N-acetylglucosamine (GlcNac) and N-acetylmuramic acid (MurNac) residues joined by stem peptide units consisting of 4 or 5 L- and D-amino acids [ 7 ]. The stem peptides may be linked to each other or they may be joined by interpeptide bridges that have a different amino acid composition than the stem peptides. The stem peptides and bridges vary between species in their length and composition. For example, B . anthracis PGN has a stem peptide consisting of three amino acids including diaminopimelic acid (DAP) [ 3 ], while S . aureus contains a lysine residue in the DAP position and has a pentapeptide stem linked by a pentaglycyl bridge [ 8 ]. The relevance of these differences in the proinflammatory properties of PGN in mammals has not been studied. In order to understand the role of PGN during the pathophysiology of bacterial infections, it is important to know the cellular receptor(s) that binds PGN. The PGN receptor is responsible for initiation of the inflammatory response and could contribute to coagulopathy in sepsis. We reported that human monocytes and neutrophils produce inflammatory cytokines [ 3 , 5 , 9 ] and that platelets exhibit a procoagulant state [ 4 ] in response to PGN. The PGN receptor on monocytes was initially reported to be a pattern recognition receptor (PRR), specifically Toll-like receptor 2 (TLR2) [ 10 , 11 ]. However, it has also been reported that highly purified PGN does not stimulate through TLR2, or that other receptors may also play a prominent role [ 12 ]. Since PGN is purified from Gram-positive bacterial cell walls, preparations of PGN may contain cell wall impurities such as lipoteichoic acids (LTA), wall teichoic acids (WTA), lipoproteins and glycolipids. These impurities can all stimulate an inflammatory response in immune cells, and may be the reported stimulant in PGN preparations. Alternative means of PGN recognition have been described. We reported evidence for IgG/Fc receptors (FcγR) as an important receptor that indirectly recognizes PGN [ 3 , 9 , 13 ]. We showed that normal human plasma contains IgG that opsonizes PGN, that plasma or anti-PGN antibodies were required for PGN binding to monocytes and neutrophils and that PGN was bound and phagocytosed by HEK293 cells transfected with FcγRIIA in the presence of human plasma. These results suggested a model in which polymeric PGN is opsonized by IgG, allowing FcγR internalization, trafficking to lysosomes, and lysosomal degradation to NOD ligands [ 3 , 9 , 13 ]. We hypothesized that the conflicting data on PGN as a TLR2 ligand may either be due to differences in purification procedures where other cell wall components might contaminate preparations or due to intrinsic structural differences within PGN from different bacterial species. Here, we purified PGN of the Lys-type from S . aureus and of the DAP-type from B . anthracis . We compared the TLR2 induction activity throughout the steps of the purification process to determine whether any steps had a significant influence on the amount of TLR2 ligands in the sample. We found that B . anthracis and S . aureus bacteria contain TLR2 ligands, but purified PGN from either source is not a TLR2 ligand. We also found that S . aureus PGN contains palmitoylated peptide(s) that cannot be removed by our PGN purification procedures. Materials and methods Materials Brefeldin A, anti-human TNFα-Allophycocyanin (APC; clone MAb11), isotype control mouse IgG1-APC (clone P3.6.2.8.1), anti-human CD14-PeCy7 (clone 61D3), anti-mouse CD11b-PE (clone M1/70), rat IgG2b K isotype control (clone eB149/10H5), anti-mouse TNFα-APC (clone MP6-T22) and rat IgG1 K Isotype Control-APC (clone eBRG1) were purchased from eBioscience. Cytochalasin D was purchased from Sigma. Low IgG fetal bovine sera (FBS) was purchased from Gibco and used throughout. Human IgG was obtained from Talecris (Immune Globulin Human GamaSTAN S/D). Proteinase K (≥20 U/mg) was obtained from Life Technologies. S . aureus strain MN8 was from BEI Resources and B . anthracis strain delta Stern was obtained from S. Kurosawa, Boston University School of Medicine, Boston MA. The Newman and ∆ lgt strains of S . aureus were provided by Dr. Dominique Missiakas of the University of Chicago, Division of Biological Sciences. Ethics statement Studies on primary human cells and blood products were performed according to protocols approved by the Oklahoma Medical Research Foundation Institutional Review. All animal studies were approved by the Oklahoma Medical Research Foundation Institutional Animal Care and Use Committee. PGN preparation Non-pyrogenic plastic labware was used throughout the procedure after harvesting the bacteria. Endotoxin-free water was used throughout for all media, solutions and washes with water. Centrifugations were at 15,000g for 10 min at room temperature, unless otherwise stated. B . anthracis Delta Sterne strain was grown to stationary phase in 4–500 ml cultures of Trypticase TM Soy Broth (BD Biosciences). The cultures were harvested by centrifugation at 4°C and washed once in water. The bacterial pellets were combined in eight 50 ml polysulfone Oak Ridge centrifuge tubes and resuspended in 30 ml 8% SDS, boiled for 30 min, cooled, centrifuged and the process was repeated. After the second SDS extraction, the cell wall extract was centrifuged and the pellets were washed 4 times in 30 ml water to remove the SDS. The pellets were resuspended in 6 mls water. Tryptic soy agar plates were streaked with 10 μl of the preparation and incubated for 24 hours at 37°C to ensure there were no viable cells before continuing the procedure. The cell wall extract was digested with DNase and RNase by resuspending each pellet in 3 ml of solution containing 331 U/ml of DNase 1 (Invitrogen), 42 U/ml RNase A and 1667 U/ml RNase T1 (Ambion RNase cocktail) in DNase 1 buffer (10 mM Tris base, 0.5 mM CaCl 2 , 2.5 mM MgCl 2 , pH 7.5) and incubating with rocking for 30 min at room temperature and centrifuged. The DNase/RNase digestion was repeated, the material was centrifuged and boiled in SDS solution as above to remove the digestive enzymes. The material was centrifuged and washed repeatedly (4–5 times) with water. After the last wash, the material was resuspend in 30 ml ice cold 48% hydrofluoric acid (HF), vortexed, incubated with rocking for 6 hours at 4°C and centrifuged. The pellets were resuspended in 30 ml of ice cold HF, incubated overnight with rocking at 4°C and centrifuged. The extract was centrifuged and washed four times with water. The pellet was resuspended in denaturing buffer (50 mM Tris base, 6M guanidine HCl, 25 mM dithiothreitol, pH 8) and incubated for 1 h at 60°C. The extract was cooled to room temperature and the denaturing buffer was removed by centrifugation. The pellet was incubated in 0.8 M Iodoacetamide, 0.8M Tris base, pH 8 in the dark for 15 min at room temperature and then washed with water. We have prepared PGN including denaturation and alkylation by iodoacetamide or not including these steps. The final products are identical in biological activity. Our current PGN purification protocol does not include denaturation and alkylation. The extract was resuspended in 30 ml proteinase K buffer (50 mM Tris base, 1 M guanidine HCl, 5 mM CaCl 2 , pH 7.5), and proteinase K was added to 0.67 mg/ml to each tube and incubated at 50°C on an orbital shaker overnight. The digested material was centrifuged and washed with water 3–4 times. The PGN was sonicated using the Misonix Ultrasonic Liquid Processor (Sonicator S-400) set at amplitude 50 with two 2.5 min pulses, separated by one minute without pulse. After sonication, the PGN was pooled, evenly distributed in Oakridge tubes, centrifuged, and resuspended in 30 ml 8% SDS. The now pure PGN was boiled for 30 min, cooled, washed with water and repeated. The PGN suspension was transferred to several pre-weighed endotoxin free microfuge tubes for drying. The suspension was dried in a speed vacuum and the microfuge tubes were weighed to obtain the weight of the PGN. The PGN was resuspend at a concentration of 10 mg/ml in water. The phosphate content of PGN preparations was measured as described [ 14 ]. The PGN was stored at -20°C for long-term storage or 4°C for short-term storage. Assay of TLR2 and TLR4 activity of PGN TLR2 and TLR4 activity assays were performed with reporter HEK293 cells, stably expressing human TLR2 and CD14 (HEK-TLR2), or human TLR4 and MD-2/CD14 (HEK-TRL4), both from Invivogen. Both cell lines also stably express a secreted embryonic alkaline phosphatase (SEAP) reporter. The SEAP reporter gene is under the control of the IFNβ minimal promoter fused to five NF-κB and AP-1-binding sites. SEAP is produced when the specific TLR receptor is stimulated by its ligand, through activation of NF-kB and AP-1. Levels of SEAP are determined by spectrophotometrically at 630 nm after chromogenic conversion of the HEK-Blue Detection Medium (Invivogen). The cells were grown to 80–90% confluence in complete Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% heat-inactivated FBS, penicillin (1,000 units/ml) and (streptomycin 100 μg/ml), 2 mM GlutaMAX (L-alanyl-L-glutamine dipeptide; Gibco) HEK-Blue selection (InvivoGen), and 100 μg/ml Normocin (InvivoGen), at 37°C in a 5% saturated CO 2 atmosphere. To perform the assay, the cells were resuspended in HEK-Blue Detection Medium and plated in 96 well non-tissue culture treated plates at a concentration of 30–40,000 cells per well for HEK-TLR2, and 50–60,000 cells per well for HEK-TLR4. The cells were incubated with control stimulants, test substances as indicated, vehicle (endotoxin-free water), or media only, at 37°C in a 5% CO 2 for 16 hours. The optical density at 630 nm was quantified on an Epoch 2 Microplate Reader (BioTek). TLR2 and TLR4 activity was expressed as a fold difference of the OD over the negative control, endotoxin-free water. For TLR4, activity that was lower than that of 10 pg/ml lipopolysaccharide (LPS) was considered nonstimulatory to monocytes as we previously showed [ 5 ]. Assay of biological activity of PGN in human monocytes Peripheral blood mononuclear cells (PBMC) were prepared from heparinized blood by centrifugation through Histopaque 1077 (Sigma) according to the manufacturers protocol. PBMCs were washed, resupended in RPMI supplemented with 1% FBS and 0.2 mg/ml human IgG, and stimulated for 12 hours in the presence of Brefeldin A (3 μg/ml) with 10 μg/ml PGN of the initial SDS extraction, after HF treatment and PGN after final purification. In some experiments, internalization-dependent signaling was inhibited by pretreatment of PBMCs with cytochalasin D (5 μM; Sigma), a pharmacologic inhibitor of phagocytosis. TNFα production by CD14+ monocytes was quantified by immunostaining after saponin permeabilization using with anti-human CD14-PeCy7 (1 μg/ml) and anti-human TNFα-APC (1 μg/ml) or isotype control, mouse IgG1-APC (1 μg/ml). Data were collected by flow cytometry on a BD LSR II or BD FACSCelesta systems and analyzed using the FlowJo software package (FlowJoLLC). A assay of TLR2 activity of PGN in bone marrow-derived mouse macrophages Mouse strains C57BL/6J (wild type (WT)) and Tlr2 null (B6.129- Tlr2tm1Kir /J, Jackson Laboratory) were used for the generation of bone marrow-derived macrophages (BMDM). Femurs were collected from mice and bone marrow was flushed with medium. Bone marrow cells were cultured in RPMI 1640 medium supplemented with 10% FBS, 50 ng/ml recombinant murine monocyte colony stimulation factor-1 (CSF-1, purchased from R&D Systems), and 200 mM Glutamax-I for 24 hours. Non-adherent cells were then harvested, and resuspended in fresh medium, and cultured for 5 additional days in the same media. Macrophages were harvested, re-suspended in culture medium and plated into wells of a 96-well plate overnight at 37°C. Cells were then stimulated for 6 hours with B . anthracis PGN (10 μg/ml), S . aureus PGN (10 μg/ml), or LPS (1μg/ml) in the presence of Brefeldin A (3 μg/ml). Following stimulation, the cells were washed in PBS with Brefeldin A and incubated with mouse Fc-block (BD Pharmingen) on ice for 10 minutes. Cells were washed with PBS/Brefeldin A and fixed with 1% formaldehyde. After fixation, cells were permeabilized with 0.5% saponin and then stained with anti-mouse CD11b-PE (2 μg/ml), anti-mouse TNFα-APC (2 μg/ml) or appropriate isotype control (2 μg/ml). Macrophages were identified as CD11b + and TNFα producing cells were quantitated by flow cytometry. Amino acid analysis PGN samples were removed for analysis at the indicated points during the extraction process from S . aureus and B . anthracis , after HF treatment, but before Proteinase K digestion. Samples were also taken from the final product that had been digested with Proteinase K, extracted with SDS and washed repeatedly to remove SDS. For the S . aureus PGN, an additional HF treatment, SDS extraction and washing was performed on the final product and a sample was obtained. The samples were dried and resuspended in water at 10 mg/ml. Amino acid analysis was performed on each sample according to the method described by [ 15 ]. Briefly, 1 mg of each PGN sample was lyophilized. 300 μl of 6M HCl with 0.1% phenol was added, the sample was divided into three replicates and the solution was heated to 110°C for 48 hours in order to hydrolyze the peptide bonds. Each sample was cooled to room temperature and vacuum-dried to remove the HCl. Each sample was resuspended in 10 mM HCl and 20 μl of the sample was added to 60 μl of 0.2M sodium borate buffer, pH 8.8. To this mixture, 20 μl of 6-aminoquinoyl-N-hydroxysuccinimidyl carbamate in acetonitrile was added to derivatize the amino acids present in the sample. The derivatized sample was heated to 55°C for 15 min to convert tyrosine byproducts to one form. High-Performance Liquid Chromatography using an Agilent 1260 series instrument and a Waters AccQ Tag 3.9x150mm column was used to separate and quantify the derivatized amino acids present in each sample. Quantification was accomplished by UV absorbance at 254nm. Statistical analysis Statistical analysis and graphic representation were done using GraphPad Prism (GraphPad Software Inc.). Where applicable, the data are expressed as the mean +/- standard error of the mean. Statistical significance was determined by analysis of variance (ANOVA) with Bonferroni post test. A p value of 0.05. In both S . aureus and B . anthracis -derived material, the TLR2 activity increased after proteinase K treatment, likely due to introduction of exogenous TLR2 ligands from the recombinant enzyme used in this step. The introduced TLR2 ligands in B . anthracis -derived PGN was not detectable using the HEK-TLR2 transfectants after the additional purification steps of SDS extractions and water washes. However, S . aureus -derived material consistently had a very low and statistically-insignificant level of TLR2 inducing activity. We subjected the S . aureus -derived PGN to an additional SDS extraction and HF treatment and found that the low TLR2 activity was not further reduced (+SDS, HF; Fig 1B ). Our previous studies showed that blocking phagocytosis of PGN inhibited the proinflammatory responses of human monocytes [ 3 ]. We tested whether the TNFα response of human monocytes to the PGN in the stages of purification were similarly reduced by cytochalasin D. Responses to B . anthracis -derived PGN at all stages of purification were reduced by cytochalasin D ( Fig 1C ). Responses to S . aureus -derived PGN were less sensitive to cytochalasin D but were still reduced ( Fig 1D ). Thus, responses to the final PGN product from both sources were highest when the material can be internalized by the responding cells but a portion of the response can occur without PGN internalization into the responding cell. We analyzed the amino acid content of the PGN preparations before and after proteinase K digestion. Before proteinase K digestion, we found significant levels of a variety of amino acids that are not predicted to be components of the peptide subunit or interpeptide bridges of purified PGN [ 7 ]. These amino acids were present in PGN preparations from both pathogens ( Fig 2A and 2D ). After proteinase K digestion, amino acid complexity was greatly reduced ( Fig 2B and 2E ), leaving only DAP, Ala, and Glx in the B . anthracis -derived material ( Fig 2E ), consistent with the DAP-type PGNs. The S . aureus PGN had significant levels of Gly, Lys, Ala, and Glx and low levels of other amino acids ( Fig 2B ). Gly, Lys Ala and Glx are known components of the S . aureus Lys-type PGN stem peptide and interpeptide bridges [ 8 ]. Our analysis was not capable of accurately detecting His or Cys. None of the trace amino acids in the S . aureus PGN were removed after additional SDS extractions and HF treatment of the final product mentioned above ( Fig 2C ). Since the two PGN preparations show the predicted amino acid content of the stem peptide and interpeptide bridges, the data indicate that contaminating proteins are absent from the B . anthracis preparation but traces of some contaminants might be present in the S . aureus preparation. 10.1371/journal.pone.0193207.g002 Fig 2 Amino acid analysis of S . aureus and B . anthracis and PGN preparations before and after proteinase K digestion. S . aureus amino acids were analyzed before (A) and after (B) Proteinase K digestion. The final S . aureus PGN product was additionally treated with HF, extracted with SDS, and analyzed (C). B . anthracis amino acids were analyzed before (D) and after (E) proteinase K digestion. Two separate PGN preparations for each species were analyzed, with three replicates of each sample. Data are expressed as the mean ± SEM of amino acid content, normalized to lysine for S . aureus PGN and to DAP for B . anthracis PGN. Biological activity is not due to LPS The PGN samples were also assayed for LPS using HEK293 cells transfected with human TLR4 and a reporter of NFKB activation. We used endotoxin-free water as a standard and compared all samples to this standard. We also used a dose of 10 pg/ml LPS as a positive control since this amount is detected in this assay and yet the amount is insufficient to stimulate TNFα production from human monocytes [ 5 ]. We found ( Fig 3 ) that the LPS content was not significantly different than 10 pg/ml of a commercially prepared endotoxin standard at any stage of the purification process. Thus, the proinflammatory activity of PGN derived from either B . anthracis or S . aureus using this protocol and described below and elsewhere is not due to LPS. 10.1371/journal.pone.0193207.g003 Fig 3 TLR4 activity of B . anthracis and S . aureus PGN. B . anthracis (open bars) and S . aureus (closed bars) PGN samples were taken after the major steps of the purification process, washed three times with endotoxin free water, dried, weighed, resuspended in water and tested for TLR4 activity at a concentration of 10 μg/ml, using the HEK-TLR4 cell assay. Two separate PGN preparations for each species were tested. The data are expressed as the mean ± SEM fold difference TLR4 activity over the negative control, endotoxin free water. Statistical significance was determined by ANOVA with Bonferroni post test. All PGN samples were not significantly different in TLR4 activity than 10 pg/ml LPS. PGN preparations have significantly lower phosphate content than whole bacteria We measured the amount of organic phosphate at representative stages of the purification process since TLR2 ligands like lipoteichoic acid have repeating glycerolphosphate units [ 16 ]. We subjected samples of PGN and heat-killed bacteria to acid hydrolysis to release phosphate from any organic compound and applied a sensitive measurement of inorganic phosphate to the acid hydrolysate. The results are shown in Fig 4 . In order to compare equivalent quantities of PGN, we used a mass of PGN that equals the mass of PGN in the heat killed bacteria. This calculation was based on our determination that 10 X 1 10 cfu contains 1.25 mg PGN. In the case of B . anthracis - and S . aureus -derived PGN, all samples taken from the purification steps showed a significantly reduced phosphate content compared to the corresponding amount present in heat killed bacteria. Thus, heat killed B . anthracis had 0.3 nmoles phosphate per μg of PGN and heat killed S . aureus had 0.5 nmoles phosphate per μg of PGN. After the initial SDS extraction, the phosphate content of PGN was reduced to 0.05 nmoles per μg in B . anthracis PGN and to 0.2 nmoles per μg S . aureus PGN. After treatment with HF, phosphate content was reduced to 0.004 nmoles/μg for B . anthracis PGN and to 0.08 nmoles/μg for S . aureus PGN. Subsequent purification steps, treatment with proteinase K and SDS, did not significantly reduce the phosphate content of either PGN in comparison with the HF treatment (P>0.05). The final PGN products from B . anthracis and S . aureus had a phosphate content of 0.03 and 0.002 nmoles phosphate/μg PGN, respectively. These results are consistent with a low level or absence of organic phosphates including forms of teichoic acids that are TLR2 ligands in the PGN final products. 10.1371/journal.pone.0193207.g004 Fig 4 Phosphate content of B . anthracis and S . aureus PGN. B . anthracis (A) and S . aureus (B) PGN samples were taken after the major steps of the purification process, washed three times with endotoxin free water, dried, weighed, resuspended tested for phosphate content. Two separate PGN preparations for each species were tested, with two replicates for each species. Data are expressed as the mean ± SEM of the nmoles of phosphate per μg/PGN. Statistical significance was determined by ANOVA with Bonferroni post test. *p0.05. In both S . aureus and B . anthracis -derived material, the TLR2 activity increased after proteinase K treatment, likely due to introduction of exogenous TLR2 ligands from the recombinant enzyme used in this step. The introduced TLR2 ligands in B . anthracis -derived PGN was not detectable using the HEK-TLR2 transfectants after the additional purification steps of SDS extractions and water washes. However, S . aureus -derived material consistently had a very low and statistically-insignificant level of TLR2 inducing activity. We subjected the S . aureus -derived PGN to an additional SDS extraction and HF treatment and found that the low TLR2 activity was not further reduced (+SDS, HF; Fig 1B ). Our previous studies showed that blocking phagocytosis of PGN inhibited the proinflammatory responses of human monocytes [ 3 ]. We tested whether the TNFα response of human monocytes to the PGN in the stages of purification were similarly reduced by cytochalasin D. Responses to B . anthracis -derived PGN at all stages of purification were reduced by cytochalasin D ( Fig 1C ). Responses to S . aureus -derived PGN were less sensitive to cytochalasin D but were still reduced ( Fig 1D ). Thus, responses to the final PGN product from both sources were highest when the material can be internalized by the responding cells but a portion of the response can occur without PGN internalization into the responding cell. We analyzed the amino acid content of the PGN preparations before and after proteinase K digestion. Before proteinase K digestion, we found significant levels of a variety of amino acids that are not predicted to be components of the peptide subunit or interpeptide bridges of purified PGN [ 7 ]. These amino acids were present in PGN preparations from both pathogens ( Fig 2A and 2D ). After proteinase K digestion, amino acid complexity was greatly reduced ( Fig 2B and 2E ), leaving only DAP, Ala, and Glx in the B . anthracis -derived material ( Fig 2E ), consistent with the DAP-type PGNs. The S . aureus PGN had significant levels of Gly, Lys, Ala, and Glx and low levels of other amino acids ( Fig 2B ). Gly, Lys Ala and Glx are known components of the S . aureus Lys-type PGN stem peptide and interpeptide bridges [ 8 ]. Our analysis was not capable of accurately detecting His or Cys. None of the trace amino acids in the S . aureus PGN were removed after additional SDS extractions and HF treatment of the final product mentioned above ( Fig 2C ). Since the two PGN preparations show the predicted amino acid content of the stem peptide and interpeptide bridges, the data indicate that contaminating proteins are absent from the B . anthracis preparation but traces of some contaminants might be present in the S . aureus preparation. 10.1371/journal.pone.0193207.g002 Fig 2 Amino acid analysis of S . aureus and B . anthracis and PGN preparations before and after proteinase K digestion. S . aureus amino acids were analyzed before (A) and after (B) Proteinase K digestion. The final S . aureus PGN product was additionally treated with HF, extracted with SDS, and analyzed (C). B . anthracis amino acids were analyzed before (D) and after (E) proteinase K digestion. Two separate PGN preparations for each species were analyzed, with three replicates of each sample. Data are expressed as the mean ± SEM of amino acid content, normalized to lysine for S . aureus PGN and to DAP for B . anthracis PGN. Biological activity is not due to LPS The PGN samples were also assayed for LPS using HEK293 cells transfected with human TLR4 and a reporter of NFKB activation. We used endotoxin-free water as a standard and compared all samples to this standard. We also used a dose of 10 pg/ml LPS as a positive control since this amount is detected in this assay and yet the amount is insufficient to stimulate TNFα production from human monocytes [ 5 ]. We found ( Fig 3 ) that the LPS content was not significantly different than 10 pg/ml of a commercially prepared endotoxin standard at any stage of the purification process. Thus, the proinflammatory activity of PGN derived from either B . anthracis or S . aureus using this protocol and described below and elsewhere is not due to LPS. 10.1371/journal.pone.0193207.g003 Fig 3 TLR4 activity of B . anthracis and S . aureus PGN. B . anthracis (open bars) and S . aureus (closed bars) PGN samples were taken after the major steps of the purification process, washed three times with endotoxin free water, dried, weighed, resuspended in water and tested for TLR4 activity at a concentration of 10 μg/ml, using the HEK-TLR4 cell assay. Two separate PGN preparations for each species were tested. The data are expressed as the mean ± SEM fold difference TLR4 activity over the negative control, endotoxin free water. Statistical significance was determined by ANOVA with Bonferroni post test. All PGN samples were not significantly different in TLR4 activity than 10 pg/ml LPS. PGN preparations have significantly lower phosphate content than whole bacteria We measured the amount of organic phosphate at representative stages of the purification process since TLR2 ligands like lipoteichoic acid have repeating glycerolphosphate units [ 16 ]. We subjected samples of PGN and heat-killed bacteria to acid hydrolysis to release phosphate from any organic compound and applied a sensitive measurement of inorganic phosphate to the acid hydrolysate. The results are shown in Fig 4 . In order to compare equivalent quantities of PGN, we used a mass of PGN that equals the mass of PGN in the heat killed bacteria. This calculation was based on our determination that 10 X 1 10 cfu contains 1.25 mg PGN. In the case of B . anthracis - and S . aureus -derived PGN, all samples taken from the purification steps showed a significantly reduced phosphate content compared to the corresponding amount present in heat killed bacteria. Thus, heat killed B . anthracis had 0.3 nmoles phosphate per μg of PGN and heat killed S . aureus had 0.5 nmoles phosphate per μg of PGN. After the initial SDS extraction, the phosphate content of PGN was reduced to 0.05 nmoles per μg in B . anthracis PGN and to 0.2 nmoles per μg S . aureus PGN. After treatment with HF, phosphate content was reduced to 0.004 nmoles/μg for B . anthracis PGN and to 0.08 nmoles/μg for S . aureus PGN. Subsequent purification steps, treatment with proteinase K and SDS, did not significantly reduce the phosphate content of either PGN in comparison with the HF treatment (P>0.05). The final PGN products from B . anthracis and S . aureus had a phosphate content of 0.03 and 0.002 nmoles phosphate/μg PGN, respectively. These results are consistent with a low level or absence of organic phosphates including forms of teichoic acids that are TLR2 ligands in the PGN final products. 10.1371/journal.pone.0193207.g004 Fig 4 Phosphate content of B . anthracis and S . aureus PGN. B . anthracis (A) and S . aureus (B) PGN samples were taken after the major steps of the purification process, washed three times with endotoxin free water, dried, weighed, resuspended tested for phosphate content. Two separate PGN preparations for each species were tested, with two replicates for each species. Data are expressed as the mean ± SEM of the nmoles of phosphate per μg/PGN. Statistical significance was determined by ANOVA with Bonferroni post test. *p<0.05, **p<0.01, ***p<0.001 versus heat killed bacteria. TLR2 activity in PGN preparations Although we did not detect a response to the PGN preparations using HEK 293 cells expressing sensing elements for TLR2 ligands, we thought mouse bone marrow-derived macrophages might be a more sensitive measure of TLR contaminants. We therefore cultured bone marrow-derived macrophages from wild type and TLR2 - / - mice with 10 μg/ml S . aureus- Lys-type PGN or B . anthracis- derived DAP-type PGN in the presence of brefeldin A, for 6 hours and measured the production of TNFα by intracellular cytokine staining and flow cytometry. The results ( Fig 5A ) showed that B . anthraci s-derived PGN failed to stimulate mouse macrophages to make TNFα, consistent with our previous finding that mouse cells do not recognize B . anthracis PGN [ 3 ]. In contrast, S . aureus PGN produced a significant response (p<0.001) in the wild type mouse macrophages but not in macrophages from TLR2 -/- mice. We quantitated the percent responding macrophages in 4 individual experiments and the results are shown in Fig 5B . The ability of Lys-type PGN from S . aureus to stimulate mouse macrophages in a TLR2-dependent manner suggest that the S . aureus PGN is itself, or contains, a TLR2 ligand while the B . anthracis -derived DAP-type PGN is not a TLR2 ligand. 10.1371/journal.pone.0193207.g005 Fig 5 Mouse macrophage response to B . anthracis - and S . aureus -derived PGN. (A-B) Differentiated BMDMs from WT or TLR2 -/- mice were plated into wells of a 96-well plate (1.5 x 10 5 cells per well) and stimulated with LPS (1 μg/ml), B . anthracis PGN (10 μg/ml), or S . aureus PGN (10 μg/ml) in the presence of Brefeldin A for 6 hours at 37°C. The cells were stained for intracellular TNFα, and then analyzed by flow cytometry. Graphs are representative of 3 independent experiments (solid grey peak is unstimulated sample). Results are mean ± SEM from three independent experiments. The p values were calculated using a two-way ANOVA using Bonferroni post hoc test for multiple comparisons. # p < .0001 compared to unstimulated control. **** p < .0001 between WT and TLR2 -/- BMDMs. We noted the amino acid analysis of the S . aureus -derived PGN showed trace amounts of amino acids that cannot be attributed to the stem peptide ( Fig 2 ). These trace amino acids may be lipidated peptides that fail to be removed by the protease K digestions during purification. Accordingly, the PGN from S . aureus might be contaminated with TLR2 ligands. To test this possibility, we purified PGN using the same procedure described above and applied to the S . aureus strain Newman and its isogenic mutant lacking lipoprotein diacylglycerol transferase (Δlgt) [ 17 ]. The PGN products from these two strains were tested on mouse bone marrow-derived macrophages. We found ( Fig 6A and 6B ) that the remaining amino acid impurities are present in approximately equal amounts in the PGN material derived from either the parent or the Δ lgt mutant S . aureus strains. We applied our flow cytometry-based assay to measure TNFα production and found that only PGN derived from the parent strain having the lipoprotein transferase was able to stimulate mouse bone marrow-derived macrophages ( Fig 6C ); PGN derived from the Δlgt S . aureus mutant failed to stimulate, as did PGN derived from B . anthracis . The average and standard error of three separate experiments are shown in Fig 6D . This finding indicates that the material in the PGN that stimulates mouse macrophages is the contaminating lipopeptides and not PGN itself. The data are consistent with the result in Fig 5 , showing that PGN derived from B . anthracis is not recognized by mouse macrophages. Taken together with the results shown in Fig 6 , we conclude that PGN itself, regardless of the bacterial species of origin, is not a TLR agonist. 10.1371/journal.pone.0193207.g006 Fig 6 Lipid modification of contaminating peptides in purified S . aureus PGN is important for TLR2 signaling. PGN from Newman S . aureus (A) and Newman S . aureus Δlgt (B) was digested with mutanolysin. After digestion samples were run in triplicate for amino acid analysis as described in Fig 2 . (C-D) Differentiated BMDMs were stimulated with LPS (1 μg/ml), B . anthracis PGN (10 μg/ml), Newman S . aureus PGN (10 μg/ml), or Newman S . aureus Δlgt PGN (10 μg/ml) in the presence of Brefeldin A for 6 hours at 37°C. The cells were stained for intracellular TNFα, and then analyzed by flow cytometry. Graphs are representative of three independent experiments (solid grey peak is unstimulated sample). Results are mean ± SEM from three independent experiments. The p values were calculated using a one-way ANOVA using Bonferroni post hoc test for multiple comparisons. # p < .05 compared to unstimulated control, ** p = .001, **** p < .0001. Earlier, we showed the human responses to purified PGN is elevated by the presence of serum factors like IgG which opsonize PGN and enhance its uptake into phagocytic cells [ 3 , 13 ]. Thus, we can use the enhancement effect of human serum to test the purity of the various PGN preparations. The flow cytometry-based measurements of TNFα are shown in the left two panels of Fig 7 and a summary of three experiments is shown in the right two panels of Fig 7 . B . anthracis -derived PGN was a weak agonist in the absence of human serum and a potent agonist in the presence of human serum ( Fig 7A and 7B ). In contrast, PGN derived from the S . aureus Newman strain having the lipoprotein transferase activity was a potent agonist towards human monocytes regardless of the presence or absence of human serum ( Fig 7C and 7D ). PGN derived from the Δlgt mutant failed to stimulate human monocytes in the absence of serum and was a weak agonist in the presence of serum. These results are consistent with those shown in Fig 6 : the contaminating lipopeptides are able to stimulate human monocytes in a human serum-independent way, while PGN lacking the lipopeptides requires human serum to stimulate the cells. 10.1371/journal.pone.0193207.g007 Fig 7 Human monocyte responses to PGN from lipoprotein deficient S . aureus require serum opsonins. (A) PBMCs were plated into wells of a 96-well plated (4–8 x 10 5 cells per well) and stimulated with B . anthracis PGN (10 μg/ml) in presence of FCS (1% v/v) with or without normal human serum (1% v/v) and with Brefeldin A for 6 hours at 37°C. The cells were stained for intracellular TNFα, and then analyzed by flow cytometry. (B) PBMCs were plated into wells of a 96-well plated (4–8 x 10 5 cells per well) and stimulated with Newman S . aureus PGN (10 μg/ml), or Newman S . aureus Δlgt PGN (10 μg/ml) in presence of FCS (1% v/v) with or without normal human serum (1% v/v) and with Brefeldin A for 6 hours at 37°C. The cells were stained for intracellular TNFα, and then analyzed by flow cytometry. Graphs are representative of experiments from three donors (solid grey peak is unstimulated sample). Results are mean ± SEM from three donors. The p values were calculated using either a two-way (B) or one-way (D) ANOVA using Bonferroni post hoc test for multiple comparisons. # p ≤ .0001 compared to unstimulated control, ** p = .001, *** p = 0001, **** p < .0001. Discussion We have performed a series of experiments to determine whether the DAP-type B . anthracis and/or Lys-type S . aureus PGNs are activating TLR2 ligands and/or to determine at what point in the purification process activating TLR ligands are removed. We found, in the case of both bacterial types, that heat killed bacteria stimulate HEK293 cells strongly through TLR2, indicating that TLR2 ligands are present in the unprocessed bacteria. However, we found differences in the amount of TLR2 activity retained during purification of PGN from these bacteria. For B . anthracis , the PGN signaling through TLR2 was lost after the initial extraction from bacterial cultures with boiling SDS. This finding suggests that TLR2 ligands might be embedded in the PGN layer of whole bacteria but are easily removed by the detergent. The TLR2 activity remained below the level of significance throughout the extraction process, except that there was a small increase through the addition of an exogenous processing agent, proteinase K. This TLR2 activity was easily removed by additional SDS extraction and washing to produce a TLR2 ligand-free final product. In contrast, for the S . aureus PGN, significant TLR2-stimulating activity was present after the initial SDS extraction as detected by HEK293-TLR2 cells and mouse bone marrow-derived macrophages. The TLR2 activity was reduced after HF treatment but largely remained throughout processing. Additional HF treatment, SDS extractions and washing did not further remove the TLR2 activity. Lastly, we showed that the TLR2 activity in S . aureus -derived PGN was not PGN itself but caused by contaminating lipopeptides. We conclude that neither DAP-type B . anthracis PGN nor Lys-type S . aureus PGN is a TLR2 ligand. However, it is clear that insufficient purification of PGN could lead to introduction of or contamination by existing TLR2 ligands. Insufficient PGN purification may have been a factor in past studies leading to the notion that PGN is a TLR2 ligand. We were unable to remove the TLR2 activity from PGN derived from S . aureus except from genetically manipulated strains. Our data also show that mouse bone marrow-derived macrophages are a more sensitive indicator of TLR2 activity than are HEK293 transfected with the TLR2 receptor and a NF-kB reporter system. Mouse macrophages produced a significant inflammatory response to S . aureus PGN through TLR2 activation ( Fig 3 ), whereas the HEK293 cells detected a very low and statistically insignificant level of TLR2 activity in the identical sample ( Fig 1 ). The response of human monocytes to the Lys-type PGN from S . aureus Δ lgt mutant was considerably weaker in the presence of human serum compared to the response to B . anthracis -derived PGN. We attribute this difference to the fact that B . anthracis -derived PGN has DAP-containing stem peptides that are lacking in S . aureus PGN. After lysosomal digestion, S . aureus PGN can form the NOD2 ligand muramyl dipeptide to stimulate innate immune cells [ 18 ]. Likewise, after lysosomal digestion, B . anthracis PGN can form muramyl dipeptide and several DAP-containing peptides that are also ligands for NOD1 [ 19 ]. Activation of both NOD1 and NOD2 by B . anthracis PGN might account for the relative potency of these forms of PGN to stimulate TNFα production by monocytes. The poly-disaccharide nature of PGN does not chemically resemble other known TLR2 ligands. Natural TLR2 ligands each have a hydrophobic component, such as present in glycolipids or lipopeptides [ 16 , 20 ]. Other TLR2 ligands are amphiphilic such as LTA [ 16 , 21 ], lipoarabinomannan [ 22 ] and glycosylphosphatidylinositol-anchored lipids [ 23 ]. However, PGN, as a poly-disaccharide linked by largely polar amino acids, is not hydrophobic. This has led to questions about TLR2's ability to recognize such a chemically diverse set of PAMPs while still maintaining the sensitivity expected of a PRR of the innate immune system [ 16 , 24 ]. Our study explains that TLR2 does not recognize the PGN poly-disaccharide and so the recognition of diverse PAMPs by TLR2 may no longer be a relevant inconsistency. We have used the need for phagocytosis of PGN as a criterion for its purity. Although most of the proinflammatory response is sensitive to agents that block phagocytosis like cytochalasin D, not all of the response is blocked. It may be that some of the PGN particles can enter cells using a process that is not sensitive to agents that disrupt actin reorganization like cytochalasin D. However, it may be that PGN is able to weakly activate immune cells through a surface receptor that is as yet unidentified. We have also used the lack of a response by mouse macrophages a criterion for PGN purification. It is unclear why mouse cells fail to respond to PGN with inflammatory cytokine production. Earlier, we showed that mice lack the PGN serum opsonins like anti-PGN antibodies that are present in humans [ 13 ]. The lack of these opsonins could prevent mouse macrophages from taking up and digesting PGN to form ligands that can be sensed by cytoplasmic NOD receptors. However, in unpublished studies, we found that experimentally-generated mouse anti-PGN IgG fails to support a response by mouse macrophages to B . anthracis -derived PGN. Likewise, we found that human IgG containing anti-PGN IgG does not support an inflammatory response in macrophages derived from mice that express human FcγRs as transgenes [ 25 ]. Mouse macrophages have been reported to respond to PGN but the activation in these reports require that the macrophages are first primed with agents like LPS [see, e.g., [ 26 ]]. Priming with such agents could alter macrophage phagocytic ability, the need for serum opsonins, or NOD expression [ 27 ]. Mice lacking NOD1 or NOD2 show a phenotype when challenged with bacterial pathogens [reviewed in [ 28 ]], including B . anthracis [ 29 , 30 ], suggesting that the PGN is recognized in some way. In these examples using live bacterial challenges in NOD-deficient animals, it may be that other bacterial PAMPS serve the priming function and allow PGN to be taken up in ways distinct from human cells. However, it may also be that some of the inflammation and accompanying pathology caused by bacterial pathogens which occurs in humans is not represented in mouse models because of the lack of PGN responses that are similar to humans.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9254050/

Persistence and shedding of senecavirus A in naturally infected boars

Senecavirus A (SVA) infection in pigs causes vesicular disease and results in a short viremia and transient shedding of the virus, mainly in oral fluids and feces. Here we describe the consistent prolonged shedding of SVA in the semen of 2 boars, and persistence of SVA within the tonsils and testes of 3 adult boars. Two SVA-infected boars that were identified on a Minnesota sow farm in 2017 shed SVA RNA in semen for >3 mo after an outbreak of vesicular disease had occurred on the farm. SVA was isolated from 1 semen sample collected 9 d after clinical disease began on the farm. The third SVA-infected boar was identified on an Indiana sow farm in 2020. All boars had SVA RNA detected in the testes and tonsils by RT-rtPCR, with lower Ct values obtained for the testes than from the tonsils. All boars had multifocal lymphocytic orchitis with segmental degeneration and atrophy of the germinal epithelium within the seminiferous tubules. One boar also had areas of seminiferous tubule collapse and interstitial fibrosis within the testes. In all boars, in situ hybridization demonstrated the presence of SVA mRNA within cells located basally in the seminiferous tubules of the testes, and within the basal surface epithelial cells, crypt epithelial cells, and subepithelial and parafollicular lymphocytes and histiocytes of the tonsil.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9024341/

Evaluation of anthrax vaccine safety in 18 to 20 year olds: A first step towards age de-escalation studies in adolescents ✩

Background/objectives: Anthrax vaccine adsorbed (AVA, BioThrax®) is recommended for post-exposure prophylaxis administration for the US population in response to large-scale Bacillus anthracis spore exposure. However, no information exists on AVA use in children and ethical barriers exist to performing pre-event pediatric AVA studies. A Presidential Ethics Commission proposed a potential pathway for such studies utilizing an age de-escalation process comparing safety and immunogenicity data from 18 to 20 year-olds to older adults and if acceptable proceeding to evaluations in younger adolescents. We conducted exploratory summary re-analyses of existing databases from 18 to 20 year-olds ( n = 74) compared to adults aged 21 to 29 years ( n = 243) who participated in four previous US government funded AVA studies. Methods: Data extracted from studies included elicited local injection-site and systemic adverse events (AEs) following AVA doses given subcutaneously at 0, 2, and 4 weeks. Additionally, proportions of subjects with ≥4-fold antibody rises from baseline to post-second and post-third AVA doses (seroresponse) were obtained. Results: Rates of any elicited local AEs were not significantly different between younger and older age groups for local events (79.2% vs. 83.8%, P = 0.120) or systemic events (45.4% vs. 50.5%, P = 0.188). Robust and similar proportions of seroresponses to vaccination were observed in both age groups. Conclusions: AVA was safe and immunogenic in 18 to 20 year-olds compared to 21 to 29 year-olds. These results provide initial information to anthrax and pediatric specialists if AVA studies in adolescents are required. Background/objectives: Anthrax vaccine adsorbed (AVA, BioThrax®) is recommended for post-exposure prophylaxis administration for the US population in response to large-scale Bacillus anthracis spore exposure. However, no information exists on AVA use in children and ethical barriers exist to performing pre-event pediatric AVA studies. A Presidential Ethics Commission proposed a potential pathway for such studies utilizing an age de-escalation process comparing safety and immunogenicity data from 18 to 20 year-olds to older adults and if acceptable proceeding to evaluations in younger adolescents. We conducted exploratory summary re-analyses of existing databases from 18 to 20 year-olds ( n = 74) compared to adults aged 21 to 29 years ( n = 243) who participated in four previous US government funded AVA studies. Methods: Data extracted from studies included elicited local injection-site and systemic adverse events (AEs) following AVA doses given subcutaneously at 0, 2, and 4 weeks. Additionally, proportions of subjects with ≥4-fold antibody rises from baseline to post-second and post-third AVA doses (seroresponse) were obtained. Results: Rates of any elicited local AEs were not significantly different between younger and older age groups for local events (79.2% vs. 83.8%, P = 0.120) or systemic events (45.4% vs. 50.5%, P = 0.188). Robust and similar proportions of seroresponses to vaccination were observed in both age groups. Conclusions: AVA was safe and immunogenic in 18 to 20 year-olds compared to 21 to 29 year-olds. These results provide initial information to anthrax and pediatric specialists if AVA studies in adolescents are required.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1464026/

The Cost of Doing Business: Cost Structure of Electronic Immunization Registries

Objective To predict the true cost of developing and maintaining an electronic immunization registry, and to set the framework for developing future cost-effective and cost-benefit analysis. Data Sources/Study Setting Primary data collected at three immunization registries located in California, accounting for 90 percent of all immunization records in registries in the state during the study period. Study Design A parametric cost analysis compared registry development and maintenance expenditures to registry performance requirements. Data Collection/Extraction Methods Data were collected at each registry through interviews, reviews of expenditure records, technical accomplishments development schedules, and immunization coverage rates. Principal Findings The cost of building immunization registries is predictable and independent of the hardware/software combination employed. The effort requires four man-years of technical effort or approximately $250,000 in 1998 dollars. Costs for maintaining a registry were approximately $5,100 per end user per three-year period. Conclusions There is a predictable cost structure for both developing and maintaining immunization registries. The cost structure can be used as a framework for examining the cost-effectiveness and cost-benefits of registries. The greatest factor effecting improvement in coverage rates was ongoing, user-based administrative investment. Objective To predict the true cost of developing and maintaining an electronic immunization registry, and to set the framework for developing future cost-effective and cost-benefit analysis. Data Sources/Study Setting Primary data collected at three immunization registries located in California, accounting for 90 percent of all immunization records in registries in the state during the study period. Study Design A parametric cost analysis compared registry development and maintenance expenditures to registry performance requirements. Data Collection/Extraction Methods Data were collected at each registry through interviews, reviews of expenditure records, technical accomplishments development schedules, and immunization coverage rates. Principal Findings The cost of building immunization registries is predictable and independent of the hardware/software combination employed. The effort requires four man-years of technical effort or approximately $250,000 in 1998 dollars. Costs for maintaining a registry were approximately $5,100 per end user per three-year period. Conclusions There is a predictable cost structure for both developing and maintaining immunization registries. The cost structure can be used as a framework for examining the cost-effectiveness and cost-benefits of registries. The greatest factor effecting improvement in coverage rates was ongoing, user-based administrative investment.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3783559/

Engineering unnatural variants of plantazolicin through codon reprogramming

Plantazolicin (PZN) is a polyheterocyclic natural product derived from a ribosomal peptide that harbors remarkable antibiotic selectivity for the causative agent of anthrax, Bacillus anthracis . To simultaneously establish the structure-activity relationship of PZN and the substrate tolerance of the biosynthetic pathway, an Escherichia coli expression strain was engineered to heterologously produce PZN analogs. Variant PZN precursor genes were produced by site-directed mutagenesis and later screened by mass spectrometry to assess posttranslational modification and export by E. coli . From a screen of 72 precursor peptides, 29 PZN variants were detected. This analog collection provided insight into the selectivity of the posttranslational modifying enzymes and established the boundaries of the natural biosynthetic pathway. Unlike other studied thiazole/oxazole-modified microcins, the biosynthetic machinery appeared to be finely tuned towards the production of PZN, such that the cognate enzymes did not process even other naturally occurring sequences from similar biosynthetic clusters. The modifying enzymes were exquisitely selective, installing heterocycles only at pre-defined positions within the precursor peptides while leaving neighboring residues unmodified. Nearly all substitutions at positions normally harboring heterocycles prevented maturation of a PZN variant, though some exceptions were successfully produced lacking a heterocycle at the penultimate residue. No variants containing additional heterocycles were detected, although several peptide sequences yielded multiple PZN variants as a result of varying oxidation states of select residues. Eleven PZN variants were produced in sufficient quantity to facilitate purification and assessment of their antibacterial activity, providing insight into the structure-activity relationship of PZN.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3366964/

Quantitative RT-PCR profiling of the Rabbit Immune Response: Assessment of Acute Shigella flexneri Infection

Quantitative reverse transcription PCR analysis is an important tool to monitor changes in gene expression in animal models. The rabbit is a widely accepted and commonly used animal model in the study of human diseases and infections by viral, fungal, bacterial and protozoan pathogens. Only a limited number of rabbit genes have, however, been analyzed by this method as the rabbit genome sequence remains unfinished. Recently, increasing coverage of the genome has permitted the prediction of a growing number of genes that are relevant in the context of the immune response. We hereby report the design of twenty-four quantitative PCR primer pairs covering common cytokines, chemoattractants, antimicrobials and enzymes for a rapid, sensitive and quantitative analysis of the rabbit immune response. Importantly, all primer pairs were designed to be used under identical experimental conditions, thereby enabling the simultaneous analysis of all genes in a high-throughput format. This tool was used to analyze the rabbit innate immune response to infection with the human gastrointestinal pathogen Shigella flexneri . Beyond the known inflammatory mediators, we identified IL-22, IL-17A and IL-17F as highly upregulated cytokines and as first responders to infection during the innate phase of the host immune response. This set of qPCR primers also provides a convenient tool for monitoring the rabbit immune response during infection with other pathogens and other inflammatory conditions. Introduction The New Zealand white rabbit ( Oryctolagus cuniculus ) is a widely used animal model for basic research of human diseases and pathogens, translational biomedical research and the development of vaccines or therapeutics. Rabbits are phylogenetically more closely related to primates than rodents are and have an anatomy and physiology that more closely resembles that of primates. In many instances this leads to a more acurate modeling of human conditions. In addition, the rabbit shares many advantages with rodents such as their small size, short gestation time, large litter sizes, ease of breeding and maintaining colonies, as well as the relatively low cost to purchase and house them. The rabbit has been used as a model organism to study a number of human diseases such as cancer [1] , [2] atherosclerosis [3] , Alzheimer's [4] as well as serving as an important model for eye research [5] . Within the infectious disease field, rabbit models have been established for a large number of pathogens including protozoa, such as Schistosoma mansoni [6] and Entamoeba histolytica [7] ; viruses, such as the papilloma virus [8] , Herpes simplex virus [9] and rotavirus [10] ; fungi, such as Candida albicans [11] and Aspergillus flavus [12] ; and a large number of bacteria, including enterotoxic Escherichia coli , Vibrio cholerae , Mycobacterium tuberculosis , Borrelia burgdorferi , Chlamydia pneumoniae , Salmonella enterica , Campylobacter jejuni, Bacillus anthracis and Shigella flexneri [13] , [14] , [15] , [16] , [17] , [18] . These pathogen-rabbit models include ocular, dermal, visceral, pulmonary, systemic and intestinal diseases. A hallmark of infection is the activation of the host immune response to combat the intruder. The use of the rabbit model is, however, hampered by the limited availability of species-specific products, such as antibodies. For now, the host response may be analyzed by measuring production of the limited number of cytokines for which an antibody or a functional assay is available, assaying for enzyme activity, such as MPO from neutrophils, monitoring transcriptional changes using quantitative PCR (qPCR), and histological scoring of morphological changes of the tissue and infiltration of leukocytes and lymphocytes. Indeed, due to the sparse availability of reagents, histological examination remains a gold standard for the analysis of the host immune response. However, quantitation of this response is highly laborious as it involves embedding samples, sectioning, staining and manually measuring morphological changes or cell infiltrates. In addition, early events of immune activation that precede gross morphological changes are difficult if not impossible to identify [19] . There therefore is a need to develop tools to more easily assess the immune response of the host in a quantitative manner. Quantitative RT-PCR is a sensitive and rapid means to assay the host response to infection and, as microarrays, macroarrays or other arrays are not available for the rabbit, remains the method of choice for transcriptional analysis for the rabbit. In addition, qPCR offers a cost-effective and simple way to analyze specific target genes of interest using widely available qPCR machines [20] . While qPCR lacks the large-scale throughput of microarrays, it is considered the gold standard for gene expression analysis due to its ease of use, high detection sensitivity and wide linear dynamic range. Currently, the two most popular qPCR techniques are the TaqMan and SYBR ® Green technologies. While TaqMan ® uses gene-specific fluorescent probes, the SYBR technology utilizes a dye that intercalates into double-stranded DNA and allows the use of a single reagent with many different primer pairs, thereby also reducing running costs. The number of genes available for analysis in the rabbit has, thus far, been limited to cloned transcripts as the rabbit genome remains unfinished. In addition, the optimal amplification conditions for primers designed to popular targets are not uniform, generally cannot be run together, and are not optimized for newer qPCR machines [21] . Now, increasing coverage of the rabbit genome has permitted the prediction of more rabbit genes including those that code for cytokines (such as IL-17A, IL-17F, IL-21, and IL-22) that recently were discovered to be highly relevant in the innate immune response to pathogen challenge and in a number of inflammatory diseases. We therefore sought to increase the number of available rabbit target genes for qPCR analysis and to standarize the amplification conditions to allow simultaneous analysis of all targets. As a test system, we assessed the inflammatory response elicited by infection of rabbits with Shigella flexneri . Shigella is a gram-negative facultative anaerobic and intracellular human pathogen and the causative agent of bacillary dysentery or shigellosis [22] . Shigella infects via the fecal-oral route, crosses the colonic mucosa via M-cells and is taken up by macrophages and dendritic cells. Shigella escapes these immune sentinels by inducing a pro-inflammatory cell death called pyroptosis and invades the epithelial layer basolaterally to replicate and spread cell-to-cell to neighboring cells via actin-based motility [22] . The Shigella flexneri -rabbit ileal loop model is commonly used to perform Shigella vaccine research, study the interaction of Shigella with the intestinal barrier, investigate the induction of the innate immune response and evaluate the virulence of S. flexneri mutants. Infection of rabbit ligated ileal loops with S. flexneri leads to a strong inflammatory response that is characterized by bacterial invasion of the epithelium, massive recruitment of polymorphonuclear cells and extensive destruction of the epithelial lining at its acute phase [19] . This animal model thereby accurately recapitulates the symptoms seen in rectal biopsies of humans suffering from shigellosis [19] , [23] . However, during shigellosis, Shigella infection foci are focally distributed [24] . This renders detection of affected areas by histology difficult even in severely sick patients and requires several biopsies to obtain a representative analysis [23] . Although shigellosis in the rabbit is robust, this issue also arises in this model, making a quantitative analysis using relatively large pieces of intestine an important complementary method to histological analysis of thin sections. We hereby report the design of a set of twenty-four primer pairs for simultaneous qPCR analysis to highly relevant rabbit immune genes for a rapid, sensitive and quantitative analysis of the rabbit immune response. This tool was able to clearly detect the activation of the innate immune system at early stages of Shigella infection before easily visible morphological alterations had taken place within the intestinal mucosa. Materials and Methods Bacterial Propagation and Infection The streptomycin-resistant invasive wild-type Shigella flexneri strain M90T-Sm of serotype 5a and its virulence plasmid-cured non-invasive derivative BS176 were thawed from –80°C stocks, grown on Tryptic Soy (TCS) Agar-Congo red (CR) plates (3% BBL-trypticase soy agar (BD Biosciences), 0.1% CR) for 16 hrs at 37°C and kept at 4°C. For infection, one colony of each strain from freshly streaked plates was grown in TCS medium at 37°C for 16 hours with agitation two days before the experiment and kept at room temperature for 8 hrs the next day. In the evening before the experiment, 50 µl of culture of each strain was plated onto TCS plates (without CR) and incubated at 37°C for 16 hrs to form a bacterial lawn. The morning of the experiment, physiological saline (0.9% NaCl) was added to each plate and plates were scraped to resuspend the bacteria. Bacterial suspensions were diluted to OD 600 = 4 (equivalent to 3×10 9 bacteria ml −1 ). As a control, the bacterial suspension was diluted by 10 6 fold and 25 µl were plated in triplicate on TCS-CR plates. After incubation at 37°C for 16 hrs, bacterial colonies were enumerated (approximately 75 colonies are expected) and checked for proper colony morphology (M90T colonies appear red and are smaller than those of BS176). Rabbit Illeal Loop Infections The rabbit illeal loop model protocol was approved by the Comite Regional d'Ethique pour l'Experimentation Animale in Paris 1 (protocol #20070004). Surgery was performed essentially as described previously with minor modifications [19] . Twelve male New Zealand White rabbits weighing 2.4–2.6 kg (Charles River Laboratories, St. Aubin, France) were split into 2 groups for short (4 to 5.5 hrs) and long (8 hrs) time points of infection. Animals were received 10 days prior to surgery and treated with Mucoxid (2.8 g L −1 sodium sulfadimethoxine salt in the drinking water) (CEVA Sante Animale, France) until the day prior to surgery to minimize infection with Coccidia parasites. Three independent experiments using two rabbits each for early and late time points were performed. Rabbits were fasted 24 hrs before infection, sedated intravenously by ear vein injection with 0.05 ml kg −1 Calmivet (Vétoquinol) containing 0.5% acepromazine and anesthesized by the same route with 0.2 ml kg −1 of Imalgene® 1000 (containing 10% ketamine HCl) (Merial, France). Prior to laparotomy, 2 ml Xylovet (CEVA Sante Animale, France) containing 2.1% lidocaine, was injected intradermally in the abdomen along the site of incision. The small intestine was exteriorized and the cecum was localized. Twelve loops of five cm segments of ileum starting at the ileum-cecum transition were ligated, avoiding all Peyer's patches, while maintaining the existing vasculature. Every other loop was injected with 0.5 ml of bacterial suspension or the saline control using a 26-gauge needle. The injection order of control, wild-type and avirulent Shigella -injected loops were randomized for each rabbit and performed in duplicate within an animal. Loops were returned into the abdominal cavity, the abdomen was closed and the animals were returned to their cage for 4–8 hrs. Animals were sacrificed by intravenous injection of 120 mg kg −1 sodium pentobarbital (Doléthal, Vétoquinol, France). The exudate of each loop was suctioned using an 18-gauge needle and measured before loops were dissected and processed for RNA extraction and histology. Tissue Preparation, RNA Extraction and cDNA Synthesis For immunohistochemical staining of Shigella , 1–2 cm-long pieces of rabbit ileal loop were fixed for 2 days at 4°C in 10 ml of 4% paraformaldehyde in phosphate-buffered saline, embedded in paraffin and sectioned into 5 µm sections (3–4 sections/slide) using a microtome. Sections were deparaffinated, rehydrated and split into antibody control and test groups. Sections were permeabilized for 15 min with 0.1% Triton-X100, treated with 3.3% H 2 O 2 for 15 min and washed. Samples were blocked for 30 min with Ultra V block (Lab Vision Corp; Thermo Fisher) and incubated overnight with an in-house mouse polyclonal anti- S. flexneri serotype 5 serum or without serum as a control. Samples were then incubated with horseradish peroxidase conjugated anti-rabbit antibody (K4002, DAKO) for 1 hr, revealed with 3-amino-9-ethylcarbzole (AEC+, K3461, DAKO), counterstained with hematoxylin (Thermo Shandon) and mounted with aqueous mounting medium (Merck). For RNA extraction, 1 cm long tissue samples were immediately submerged in 4 ml Trizol (Invitrogen), homogenized for 1 min with a tissue homogenizer and stored at –80°C until further processed. A volumen of 1.2 ml of each sample was spun at 4°C for 15 min at 12,000×g to remove debris and DNA, 1 ml of supernatant was mixed with 200 µl chloroform, shaken for 154 seconds, incubated at RT for 2–3 minutes and spun for 10 min at 12,000×g at 4°C. RNA was precipitated by adding 500 µl of the aqueous phase to an equal volume of isopropanol and spun at 14,000×g at RT for 10 min. RNA was washed with 75% ethanol, spun at 14,000×g at 4°C for 10 min, dried and resuspended in 60 µl DEPC-treated H 2 O (Ambion). RNA was quantitated on a spectrophotometer (Nanodrop 2000) and 40 µg of RNA was used for a second RNA purification using the Nucleospin RNA II kit (Macherey-Nagel GmbH) including the on-column digestion of DNA. The final RNA concentration was determined using a spectrophotometer (Nanodrop 2000) and the purity was assessed by agarose gel electrophoresis or the Agilent's 2100 Bioanalyzer. CDNA synthesis was performed on 3 µg of RNA in a 10 µl sample volume using SuperScript II reverse transcriptase (Invitrogen) as recommended by the manufacturer. The RNA was incubated with 0.5 µg of oligo(dT)12–18mers primers (Invitrogen) for 7 min at 70°C and then transferred onto ice. Then, 9 µl of a master mix containing 4 µl of SuperScript II buffer, 2 µl of 0.1 M DTT (Invitrogen), and 1 µl each of dNTPs stock (10 mM) (Invitrogen), Rnasin (40 UI) (Promega) and SuperScript II (Invitrogen) were added to the RNA sample, spun and incubated at 42°C for 60 min followed by 5 min at 70°C to inactivate the enzyme. CDNA was stored at −20°C. Primer Design and Quality Control Primers were designed using the online program Primer3Input ( www.fokker.wi.mit.edu/primer3/input.htm ). Primer selection parameters were set to primer size: 20–26 nts; primer melting temperature: 62 to 64°C; GC clamp: 1; and product size range: generally 120–140 bp but down to 80 bp if no appropriate primers could be identified. Primers were ordered from Invitrogen and tested using serial 2-fold dilutions of cDNA, from 1 in 40 to 1 in 2560, from 8 hour S. flexneri -infected rabbit ileal tissue to ensure efficient and comparable linear amplification of the amplicons across a wide range of target abundance. Cytokine Expression Analysis by Quantitative Real-time PCR QRT-PCR was performed in a total volume of 15 µl including 300 ng of cDNA, primers (0.2 µM each) and 7.5 µl of Power SYBR Green mix (Applied Biosystems). Reactions were run in duplicate on an ABI 7900HT (Applied Biosystems) using the universal thermal cycling parameters (2 min 60°C, 95°C 10 min, 40 cycles of 15 sec at 95°C and 60 sec at 60°C; dissociation curve: 15 sec at 95°C, 15 sec at 60°C and 15 sec at 95°C). Results were obtained using the sequence detection software ABI 7900HT SDS2.2 and analyzed using Microsoft Excel. For all samples, dissociation curves were acquired for quality control purposes. In addition, amplification products were visualized by agarose gel electrophoresis. For gene expression quantification, we used the comparative Ct method. First, gene expression levels for each sample were normalized to the expression level of the housekeeping gene encoding Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) within a given sample (Δ -Ct); the difference between the infected loops compared to the control uninfected (saline only) loop was used to determine the Δ -Δ-C t . The log 2 Δ -Δ-C t ) gave the relative fold increase in gene expression of the test versus the control condition. Statistical significance of the fold difference between the infected and the uninfected samples was calculated with the Wilcoxon Signed Rank test using PRISM software. Bacterial Propagation and Infection The streptomycin-resistant invasive wild-type Shigella flexneri strain M90T-Sm of serotype 5a and its virulence plasmid-cured non-invasive derivative BS176 were thawed from –80°C stocks, grown on Tryptic Soy (TCS) Agar-Congo red (CR) plates (3% BBL-trypticase soy agar (BD Biosciences), 0.1% CR) for 16 hrs at 37°C and kept at 4°C. For infection, one colony of each strain from freshly streaked plates was grown in TCS medium at 37°C for 16 hours with agitation two days before the experiment and kept at room temperature for 8 hrs the next day. In the evening before the experiment, 50 µl of culture of each strain was plated onto TCS plates (without CR) and incubated at 37°C for 16 hrs to form a bacterial lawn. The morning of the experiment, physiological saline (0.9% NaCl) was added to each plate and plates were scraped to resuspend the bacteria. Bacterial suspensions were diluted to OD 600 = 4 (equivalent to 3×10 9 bacteria ml −1 ). As a control, the bacterial suspension was diluted by 10 6 fold and 25 µl were plated in triplicate on TCS-CR plates. After incubation at 37°C for 16 hrs, bacterial colonies were enumerated (approximately 75 colonies are expected) and checked for proper colony morphology (M90T colonies appear red and are smaller than those of BS176). Rabbit Illeal Loop Infections The rabbit illeal loop model protocol was approved by the Comite Regional d'Ethique pour l'Experimentation Animale in Paris 1 (protocol #20070004). Surgery was performed essentially as described previously with minor modifications [19] . Twelve male New Zealand White rabbits weighing 2.4–2.6 kg (Charles River Laboratories, St. Aubin, France) were split into 2 groups for short (4 to 5.5 hrs) and long (8 hrs) time points of infection. Animals were received 10 days prior to surgery and treated with Mucoxid (2.8 g L −1 sodium sulfadimethoxine salt in the drinking water) (CEVA Sante Animale, France) until the day prior to surgery to minimize infection with Coccidia parasites. Three independent experiments using two rabbits each for early and late time points were performed. Rabbits were fasted 24 hrs before infection, sedated intravenously by ear vein injection with 0.05 ml kg −1 Calmivet (Vétoquinol) containing 0.5% acepromazine and anesthesized by the same route with 0.2 ml kg −1 of Imalgene® 1000 (containing 10% ketamine HCl) (Merial, France). Prior to laparotomy, 2 ml Xylovet (CEVA Sante Animale, France) containing 2.1% lidocaine, was injected intradermally in the abdomen along the site of incision. The small intestine was exteriorized and the cecum was localized. Twelve loops of five cm segments of ileum starting at the ileum-cecum transition were ligated, avoiding all Peyer's patches, while maintaining the existing vasculature. Every other loop was injected with 0.5 ml of bacterial suspension or the saline control using a 26-gauge needle. The injection order of control, wild-type and avirulent Shigella -injected loops were randomized for each rabbit and performed in duplicate within an animal. Loops were returned into the abdominal cavity, the abdomen was closed and the animals were returned to their cage for 4–8 hrs. Animals were sacrificed by intravenous injection of 120 mg kg −1 sodium pentobarbital (Doléthal, Vétoquinol, France). The exudate of each loop was suctioned using an 18-gauge needle and measured before loops were dissected and processed for RNA extraction and histology. Tissue Preparation, RNA Extraction and cDNA Synthesis For immunohistochemical staining of Shigella , 1–2 cm-long pieces of rabbit ileal loop were fixed for 2 days at 4°C in 10 ml of 4% paraformaldehyde in phosphate-buffered saline, embedded in paraffin and sectioned into 5 µm sections (3–4 sections/slide) using a microtome. Sections were deparaffinated, rehydrated and split into antibody control and test groups. Sections were permeabilized for 15 min with 0.1% Triton-X100, treated with 3.3% H 2 O 2 for 15 min and washed. Samples were blocked for 30 min with Ultra V block (Lab Vision Corp; Thermo Fisher) and incubated overnight with an in-house mouse polyclonal anti- S. flexneri serotype 5 serum or without serum as a control. Samples were then incubated with horseradish peroxidase conjugated anti-rabbit antibody (K4002, DAKO) for 1 hr, revealed with 3-amino-9-ethylcarbzole (AEC+, K3461, DAKO), counterstained with hematoxylin (Thermo Shandon) and mounted with aqueous mounting medium (Merck). For RNA extraction, 1 cm long tissue samples were immediately submerged in 4 ml Trizol (Invitrogen), homogenized for 1 min with a tissue homogenizer and stored at –80°C until further processed. A volumen of 1.2 ml of each sample was spun at 4°C for 15 min at 12,000×g to remove debris and DNA, 1 ml of supernatant was mixed with 200 µl chloroform, shaken for 154 seconds, incubated at RT for 2–3 minutes and spun for 10 min at 12,000×g at 4°C. RNA was precipitated by adding 500 µl of the aqueous phase to an equal volume of isopropanol and spun at 14,000×g at RT for 10 min. RNA was washed with 75% ethanol, spun at 14,000×g at 4°C for 10 min, dried and resuspended in 60 µl DEPC-treated H 2 O (Ambion). RNA was quantitated on a spectrophotometer (Nanodrop 2000) and 40 µg of RNA was used for a second RNA purification using the Nucleospin RNA II kit (Macherey-Nagel GmbH) including the on-column digestion of DNA. The final RNA concentration was determined using a spectrophotometer (Nanodrop 2000) and the purity was assessed by agarose gel electrophoresis or the Agilent's 2100 Bioanalyzer. CDNA synthesis was performed on 3 µg of RNA in a 10 µl sample volume using SuperScript II reverse transcriptase (Invitrogen) as recommended by the manufacturer. The RNA was incubated with 0.5 µg of oligo(dT)12–18mers primers (Invitrogen) for 7 min at 70°C and then transferred onto ice. Then, 9 µl of a master mix containing 4 µl of SuperScript II buffer, 2 µl of 0.1 M DTT (Invitrogen), and 1 µl each of dNTPs stock (10 mM) (Invitrogen), Rnasin (40 UI) (Promega) and SuperScript II (Invitrogen) were added to the RNA sample, spun and incubated at 42°C for 60 min followed by 5 min at 70°C to inactivate the enzyme. CDNA was stored at −20°C. Primer Design and Quality Control Primers were designed using the online program Primer3Input ( www.fokker.wi.mit.edu/primer3/input.htm ). Primer selection parameters were set to primer size: 20–26 nts; primer melting temperature: 62 to 64°C; GC clamp: 1; and product size range: generally 120–140 bp but down to 80 bp if no appropriate primers could be identified. Primers were ordered from Invitrogen and tested using serial 2-fold dilutions of cDNA, from 1 in 40 to 1 in 2560, from 8 hour S. flexneri -infected rabbit ileal tissue to ensure efficient and comparable linear amplification of the amplicons across a wide range of target abundance. Cytokine Expression Analysis by Quantitative Real-time PCR QRT-PCR was performed in a total volume of 15 µl including 300 ng of cDNA, primers (0.2 µM each) and 7.5 µl of Power SYBR Green mix (Applied Biosystems). Reactions were run in duplicate on an ABI 7900HT (Applied Biosystems) using the universal thermal cycling parameters (2 min 60°C, 95°C 10 min, 40 cycles of 15 sec at 95°C and 60 sec at 60°C; dissociation curve: 15 sec at 95°C, 15 sec at 60°C and 15 sec at 95°C). Results were obtained using the sequence detection software ABI 7900HT SDS2.2 and analyzed using Microsoft Excel. For all samples, dissociation curves were acquired for quality control purposes. In addition, amplification products were visualized by agarose gel electrophoresis. For gene expression quantification, we used the comparative Ct method. First, gene expression levels for each sample were normalized to the expression level of the housekeeping gene encoding Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) within a given sample (Δ -Ct); the difference between the infected loops compared to the control uninfected (saline only) loop was used to determine the Δ -Δ-C t . The log 2 Δ -Δ-C t ) gave the relative fold increase in gene expression of the test versus the control condition. Statistical significance of the fold difference between the infected and the uninfected samples was calculated with the Wilcoxon Signed Rank test using PRISM software. Results We designed and tested twenty-four real-time PCR primer pairs for a quantitative gene expression analysis of key rabbit genes involved in the innate and adaptive immune response to infection ( Table 1 ). We assayed the expression of three chemokines [IL-8, chemokine (C-C motif) ligand (CCL)-4, and CCL20], sixteen cytokines [Interleukin (IL)-1β IL-2, IL-4, IL-6, Il-10, IL-12p35, IL12/23p40, IL-17A, IL-17F, IL-18, IL-21, IL-22, Interferon (IFN)- β, IFN-γ, Transforming growth factor (TGF)-β and Tumor necrosis factor (TNF)-α], three antimicrobials [Leukocyte Protein (LeukoP) p15, neutrophil defensin NP-3α, and the cathelicidin CAP-18] and two enzymes [inducible Nitric oxide synthase (iNOS) and cyclo-oxygenase (COX)-2]. This set of twenty-four genes exhibited a wide range of basal expression levels in uninfected loops (5.9 to 14.8 fold Ct over the expression level of the housekeeping gene GAPDH) ( Figure 1 ). The most abundant transcript at the steady state level was IL-8, followed by CCL4, IL-6, IL-1β and IL-18, while rather low levels were detected for IL-17A, COX-2, IL-4, IFN-β, IL-12/23p40, IL-12p35, IL-2, iNOS and IL-22. 10.1371/journal.pone.0036446.g001 Figure 1 Gene expression changes in rabbit illeal loops infected with avirulent or virulent Shigella . Transcriptional response at A) early (4 to 5.5 hrs) and B) late (8 hrs) time points post infection with the avirulent BS176 strain or the virulent M90T strain. The fold induction in gene expression over uninfected (UI) control loops (Δ-Δ-Ct) are given as the median of twelve samples obtained from six rabbits in three independent experiments. Basal level expressions are listed as the median Ct value above the housekeeping gene GAPDH (Δ-Ct) from twelve uninfected loops at the early time point from six rabbits. As GAPDH is an abundant transcript, smaller changes in fold Ct compared to GAPDH signify more abundant gene transcripts. Increased shading highlights increasing upregulation of statistically significan values while stripes signify downregulation. The statistical significance (p-value) was calculated using the Wilcoxon Signed Rank Test. 10.1371/journal.pone.0036446.t001 Table 1 Rabbit primer pairs for quantitative RT-PCR analysis. Rabbit target gene Forward and reverse primer sequences Primer location within CDS Target size NCBI Accession # Control gene GAPDH PS182:TGACGACATCAAGAAGGTGGTG; PS183:GAAGGTGGAGGAGTGGGTGTC Exon 1 of 1 120 nts NM_001082253 Cytokines & Chemokines CCL4 PS343:GAGACCACCAGCCTCTGCTC; PS344:TCAGTTCAGTTCCAAGTCATCCAC Exon 2 and 3 from 3 123 nts NM_001082196 CCL20 PS565:TATCGTGGGCTTCACACAGC; PS566:CCATTCCTTCTTCGGATCTGC Exon 2 and 3 from UD 115 nts Trace Archive IFN-β PS291:TCCAACTATGGCACGGAAGTCT; PS292:TTCTGGAGCTGTTGTGGTTCCT Exon 1 from 1 133 nts XM_002707968 IFN-γ PS186:TGCCAGGACACACTAACCAGAG; PS187:TGTCACTCTCCTCTTTCCAATTCC Exon 1 and 2/3 from 4 127 nts NM_001081991 IL-1β PS168:TTGAAGAAGAACCCGTCCTCTG; PS169:CTCATACGTGCCAGACAACACC Exon 3/4 and 4 from∼6 128 nts NM_001082201 IL-2 PS275:GCCCAAGAAGGTCACAGAATTG; PS276:TGCTGATTGATTCTCTGGTATTTCC Exon 2/3 and 3 from 4 128 nts NM_001163180 IL-4 PS267:CGACATCATCCTACCCGAAGTC; PS268:CCTCTCTCTCGGTTGTGTTCTTG Exon 1 and 2/3 from 4 122 nts NM_001163177 IL-6 PS170:CTACCGCTTTCCCCACTTCAG; PS171:TCCTCAGCTCCTTGATGGTCTC Exon 2 from ∼5 135 nts NM_001082064 IL-8 PS287:CCACACCTTTCCATCCCAAAT; PS288:CTTCTGCACCCACTTTTCCTTG Exon 2 and 3 from 4 122 nts NM_001082293 IL-10 PS281:CTTTGGCAGGGTGAAGACTTTC; PS282:AACTGGATCATCTCCGACAAGG Exon 1 and 3 from 5 126 nts NM_001082045 IL-12p35 PS214:AAGGCCAGACAAACTCTAGAATTC; PS215:TTGGTTAACTCCAGTGGTAAACAGG Exon 3/4 and 4/5 from ∼8 116 nts XM_002716291 IL-12/IL-23p40 PS211:CTCCGAAGAAGATGGCATTACC; PS212:TCTCCTTTGTGGCAGGTGTATTG Exon 2 from 6 126 nts XM_002710347 IL-17A PS591:CCAGCAAGAGATCCTGGTCCTA; PS592:ATGGATGATGGGGGTTACACAG Exon 3 from 3 112 nts XM_002714498 IL-17F PS589:AAAATCCCAAAGTGGAGGATGC; PS590:AGCGGTTCTGGAAGTCATGTGT Exon 2 from 3 138 nts XM_002714499 IL-18 PS575:ACCAAGGACAGCAACCTGTGTT; PS576:ACAGAGAGGCTTACAGCCATGC Exon 3 and 4 from 5 120 nts NM_001122940 IL-21 PS579:GCTGGCAACATGGAAAGGATAG; PS580:TTGCCCTTTGGAGCTTGATTTA Exon 4 from 8 84 nts XM_002717257 IL-22 PS567:ACCTCACCTTCATGCTGGCTAA; PS568:CATGGAACAGCTCATTCCCAAT Exon 1 and 2 from 5 84 nts XM_002711248 TGF-β PS199:CAGTGGAAAGACCCCACATCTC; PS200GACGCAGGCAGCAATTATCC Exon 6 and 7 from ∼8 140 nts NM_001082660 TNF-α PS174:CTGCACTTCAGGGTGATCG;PS175:CTACGTGGGCTAGAGGCTTG Exon 1 and 3 from ∼4 133 nts NM_001082263 Antimicrobials CAP-18 PS176:CCCAAGAGTCCCCAGAACCTAC; PS177:TCTGTCCTGGGTGCAAGTTTC Exon 3/4 and 4 from ∼4 130 nts NM_001082305 LeukoP PS225:GTCGCCGTCTGAGATATGAGGA; PS226:GTTGAGTGGGATCCTGGATTTG Exon 1 and 2 from 2 140 nts NM_001082325 NP3α PS205:ACCTTACAGGGGAGGAAAGCTC; PS206:GTACATAGCGGGCTCCATTGAC Exon 1 and 2 from 2 132 nts NM_001082298 Enzymes COX-2 PS329:CGGATTCTACGGTGAAAACTGC; PS330:GACGATGTTCCAGACTCCCTTG Exon 1 and 2 from 10 124 nts NM_001082388 iNOS PS573:GACGTCCAGCGCTACAATATCC; PS374:GATCTCTGTGACGGCCTGATCT Undetermined 102 nts XM_002718780 Rabbit primer pairs were designed within the coding sequence (CDS) of rabbit genes identified either by homology to the mouse or human gene using the NCBI rabbit. Trace Archive database, or predicted or experimentally-determined CDSs available at NCBI. All primers were designed using identical design parameters (see Materials and Methods ) and made to span exon junctions when possible. CAP: cationic animicrobial protien (a cathelicidin (LL-37 in humans); CCL: cheomokine (C-C motif) ligand; COX: cyclo-oxygenase; iNOS: inducible Nitric oxide synthase; IFN: Interferon; IL: Interleukin; LeukoP: Leukocyte protein (cationic antimicrobial peptide); NP: Neutrophil protein (a defensin); TGF: Transforming growth factor. We then used the established rabbit ileal loop model of shigellosis [19] to characterize the host response to acute S. flexneri infection at both early (4 to 5.5 hrs) and late (8 hrs) time points after infection. For each time point, six rabbits from three independent experiments were utilized. Duplicate ileal loops for each rabbit were injected with the virulent, invasive S. flexneri strain M90T, the avirulent and noninvasive S. flexneri strain BS176, and physiological saline alone for a total of twelve samples per time point for each condition. At indicated times after infection, animals were sacrificed and portions of the ligated loops were processed for RNA extraction and histology. At the early time point of infection (4 to 5.5 hrs), the avirulent strain did not induce any structural changes in the epithelium (data not shown), while only minor morphological changes in the intestinal structure and invasion of the villi could be detected by histology for the virulent strain by 5.5 hours ( Figure 2 ). Villi were slightly swollen and showed early signs of leukocyte infiltration. In contrast, villi destruction, leukocyte infiltration and intracellular bacterial localization were clearly visible for the virulent strain by 8 hours of infection. Loops infected with the avirulent strain for 8 hours showed no apparent morphological alterations as compared to the non-infected control ( Figure 2 ). 10.1371/journal.pone.0036446.g002 Figure 2 Immunohistochemistry of rabbit ileal loops left uninfected or infected with Shigella strains. Tissue sections of rabbit illeal loops taken at either 4–5.5 or 8 hours post surgery were stained for Shigella (brown) with a murine polyclonal anti- S. flexneri 5a LPS serum and counterstained with hematoxylin. Representative images are shown. Images are taken at 100x (top) and 400x (bottom) magnification. Destruction of villi and intracellular localization of the invasive strain M90T can clearly be seen only by 8 hrs and not by 5.5 hrs, whereas clear morphological changes were not observed even after 8 hrs of infection with the avirulent BS176 strain. Arrows point to epithelial cells that also stain for virulent S. flexneri . The gene expression profile at early times of infection with virulent S. flexneri revealed a substantial increase in pro-inflammatory cytokines and an early antimicrobial response as compared to uninfected (UI) control loops ( Figure 1 ). This early response was dominated by IL-22, TNF-α, iNOS and IL-1β expression (3.7 to 17 fold increase), while IL-17A, IL-17F and IL-4 expression were strongly induced in some but not all samples ( Figure 3A ). Indeed, infection with the virulent strain also led to an increase in expression (1.5 to 2.3 fold) for most of the other pro-inflammatory mediators tested (CCL20, IL-6, IL-12p40, and IFN-γ) as well as to an upregulation (1.7 fold) of the antimicrobial LeukoP as compared to the UI control loops. Infection with avirulent S. flexneri led to a substantial upregulation of only IL-22 (3.3 fold), although upregulation (1.4 fold) could also be detected for IL-17F and TNF-α. 10.1371/journal.pone.0036446.g003 Figure 3 Gene expression changes in rabbit ileal loops infected with avirulent or virulent Shigella . Transcriptional response at A) early (4 to 5.5 hrs) and B) late (8 hrs) time points post infection with the avirulent BS176 strain or the virulent M90T strain. The median fold induction over uninfected controls for each gene is indicated in red while symbols represent values from individual loops. A total of twelve loops from six animals obtained in three independent experiments are shown. Note the difference in scale for the early and late time points. After 8 hours of infection, at the more acute phase, the virulent strain had induced an extensive increase in many cytokines, particularly IL-22, IL-17A and IL-17F (132 to 233 fold) ( Figure 2 ). These cytokines were also highly upregulated (8 to 28 fold) by infection with the avirulent strain. While the expression of many pro-inflammatory cytokines (IL-6, TNF-α, IL-1β, IFN-γ and IL12p40) was also strongly induced (18–69 fold) by the virulent strain, these cytokines showed either no increase (IFN-γ and IL12p40) or only slightly elevated levels (1.4 to 1.9 fold) for the avirulent strain. Interestingly, infection by virulent S. flexneri also led to a significant decrease (2.0 fold, p<0.01) in IL-18 at the late time point ( Figure 1 and Figure 3B ). Furthermore, all chemoattractants (IL-8, CCL4 and CCL20) showed large increases (24 to 32 fold) in expression in loops infected with virulent S. flexneri, while remaining mostly unresponsive when challanged with the avirulent mutant, although some samples showed high CCL20 expression ( Figure 3B ). Likewise, the antimicrobials CAP-18, LeukoP and iNOS showed a strong induction (11 to 52 fold) after infection with the virulent strain at the late time point while all, except iNOS (3.7 fold increase), remained largely at the basal level during challenge with the avirulent strain. The potent pro-inflammatory response elicited was not offset by an increased abundance in anti-inflammatory cytokine transcripts as TGF-α expression remained largely unchanged under all conditions tested and IL-10 was only slightly (1.9 fold), albeit significantly (p<0.01), increased by the virulent strain after 8 hours of infection. IL-4 expression continued to increase over time (from 2.3 to 4.4 fold) for the wild-type strain but remained unchanged for the avirulent mutant while IL-21 transcript showed a significant increase (4.0 fold) only at 8 hours post infection with the virulent strain. Finally, IL-2 and IFN-β remained largely unchanged in all conditions tested. We hereby identified four classes of gene expression profiles: i. those that are highly upregulated only by infection with virulent S. flexneri (CCL4, CCL20, IFN-γ, IL-1β, IL-4, IL-6, IL-8, IL-12/23p40, IL-21, TNF-α, CAP-18, LeukoP and COX-2); ii. those that are down-regulated by infection with virulent S. flexneri (IL-18); iii. those that are highly upregulated by infection of both the virulent and avirulent strains (IL-22, IL-17A, IL-17F and iNOS); and iv. those that remain largely unmodulated by infection with either the avirulent or the virulent strain (IL-2, Il-12p35, TGF-β and IFN-β). In addition, our results point to IL-22, IL-17A and IL-17F as key innate responders to bacterial challenge in the intestine. Indeed, infection with virulent Shigella induced a rapid and strong (133 to 233 fold) induction by 8 hours while the eight next most highly expressed cytokines averaged 32 fold (18 to 69 fold) induction. Interestingly, a significant induction (8 to 28 fold) of IL-22, IL-17A and IL-17F was also observed for the avirulent strain, while induction of the eight next most expressed cytokines during infection with the virulent strain were either not induced or induced weakly (<2-fold). Discussion Shigella flexneri infection in humans leads to an acute intestinal infection with symptoms ranging from watery diarrhea to severe bloody mucoid diarrhea accompanied by fever and intestinal cramps [25] . Immunohistological analysis of rectal biopsies from Shigella -infected patients revealed extensive synthesis of a number of proinflammatory cytokines (IL-1α, IL-1β, IL-6, TNF-α and IFN-γ) at the local site of infection during the acute phase and increasing frequencies of cytokine-producing cells correlated with increasing severity of the disease. In addition, Shigella also induced the local production of a large number of other cytokines (TNF-α, IL-4, IL-10, TGF-β, IL-1ra and IL-8) [23] . Marked inflammation was accompanied by infiltration of granulocytes, T lymphocytes, macrophages and natural killer (NK) cells [26] . As seen during the natural infection, Shigella infection of rabbit ileal loops led to a large increase in mRNA abundance for the pro-inflammatory cytokines IL-1β, TNF- α, IL-6, IL-4, IFN-γ and IL-8, highlighting the similar cytokine response elicited in the two systems. In addition, we observed a large induction for IL-12/23p40 and a moderate induction of IL-21. IL-12 and IL-21 activate NK cells to promote a Th1 response while IL-23 is critical for IL-22 expression and sustaining Th17 responses. In vitro studies characterizing the colonic epithelial cell response to invasive S. flexneri infection using microarrays revealed a strong upregulation of genes encoding chemokines (IL-8, CCL20, CXCL1 and 2), cytokines (GM-CSF and TNF-α) and adherence molecules (ICAM-1) [26] that together likely induce the strong recruitment of PMNs, NK cells, lymphocytes and dendritic cells that is observed during the natural infection of the human colon. Similarly, we observed a strong induction of mRNA abundance of the tested chemokines (CCL4, CCL20 and IL-8) and cytokines (TNF-α and others) in the rabbit model. The transcriptional regulation for IL-8 also showed high concordance with relative IL-8 protein levels detected in infected rabbits at various times after infection [27] . In the rabbit ileum, most IL-8 is produced by intestinal epithelial cells [27] ; however, in human rectal biopsies of shigellosis patients, IL-8 is confined mainly to the crypt lumen while in healthy biopsies, IL-8 immunostaining gives strong labeling in the crypts [23] , suggesting a localized storage depot at this location. Whether this kind of IL-8 storage is present in the rabbit ileum is unclear although it could explain the highly abundant IL-8 transcript level detected in our uninfected rabbit control loops. The pro-inflammatory cytokine IL-18 was the only cytokine that decreased (2 fold) in expression levels in the rabbit model after infection with the virulent Shigella strain. IL-18 is produced by macrophages, which are efficently killed when infected with virulent Shigella [22] . IL-18 mRNA down-regulation was also previously observed in an IL-8 supplemented mouse model of shigellosis [28] , although rectal tissues from patients infected with S. dysenteriae show an influx (3 fold) of IL-18-expressing cells at the acute stage compared to healthy controls [29] . As no transcriptome data is available for IL-18 in humans, these discrepant findings may be due to the method used to analyze IL-18 abundance or due to differences in the time points analyzed after the onset of disease. For anti-inflammatory markers, we included IL-10 and TGF-β in our analysis. Human rectal biopsies from Shigella -infected patients had shown increased production of TGF-β and IL-10 cytokine expression as compared to healthy controls. However, our transcriptional analysis in the rabbit did not identify an upregulation for TGF-β and only a small increase (1.9 fold) was observed for IL-10 [23] . Notably, increased levels of TGF-β and IL-10 were detected at the convalescent stage for the humans while our analysis in the rabbit focuses on the early acute phase of the disease, possibly explaining the differences observed. Differences in the anatomical location may also play a role. In addition to the extensive pro-inflammatory response elicited during infection with Shigella , Shigella was also found to strikingly down-regulate the production of the cathelicidin LL-37, the human homolog of CAP-18, in the epithelial lining of rectal biopsies from Shigella -infected patients [30] while LL-37 could be detected in infiltrating granulocytes and macrophages. Shigella could also down-regulate LL-37 and human beta defensin 3 in an in vitro infection models using human colonic epithelial cells [31] . As seen in the human biopsies, CAP-18 is exclusively localized to the surface epithelium in healthy rabbit rectal samples and Shigella infection also led to an almost complete loss of this pool with a concomittant infiltration of inflammatory cells (macrophages and neutrophils) expressing CAP-18 [32] . Notably, CAP-18 was not down-regulated in the proximal colon and its expression pattern in the ileum has not been explored. In our rabbit ileal loop model, CAP-18 transcript levels were upregulated only at the eight hour time point when abundant inflammatory cell infiltration into the villi was observed and levels of the neutrophil antimicrobial LeukoP transcript was also strongly enriched. These observations suggest that the observed increase in CAP-18 mRNA levels may be due to inflammatory cell influx. iNOS expression was markedly upregulated even at the early time point, revealing a first responder type of response of this enzyme. These results also accurately recapitulate the natural infection. Strong iNOS upregulation in the surface epithelium and infiltration of iNOS positive cells, accompanied by increased iNOS mRNA levels, was observed in rectal biopsies from human patients with acute shigellosis but not in healthy controls [33] , [34] . Furthermore, upregulation of COX-2 transcripts in the rabbit model echoes the increased prostraglandin levels observed in stool samples of shigellosis patients [34] . In addition to the previously established activation of the classical pro-inflammatory cytokines and chemokines, we observed that the early transcriptional response in rabbits infected with Shigella was dominated by the cytokines IL-22, IL-17A and IL-17F. Interestingly, unlike any of the other cytokines, these cytokines were quite strongly induced during infection with both the avirulent non-invasive strain as well as with the invasive wild-type Shigella strain. These results point to IL-22, IL-17A and IL-17F as first responders to not only pathogenic bacteria but also to changes in gut bacterial composition. In the adaptive immune system, Th17 cells are a relatively newly discovered Th subset that produce IL-17A, IL-17F, IL-22 and IL-21 and are important for the host defense to bacterial pathogens at mucosal surfaces. Th17 cells are the primary Th cell type primed in a murine lung model of shigellosis and are important for mediating protective immunity [35] . However, in this mouse model, IL-17A and IL-22 could be detected above uninfected controls only by 6 days after infection. In the rabbit illeal loop model, the rapid induction (4 to 8 hrs) of IL-22, IL-17A and IL-17F is incompatible with Th17 stimulation and points to IL-17 and IL-22 production by innate immune cells. Indeed, the IL-22-IL-17 axis of innate immune cells is a very recent, yet rapidly expanding, field of research [36] , [37] , [38] . IL-17A and IL-17F are able to induce proinflammatory cytokines, chemokines and antimicrobial peptide expression, as well as to promote neutrophil recruitment [39] , [40] and provide mucosal host defenses against fungal and bacterial infections [39] , [41] . As with Shigella , rapid production (4 to 24 hours) of IL-17A by the host was observed after infection with a number of bacterial pathogens ( S. enterica serovar Typhimurium in the intestine, L. monocytogenes in the liver, Klebsiella pneumoniae in the lung and with Escherichia coli after intraperitonneal challenge) [42] , [43] , [44] , [45] , [46] . Conversely, IL-22, an IL-10-related cytokine, induces gene expression of molecules involved in tissue inflammation, antimicrobial defense and tissue repair and is important for regulating homeostasis of epithelial cells at barrier surfaces [37] . Interestingly, IL-22 promotes innate immunity to bacterial pathogens that directly interact with the host epithelium and induce functional or pathological changes in the epithelium. IL-22 was thus shown to be important for inducing a protective host immunity to the extracellular gram-negative pathogens K. pneumoniae in the lung and C. rodentium in the intestine [41] , [47] but did not have an important role in the host defense against infection with M. tuberculosis , Mycobacterium avium , L. monocytogenes and S. mansoni [48] , [49] . In conclusion, the development of this set of qPCR primers targeted to key genes involved in the immune response to infection showed that the rabbit ileal loop model of infection by Shigella is an accurate model of the disease not only morphologically but also with respect to the gene activation underlying the response; it revealed a clear immunological response at early time points of the infection and it identified the activation of first responder genes and previously unexplored cytokines involved in the innate mucosal immunity to Shigella . Thus, this primer set will be highly valuable to characterize the virulence of mutants of Shigella and other pathogens in a quantitative manner in this in vivo model, especially since many bacterial effectors specifically target the innate immune response of the host [50] , [51] . Finally, these twenty-five primer pairs should prove useful to investigators using the rabbit as an animal model for any type of system that elicits an immune response.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6729954/

Switching Mature Retinal Ganglion Cells to a Robust Growth State In Vivo : Gene Expression and Synergy with RhoA Inactivation

The inability of mature CNS neurons to regenerate injured axons has been attributed to a loss of inherent growth potential of cells and to inhibitory signals associated with myelin and the glial scar. The present study investigated two complementary issues: (1) whether mature CNS neurons can be stimulated to alter their gene expression profile and switch into a strong growth state; and (2) whether inactivating RhoA, a convergence point for multiple inhibitory signals, is sufficient to produce strong regeneration even without activating the growth state of neurons. In the mature rat, retinal ganglion cells (RGCs) normally fail to regenerate axons through the injured optic nerve but can be stimulated to do so by activating macrophages in the eye (e.g., by lens injury). To investigate underlying changes in gene expression, we retrogradely labeled RGCs with a fluorescent dye, performed optic nerve surgery with or without lens injury, and 4 d later, dissociated retinas, isolated RGCs by fluorescence-activated cell sorting, and examined their profiles of gene expression using microarrays. To investigate the effects of inactivating RhoA, we transfected RGCs with adeno-associated viruses carrying a gene for C3 ribosyltransferase. Our results show that, with appropriate stimulation, mature CNS neurons can undergo dramatic changes in gene expression comparable with those seen in regenerating neurons of the PNS, and that RhoA inactivation by itself results in moderate regeneration, and strongly potentiates axon regeneration through the mature optic nerve when the growth state of neurons is activated.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2676634/

Influence of biologically inspired nanometer surface roughness on antigen–antibody interactions for immunoassay–biosensor applications

Current research efforts to improve immunoassay–biosensor functionality have centered on detection through the optimal design of microfluidic chambers, electrical circuitry, optical sensing elements, and so on. To date, little attention has been paid to the immunoassay–biosensor membrane surface on which interactions between antibodies and antigens must occur. For this reason, the objective of the present study was to manipulate the nanometer surface roughness of a model immunoassay–biosensor membrane to determine its role on sensitivity and specificity. It was hypothesized that surface roughness characteristics similar to those used by the body's own immune system with B-lymphocyte cell membranes would promote antigen-antibody interactions and minimize non-specific binding. To test this hypothesis, polystyrene 96-well plate surfaces were modified to possess similar topographies as those of B-lymphocyte cell membranes. This was accomplished by immobilizing Protein A conjugated gold particles and Protein A conjugated polystyrene particles ranging in sizes from 40 to 860 nm to the bottom of polystyrene wells. Atomic force microscopy results provided evidence of well-dispersed immunoassay–biosensor surfaces for all particles tested with high degrees of biologically inspired nanometer roughness. Testing the functionality of these immunosurfaces using antigenic fluorescent microspheres showed that specific antigen capture increased with greater nanometer surface roughness while nonspecific antigen capture did not correlate with surface roughness. In this manner, results from this study suggest that large degrees of biologically inspired nanometer surface roughness not only increases the amount of immobilized antibodies onto the immunosurface membrane, but it also enhances the functionality of those antibodies for optimal antigen capture, criteria critical for improving immunoassay–biosensor sensitivity and specificity. Introduction There are numerous immunoassay–biosensor applications necessitating highly sensitive pathogen detection. For example, aerosolized Bacillus anthracis spores are odorless, invisible to the naked eye, have the potential to travel many kilometers, and can survive for decades in ambient conditions. Extrapolation from primate studies have shown that between 1 and 3 of these spores may be sufficient for an infection ( Inglesby 2002 ). Unfortunately, current immunoassay–biosensor limitations lack the sensitivity and specificity for proper B. anthracis spore detection ( Service 2005 ). Hence, device improvement for the detection of such pathogens is of paramount importance. Although there are a number of different designs to improve immunoassay–biosensor capabilities, one approach that has not received much attention to date is to mimic the nanostructure surface roughness of cells from our own immune system. Clearly, our own immune system has been optimized for antigen–antibody capture. For example, the avidity of the non-covalent interactions on a B-lymphocyte's membrane suggests that many properties (such as flexibility, charge, and roughness) may promote antigen capture. Several studies have observed and reported the nanometer membrane topography of a lymphoid cell using atomic force microscopy ( Damjanovich et al 1995 ; Cricenti et al 1999 ; Sakaue and Taniguchi 2001 ) or scanning electron microscopy ( Setum et al 1993 ). It should not be surprising that our own immune cells have highly nanostructured membranes due to the presence of membrane-linked proteins, phospholipid bilayers, and other bioactive molecules. Thus, it should also not be surprising that computational modeling has proposed that promoting surface roughness may be one way to enhance antigen capture on immunoassay–biosensor surfaces through enlarged antigen contact surface area ( Zheng and Rundell 2003 ). In addition to increased surface area, nanoscale roughness on materials allows for unique energetics through greater portions of surface defects and altered electron delocalizations. Because of this, nanometer surface roughness has been shown to influence the behavior of many cell types. For example, studies have demonstrated increased adhesion and growth of endothelial cells ( Miller et al 2004 ), smooth muscle cells ( Miller et al 2004 ), neurons ( Ejiofor et al 2004 ), osteoblasts ( Price et al 2003 ), and leukocytes ( Eriksson et al 2001 ) on nanometer compared with micron rough surfaces. Surface topography causes modulation of chemokines and cytokines in macrophages ( Refai et al 2004 ), activation of platelets and monocytes ( Hsu et al 2004 ), and changes in the locomotion of different T cell types ( Mello et al 2003 ). Although showing promise for implant/tissue engineering applications, the use of nanometer surface roughness on immunoassay–biosensor membranes for enhancing antigen-antibody capture remains largely uninvestigated. For all of the above reasons, the objective of the present study was to investigate antigen capture on model immunoassay–biosensor surfaces of varying degrees of nanometer roughness. It is proposed that this biologically inspired nanometer surface roughness is one factor that naturally promotes antigen–antibody interactions which has yet to be explored in current immunoassay–biosensor designs. Materials and methods Immunosurface preparation To determine the size of particles that should be used to model the surface roughness of B-lymphocytes, imaging software (ImageJ) was employed to evaluate the change in surface area (that is, the ratio of the outlined surface area of the cell membrane to that of a circle) of a B-lymphocyte from an image provided in the literature ( Roitt et al 1993 ). The percentage change in surface area was calculated to be 1.851431 ± 0.034405 (average ± SEM) This value was close to what could be obtained by using 860-nm diameter particles placed on flat immunosurfaces; particles of two additional sizes (specifically, 40 and 460 nm) were added in this study for comparison purposes. The model immunosurfaces were constructed in three layers through physisorption. IgG antibodies comprised the first layer, the second layer consisted of either Protein A (PA) or PA conjugated particles, and the third layer contained the second antibody layer ( Figure 1 ). The surface roughness was controlled by the size and surface density of the immobilized PA conjugated particles. All layers were constructed at room temperature (27°C). Throughout this immunosurface construction procedure, the samples were not allowed to dry to avoid non-uniformity ( Hayes 1998 ). Construction of layer 1 In the present study, mouse IgG2aαDNP (anti-2,4-dinitrophenol, kappa specific; Accurate Chemical) antibodies were used for the active surfaces and non-specific mouse IgG2a (kappa specific; Sigma) antibodies were used for the inactive surfaces. These antibodies were diluted in 1X phosphate-buffered saline (PBS) without ions (4.0 g NaCl, 0.1 g KCl, 0.75 g Na 2 HPO 4 , and 0.1 g KH 2 PO 4 added to 500 mL Milli-Q water, pH adjusted to 7.4) to a final concentration of either 11.1, 6.7, or 3.3 μg/mL. 50 μL of this antibody solution was aliquoted to each sample well of a 96-well plate (poly-styrene, flat bottom without lid, high binding, non-sterile, EIA/RIA; Corning). The antibodies were adsorbed to the polystyrene surface for 1 hour. Samples were gently rinsed 5 times with 200 μL of PBS in each well. The adsorption of the initial antibody layer results in a random orientation ( Lu et al 1996 ), leaving some correctly oriented accessible Fc regions for the secondary attachment of PA. Construction of layer 2 The second layer consisted of either PA (Sigma) or PA conjugated particles. The PA for the conventional surface was diluted in PBS to a final concentration of 4 μg/mL. 50 μL of the PA solution was adsorbed for 20 min and was gently washed 5 times with 200 μL of PBS in each well. The PA conjugated particles included 40-nm gold particles (OD = 10.1; Ted Pella, Inc), 460-nm yellow fluorescent polystyrene particles (0.1% w/v; Spherotech, Inc), and 860-nm polystyrene particles (1.0% w/v; Spherotech, Inc). PA conjugated particle concentrations were used at stock solution concentrations (8.7 × 10 11 particles/mL, 18.0 × 10 9 particles/mL, and 14.0 × 10 9 particles/mL for the 40, 460, and 860 nm particles, respectively). To ensure that the only protein in the solution was PA conjugated to the particles, each solution was microcentrifuged (Model V, VWR) at 5585 × g at 4°C for 12 min, whereby the supernatant was removed and replaced with an equal volume of PBS. After replacement of the supernatant, the preparation was vortexed at 3200 rpm (speed setting 10) (Mini Vortexer, VWR) and was sonicated for 2 min (Aquasonic 75T, 90 W, VWR) to resuspend the particles in solution. This procedure was repeated twice. 50 μL of the final particle solution was adsorbed onto the first layer of antibodies overnight. Alternatively, 50 μL of the final particle solution was further diluted from this stock solution in PBS to the desired dilution factor (1:0, 1:10, 1:100, 1:1000, and 1:10,000) which was then adsorbed overnight. The 96-well plate was covered with a lid as much as possible to avoid evaporation and contamination. After the particles adsorbed overnight, the samples were gently washed 5 times with 200 μL PBS in each well. Construction of layer 3 The second antibody layer, either specific or non-specific, was adsorbed to form the third and final layer of the model immunoassay surface. The procedure for the third layer was identical to that for the first layer. The PA on the second layer served to correctly orient the antibodies through binding of PA to the Fc region of the antibody, allowing the Fab regions to be accessible to the antigen. Immunosurface characterization AFM measurements of surface roughness Atomic force microscopy (AFM) was used to determine the topology and roughness of the various surfaces created in the present study. Specifically, samples were prepared for AFM by constructing only the first two layers of the immunoassay membrane (as described above). Samples were mounted using double-stick tape (3M) onto 12-mm mica specimen discs (Ted Pella, Inc.) and were imaged using a Veeco Nanoscope IIIa MultiMode scanning probe microscopy (SPM) in Tapping™ mode. Standard AFM tips and measuring conditions (thickness: 4 μm, width: 30 μm, length: 125 μm, force constant: 42 N/m, resonance frequency: 320 kHz) were used in non-contact/Tapping ™ mode with a reflexive coating (Pacific Nanotechnology). Three AFM 5 × 5 μm scans at 512 × 512 lines per image were taken for each sample, for 3 different samples of each surface type, at a rate of 1 Hz and a velocity of 10 μm/s. The average room temperature during imaging was 24°C with an average humidity of 30%–40%. Three types of roughness measurements were calculated using Nanoscope IIIa 4.43r8 software in this study: root-mean-square roughness (RMS), average surface roughness (Ra), and the change in surface area (ΔSA) or Wenzel ratio. Three-dimensional topographic images were constructed using WSxM 4.0 Develop 4.4 software (Nanotec Electronica S.L.). Hydrophilicity measurements To determine hydrophilicity of the model immunoassay membranes, static contact angles were measured (Cam-Plus Contact Angle Reader, Tantec) in triplicate. In this manner, contact angles on each of the surfaces of interest to the present study (including the conventional surfaces) were measured immediately after adding 1 μL of Milli-q water using the Sessile Drop, Half-Angle™ Tangent line method (Tantec). Protein A quantification An estimate of the PA surface density was obtained for the different immunosurface constructs by multiplying the amount of PA determined per particle times the surface density of the particles as determined from the above mentioned 5 × 5 μm AFM scans. A well-established BCA ™ (Pierce) commercially available technique was used to determine the amount of PA per particle from stock solutions of a known concentration of particles. Functionality of immunosurfaces: antigen binding Undoubtedly, the most important characteristic of an immunoassay–sensor is its ability to bind antigen. To test this, the functionality of the immunosurface was assessed by considering both specific and non-specific antigen capture. Antigens were modeled as microspheres (MSs) in the present study. Specific antigen interactions were quantified by the number of bound active MSs (DNP-BSA) to the immuno-surfaces constructed with specific antibodies (IgG). Three types of non-specific interactions were investigated: (1) non-specific antibody (IgG (NS)) surfaces exposed to active and (2) control MSs, and (3) specific antibody (IgG) surfaces exposed to control MSs. Only active MS non-specific binding was reported due to space limitations as the other two types of non-specific binding were always less than this value. The various surface types were constructed as described above and the antigens were constructed and exposed to the immunosurfaces as described in the sections that follow. Antigen preparation The antigen used for testing the functionality of the immunosurfaces for both specific and non-specific binding was an antigenic complex of DNP conjugated to bovine serum albumin (BSA) (Molecular Probes with a valency of 22.8) adsorbed onto carboxylate-modified, 1 μm polystyrene Nile red fluorescent MSs (Excitation: 535 nm, Emission: 575 nm; FluoSpheres, Molecular Probes); this created active MSs for testing specific binding to the various immunosurfaces. For this purpose, lyophilized DNP-BSA powder was reconstituted in PBS to form a DNP-BSA solution concentration of 1.25 mg/mL, where the fluorescent MSs (2.0% solids stock) were added at a ratio of 1:20, vortexed at 3200 rpm, and allowed to sit in the dark at 4°C for 2 days to allow for the adsorption of the antigen (DNP-BSA) onto the fluorescent MSs. The MS/DNP-BSA solution was then microcentrifuged at 5585 × g at 4°C for 20 min, whereby the derivatized MS pellet was washed and resuspended by vortexing with PBS. The resuspended MS/DNP-BSA solution was allowed to sit and cure for at least 3 days at 4°C in the dark to stabilize the MS/DNP-BSA interaction. Control microspheres for testing non-specific binding were constructed from BSA adsorbed to the MSs, prepared using the procedure mentioned above with 1.25 mg/mL BSA replacing the DNP-BSA solution. Quantifying antigen capture on immunosurfaces Following the construction of the various immunosurfaces, 50 μL (2.4 ×10 6 MS) of active (DNP-BSA) or control (BSA) MSs were exposed to the various immunosurfaces in the 96 well plate for 10 min. 50 μL of each solution of active and control MSs were produced from an overall 1.15 mL solution that consisted of 30 μL of the respective stock solution in 1 mL of PBS and 120 μL of BlockAid ™ (Molecular Probes). BlockAid ™ is a proprietary protein solution which limits non-specific binding when using Molecular Probes MSs (Molecular Probes Product Information, 2001). After the MSs were allowed to interact with the immunosurfaces for 10 min, samples were gently rinsed 5 times with 200 μL of PBS in each well. One image per well (within the center) was collected using an inverted fluorescent microscope (Leica DM IRB) equipped with Texas red isothiocyanate (TRITC)-compatible filters (BP 515–560, LP 590, Leica) and a capacitive coupled display (CCD) camera at 10x magnification. Bound MSs were counted using ImagePro Plus software. Statistical analyses Analysis of variance (ANOVA) with a variance stabilization procedure was performed to determine differences between the surfaces of interest to the present study. Experimental data (reported as average ± SEM)) were analyzed using Tukey's procedure (the T method). A two-sample t-test was used in a few cases in which unequal sample sizes existed. Immunosurface preparation To determine the size of particles that should be used to model the surface roughness of B-lymphocytes, imaging software (ImageJ) was employed to evaluate the change in surface area (that is, the ratio of the outlined surface area of the cell membrane to that of a circle) of a B-lymphocyte from an image provided in the literature ( Roitt et al 1993 ). The percentage change in surface area was calculated to be 1.851431 ± 0.034405 (average ± SEM) This value was close to what could be obtained by using 860-nm diameter particles placed on flat immunosurfaces; particles of two additional sizes (specifically, 40 and 460 nm) were added in this study for comparison purposes. The model immunosurfaces were constructed in three layers through physisorption. IgG antibodies comprised the first layer, the second layer consisted of either Protein A (PA) or PA conjugated particles, and the third layer contained the second antibody layer ( Figure 1 ). The surface roughness was controlled by the size and surface density of the immobilized PA conjugated particles. All layers were constructed at room temperature (27°C). Throughout this immunosurface construction procedure, the samples were not allowed to dry to avoid non-uniformity ( Hayes 1998 ). Construction of layer 1 In the present study, mouse IgG2aαDNP (anti-2,4-dinitrophenol, kappa specific; Accurate Chemical) antibodies were used for the active surfaces and non-specific mouse IgG2a (kappa specific; Sigma) antibodies were used for the inactive surfaces. These antibodies were diluted in 1X phosphate-buffered saline (PBS) without ions (4.0 g NaCl, 0.1 g KCl, 0.75 g Na 2 HPO 4 , and 0.1 g KH 2 PO 4 added to 500 mL Milli-Q water, pH adjusted to 7.4) to a final concentration of either 11.1, 6.7, or 3.3 μg/mL. 50 μL of this antibody solution was aliquoted to each sample well of a 96-well plate (poly-styrene, flat bottom without lid, high binding, non-sterile, EIA/RIA; Corning). The antibodies were adsorbed to the polystyrene surface for 1 hour. Samples were gently rinsed 5 times with 200 μL of PBS in each well. The adsorption of the initial antibody layer results in a random orientation ( Lu et al 1996 ), leaving some correctly oriented accessible Fc regions for the secondary attachment of PA. Construction of layer 2 The second layer consisted of either PA (Sigma) or PA conjugated particles. The PA for the conventional surface was diluted in PBS to a final concentration of 4 μg/mL. 50 μL of the PA solution was adsorbed for 20 min and was gently washed 5 times with 200 μL of PBS in each well. The PA conjugated particles included 40-nm gold particles (OD = 10.1; Ted Pella, Inc), 460-nm yellow fluorescent polystyrene particles (0.1% w/v; Spherotech, Inc), and 860-nm polystyrene particles (1.0% w/v; Spherotech, Inc). PA conjugated particle concentrations were used at stock solution concentrations (8.7 × 10 11 particles/mL, 18.0 × 10 9 particles/mL, and 14.0 × 10 9 particles/mL for the 40, 460, and 860 nm particles, respectively). To ensure that the only protein in the solution was PA conjugated to the particles, each solution was microcentrifuged (Model V, VWR) at 5585 × g at 4°C for 12 min, whereby the supernatant was removed and replaced with an equal volume of PBS. After replacement of the supernatant, the preparation was vortexed at 3200 rpm (speed setting 10) (Mini Vortexer, VWR) and was sonicated for 2 min (Aquasonic 75T, 90 W, VWR) to resuspend the particles in solution. This procedure was repeated twice. 50 μL of the final particle solution was adsorbed onto the first layer of antibodies overnight. Alternatively, 50 μL of the final particle solution was further diluted from this stock solution in PBS to the desired dilution factor (1:0, 1:10, 1:100, 1:1000, and 1:10,000) which was then adsorbed overnight. The 96-well plate was covered with a lid as much as possible to avoid evaporation and contamination. After the particles adsorbed overnight, the samples were gently washed 5 times with 200 μL PBS in each well. Construction of layer 3 The second antibody layer, either specific or non-specific, was adsorbed to form the third and final layer of the model immunoassay surface. The procedure for the third layer was identical to that for the first layer. The PA on the second layer served to correctly orient the antibodies through binding of PA to the Fc region of the antibody, allowing the Fab regions to be accessible to the antigen. Construction of layer 1 In the present study, mouse IgG2aαDNP (anti-2,4-dinitrophenol, kappa specific; Accurate Chemical) antibodies were used for the active surfaces and non-specific mouse IgG2a (kappa specific; Sigma) antibodies were used for the inactive surfaces. These antibodies were diluted in 1X phosphate-buffered saline (PBS) without ions (4.0 g NaCl, 0.1 g KCl, 0.75 g Na 2 HPO 4 , and 0.1 g KH 2 PO 4 added to 500 mL Milli-Q water, pH adjusted to 7.4) to a final concentration of either 11.1, 6.7, or 3.3 μg/mL. 50 μL of this antibody solution was aliquoted to each sample well of a 96-well plate (poly-styrene, flat bottom without lid, high binding, non-sterile, EIA/RIA; Corning). The antibodies were adsorbed to the polystyrene surface for 1 hour. Samples were gently rinsed 5 times with 200 μL of PBS in each well. The adsorption of the initial antibody layer results in a random orientation ( Lu et al 1996 ), leaving some correctly oriented accessible Fc regions for the secondary attachment of PA. Construction of layer 2 The second layer consisted of either PA (Sigma) or PA conjugated particles. The PA for the conventional surface was diluted in PBS to a final concentration of 4 μg/mL. 50 μL of the PA solution was adsorbed for 20 min and was gently washed 5 times with 200 μL of PBS in each well. The PA conjugated particles included 40-nm gold particles (OD = 10.1; Ted Pella, Inc), 460-nm yellow fluorescent polystyrene particles (0.1% w/v; Spherotech, Inc), and 860-nm polystyrene particles (1.0% w/v; Spherotech, Inc). PA conjugated particle concentrations were used at stock solution concentrations (8.7 × 10 11 particles/mL, 18.0 × 10 9 particles/mL, and 14.0 × 10 9 particles/mL for the 40, 460, and 860 nm particles, respectively). To ensure that the only protein in the solution was PA conjugated to the particles, each solution was microcentrifuged (Model V, VWR) at 5585 × g at 4°C for 12 min, whereby the supernatant was removed and replaced with an equal volume of PBS. After replacement of the supernatant, the preparation was vortexed at 3200 rpm (speed setting 10) (Mini Vortexer, VWR) and was sonicated for 2 min (Aquasonic 75T, 90 W, VWR) to resuspend the particles in solution. This procedure was repeated twice. 50 μL of the final particle solution was adsorbed onto the first layer of antibodies overnight. Alternatively, 50 μL of the final particle solution was further diluted from this stock solution in PBS to the desired dilution factor (1:0, 1:10, 1:100, 1:1000, and 1:10,000) which was then adsorbed overnight. The 96-well plate was covered with a lid as much as possible to avoid evaporation and contamination. After the particles adsorbed overnight, the samples were gently washed 5 times with 200 μL PBS in each well. Construction of layer 3 The second antibody layer, either specific or non-specific, was adsorbed to form the third and final layer of the model immunoassay surface. The procedure for the third layer was identical to that for the first layer. The PA on the second layer served to correctly orient the antibodies through binding of PA to the Fc region of the antibody, allowing the Fab regions to be accessible to the antigen. Immunosurface characterization AFM measurements of surface roughness Atomic force microscopy (AFM) was used to determine the topology and roughness of the various surfaces created in the present study. Specifically, samples were prepared for AFM by constructing only the first two layers of the immunoassay membrane (as described above). Samples were mounted using double-stick tape (3M) onto 12-mm mica specimen discs (Ted Pella, Inc.) and were imaged using a Veeco Nanoscope IIIa MultiMode scanning probe microscopy (SPM) in Tapping™ mode. Standard AFM tips and measuring conditions (thickness: 4 μm, width: 30 μm, length: 125 μm, force constant: 42 N/m, resonance frequency: 320 kHz) were used in non-contact/Tapping ™ mode with a reflexive coating (Pacific Nanotechnology). Three AFM 5 × 5 μm scans at 512 × 512 lines per image were taken for each sample, for 3 different samples of each surface type, at a rate of 1 Hz and a velocity of 10 μm/s. The average room temperature during imaging was 24°C with an average humidity of 30%–40%. Three types of roughness measurements were calculated using Nanoscope IIIa 4.43r8 software in this study: root-mean-square roughness (RMS), average surface roughness (Ra), and the change in surface area (ΔSA) or Wenzel ratio. Three-dimensional topographic images were constructed using WSxM 4.0 Develop 4.4 software (Nanotec Electronica S.L.). Hydrophilicity measurements To determine hydrophilicity of the model immunoassay membranes, static contact angles were measured (Cam-Plus Contact Angle Reader, Tantec) in triplicate. In this manner, contact angles on each of the surfaces of interest to the present study (including the conventional surfaces) were measured immediately after adding 1 μL of Milli-q water using the Sessile Drop, Half-Angle™ Tangent line method (Tantec). Protein A quantification An estimate of the PA surface density was obtained for the different immunosurface constructs by multiplying the amount of PA determined per particle times the surface density of the particles as determined from the above mentioned 5 × 5 μm AFM scans. A well-established BCA ™ (Pierce) commercially available technique was used to determine the amount of PA per particle from stock solutions of a known concentration of particles. Functionality of immunosurfaces: antigen binding Undoubtedly, the most important characteristic of an immunoassay–sensor is its ability to bind antigen. To test this, the functionality of the immunosurface was assessed by considering both specific and non-specific antigen capture. Antigens were modeled as microspheres (MSs) in the present study. Specific antigen interactions were quantified by the number of bound active MSs (DNP-BSA) to the immuno-surfaces constructed with specific antibodies (IgG). Three types of non-specific interactions were investigated: (1) non-specific antibody (IgG (NS)) surfaces exposed to active and (2) control MSs, and (3) specific antibody (IgG) surfaces exposed to control MSs. Only active MS non-specific binding was reported due to space limitations as the other two types of non-specific binding were always less than this value. The various surface types were constructed as described above and the antigens were constructed and exposed to the immunosurfaces as described in the sections that follow. Antigen preparation The antigen used for testing the functionality of the immunosurfaces for both specific and non-specific binding was an antigenic complex of DNP conjugated to bovine serum albumin (BSA) (Molecular Probes with a valency of 22.8) adsorbed onto carboxylate-modified, 1 μm polystyrene Nile red fluorescent MSs (Excitation: 535 nm, Emission: 575 nm; FluoSpheres, Molecular Probes); this created active MSs for testing specific binding to the various immunosurfaces. For this purpose, lyophilized DNP-BSA powder was reconstituted in PBS to form a DNP-BSA solution concentration of 1.25 mg/mL, where the fluorescent MSs (2.0% solids stock) were added at a ratio of 1:20, vortexed at 3200 rpm, and allowed to sit in the dark at 4°C for 2 days to allow for the adsorption of the antigen (DNP-BSA) onto the fluorescent MSs. The MS/DNP-BSA solution was then microcentrifuged at 5585 × g at 4°C for 20 min, whereby the derivatized MS pellet was washed and resuspended by vortexing with PBS. The resuspended MS/DNP-BSA solution was allowed to sit and cure for at least 3 days at 4°C in the dark to stabilize the MS/DNP-BSA interaction. Control microspheres for testing non-specific binding were constructed from BSA adsorbed to the MSs, prepared using the procedure mentioned above with 1.25 mg/mL BSA replacing the DNP-BSA solution. Quantifying antigen capture on immunosurfaces Following the construction of the various immunosurfaces, 50 μL (2.4 ×10 6 MS) of active (DNP-BSA) or control (BSA) MSs were exposed to the various immunosurfaces in the 96 well plate for 10 min. 50 μL of each solution of active and control MSs were produced from an overall 1.15 mL solution that consisted of 30 μL of the respective stock solution in 1 mL of PBS and 120 μL of BlockAid ™ (Molecular Probes). BlockAid ™ is a proprietary protein solution which limits non-specific binding when using Molecular Probes MSs (Molecular Probes Product Information, 2001). After the MSs were allowed to interact with the immunosurfaces for 10 min, samples were gently rinsed 5 times with 200 μL of PBS in each well. One image per well (within the center) was collected using an inverted fluorescent microscope (Leica DM IRB) equipped with Texas red isothiocyanate (TRITC)-compatible filters (BP 515–560, LP 590, Leica) and a capacitive coupled display (CCD) camera at 10x magnification. Bound MSs were counted using ImagePro Plus software. AFM measurements of surface roughness Atomic force microscopy (AFM) was used to determine the topology and roughness of the various surfaces created in the present study. Specifically, samples were prepared for AFM by constructing only the first two layers of the immunoassay membrane (as described above). Samples were mounted using double-stick tape (3M) onto 12-mm mica specimen discs (Ted Pella, Inc.) and were imaged using a Veeco Nanoscope IIIa MultiMode scanning probe microscopy (SPM) in Tapping™ mode. Standard AFM tips and measuring conditions (thickness: 4 μm, width: 30 μm, length: 125 μm, force constant: 42 N/m, resonance frequency: 320 kHz) were used in non-contact/Tapping ™ mode with a reflexive coating (Pacific Nanotechnology). Three AFM 5 × 5 μm scans at 512 × 512 lines per image were taken for each sample, for 3 different samples of each surface type, at a rate of 1 Hz and a velocity of 10 μm/s. The average room temperature during imaging was 24°C with an average humidity of 30%–40%. Three types of roughness measurements were calculated using Nanoscope IIIa 4.43r8 software in this study: root-mean-square roughness (RMS), average surface roughness (Ra), and the change in surface area (ΔSA) or Wenzel ratio. Three-dimensional topographic images were constructed using WSxM 4.0 Develop 4.4 software (Nanotec Electronica S.L.). Hydrophilicity measurements To determine hydrophilicity of the model immunoassay membranes, static contact angles were measured (Cam-Plus Contact Angle Reader, Tantec) in triplicate. In this manner, contact angles on each of the surfaces of interest to the present study (including the conventional surfaces) were measured immediately after adding 1 μL of Milli-q water using the Sessile Drop, Half-Angle™ Tangent line method (Tantec). Protein A quantification An estimate of the PA surface density was obtained for the different immunosurface constructs by multiplying the amount of PA determined per particle times the surface density of the particles as determined from the above mentioned 5 × 5 μm AFM scans. A well-established BCA ™ (Pierce) commercially available technique was used to determine the amount of PA per particle from stock solutions of a known concentration of particles. Functionality of immunosurfaces: antigen binding Undoubtedly, the most important characteristic of an immunoassay–sensor is its ability to bind antigen. To test this, the functionality of the immunosurface was assessed by considering both specific and non-specific antigen capture. Antigens were modeled as microspheres (MSs) in the present study. Specific antigen interactions were quantified by the number of bound active MSs (DNP-BSA) to the immuno-surfaces constructed with specific antibodies (IgG). Three types of non-specific interactions were investigated: (1) non-specific antibody (IgG (NS)) surfaces exposed to active and (2) control MSs, and (3) specific antibody (IgG) surfaces exposed to control MSs. Only active MS non-specific binding was reported due to space limitations as the other two types of non-specific binding were always less than this value. The various surface types were constructed as described above and the antigens were constructed and exposed to the immunosurfaces as described in the sections that follow. Antigen preparation The antigen used for testing the functionality of the immunosurfaces for both specific and non-specific binding was an antigenic complex of DNP conjugated to bovine serum albumin (BSA) (Molecular Probes with a valency of 22.8) adsorbed onto carboxylate-modified, 1 μm polystyrene Nile red fluorescent MSs (Excitation: 535 nm, Emission: 575 nm; FluoSpheres, Molecular Probes); this created active MSs for testing specific binding to the various immunosurfaces. For this purpose, lyophilized DNP-BSA powder was reconstituted in PBS to form a DNP-BSA solution concentration of 1.25 mg/mL, where the fluorescent MSs (2.0% solids stock) were added at a ratio of 1:20, vortexed at 3200 rpm, and allowed to sit in the dark at 4°C for 2 days to allow for the adsorption of the antigen (DNP-BSA) onto the fluorescent MSs. The MS/DNP-BSA solution was then microcentrifuged at 5585 × g at 4°C for 20 min, whereby the derivatized MS pellet was washed and resuspended by vortexing with PBS. The resuspended MS/DNP-BSA solution was allowed to sit and cure for at least 3 days at 4°C in the dark to stabilize the MS/DNP-BSA interaction. Control microspheres for testing non-specific binding were constructed from BSA adsorbed to the MSs, prepared using the procedure mentioned above with 1.25 mg/mL BSA replacing the DNP-BSA solution. Quantifying antigen capture on immunosurfaces Following the construction of the various immunosurfaces, 50 μL (2.4 ×10 6 MS) of active (DNP-BSA) or control (BSA) MSs were exposed to the various immunosurfaces in the 96 well plate for 10 min. 50 μL of each solution of active and control MSs were produced from an overall 1.15 mL solution that consisted of 30 μL of the respective stock solution in 1 mL of PBS and 120 μL of BlockAid ™ (Molecular Probes). BlockAid ™ is a proprietary protein solution which limits non-specific binding when using Molecular Probes MSs (Molecular Probes Product Information, 2001). After the MSs were allowed to interact with the immunosurfaces for 10 min, samples were gently rinsed 5 times with 200 μL of PBS in each well. One image per well (within the center) was collected using an inverted fluorescent microscope (Leica DM IRB) equipped with Texas red isothiocyanate (TRITC)-compatible filters (BP 515–560, LP 590, Leica) and a capacitive coupled display (CCD) camera at 10x magnification. Bound MSs were counted using ImagePro Plus software. Statistical analyses Analysis of variance (ANOVA) with a variance stabilization procedure was performed to determine differences between the surfaces of interest to the present study. Experimental data (reported as average ± SEM)) were analyzed using Tukey's procedure (the T method). A two-sample t-test was used in a few cases in which unequal sample sizes existed. Results and discussion Immunosurface characterization: roughness As expected, large differences in surface roughness were observed through the use of the various particle sizes in the present study ( Figure 2 and Table 1 ). Most importantly, results demonstrated increased surface roughness through the addition of particles to the polystyrene wells. Specifically, the highest degree of roughness was observed with the use of 860-nm followed by 460-nm followed by 40-nm diameter particles; all had greater surface roughness than polystyrene alone. As a reminder, the type of surface roughness that best modeled that of a B-lymphocyte from an image provided in the literature ( Roitt et al 1993 ) was through the use of 860-nm diameter particles. Immunosurface characterization: antigen binding Results of this study provided evidence that specific antigen capture increased as the surface roughness increased, while non-specific antigen capture was independent of surface roughness ( Figure 3 ). The change in surface roughness data within each surface type reflects the serial dilutions of the particles (1:0, 1:10, 1:100, 1:1000, and 1:10000). These data showed that even though surface roughness may be a factor, there are likely other factors involved in mediating antigen capture, such as surface chemistry changes or alterations in PA surface density. This is likely since at the 100-nm surface roughness regime there is a difference in specific and non-specific antigen capture between surface types. This can be further confirmed in Figure 4 where at the 100-nm surface roughness regime the surface types have different energetics, as determined through contact angle measurements. This may be due to a change in surface chemistry from the different types of polystyrene particles used and/or the different amounts of PA conjugated on those particles. Moreover, in Figure 4 it can be seen that within each immobilized particle type, the surface energetics change as the surface roughness changes. It is important to note that Figure 4 complements Figure 3 since collectively they show that when either surface roughness or surface energetics (that is, contact angles) were held constant, there was still a change in specific and non-specific antigen capture. That is, specific and non-specific antigen capture is most likely due to a combination of a change in surface chemistry and surface roughness, with surface roughness playing a larger role in specific antigen capture than non-specific antigen capture. To investigate further what properties may have changed when using the more rough surfaces (that is, those with 860-nm particles), surface chemistry differences possibly due to various amounts of PA were evaluated. Results provided evidence that the density of PA was not correlated with surface roughness and that specific and non-specific antigen capture did not depend solely on the density of PA on the surface ( Figure 5 ). Specifically, the 460- and 860-nm particle surfaces had similar amounts of PA yet the 860-nm particle surface enhanced antigen capture. As seen in Figures 3 and 4 , this again suggests that the surface roughness created in this study through the use of various nanometer particle sizes did not cause different surface energetics that manipulated protein density. When analyzing different physisorbed antibody concentrations, the disproportionate increases in specific antigen capture on the 860-nm particle surface may be explained due to a greater amount of antibody immobilized and/or a greater functionality of those immobilized antibodies ( Figure 6 ). Comparison of Figures 3 , 4 , and 5 highlights the possibility that it may not just be the amount of IgG adsorbed but the functionality of those adsorbed antibodies, since the most specific binding was not found on the surface with the most PA. Reasons for altered antigen binding When surface roughness and surface chemistry are both changed, as was done here, it is difficult to determine to what extent each variable is contributing to specific and nonspecific antigen capture. However, the results of the present study demonstrated that surface roughness is an important parameter for increasing antigen capture. Moreover, the largest specific antigen capture was seen when 860-nm particles were used on the immunoassay surfaces, which corresponded to the aforementioned estimate of the surface roughness of a B-lymphocyte membrane. In this manner, one explanation of why antibody functionality may be enhanced when immobilized on increasingly roughened immunosurfaces may lie with the probabilistic geometrical nature of a roughened surface. That is, various non-covalent forces govern every antigen–antibody binding event. These different forces vary in strength depending on the distance between an antigen and an antibody. It can then be rationalized that the surface area of an immunoassay–sensor is mirrored in solution due to these varying forces. Thus, a more rough immunoassay surface may result in a more rough mirrored antigen contact area created by the non-covalent forces within the solution. Probabilistically, antigens in solution will be more likely to interact with these non-covalent forces to result in specific antigen capture. In other words, the functionality of the individual antibodies may not be intrinsically enhanced, but the collective whole of the antibodies on the immunosurface may be enhanced. Immunosurface characterization: roughness As expected, large differences in surface roughness were observed through the use of the various particle sizes in the present study ( Figure 2 and Table 1 ). Most importantly, results demonstrated increased surface roughness through the addition of particles to the polystyrene wells. Specifically, the highest degree of roughness was observed with the use of 860-nm followed by 460-nm followed by 40-nm diameter particles; all had greater surface roughness than polystyrene alone. As a reminder, the type of surface roughness that best modeled that of a B-lymphocyte from an image provided in the literature ( Roitt et al 1993 ) was through the use of 860-nm diameter particles. Immunosurface characterization: antigen binding Results of this study provided evidence that specific antigen capture increased as the surface roughness increased, while non-specific antigen capture was independent of surface roughness ( Figure 3 ). The change in surface roughness data within each surface type reflects the serial dilutions of the particles (1:0, 1:10, 1:100, 1:1000, and 1:10000). These data showed that even though surface roughness may be a factor, there are likely other factors involved in mediating antigen capture, such as surface chemistry changes or alterations in PA surface density. This is likely since at the 100-nm surface roughness regime there is a difference in specific and non-specific antigen capture between surface types. This can be further confirmed in Figure 4 where at the 100-nm surface roughness regime the surface types have different energetics, as determined through contact angle measurements. This may be due to a change in surface chemistry from the different types of polystyrene particles used and/or the different amounts of PA conjugated on those particles. Moreover, in Figure 4 it can be seen that within each immobilized particle type, the surface energetics change as the surface roughness changes. It is important to note that Figure 4 complements Figure 3 since collectively they show that when either surface roughness or surface energetics (that is, contact angles) were held constant, there was still a change in specific and non-specific antigen capture. That is, specific and non-specific antigen capture is most likely due to a combination of a change in surface chemistry and surface roughness, with surface roughness playing a larger role in specific antigen capture than non-specific antigen capture. To investigate further what properties may have changed when using the more rough surfaces (that is, those with 860-nm particles), surface chemistry differences possibly due to various amounts of PA were evaluated. Results provided evidence that the density of PA was not correlated with surface roughness and that specific and non-specific antigen capture did not depend solely on the density of PA on the surface ( Figure 5 ). Specifically, the 460- and 860-nm particle surfaces had similar amounts of PA yet the 860-nm particle surface enhanced antigen capture. As seen in Figures 3 and 4 , this again suggests that the surface roughness created in this study through the use of various nanometer particle sizes did not cause different surface energetics that manipulated protein density. When analyzing different physisorbed antibody concentrations, the disproportionate increases in specific antigen capture on the 860-nm particle surface may be explained due to a greater amount of antibody immobilized and/or a greater functionality of those immobilized antibodies ( Figure 6 ). Comparison of Figures 3 , 4 , and 5 highlights the possibility that it may not just be the amount of IgG adsorbed but the functionality of those adsorbed antibodies, since the most specific binding was not found on the surface with the most PA. Reasons for altered antigen binding When surface roughness and surface chemistry are both changed, as was done here, it is difficult to determine to what extent each variable is contributing to specific and nonspecific antigen capture. However, the results of the present study demonstrated that surface roughness is an important parameter for increasing antigen capture. Moreover, the largest specific antigen capture was seen when 860-nm particles were used on the immunoassay surfaces, which corresponded to the aforementioned estimate of the surface roughness of a B-lymphocyte membrane. In this manner, one explanation of why antibody functionality may be enhanced when immobilized on increasingly roughened immunosurfaces may lie with the probabilistic geometrical nature of a roughened surface. That is, various non-covalent forces govern every antigen–antibody binding event. These different forces vary in strength depending on the distance between an antigen and an antibody. It can then be rationalized that the surface area of an immunoassay–sensor is mirrored in solution due to these varying forces. Thus, a more rough immunoassay surface may result in a more rough mirrored antigen contact area created by the non-covalent forces within the solution. Probabilistically, antigens in solution will be more likely to interact with these non-covalent forces to result in specific antigen capture. In other words, the functionality of the individual antibodies may not be intrinsically enhanced, but the collective whole of the antibodies on the immunosurface may be enhanced. Conclusions In conclusion, this study provided evidence that nanometer surface roughness may be a critical design parameter for future immunoassays–sensors since the sensitivity of an antibody–antigen capture may be enhanced without an increase in non-specific binding. An increase in nanometer surface roughness was observed in this study through the use of larger immobilized particles. It then followed that a decrease in surface roughness was created through serial dilutions of those immobilized particles. Greater specific antigen capture correlated with increased surface roughness and physisorbed antibody concentrations, while nonspecific antigen capture was independent of surface roughness. Surface energetic experiments involving contact angles suggested that there might be a change in surface chemistry between the different immunoassay surface types. However, even amidst a possible surface chemistry change, results of this study implied that nanometer surface roughness was the dominating factor that contributed to greater specific antigen capture. Such knowledge of the use of nanostructured surface roughness should be used to design better immunoassays/biosensors.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5974032/

CH2 Domain of Mouse IgG3 Governs Antibody Oligomerization, Increases Functional Affinity to Multivalent Antigens and Enhances Hemagglutination

Mouse IgG3 is highly protective against several life-threatening bacteria. This isotype is the only one among mouse IgGs that forms non-covalent oligomers, has increased functional affinity to polyvalent antigens, and efficiently agglutinates erythrocytes. IgG3 also triggers the complement cascade. The high efficacy of protection after passive immunization with IgG3 is correlated with the unique properties of this isotype. Although the features of IgG3 are well documented, their molecular basis remains elusive. Based on functional analyses of IgG1/IgG3 hybrid molecules with swapped constant domains, we identified IgG3-derived CH2 domain as a major determinant of antibody oligomerization and increased functional affinity to a multivalent antigen. The CH2 domain was also crucial for efficient hemagglutination triggered by IgG3 and was indispensable for complement cascade activation. This domain is glycosylated and atypically charged. A mutational analysis based on molecular models of CH2 domain charge distribution indicated that the functional affinity was influenced by the specific charge location. N-glycans were essential for CH2-dependent enhancement of hemagglutination and complement activation. Oligomerization was independent of CH2 charge and glycosylation. We also verified that known C1q-binding motifs are functional in mouse IgG3 but not in IgG1 framework. We generated for the first time a gain-of-function antibody with properties transferred from IgG3 into IgG1 by replacing the CH2 domain. Finding that the CH2 domain of IgG3 governs unique properties of this isotype is likely to open an avenue toward the generation of IgG3-inspired antibodies that will be protective against existing or emerging lethal pathogens. Introduction There are four subclasses of mouse IgGs: IgG1, IgG2a, IgG2b, and IgG3. Although structurally very similar, they significantly differ in their functions ( 1 ). Mouse IgG3s are particularly interesting, because they are able to form oligomers, which strongly influences their biological activities ( 2 ). Mouse IgG3 was described for the first time almost 50 years ago ( 3 ) and different aspects of its biology have been investigated by several groups. The propensity of IgG3 oligomerization was noticed already by its discoverers ( 3 ). Then, other researchers reported cooperative binding of IgG3 to a multivalent antigen ( 4 , 5 ). Although the initial report on IgG3 oligomerization concerned molecules in solution ( 3 ), later studies revealed that binding to multivalent antigens promoted IgG3 intermolecular interactions, which in turn resulted in its increased functional affinity to the antigen ( 4 ). This phenomenon depended on Fc, because IgG3 F(ab') 2 fragments did not bind to the antigen cooperatively ( 4 ). However, the exact molecular mechanism of IgG3 oligomerization remains unknown. IgG3 is a major component of cryoglobulins in mice ( 2 ). Cryoglobulins are plasma proteins that reversibly precipitate at low temperatures or at high concentrations ( 6 ). Cryogenic activity of IgG3 was shown to correlate with its ability to oligomerize, with the presence of charged residues in the variable region and the level of sialylation ( 7 ). Also, IgG3 was reported as exceptionally effective in preventing or fighting several life-threatening microbial infections, e.g., with Neisseria meningitidis ( 8 ) or Bacillus anthracis ( 9 ). A comparison between the four mouse IgG subclasses with the same variable region specific to B. anthracis capsule proved that only IgG3 is protective against pulmonary anthrax in a mouse model ( 9 ). Importantly, mouse–human chimeric antibodies containing a constant region of any human IgG subclass were not effective, although they had the same variable region as the protective mouse IgG3 ( 10 ). These reports indicated that mouse IgG3 constant region has unique properties, but the authors only speculated about possible molecular mechanisms behind the observed phenomenon. The exceptional characteristic of IgG3 was also confirmed by our previous report that among mouse IgGs recognizing a surface antigen of erythrocytes only IgG3s are able to agglutinate the cells ( 11 ). We rejected the hypothesis that IgG3-mediated hemagglutination results from oligomerization of the antibodies, because IgG3 F(ab') 2 was sufficient to trigger hemagglutination. Molecular modeling indicated that IgG3 has a larger span of Fab arms than other IgG subclasses. IgG3 has a long-upper hinge that may extend the Fab range, but whether this may account for the ability to hemagglutinate was not verified experimentally. Here, we present the results of our attempts to find molecular determinants of the properties of mouse IgG3. Our experimental model comprised two antibodies (M18 and O10) specific to antigen B of the ABO blood group system ( 11 ). Antigen B is a pentasaccharide O-glycan attached to numerous proteins and lipids on the erythrocyte surface. The large quantity and high density of the antigen, as well as a strong negative charge of the erythrocyte surface, could be considered as a safe and easy-to-handle model of a pathogen surface. We generated many muteins of IgG3 molecules and analyzed their functional affinity to the antigen as well as their ability to hemagglutinate and oligomerize. The results showed that IgG3-derived CH2 domain determines antibody oligomerization and increases its affinity to the antigen. This domain also strongly enhances agglutination of erythrocytes bearing B antigen. Moreover, we investigated complement activation by the muteins, and we confirmed that known C1q-binding motifs are functional in mouse IgG3 but not in the mouse IgG1 framework. Materials and Methods Generation of Vectors Coding for Antibody Muteins Expression vectors coding for heavy chain muteins with swapped domains or mutated CH2 were generated using synthetic nucleic acids (Gene Art, Germany) cloned into pFUSEss-CHIg-mG1_M18 (Addgene #82357) or pFUSEss-CHIg-mG3_M18 (Addgene #82356) plasmids ( 11 ). Only endogenous restriction sites present in the ORFs were used for cloning. Vectors coding for other muteins were prepared with Q5-based site directed mutagenesis kit (NEB). All plasmids coding for heavy chain variants of M18 antibody are available via Addgene repository along with their full sequences and maps, accession numbers: 105849–105863. Plasmids encoding O10 antibody variants were obtained by replacing the sequence coding for M18 variable fragment with a corresponding O10-derived cDNA using EcoR I and Afe I restriction sites in the vectors coding for M18 heavy chain variants. The sequence of O10 antibody is proprietary and cannot be disclosed. All plasmids were verified using Sanger sequencing (Genomed, Poland) or NGS (Addgene). Production of Antibodies The recombinant antibodies were transiently expressed in HEK293T cells cultured in DMEM with 4.5 g/l glucose (Lonza) supplemented with 10% FBS (Biowest). The cells were co-transfected with plasmids coding for a heavy chain and a cognate light chain using Lipofectamine2000 (Thermo) or PEI MAX (Polysciences, MW 40,000). In the case of M18 variants, the plasmid pFUSE2ss-CLIg-mK_M18 (Addgene, #82358) ( 11 ) coding for M18 light chain was used. A similar plasmid coding for O10 light chain was used for expression of O10 variants. Hybridoma-derived M18 IgG3 was produced as described previously ( 11 ). The antibodies were purified using CaptureSelect LC-kappa (mur) affinity matrix (Thermo) according to instructions of the supplier. Glycine–HCl (100 mM, pH 2.0) was used for elution. Production of IgG3 F(ab') 2 A sequence encoding IgG2b core hinge, HA tag and a stop codon was cloned into pFUSEss-CHIg-mG3_M18 downstream of the sequence coding for the upper hinge of M18. Sequences of the upper and core hinges of the antibodies are described by Dangl et al ( 12 ). Mouse IgG2b core hinge contains four cysteine residues that allow an efficient association of Fab' fragments into F(ab') 2 ( 13 ). Recombinant F(ab') 2 was expressed as described for other antibodies. Alternatively, F(ab') 2 was produced by enzymatic digestion with IdeZ (NEB) according to the manufacturer's protocol. Measurements of Antibody Concentration Antibody concentrations in cell culture media were measured using a standard sandwich ELISA on plates coated with sheep anti-mouse Fab polyclonal antibody (Jackson Laboratory, cat. #515-005-072, lot #105461). Goat anti-mouse kappa polyclonal antibody (1:3,000, BioRad, cat. #105008, batch #160617), HRP-labeled streptavidin (1:40,000, Sigma), and the TMB substrate for HRP (BD Bioscience) were used for detection. The HRP-dependent colorogenic reaction was stopped with 1 M HCl, and the absorbance at 450 nm was read using the microplate spectrophotometer Synergy H1 operated by Gen5 2.00 Software (BioTek). Purified M18 (IgG3) and MCP21 (IgG1, Sigma) were used as standards. Concentrations of muteins were calculated based on their Fab type (IgG1- or IgG3-type Fab). BCA assay (Sigma) was used for measurements of purified antibody concentrations with bovine γ-globulin (Thermo) as a reference. Antibody Binding to Immobilized Erythrocytes Polystyrene plates were coated with 50 µg/ml poly- l -Lys (Sigma) for 1 h at room temperature (RT). Then, 100 µl of 0.1% (hematocrit) suspension of red blood cells in PBS were added to the wells and the cells were allowed to settle for 1 h at RT. After gentle aspiration of the solution, the cells were fixed using 0.025% glutaraldehyde for 40 min. Endogenous peroxidase activity was blocked with 3% H 2 O 2 for 1 h. The plates were blocked overnight with 0.2% gelatin in PBS containing 0.05% Tween-20 at 4°C. Then, the cells were incubated with serial dilutions of cell culture media containing analyzed antibodies followed by detection with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad) and HRP-labeled streptavidin (1:40,000, Sigma). All reagents were diluted in the blocking buffer. The colorogenic reaction was performed, and the absorbance at 450 nm was measured as described above. C1q Binding Duplicates of polystyrene plates were coated with 6 µg/ml of BSA conjugated with the discriminating trisaccharide of the B antigen (Dextra Laboratories, cat #NGP6323, batch #ATDX232-039) overnight at 4°C, blocked with 1% BSA in PBS (1 h) and incubated with serial dilutions of cell culture media containing O10 muteins (2 h). Then, one set of plates was used for evaluation of C1q binding by analyzed antibodies; and the plates were incubated with 2 µg/ml of C1q purified from human serum (Biorad, cat. #22215504, batch #130815; 2 h) and next with HRP-labeled anti-human C1q polyclonal antibody (1:400, Abcam, cat. #ab46191, lot #GR205436-5; 1 h). The second set of plates was used to analyze quantities of the muteins bound to the immobilized antigen; and the plates were incubated with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad, 2 h) and HRP-labeled streptavidin (1:40,000, Sigma; 1 h). Absorbance of HRP product was measured as described above. Each incubation step was preceded by extensive washing with 0.05% Tween-20 in PBS. The antibodies, streptavidin, and C1q were diluted in 0.1% BSA in PBS. All incubations, except plate coating, were at RT. The normalized C1q binding ( a ) was calculated by dividing the signal corresponding to C1q binding ( b ) by the signal corresponding to bound antibody ( c ). normalized C 1 q binding ( a ) = C 1 q binding signal ( b ) bound antibody signal ( c ) Uncertainty of a (Δ a ) was calculated by exact differential. Uncertainties of b and c (Δ b and Δ c ) equaled to standard deviations of the absorbance measurements. uncertainty of a ​ ( Δ a ) = | ∂ a ∂ b Δ b | + | ∂ a ∂ c Δ c | = | Δ b c | + | b c 2 Δ c | Complement Cascade Activation Washed human red blood cells suspended to a hematocrit of 2% were coated with 3 µg/ml or 1.5 µg/ml of the analyzed antibodies for 1.5 h at RT, washed twice with PBS and resuspended in PBS with Ca 2+ and Mg 2+ . Then, the same volume of human complement serum (Sigma, cat. #S1764, lot #SLBS5471V, #SLBQ0752V or #SLBP0461V) diluted to 7 CH 50 U/ml in PBS with Ca 2+ and Mg 2+ was added to the coated red blood cells. The samples were centrifuged after 2 h of incubation at 37°C, and the absorbance of released hemoglobin was measured in the supernatants at 540 nm. Red Blood Cells and Agglutination Standard human red blood cells were purchased from Regional Centre of Blood Donation and Blood Treatment in Katowice, Poland. Agglutination tests were performed in 96-flat bottom plates. Serially diluted solutions of analyzed antibodies (100 µl) were gently mixed with 0.45% (hematocrit) suspension of red blood cells. The level of agglutination was analyzed using a phase-contrast microscope after 20 min of moderate shaking. A six-point scale was used for evaluation of agglutination intensity: from 4+ (complete cell aggregation), to 3+, 2+, 1+, ± to a negative score. The agglutination score reflects both the size of aggregates and quantity of non-agglutinated cells. IgG3 Self-Association Assay Oligomerization of IgG3 was analyzed similarly to the method described by Abdelmoula et al ( 2 ). Three different concentrations of the purified domain muteins with M18 variable region (150, 100, and 20 µg/ml) were incubated with non-mutated, biotinylated IgG3 M18 (100 ng/ml) for 72–96 h at 4°C in the presence of 4% BSA (Sigma, cat. #A9576). Then, the antibody complexes were precipitated by adding 50% PEG-6000 (Sigma) to the final concentration of 7.5%. After 1 h of incubation on ice, the samples were centrifuged (30 min, 3,000 × g , 4°C). The supernatants were preserved for further analysis, and the precipitates were washed with ice-cold 7.5% PEG-6000 in PBS and centrifuged again (30 min, 3,000 × g , 4°C). Then, the precipitates were dissolved in PBS with 0.1% BSA by pipetting at 37°C. Biotinylated IgG3 in the precipitates and supernatants was quantified using ELISA on polystyrene plates coated with streptavidin (8 µg/ml, Thermo, cat. #434301, lot #RB233354). Bound biotinylated IgG3 was detected using rabbit monoclonal antibody M111-2 (1:1,000, Abcam, cat. #ab125904, lot #C050311, #GR157092-1) and HRP-labeled goat anti-rabbit polyclonal antibody (1:3,000, Sigma, cat #A6667). The absorbance of HRP product was measured as described above. SDS-PAGE and Western Blotting Samples were resolved in 8 or 12% polyacrylamide gels under non-reducing or reducing conditions according to the protocol of Laemmli ( 14 ). After wet electrotransfer onto PVDF membrane and blocking with 4% skim milk in PBS, the samples were probed with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad) and HRP-labeled streptavidin (1:40,000, Sigma). Alternatively, rabbit anti-HA tag polyclonal antibody (1:10,000, Abcam, cat. #ab9110) and HRP-labeled goat anti-rabbit F(ab') 2 polyclonal antibody (1:10,000, Sigma, cat. #A6667, lot #SLBG3029) were used. Mouse IgG3 heavy chain was detected using goat antiserum to mouse IgG3 (1:500, Sigma, cat. #ISO2) and rabbit anti-goat polyclonal antibody (1:5,000, Sigma, cat. #A4174). Bands were visualized using Immobilon Western Chemiluminescent Substrate for HRP (Millipore). The images were captured and analyzed using Fusion Fx apparatus with the Fusion Capt Advance Fx5 program (Vilbert Lourmat, France). Generation of Vectors Coding for Antibody Muteins Expression vectors coding for heavy chain muteins with swapped domains or mutated CH2 were generated using synthetic nucleic acids (Gene Art, Germany) cloned into pFUSEss-CHIg-mG1_M18 (Addgene #82357) or pFUSEss-CHIg-mG3_M18 (Addgene #82356) plasmids ( 11 ). Only endogenous restriction sites present in the ORFs were used for cloning. Vectors coding for other muteins were prepared with Q5-based site directed mutagenesis kit (NEB). All plasmids coding for heavy chain variants of M18 antibody are available via Addgene repository along with their full sequences and maps, accession numbers: 105849–105863. Plasmids encoding O10 antibody variants were obtained by replacing the sequence coding for M18 variable fragment with a corresponding O10-derived cDNA using EcoR I and Afe I restriction sites in the vectors coding for M18 heavy chain variants. The sequence of O10 antibody is proprietary and cannot be disclosed. All plasmids were verified using Sanger sequencing (Genomed, Poland) or NGS (Addgene). Production of Antibodies The recombinant antibodies were transiently expressed in HEK293T cells cultured in DMEM with 4.5 g/l glucose (Lonza) supplemented with 10% FBS (Biowest). The cells were co-transfected with plasmids coding for a heavy chain and a cognate light chain using Lipofectamine2000 (Thermo) or PEI MAX (Polysciences, MW 40,000). In the case of M18 variants, the plasmid pFUSE2ss-CLIg-mK_M18 (Addgene, #82358) ( 11 ) coding for M18 light chain was used. A similar plasmid coding for O10 light chain was used for expression of O10 variants. Hybridoma-derived M18 IgG3 was produced as described previously ( 11 ). The antibodies were purified using CaptureSelect LC-kappa (mur) affinity matrix (Thermo) according to instructions of the supplier. Glycine–HCl (100 mM, pH 2.0) was used for elution. Production of IgG3 F(ab') 2 A sequence encoding IgG2b core hinge, HA tag and a stop codon was cloned into pFUSEss-CHIg-mG3_M18 downstream of the sequence coding for the upper hinge of M18. Sequences of the upper and core hinges of the antibodies are described by Dangl et al ( 12 ). Mouse IgG2b core hinge contains four cysteine residues that allow an efficient association of Fab' fragments into F(ab') 2 ( 13 ). Recombinant F(ab') 2 was expressed as described for other antibodies. Alternatively, F(ab') 2 was produced by enzymatic digestion with IdeZ (NEB) according to the manufacturer's protocol. Measurements of Antibody Concentration Antibody concentrations in cell culture media were measured using a standard sandwich ELISA on plates coated with sheep anti-mouse Fab polyclonal antibody (Jackson Laboratory, cat. #515-005-072, lot #105461). Goat anti-mouse kappa polyclonal antibody (1:3,000, BioRad, cat. #105008, batch #160617), HRP-labeled streptavidin (1:40,000, Sigma), and the TMB substrate for HRP (BD Bioscience) were used for detection. The HRP-dependent colorogenic reaction was stopped with 1 M HCl, and the absorbance at 450 nm was read using the microplate spectrophotometer Synergy H1 operated by Gen5 2.00 Software (BioTek). Purified M18 (IgG3) and MCP21 (IgG1, Sigma) were used as standards. Concentrations of muteins were calculated based on their Fab type (IgG1- or IgG3-type Fab). BCA assay (Sigma) was used for measurements of purified antibody concentrations with bovine γ-globulin (Thermo) as a reference. Antibody Binding to Immobilized Erythrocytes Polystyrene plates were coated with 50 µg/ml poly- l -Lys (Sigma) for 1 h at room temperature (RT). Then, 100 µl of 0.1% (hematocrit) suspension of red blood cells in PBS were added to the wells and the cells were allowed to settle for 1 h at RT. After gentle aspiration of the solution, the cells were fixed using 0.025% glutaraldehyde for 40 min. Endogenous peroxidase activity was blocked with 3% H 2 O 2 for 1 h. The plates were blocked overnight with 0.2% gelatin in PBS containing 0.05% Tween-20 at 4°C. Then, the cells were incubated with serial dilutions of cell culture media containing analyzed antibodies followed by detection with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad) and HRP-labeled streptavidin (1:40,000, Sigma). All reagents were diluted in the blocking buffer. The colorogenic reaction was performed, and the absorbance at 450 nm was measured as described above. C1q Binding Duplicates of polystyrene plates were coated with 6 µg/ml of BSA conjugated with the discriminating trisaccharide of the B antigen (Dextra Laboratories, cat #NGP6323, batch #ATDX232-039) overnight at 4°C, blocked with 1% BSA in PBS (1 h) and incubated with serial dilutions of cell culture media containing O10 muteins (2 h). Then, one set of plates was used for evaluation of C1q binding by analyzed antibodies; and the plates were incubated with 2 µg/ml of C1q purified from human serum (Biorad, cat. #22215504, batch #130815; 2 h) and next with HRP-labeled anti-human C1q polyclonal antibody (1:400, Abcam, cat. #ab46191, lot #GR205436-5; 1 h). The second set of plates was used to analyze quantities of the muteins bound to the immobilized antigen; and the plates were incubated with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad, 2 h) and HRP-labeled streptavidin (1:40,000, Sigma; 1 h). Absorbance of HRP product was measured as described above. Each incubation step was preceded by extensive washing with 0.05% Tween-20 in PBS. The antibodies, streptavidin, and C1q were diluted in 0.1% BSA in PBS. All incubations, except plate coating, were at RT. The normalized C1q binding ( a ) was calculated by dividing the signal corresponding to C1q binding ( b ) by the signal corresponding to bound antibody ( c ). normalized C 1 q binding ( a ) = C 1 q binding signal ( b ) bound antibody signal ( c ) Uncertainty of a (Δ a ) was calculated by exact differential. Uncertainties of b and c (Δ b and Δ c ) equaled to standard deviations of the absorbance measurements. uncertainty of a ​ ( Δ a ) = | ∂ a ∂ b Δ b | + | ∂ a ∂ c Δ c | = | Δ b c | + | b c 2 Δ c | Complement Cascade Activation Washed human red blood cells suspended to a hematocrit of 2% were coated with 3 µg/ml or 1.5 µg/ml of the analyzed antibodies for 1.5 h at RT, washed twice with PBS and resuspended in PBS with Ca 2+ and Mg 2+ . Then, the same volume of human complement serum (Sigma, cat. #S1764, lot #SLBS5471V, #SLBQ0752V or #SLBP0461V) diluted to 7 CH 50 U/ml in PBS with Ca 2+ and Mg 2+ was added to the coated red blood cells. The samples were centrifuged after 2 h of incubation at 37°C, and the absorbance of released hemoglobin was measured in the supernatants at 540 nm. Red Blood Cells and Agglutination Standard human red blood cells were purchased from Regional Centre of Blood Donation and Blood Treatment in Katowice, Poland. Agglutination tests were performed in 96-flat bottom plates. Serially diluted solutions of analyzed antibodies (100 µl) were gently mixed with 0.45% (hematocrit) suspension of red blood cells. The level of agglutination was analyzed using a phase-contrast microscope after 20 min of moderate shaking. A six-point scale was used for evaluation of agglutination intensity: from 4+ (complete cell aggregation), to 3+, 2+, 1+, ± to a negative score. The agglutination score reflects both the size of aggregates and quantity of non-agglutinated cells. IgG3 Self-Association Assay Oligomerization of IgG3 was analyzed similarly to the method described by Abdelmoula et al ( 2 ). Three different concentrations of the purified domain muteins with M18 variable region (150, 100, and 20 µg/ml) were incubated with non-mutated, biotinylated IgG3 M18 (100 ng/ml) for 72–96 h at 4°C in the presence of 4% BSA (Sigma, cat. #A9576). Then, the antibody complexes were precipitated by adding 50% PEG-6000 (Sigma) to the final concentration of 7.5%. After 1 h of incubation on ice, the samples were centrifuged (30 min, 3,000 × g , 4°C). The supernatants were preserved for further analysis, and the precipitates were washed with ice-cold 7.5% PEG-6000 in PBS and centrifuged again (30 min, 3,000 × g , 4°C). Then, the precipitates were dissolved in PBS with 0.1% BSA by pipetting at 37°C. Biotinylated IgG3 in the precipitates and supernatants was quantified using ELISA on polystyrene plates coated with streptavidin (8 µg/ml, Thermo, cat. #434301, lot #RB233354). Bound biotinylated IgG3 was detected using rabbit monoclonal antibody M111-2 (1:1,000, Abcam, cat. #ab125904, lot #C050311, #GR157092-1) and HRP-labeled goat anti-rabbit polyclonal antibody (1:3,000, Sigma, cat #A6667). The absorbance of HRP product was measured as described above. SDS-PAGE and Western Blotting Samples were resolved in 8 or 12% polyacrylamide gels under non-reducing or reducing conditions according to the protocol of Laemmli ( 14 ). After wet electrotransfer onto PVDF membrane and blocking with 4% skim milk in PBS, the samples were probed with anti-mouse kappa polyclonal antibody conjugated with biotin (1:3,000, BioRad) and HRP-labeled streptavidin (1:40,000, Sigma). Alternatively, rabbit anti-HA tag polyclonal antibody (1:10,000, Abcam, cat. #ab9110) and HRP-labeled goat anti-rabbit F(ab') 2 polyclonal antibody (1:10,000, Sigma, cat. #A6667, lot #SLBG3029) were used. Mouse IgG3 heavy chain was detected using goat antiserum to mouse IgG3 (1:500, Sigma, cat. #ISO2) and rabbit anti-goat polyclonal antibody (1:5,000, Sigma, cat. #A4174). Bands were visualized using Immobilon Western Chemiluminescent Substrate for HRP (Millipore). The images were captured and analyzed using Fusion Fx apparatus with the Fusion Capt Advance Fx5 program (Vilbert Lourmat, France). Results Comparison of Hemagglutination Induced by IgG3 and IgG3 F(ab') 2 In our previous work, we reported that F(ab') 2 obtained from IgG3 induces agglutination of erythrocytes bearing a cognate antigen ( 11 ). However, as shown below, complete IgG3 agglutinates erythrocytes more efficiently than its F(ab') 2 , i.e., a much higher concentration of F(ab') 2 than that of the intact molecule is required for agglutination. We compared hemagglutination triggered by: (i) native, purified, full-length IgG3, and its F(ab') 2 obtained by IdeZ protease digestion and (ii) culture media of cells producing recombinant IgG3 or recombinant F(ab') 2 (Table 1 ). Concentrations of IgG3 and F(ab') 2 in the media were measured using ELISA and equalized for the comparative tests. We verified the quality of analyzed proteins and confirmed that recombinant IdeZ protease, which cleaves IgG3 at a single site in its hinge region, generates homogeneous F(ab') 2 (Figures 1 A,B). Based on this analysis, we estimated that IgG3 was about 32 to 64-times more potent than its F(ab') 2 in hemagglutination (Table 1 ; Figure 1 B). The results indicate that the Fc of IgG3 strongly enhances hemagglutination induced by this isotype. Table 1 M18 full-length IgG3 and M18 IgG3 F(ab') 2 induced hemagglutination with different efficacy. Concentration (nM) Score of hemagglutination a Concentration (nM) Score of hemagglutination Native IgG3 F(ab') 2 obtained using IdeZ Recombinant IgG3 Recombinant F(ab') 2 82.5 ++++ + 93 ++++ ± 41.3 ++++ ± 46.5 ++++ – 20.6 ++++ – 23.3 ++++ – 10.3 ++++ – 11.6 +++ – 5.2 +++ – 5.8 +++ – 2.6 +++ – 2.9 + – 1.3 ++ – 1.5 ± – 0.6 ± – 0.7 – – 0.3 – – 0.4 – – a Representative results of three independent experiments . Figure 1 Hemagglutination induced by M18 IgG3 and its F(ab') 2 . (A) Integrity of generated F(ab') 2 was verified using SDS-PAGE. In the case of the purified antibody digested with IdeZ, the gels were stained with Coomassie Brilliant Blue (CBB). Recombinant F(ab') 2 was equipped with HA tag and its integrity was confirmed using Western blotting with anti-HA tag antibody. The molecular mass of non-reduced F(ab') 2 is about 120 kDa. HC, heavy chain; HC', heavy chain fragments generated after IdeZ cleavage; LC, light chain; (B) Microscopic images of erythrocytes agglutinated by equal molar concentrations of IgG3 antibody and its F(ab') 2 . The antibody fragment was obtained from native IgG3 using IdeZ digestion. The CH2 Domain Derived From IgG3 Enhanced Hemagglutination Efficacy of an Antibody Our previous attempts to explain the mechanism of IgG3-dependent hemagglutination brought us to the hypothesis that the elongated hinge of IgG3 determines its hemagglutination ability ( 11 ). In light of the new results, the hypothesis required revision. To elucidate which domains of IgG3 are crucial for its hemagglutination ability, we generated a panel of domain muteins of agglutinating IgG3 and non-agglutinating IgG1 isotypes (Figure 2 A). We generated pairs of IgG1 and IgG3 molecules with the same variable regions and with swapped: (i) hinge regions; (ii) hinge regions + CH1 domains; (iii) CH2 domains, and (iv) CH3 domains and searched for muteins of two types: loss-of-function in the case of IgG3 and gain-of-function in the case of IgG1. Figure 2 Hemagglutination induced by IgG1 and IgG3 muteins. (A) Generated domain muteins and their nomenclature. (B) Microscopic images of hemagglutination induced by the domain muteins. All antibodies were used at a concentration of 1.5 µg/ml, except of O10 IgG1_CH2-3 that was used at 3 µg/ml. Scale bar—100 µm; (C) Hemagglutination induced by selected O10 muteins used at 10 µg/ml. Preliminary experiments showed that hinge swapping between IgG1 and IgG3 hinders disulfide bonds formation between chains of the muteins (Figure S1 in Supplementary Material). However, the IgG1_h-3 and IgG3_h-1 variants were functional and their affinity to the antigen was similar to that of the parental molecules (shown in Figure 3 ). As demonstrated by Dall'Acqua et al., immunoglobulins with modified hinges frequently form functional heterotetramers (HC) 2 (LC) 2 despite the lack of disulfide bonds between the chains ( 15 ). Gel filtration confirmed that IgG1_h-3 and IgG3_h-1 have molecular mass greater than 150 kDa and form stable (HC) 2 (LC) 2 heterotetramers (data not shown). To make sure that the results of the following experiments are not the consequence of incorrect assemblies of the hinge-swapped muteins, we also generated muteins with swapped fragments comprising hinge regions and CH1 domains. All muteins were successfully expressed and their integrity was verified using SDS-PAGE and Western blotting (Figure S1 in Supplementary Material). Figure 3 Antigen binding by the domain muteins. M18 and O10 antibodies are specific to B-antigen present on human erythrocytes. B-antigen is a pentasaccharide O-glycan. O10 antibody, but not M18, binds terminal fragment of the antigen, called B-trisaccharide. Antigen–antibody interaction was analyzed using ELISA on immobilized erythrocytes. In the case of O10, plates coated with BSA conjugated with the synthetic B-trisaccharide were also used. The plots present mean values from two independent experiments performed in duplicates. Results obtained for IgG1 and IgG3, the parental molecules, are presented on each plot to allow convenient comparisons. Then, we compared hemagglutination induced by the muteins (Tables 2 and 3 ; Figures 2 B,C). The results showed that neither the CH1 domain nor the hinge determined the agglutination ability of IgG3. The introduction of the CH3 domain from IgG1 into IgG3 resulted in a molecule with slightly reduced hemagglutination score but CH3 from IgG3 did not translate into IgG1 ability of hemagglutination. In contrast, CH2 swapping led to the generation of IgG1 mutein (IgG1_CH2-3) that gained the ability of hemagglutination (Tables 2 and 3 ). Moreover, the paired IgG3 mutein (IgG3_CH2-1) had about 16-times reduced hemagglutination score in comparison to the parental IgG3. Table 2 Scores of hemagglutination induced by M18 variants. Conc. (μg/ml) Parental IgGs Swap of hinge regions Swap of CH1 + hinge domains Swap of CH2 domains Swap of CH3 domains IgG1 IgG3 IgG1_h-3 IgG3_h-1 IgG1_CH1h-3 IgG3_CH1h-1 IgG1_CH2-3 IgG3_CH2-1 IgG1_CH3-3 IgG3_CH3-1 I a II I II I II I II I II I II I II I II I II I II 1.500 – – +++ ++ – – +++ +++ – – +++ +++ + + – + – ± ++ +++ 0.750 – – +++ +++ – – + +++ – – ++ +++ – ± – – – – + ++ 0.375 – – ++ +++ – – ± + – – + ++ – – – – – – ± ++ 0.188 – – + ++ – – – – – – ± + – – – – – – – – 0.094 – – ± ± – – – – – – – – – – – – – – – – 0.047 – – – – – – – – – – – – – – – – – – – – a Results of two independent experiments designated as I and II . Table 3 Scores of hemagglutination induced by O10 variants. Conc. (μg/ml) Parental IgGs Swap of hinge regions Swap of hinge + CH1 domains Swap of CH2 domains Swap of CH3 domains IgG1 IgG3 IgG1_h-3 IgG3_h-1 IgG1_CH1h-3 IgG3_CH1h-1 IgG1_CH2-3 IgG3_CH2-1 IgG1_CH3-3 IgG3_CH3-1 I a II I II I II I II I II I II I II I II I II I II 3.000 – – ++ ++ – – +++ ++ – – +++ ++ + + – ± – + +++ +++ 1.500 – – +++ +++ – – +++ ++ – – +++ ++ ± – – – – – + ++ 0.750 – – ++ ++ – – ++ ++ – – + ++ – – – – – – ± ± 0.375 – – + + – – + + – – ± ± – – – – – – – – 0.188 – – ± ± – – ± ± – – – + – – – – – – – – 0.094 – – – – – – – – – – – – – – – – – – – – a Results of two independent experiments designated as I and II . IgG1_CH2-3 as a gain-of-function mutein was particularly interesting, because it indicated that the CH2 domain of IgG3 is the one critical for hemagglutination. However, IgG1_CH2-3 agglutinated erythrocytes with considerably lower score than native IgG3. We also compared hemagglutination efficacy of native M18 IgG3 and its deglycosylated form. Deglycosylated IgG3 agglutinated erythrocytes about 16-times weaker than the native molecule (Table S1 in Supplementary Material). To sum up, the ability of IgG3 to agglutinate erythrocytes results from its unique structure, in which the CH2 domain is especially important and strongly enhances the efficacy of the process. Although the IgG3 F(ab') 2 is sufficient to trigger hemagglutination, its efficacy is much lower in comparison to full-length IgG3, probably just due to the lack of the CH2 domain. The hinge region seems to have little influence on agglutination ability, because IgG3 with the IgG1-derived hinge agglutinated erythrocytes only slightly less effectively than the parental molecule. IgG3 Constant Domains Modify Functional Affinity to an Antigen There is a general agreement that the increased functional affinity of IgG3 results from an avidity effect caused by the interactions between the Fc fragments of the molecules ( 16 ). In line with that, we observed that IgG3 binds to erythrocytes much more efficiently than IgG1 with the same variable region and much more efficiently than IgG3-derived F(ab') 2 (Figure S2 in Supplementary Material). Some authors discussed also the potential role of N -glycans in IgG3 unique properties ( 17 , 18 ), but we did not observe any differences in antigen binding between control and deglycosylated antibody (Figure S2 in Supplementary Material). Aiming to understand why IgG3 has increased functional affinity, we analyzed antigen binding by the domain muteins (Figure 3 ). The results showed that the hinge region of IgG3 does not influence the functional affinity of the antibody, but muteins with the swapped CH1 + hinge, CH2, or CH3 domains had changed functional affinity. The introduction of IgG3-derived CH1 + hinge or CH2 domain into the IgG1 framework enhanced antigen binding in comparison to the parental IgG1. Conversely, the paired IgG3 muteins with IgG1-derived CH1 + hinge or CH2 had reduced functional affinity. The swapping of the CH3 domains resulted in the IgG3 mutein with decreased affinity, but in the paired IgG1 mutein, the effect was not substantial. The calculated EC 50 values of antigen binding for IgG3 muteins indicated that the CH2 domain had the strongest influence on IgG3-antigen interaction (Table 4 ). CH2 swapping resulted in IgG3 muteins with 3–12 times decreased functional affinity. Table 4 EC 50 of mutein binding to the antigen calculated using data from Figure 3 . Variable region of the mutein and type of the antigen EC 50 of mutein binding (nM) IgG3 IgG3_h-1 IgG3_CH1h-1 IgG3_CH2-1 IgG3_CH3-1 M18 (erythrocytes) 0.24 ± 0.02 0.38 ± 0.01 0.61 ± 0.03 0.86 ± 0.04 0.87 ± 0.07 O10 (erythrocytes) 0.81 ± 0.01 0.88 ± 0.02 2.05 ± 0.06 9.75 ± 2.04 3.60 ± 0.29 O10 (B-trisaccharide conjugated to BSA) 0.50 ± 0.04 0.38 ± 0.02 1.15 ± 0.06 1.40 ± 0.03 0.79 ± 0.06 Overall, the results indicate that the higher (in comparison to IgG1) functional affinity of IgG3 to its antigen does not depend on a separate constant domain of this isotype, but rather is an additive result of discrete properties of the all three constant domains CH1, CH2, and CH3, but not the hinge region. Of all the constant domains, CH2 contributes the most to the high functional affinity of IgG3. Fc-Dependent Oligomerization of the Domain Muteins The hallmark of mouse IgG3 is its ability to oligomerize. The process depends on Fc fragment, but its exact molecular mechanism is unknown. We analyzed whether the domain muteins form non-covalent complexes using polyethylene glycol (PEG) precipitation with a labeled IgG3 probe ( 2 ). In comparison to the original method, we used biotinylated IgG3 instead of radiolabeled IgG3. The IgG3-biotin interacted with oligomerizing muteins and the complexes comprised the mutein and the probe. The complexes were precipitated using PEG, and then, IgG3-biotin was quantified in precipitates and supernatants using ELISA. A high precipitate/supernatant ratio of IgG3-biotin quantities indicates that the mutein forms oligomers. The experiment showed that 5 out of 10 analyzed molecules form PEG-precipitable oligomers—IgG3 (control) and all IgG3 muteins but the one containing IgG1-derived CH2 domain and none of IgG1 muteins but the one with IgG3-derived CH2 domain (Figure 4 ). Oligomerization did not depend on CH2 glycosylation (Figure S5 in Supplementary Material). The results indicate that the CH2 domain is crucial for oligomerization of IgG3. Figure 4 Oligomerization of the domain muteins. The antibodies (M18 variants, 150 µg/ml) were incubated at 4°C and oligomers were precipitated using PEG. Biotinylated IgG3 was used as a probe that oligomerized with the muteins and became a part of the complexes. The charts present results from two independent experiments. The results obtained for 100 and 20 µg/ml of the muteins are shown in Figure S3 in Supplementary Material. A percentage of the total IgG3-biotin detected in the precipitates and the supernatants is presented in Figure S4 in Supplementary Material. Complement Activation by the IgG1 and IgG3 Muteins Similar to human antibodies, there are pronounced differences between mouse IgG subclasses in their ability to trigger complement cascade. Mouse IgG3 activates complement efficiently, whereas mouse IgG1 does not. Although there are many excellent reports concerning correlation between human antibody structure and its ability of complement activation, the structural determinants of mouse antibodies that allow to trigger the cascade are not precisely known. The best characterized Ig with respect to complement activation is human IgG1, in which several amino acid residues were identified as crucial for the initiation of the complement cascade (Figure S6 in Supplementary Material) ( 19 – 21 ). The sequence alignment of human IgG1, mouse IgG1, and mouse IgG3 indicated that the majority of human IgG1 amino acid residues involved in complement activation are conserved in both mouse isotypes (Figure S6 in Supplementary Material). However, it revealed two differences between mouse IgG1 and IgG3 within the regions corresponding to those involved in C1q binding by human IgG1—in the N-terminal fragment of the CH2 domain (Val231-Ser238 in IgG1 and Ile234-Pro238 in IgG3, EU numbering ( 22 )) and in the residue 322 (Figure S6 in Supplementary Material). To verify whether these motifs are involved in complement activation by mouse IgG3, we generated additional muteins in which we swapped them between IgG1 and IgG3—IgG1_ILGGP (Val231Ile Pro232Leu Glu236Gly Val237Gly Ser238Pro); IgG3_VPEVS (Ile234Val Leu235Pro Gly236Glu Gly237Val Pro238Ser); IgG1_Arg322Lys; IgG3_Lys322Arg, and a double mutein IgG1_ILGGP_Arg322Lys. The IgG3 heavy chain containing VPEVS did not associate with a light chain and was not secreted (Figure 5 A). Lys322Arg replacement completely abolished complement activation by IgG3 indicating that Lys322 is crucial for this process (Figures 5 B,C). The three muteins of IgG1 did not bind C1q nor activated complement cascade indicating that the IgG1 framework prevents activation of complement (Figures 5 B,C). The results showed that the known C1q-binding motifs are functional in the mouse IgG3 but not in the mouse IgG1 framework. Figure 5 Functionality of known C1q-bining motifs in mouse IgG1 and mouse IgG3 frameworks. (A) IgG3_VPEVS did not associate with a light chain and was not secreted by the producing cells. (B) C1q binding by the muteins. Plates coated with BSA conjugated with the antigen, B-trisaccharide, were incubated with the muteins at 3 µg/ml (O10 variants). Then purified C1q was added. The muteins bind the antigen with different functional affinity. Thus, the C1q signal was normalized to the quantity of the bound antibody. Data used for calculation of the normalized binding are shown in Figure S7 in Supplementary Material. Error bars correspond to uncertainty calculated as presented in Section " Materials and Methods ." (C) Complement cascade activation by the muteins (3 µg/ml). Erythrocytes coated with the antibodies were incubated with complement serum. Complete lysis (100%) corresponds to water-induced lysis. In (A–C) representative results of two independent experiments are shown. Some authors observed a correlation between hinge-dependent segmental flexibility of an antibody and its ability to activate complement ( 12 ). Thus, the differences between activity of mouse IgG1 and IgG3 are frequently explained on the basis of the length of their hinges. We decided to empirically verify this hypothesis using the domain muteins. First, we analyzed the binding of the complement cascade initiator (C1q) to the muteins (Figure 6 A). The results showed that the hinge modification does not affect C1q binding. The swapping of the CH2 domain from IgG1 into IgG3 abolished C1q binding by the latter. Interestingly, the paired mutein (IgG1_CH2-3) did not gain the ability to strongly interact with C1q; its binding of C1q reached ~12% of that characteristic for native IgG3. Swapping of the CH1 + hinge domains or the CH3 domains between IgG1 and IgG3 moderately diminished C1q binding by IgG3 and did not increase its binding by IgG1. Figure 6 Complement activation induced by the domain muteins. (A) C1q binding to the domain muteins (O10 variants). The data used for calculations are presented in Figure S7 in Supplementary Material. Error bars correspond to uncertainty calculated as described in Section " Materials and Methods ." (B) Complement cascade activation by the domain muteins. Erythrocytes coated with 3 µg/ml of the muteins were incubated with complement serum. 100% lysis corresponds to water-induced lysis. The bars present mean values and standard deviation of duplicates from one experiment. Results obtained with 1.5 µg/ml of the muteins are presented in Figure S8 in Supplementary Material. (A,B) Representative results of two independent experiments. We also analyzed complement activation in serum triggered by erythrocytes coated with the domain muteins (Figure 6 B). The levels of erythrocyte lysis indicated that all muteins containing the CH2 domain derived from IgG3 activate complement cascade. The observed differences in C1q binding were not reflected by the different efficacy of the cascade triggering. The muteins with low (IgG1_CH2-3) or moderate (IgG3_CH1h-1, IgG3_CH3-1) ability of C1q-binding activated complement cascade with efficacy similar to that of the parental IgG3. It seems that in the case of the antibodies comprising IgG3-derived CH2 domain, even weak interaction with C1q was sufficient to effectively activate the whole complement cascade. The results showed that both IgG1- and IgG3-derived CH1, hinge, and CH3 domains are permissive for C1q binding and complement activation. The CH2 domain of IgG1 is a non-permissive framework for the known C1q-binding motifs. Overall, the results pointed to the CH2 domain as the major determinant of mouse IgG3 functions and unique properties of this isotype. In the last part of our work, we sought for features of the IgG3-derived CH2 domain that may account for IgG3 distinctive characteristic. Properties of Muteins With Reversed Charge of the CH2 Domains The most striking difference between mouse IgG3-derived CH2 and CH2 domains of other IgG subclasses is their charges; only the former has a strong positive charge. For example, at pH 7.0, the net charge of the CH2 domain of IgG1 is −2.6 and of IgG3 is +2.6 (calculated using http://protcalc.sourceforge.net/ ). Hovenden et al. ( 9 ) found a correlation between the charge of CH2 domains of mouse IgG subclasses and their affinity to a negatively charged polyvalent antigen (poly-glutamic acid, poly-GA); and the high affinity of IgG3 to poly-GA was attributed to the charge of its CH2 domain. We analyzed spatial distribution of charged residues on the CH2 surface of IgG1 and IgG3 using previously obtained molecular models ( 11 ) and data deposited in PDB record 1IGY (Figure 7 A). We identified 29 residues that differ between CH2 domains of mouse IgG1 and IgG3, 9 of which have different charge (Figure S9 in Supplementary Material). Based on the models, we selected four basic residues (His274, Lys282, Arg315, and Lys326) that are regularly spaced on the outer surface of the CH2 domain of IgG3 (Figure 7 A; Figure S9 in Supplementary Material). The same residues in IgG1 are not charged. To verify whether CH2 charge influences IgG3 properties, we generated two muteins in which the four residues were swapped—IgG3_CH2charge (His274Gln Lys282Val Arg315Asn Lys326Ala) and IgG1_CH2charge (Gln274His Val282Lys Asn315Arg Ala326Lys). These muteins were expressed, correctly assembled, and soluble (Figure S1 in Supplementary Material). The introduced mutations reversed the charge of the CH2 domains. It was 0.6 and −0.7 at pH 7.0 for the CH2 domain of IgG1_CH2charge and IgG3_CH3charge, respectively. Figure 7 Properties of the muteins with modified charge of the CH2 domain. (A) Charge location on the CH2 domain of IgG1 and IgG3. Basic residues (Arg, His, and Lys) are faint red, acidic residues (Asp, Glu) are blue, and a site of CH2 N-glycosylation (Asn297) is green. His274, Lys282, Arg315, and Lys326 of IgG3 CH2 are dark red. These four residues were swapped between IgG1 and IgG3 to generate IgG1_CH2charge and IgG3_CH2charge muteins. The images present views obtained by 90° rotation of the domain models. (B) Antigen binding by the muteins. The charts present representative results of two independent experiments performed in duplicates or triplicates. Error bars equal to SD. (C) Oligomerization of the muteins. Results from two independent experiments with 100 µg/ml of the antibodies (M18 variants) are shown. A percentage of the total IgG3-biotin detected in precipitates and supernatants are presented in Figure S4 in Supplementary Material. Results for IgG1 and IgG3 are the same as in Figure 4 because the data were collected in the same experiments. (D) C1q binding by the muteins (O10 variants, 3 µg/ml). Data used for calculation of the normalized binding are shown in Figure S7 in Supplementary Material. The chart presents representative results of two independent experiments. Error bars correspond to uncertainty calculated as presented in Section " Materials and Methods ." (E) Complement cascade activation by the muteins (3 µg/ml). Erythrocytes coated with the antibodies were incubated with complement serum. Complete lysis (100%) corresponds to water-induced lysis. Representative results of two independent experiments are shown. We compared properties of the parental molecules and the muteins with modified CH2 charge. We observed that the charge influenced binding to erythrocytes (Figure 7 B). However, hemagglutination, oligomerization, C1q binding, and complement activation were not affected by this charge modification (Table 5 ; Figures 7 C–E). The results indicate that the four analyzed residues have only limited impact on the IgG3 properties. We cannot exclude that other charged residues within the CH2 domain of IgG3 may influence or determine properties of this isotype. Table 5 Hemagglutination induced by the muteins with modified charge of the CH2 domain. M18 variants O10 variants Conc. (μg/ml) IgG1 IgG3 IgG1_CH2charge IgG3_CH2charge Conc. (μg/ml) IgG1 IgG3 IgG1_CH2charge IgG3_CH2charge 5.00 ± a ++++ – +++ 2.00 – ++++ ± ++++ 2.50 – ++++ – +++ 1.00 – +++ – +++ 1.25 – +++ – ++ 0.50 – ++ – ++ 0.63 – ++ – ++ 0.25 – ± – ± 0.31 – + – ± 0.13 – – – – 0.16 – ± – – 0.06 – – – – 0.00 – – – – 0.00 – – – – a Representative results of two independent experiments . Comparison of Hemagglutination Induced by IgG3 and IgG3 F(ab') 2 In our previous work, we reported that F(ab') 2 obtained from IgG3 induces agglutination of erythrocytes bearing a cognate antigen ( 11 ). However, as shown below, complete IgG3 agglutinates erythrocytes more efficiently than its F(ab') 2 , i.e., a much higher concentration of F(ab') 2 than that of the intact molecule is required for agglutination. We compared hemagglutination triggered by: (i) native, purified, full-length IgG3, and its F(ab') 2 obtained by IdeZ protease digestion and (ii) culture media of cells producing recombinant IgG3 or recombinant F(ab') 2 (Table 1 ). Concentrations of IgG3 and F(ab') 2 in the media were measured using ELISA and equalized for the comparative tests. We verified the quality of analyzed proteins and confirmed that recombinant IdeZ protease, which cleaves IgG3 at a single site in its hinge region, generates homogeneous F(ab') 2 (Figures 1 A,B). Based on this analysis, we estimated that IgG3 was about 32 to 64-times more potent than its F(ab') 2 in hemagglutination (Table 1 ; Figure 1 B). The results indicate that the Fc of IgG3 strongly enhances hemagglutination induced by this isotype. Table 1 M18 full-length IgG3 and M18 IgG3 F(ab') 2 induced hemagglutination with different efficacy. Concentration (nM) Score of hemagglutination a Concentration (nM) Score of hemagglutination Native IgG3 F(ab') 2 obtained using IdeZ Recombinant IgG3 Recombinant F(ab') 2 82.5 ++++ + 93 ++++ ± 41.3 ++++ ± 46.5 ++++ – 20.6 ++++ – 23.3 ++++ – 10.3 ++++ – 11.6 +++ – 5.2 +++ – 5.8 +++ – 2.6 +++ – 2.9 + – 1.3 ++ – 1.5 ± – 0.6 ± – 0.7 – – 0.3 – – 0.4 – – a Representative results of three independent experiments . Figure 1 Hemagglutination induced by M18 IgG3 and its F(ab') 2 . (A) Integrity of generated F(ab') 2 was verified using SDS-PAGE. In the case of the purified antibody digested with IdeZ, the gels were stained with Coomassie Brilliant Blue (CBB). Recombinant F(ab') 2 was equipped with HA tag and its integrity was confirmed using Western blotting with anti-HA tag antibody. The molecular mass of non-reduced F(ab') 2 is about 120 kDa. HC, heavy chain; HC', heavy chain fragments generated after IdeZ cleavage; LC, light chain; (B) Microscopic images of erythrocytes agglutinated by equal molar concentrations of IgG3 antibody and its F(ab') 2 . The antibody fragment was obtained from native IgG3 using IdeZ digestion. The CH2 Domain Derived From IgG3 Enhanced Hemagglutination Efficacy of an Antibody Our previous attempts to explain the mechanism of IgG3-dependent hemagglutination brought us to the hypothesis that the elongated hinge of IgG3 determines its hemagglutination ability ( 11 ). In light of the new results, the hypothesis required revision. To elucidate which domains of IgG3 are crucial for its hemagglutination ability, we generated a panel of domain muteins of agglutinating IgG3 and non-agglutinating IgG1 isotypes (Figure 2 A). We generated pairs of IgG1 and IgG3 molecules with the same variable regions and with swapped: (i) hinge regions; (ii) hinge regions + CH1 domains; (iii) CH2 domains, and (iv) CH3 domains and searched for muteins of two types: loss-of-function in the case of IgG3 and gain-of-function in the case of IgG1. Figure 2 Hemagglutination induced by IgG1 and IgG3 muteins. (A) Generated domain muteins and their nomenclature. (B) Microscopic images of hemagglutination induced by the domain muteins. All antibodies were used at a concentration of 1.5 µg/ml, except of O10 IgG1_CH2-3 that was used at 3 µg/ml. Scale bar—100 µm; (C) Hemagglutination induced by selected O10 muteins used at 10 µg/ml. Preliminary experiments showed that hinge swapping between IgG1 and IgG3 hinders disulfide bonds formation between chains of the muteins (Figure S1 in Supplementary Material). However, the IgG1_h-3 and IgG3_h-1 variants were functional and their affinity to the antigen was similar to that of the parental molecules (shown in Figure 3 ). As demonstrated by Dall'Acqua et al., immunoglobulins with modified hinges frequently form functional heterotetramers (HC) 2 (LC) 2 despite the lack of disulfide bonds between the chains ( 15 ). Gel filtration confirmed that IgG1_h-3 and IgG3_h-1 have molecular mass greater than 150 kDa and form stable (HC) 2 (LC) 2 heterotetramers (data not shown). To make sure that the results of the following experiments are not the consequence of incorrect assemblies of the hinge-swapped muteins, we also generated muteins with swapped fragments comprising hinge regions and CH1 domains. All muteins were successfully expressed and their integrity was verified using SDS-PAGE and Western blotting (Figure S1 in Supplementary Material). Figure 3 Antigen binding by the domain muteins. M18 and O10 antibodies are specific to B-antigen present on human erythrocytes. B-antigen is a pentasaccharide O-glycan. O10 antibody, but not M18, binds terminal fragment of the antigen, called B-trisaccharide. Antigen–antibody interaction was analyzed using ELISA on immobilized erythrocytes. In the case of O10, plates coated with BSA conjugated with the synthetic B-trisaccharide were also used. The plots present mean values from two independent experiments performed in duplicates. Results obtained for IgG1 and IgG3, the parental molecules, are presented on each plot to allow convenient comparisons. Then, we compared hemagglutination induced by the muteins (Tables 2 and 3 ; Figures 2 B,C). The results showed that neither the CH1 domain nor the hinge determined the agglutination ability of IgG3. The introduction of the CH3 domain from IgG1 into IgG3 resulted in a molecule with slightly reduced hemagglutination score but CH3 from IgG3 did not translate into IgG1 ability of hemagglutination. In contrast, CH2 swapping led to the generation of IgG1 mutein (IgG1_CH2-3) that gained the ability of hemagglutination (Tables 2 and 3 ). Moreover, the paired IgG3 mutein (IgG3_CH2-1) had about 16-times reduced hemagglutination score in comparison to the parental IgG3. Table 2 Scores of hemagglutination induced by M18 variants. Conc. (μg/ml) Parental IgGs Swap of hinge regions Swap of CH1 + hinge domains Swap of CH2 domains Swap of CH3 domains IgG1 IgG3 IgG1_h-3 IgG3_h-1 IgG1_CH1h-3 IgG3_CH1h-1 IgG1_CH2-3 IgG3_CH2-1 IgG1_CH3-3 IgG3_CH3-1 I a II I II I II I II I II I II I II I II I II I II 1.500 – – +++ ++ – – +++ +++ – – +++ +++ + + – + – ± ++ +++ 0.750 – – +++ +++ – – + +++ – – ++ +++ – ± – – – – + ++ 0.375 – – ++ +++ – – ± + – – + ++ – – – – – – ± ++ 0.188 – – + ++ – – – – – – ± + – – – – – – – – 0.094 – – ± ± – – – – – – – – – – – – – – – – 0.047 – – – – – – – – – – – – – – – – – – – – a Results of two independent experiments designated as I and II . Table 3 Scores of hemagglutination induced by O10 variants. Conc. (μg/ml) Parental IgGs Swap of hinge regions Swap of hinge + CH1 domains Swap of CH2 domains Swap of CH3 domains IgG1 IgG3 IgG1_h-3 IgG3_h-1 IgG1_CH1h-3 IgG3_CH1h-1 IgG1_CH2-3 IgG3_CH2-1 IgG1_CH3-3 IgG3_CH3-1 I a II I II I II I II I II I II I II I II I II I II 3.000 – – ++ ++ – – +++ ++ – – +++ ++ + + – ± – + +++ +++ 1.500 – – +++ +++ – – +++ ++ – – +++ ++ ± – – – – – + ++ 0.750 – – ++ ++ – – ++ ++ – – + ++ – – – – – – ± ± 0.375 – – + + – – + + – – ± ± – – – – – – – – 0.188 – – ± ± – – ± ± – – – + – – – – – – – – 0.094 – – – – – – – – – – – – – – – – – – – – a Results of two independent experiments designated as I and II . IgG1_CH2-3 as a gain-of-function mutein was particularly interesting, because it indicated that the CH2 domain of IgG3 is the one critical for hemagglutination. However, IgG1_CH2-3 agglutinated erythrocytes with considerably lower score than native IgG3. We also compared hemagglutination efficacy of native M18 IgG3 and its deglycosylated form. Deglycosylated IgG3 agglutinated erythrocytes about 16-times weaker than the native molecule (Table S1 in Supplementary Material). To sum up, the ability of IgG3 to agglutinate erythrocytes results from its unique structure, in which the CH2 domain is especially important and strongly enhances the efficacy of the process. Although the IgG3 F(ab') 2 is sufficient to trigger hemagglutination, its efficacy is much lower in comparison to full-length IgG3, probably just due to the lack of the CH2 domain. The hinge region seems to have little influence on agglutination ability, because IgG3 with the IgG1-derived hinge agglutinated erythrocytes only slightly less effectively than the parental molecule. IgG3 Constant Domains Modify Functional Affinity to an Antigen There is a general agreement that the increased functional affinity of IgG3 results from an avidity effect caused by the interactions between the Fc fragments of the molecules ( 16 ). In line with that, we observed that IgG3 binds to erythrocytes much more efficiently than IgG1 with the same variable region and much more efficiently than IgG3-derived F(ab') 2 (Figure S2 in Supplementary Material). Some authors discussed also the potential role of N -glycans in IgG3 unique properties ( 17 , 18 ), but we did not observe any differences in antigen binding between control and deglycosylated antibody (Figure S2 in Supplementary Material). Aiming to understand why IgG3 has increased functional affinity, we analyzed antigen binding by the domain muteins (Figure 3 ). The results showed that the hinge region of IgG3 does not influence the functional affinity of the antibody, but muteins with the swapped CH1 + hinge, CH2, or CH3 domains had changed functional affinity. The introduction of IgG3-derived CH1 + hinge or CH2 domain into the IgG1 framework enhanced antigen binding in comparison to the parental IgG1. Conversely, the paired IgG3 muteins with IgG1-derived CH1 + hinge or CH2 had reduced functional affinity. The swapping of the CH3 domains resulted in the IgG3 mutein with decreased affinity, but in the paired IgG1 mutein, the effect was not substantial. The calculated EC 50 values of antigen binding for IgG3 muteins indicated that the CH2 domain had the strongest influence on IgG3-antigen interaction (Table 4 ). CH2 swapping resulted in IgG3 muteins with 3–12 times decreased functional affinity. Table 4 EC 50 of mutein binding to the antigen calculated using data from Figure 3 . Variable region of the mutein and type of the antigen EC 50 of mutein binding (nM) IgG3 IgG3_h-1 IgG3_CH1h-1 IgG3_CH2-1 IgG3_CH3-1 M18 (erythrocytes) 0.24 ± 0.02 0.38 ± 0.01 0.61 ± 0.03 0.86 ± 0.04 0.87 ± 0.07 O10 (erythrocytes) 0.81 ± 0.01 0.88 ± 0.02 2.05 ± 0.06 9.75 ± 2.04 3.60 ± 0.29 O10 (B-trisaccharide conjugated to BSA) 0.50 ± 0.04 0.38 ± 0.02 1.15 ± 0.06 1.40 ± 0.03 0.79 ± 0.06 Overall, the results indicate that the higher (in comparison to IgG1) functional affinity of IgG3 to its antigen does not depend on a separate constant domain of this isotype, but rather is an additive result of discrete properties of the all three constant domains CH1, CH2, and CH3, but not the hinge region. Of all the constant domains, CH2 contributes the most to the high functional affinity of IgG3. Fc-Dependent Oligomerization of the Domain Muteins The hallmark of mouse IgG3 is its ability to oligomerize. The process depends on Fc fragment, but its exact molecular mechanism is unknown. We analyzed whether the domain muteins form non-covalent complexes using polyethylene glycol (PEG) precipitation with a labeled IgG3 probe ( 2 ). In comparison to the original method, we used biotinylated IgG3 instead of radiolabeled IgG3. The IgG3-biotin interacted with oligomerizing muteins and the complexes comprised the mutein and the probe. The complexes were precipitated using PEG, and then, IgG3-biotin was quantified in precipitates and supernatants using ELISA. A high precipitate/supernatant ratio of IgG3-biotin quantities indicates that the mutein forms oligomers. The experiment showed that 5 out of 10 analyzed molecules form PEG-precipitable oligomers—IgG3 (control) and all IgG3 muteins but the one containing IgG1-derived CH2 domain and none of IgG1 muteins but the one with IgG3-derived CH2 domain (Figure 4 ). Oligomerization did not depend on CH2 glycosylation (Figure S5 in Supplementary Material). The results indicate that the CH2 domain is crucial for oligomerization of IgG3. Figure 4 Oligomerization of the domain muteins. The antibodies (M18 variants, 150 µg/ml) were incubated at 4°C and oligomers were precipitated using PEG. Biotinylated IgG3 was used as a probe that oligomerized with the muteins and became a part of the complexes. The charts present results from two independent experiments. The results obtained for 100 and 20 µg/ml of the muteins are shown in Figure S3 in Supplementary Material. A percentage of the total IgG3-biotin detected in the precipitates and the supernatants is presented in Figure S4 in Supplementary Material. Complement Activation by the IgG1 and IgG3 Muteins Similar to human antibodies, there are pronounced differences between mouse IgG subclasses in their ability to trigger complement cascade. Mouse IgG3 activates complement efficiently, whereas mouse IgG1 does not. Although there are many excellent reports concerning correlation between human antibody structure and its ability of complement activation, the structural determinants of mouse antibodies that allow to trigger the cascade are not precisely known. The best characterized Ig with respect to complement activation is human IgG1, in which several amino acid residues were identified as crucial for the initiation of the complement cascade (Figure S6 in Supplementary Material) ( 19 – 21 ). The sequence alignment of human IgG1, mouse IgG1, and mouse IgG3 indicated that the majority of human IgG1 amino acid residues involved in complement activation are conserved in both mouse isotypes (Figure S6 in Supplementary Material). However, it revealed two differences between mouse IgG1 and IgG3 within the regions corresponding to those involved in C1q binding by human IgG1—in the N-terminal fragment of the CH2 domain (Val231-Ser238 in IgG1 and Ile234-Pro238 in IgG3, EU numbering ( 22 )) and in the residue 322 (Figure S6 in Supplementary Material). To verify whether these motifs are involved in complement activation by mouse IgG3, we generated additional muteins in which we swapped them between IgG1 and IgG3—IgG1_ILGGP (Val231Ile Pro232Leu Glu236Gly Val237Gly Ser238Pro); IgG3_VPEVS (Ile234Val Leu235Pro Gly236Glu Gly237Val Pro238Ser); IgG1_Arg322Lys; IgG3_Lys322Arg, and a double mutein IgG1_ILGGP_Arg322Lys. The IgG3 heavy chain containing VPEVS did not associate with a light chain and was not secreted (Figure 5 A). Lys322Arg replacement completely abolished complement activation by IgG3 indicating that Lys322 is crucial for this process (Figures 5 B,C). The three muteins of IgG1 did not bind C1q nor activated complement cascade indicating that the IgG1 framework prevents activation of complement (Figures 5 B,C). The results showed that the known C1q-binding motifs are functional in the mouse IgG3 but not in the mouse IgG1 framework. Figure 5 Functionality of known C1q-bining motifs in mouse IgG1 and mouse IgG3 frameworks. (A) IgG3_VPEVS did not associate with a light chain and was not secreted by the producing cells. (B) C1q binding by the muteins. Plates coated with BSA conjugated with the antigen, B-trisaccharide, were incubated with the muteins at 3 µg/ml (O10 variants). Then purified C1q was added. The muteins bind the antigen with different functional affinity. Thus, the C1q signal was normalized to the quantity of the bound antibody. Data used for calculation of the normalized binding are shown in Figure S7 in Supplementary Material. Error bars correspond to uncertainty calculated as presented in Section " Materials and Methods ." (C) Complement cascade activation by the muteins (3 µg/ml). Erythrocytes coated with the antibodies were incubated with complement serum. Complete lysis (100%) corresponds to water-induced lysis. In (A–C) representative results of two independent experiments are shown. Some authors observed a correlation between hinge-dependent segmental flexibility of an antibody and its ability to activate complement ( 12 ). Thus, the differences between activity of mouse IgG1 and IgG3 are frequently explained on the basis of the length of their hinges. We decided to empirically verify this hypothesis using the domain muteins. First, we analyzed the binding of the complement cascade initiator (C1q) to the muteins (Figure 6 A). The results showed that the hinge modification does not affect C1q binding. The swapping of the CH2 domain from IgG1 into IgG3 abolished C1q binding by the latter. Interestingly, the paired mutein (IgG1_CH2-3) did not gain the ability to strongly interact with C1q; its binding of C1q reached ~12% of that characteristic for native IgG3. Swapping of the CH1 + hinge domains or the CH3 domains between IgG1 and IgG3 moderately diminished C1q binding by IgG3 and did not increase its binding by IgG1. Figure 6 Complement activation induced by the domain muteins. (A) C1q binding to the domain muteins (O10 variants). The data used for calculations are presented in Figure S7 in Supplementary Material. Error bars correspond to uncertainty calculated as described in Section " Materials and Methods ." (B) Complement cascade activation by the domain muteins. Erythrocytes coated with 3 µg/ml of the muteins were incubated with complement serum. 100% lysis corresponds to water-induced lysis. The bars present mean values and standard deviation of duplicates from one experiment. Results obtained with 1.5 µg/ml of the muteins are presented in Figure S8 in Supplementary Material. (A,B) Representative results of two independent experiments. We also analyzed complement activation in serum triggered by erythrocytes coated with the domain muteins (Figure 6 B). The levels of erythrocyte lysis indicated that all muteins containing the CH2 domain derived from IgG3 activate complement cascade. The observed differences in C1q binding were not reflected by the different efficacy of the cascade triggering. The muteins with low (IgG1_CH2-3) or moderate (IgG3_CH1h-1, IgG3_CH3-1) ability of C1q-binding activated complement cascade with efficacy similar to that of the parental IgG3. It seems that in the case of the antibodies comprising IgG3-derived CH2 domain, even weak interaction with C1q was sufficient to effectively activate the whole complement cascade. The results showed that both IgG1- and IgG3-derived CH1, hinge, and CH3 domains are permissive for C1q binding and complement activation. The CH2 domain of IgG1 is a non-permissive framework for the known C1q-binding motifs. Overall, the results pointed to the CH2 domain as the major determinant of mouse IgG3 functions and unique properties of this isotype. In the last part of our work, we sought for features of the IgG3-derived CH2 domain that may account for IgG3 distinctive characteristic. Properties of Muteins With Reversed Charge of the CH2 Domains The most striking difference between mouse IgG3-derived CH2 and CH2 domains of other IgG subclasses is their charges; only the former has a strong positive charge. For example, at pH 7.0, the net charge of the CH2 domain of IgG1 is −2.6 and of IgG3 is +2.6 (calculated using http://protcalc.sourceforge.net/ ). Hovenden et al. ( 9 ) found a correlation between the charge of CH2 domains of mouse IgG subclasses and their affinity to a negatively charged polyvalent antigen (poly-glutamic acid, poly-GA); and the high affinity of IgG3 to poly-GA was attributed to the charge of its CH2 domain. We analyzed spatial distribution of charged residues on the CH2 surface of IgG1 and IgG3 using previously obtained molecular models ( 11 ) and data deposited in PDB record 1IGY (Figure 7 A). We identified 29 residues that differ between CH2 domains of mouse IgG1 and IgG3, 9 of which have different charge (Figure S9 in Supplementary Material). Based on the models, we selected four basic residues (His274, Lys282, Arg315, and Lys326) that are regularly spaced on the outer surface of the CH2 domain of IgG3 (Figure 7 A; Figure S9 in Supplementary Material). The same residues in IgG1 are not charged. To verify whether CH2 charge influences IgG3 properties, we generated two muteins in which the four residues were swapped—IgG3_CH2charge (His274Gln Lys282Val Arg315Asn Lys326Ala) and IgG1_CH2charge (Gln274His Val282Lys Asn315Arg Ala326Lys). These muteins were expressed, correctly assembled, and soluble (Figure S1 in Supplementary Material). The introduced mutations reversed the charge of the CH2 domains. It was 0.6 and −0.7 at pH 7.0 for the CH2 domain of IgG1_CH2charge and IgG3_CH3charge, respectively. Figure 7 Properties of the muteins with modified charge of the CH2 domain. (A) Charge location on the CH2 domain of IgG1 and IgG3. Basic residues (Arg, His, and Lys) are faint red, acidic residues (Asp, Glu) are blue, and a site of CH2 N-glycosylation (Asn297) is green. His274, Lys282, Arg315, and Lys326 of IgG3 CH2 are dark red. These four residues were swapped between IgG1 and IgG3 to generate IgG1_CH2charge and IgG3_CH2charge muteins. The images present views obtained by 90° rotation of the domain models. (B) Antigen binding by the muteins. The charts present representative results of two independent experiments performed in duplicates or triplicates. Error bars equal to SD. (C) Oligomerization of the muteins. Results from two independent experiments with 100 µg/ml of the antibodies (M18 variants) are shown. A percentage of the total IgG3-biotin detected in precipitates and supernatants are presented in Figure S4 in Supplementary Material. Results for IgG1 and IgG3 are the same as in Figure 4 because the data were collected in the same experiments. (D) C1q binding by the muteins (O10 variants, 3 µg/ml). Data used for calculation of the normalized binding are shown in Figure S7 in Supplementary Material. The chart presents representative results of two independent experiments. Error bars correspond to uncertainty calculated as presented in Section " Materials and Methods ." (E) Complement cascade activation by the muteins (3 µg/ml). Erythrocytes coated with the antibodies were incubated with complement serum. Complete lysis (100%) corresponds to water-induced lysis. Representative results of two independent experiments are shown. We compared properties of the parental molecules and the muteins with modified CH2 charge. We observed that the charge influenced binding to erythrocytes (Figure 7 B). However, hemagglutination, oligomerization, C1q binding, and complement activation were not affected by this charge modification (Table 5 ; Figures 7 C–E). The results indicate that the four analyzed residues have only limited impact on the IgG3 properties. We cannot exclude that other charged residues within the CH2 domain of IgG3 may influence or determine properties of this isotype. Table 5 Hemagglutination induced by the muteins with modified charge of the CH2 domain. M18 variants O10 variants Conc. (μg/ml) IgG1 IgG3 IgG1_CH2charge IgG3_CH2charge Conc. (μg/ml) IgG1 IgG3 IgG1_CH2charge IgG3_CH2charge 5.00 ± a ++++ – +++ 2.00 – ++++ ± ++++ 2.50 – ++++ – +++ 1.00 – +++ – +++ 1.25 – +++ – ++ 0.50 – ++ – ++ 0.63 – ++ – ++ 0.25 – ± – ± 0.31 – + – ± 0.13 – – – – 0.16 – ± – – 0.06 – – – – 0.00 – – – – 0.00 – – – – a Representative results of two independent experiments . Discussion We summarized the results of the experiments in Table 6 . We observed that molecular determinants of the unique features of IgG3 are present in the CH2 domain. However, the modifications of CH2 differently affected the features suggesting that their molecular bases are different. Table 6 Summary of experimental results. IgG3 feature/function Influence by the CH2 domain Presence Net charge a Glycosylation Hemagglutination Strong enhancement No effect Enhancement Functional affinity to polyvalent antigens Strong enhancement Weak to moderate effect No effect Oligomerization in solution Dependence No effect No effect Activation of complement cascade Dependence No effect Dependence b a Associated with the presence of His274, Lys282, Arg315, Lys326 . b Data not shown . The prominent role of the CH2 domain in IgG3 biology was originally reported by Hovenden et al. ( 9 ). The authors investigated highly protective IgG3 antibodies against the capsular antigen of B. anthracis . They generated an IgG3 mutein with CH2 swapped from non-protective IgG2b. The mutein lost protective activity of the parental molecule and had reduced affinity to the antigen. In contrast to the work of Hovenden et al., we generated, for the first time, an antibody mutein that gained the unique properties of IgG3. We swapped IgG3-derived CH2 into IgG1, and the obtained molecule (IgG1_CH2-3) had properties typical for IgG3—it agglutinated erythrocytes, oligomerized, had increased functional affinity to a polyvalent antigen, and activated the complement cascade. Thus, we proved that these unique features of mouse IgG3 could be transferred into a new antibody framework. The mechanism of IgG3-dependent hemagglutination is still not completely understood. We previously reported that F(ab') 2 of IgG3 is sufficient to agglutinate erythrocytes ( 11 ). Here, we show that the presence of the CH2 domain in the IgG3 molecule profoundly diminishes the antibody concentration required for the F(ab') 2 -mediated process. Moreover, the introduction of IgG3-derived CH2 into IgG1 framework resulted in the IgG1_CH2-3 mutein that agglutinates erythrocytes. The results indicate that efficient hemagglutination is triggered only by the antibodies equipped with the IgG3-derived CH2 domain. The CH2 domain of IgG3 is positively charged at neutral pH. In contrast, the CH2 domains of other IgG subclasses are negatively charged under the same condition. Considering that erythrocyte surface has a strong negative charge and high zeta potential, it was likely that a positive charge of the IgG3-derived CH2 domain reduces the zeta potential and as a consequence enhances hemagglutination. Unexpectedly, net charge modification of the CH2 domains in IgG1 and IgG3 did not change hemagglutination potential of these isotypes, and we had to reject the hypothesis linking the CH2 net charge with the efficiency of hemagglutination. Alternatively, antibody oligomerization may explain hemagglutination enhancement by the CH2 domain of IgG3. We showed that this domain solely determined antibody oligomerization in solution and thus most probably also on a multi-epitope surface. It is possible that oligomerization between antibodies bound to separate erythrocytes occurs parallel to a sensitization phase of hemagglutination. Thus, antibody oligomerization may lead to the formation of zipper-like structures that stabilize cell aggregates and increase a hemagglutination score. Moreover, the CH2 domain of IgG3 increased functional affinity of an antibody to erythrocyte surface. Thus, hemagglutination enhancement may at least partially depend on the increased affinity. However, the observed enhancement of hemagglutination by the CH2 domain of IgG3 was affected by enzymatic deglycosylation. In contrast, oligomerization in solution and increased functional affinity to polyvalent antigen were independent of CH2 glycosylation. This difference indicates that antibody oligomerization does not fully account for the CH2 domain-mediated enhancement of hemagglutination. Mouse IgG3 has a putative site of N-glycosylation in its CH3 domain on Asn471. Panka reported that the mutation of this Asn residue into Ser diminished the self-association of IgG3 ( 17 ). This finding was later contradicted by Kuroki et al., who provided evidence that this putative N-glycosylation site in the CH3 domain is not occupied and the mutation Asn471Thr does not influence IgG3 self-association or cryoglobulin activity ( 18 ). Our observations are in line with the findings of Kuroki et al. We did not observe any differences between oligomerization of IgG3 and its enzymatically deglycosylated variant. It is important to note that we and Kuroki et al. used PEG-precipitation for oligomerization analyses. Panka used different methods, ELISA and native electrophoresis, which may account for the discrepancies. Greenspan et al. showed that Fc-dependent oligomerization increases functional affinity of IgG3 to polyvalent antigens ( 5 ). Our results confirm that finding, but we showed that the relation between oligomerization and increased functional affinity is more complex than previously thought. First, functional affinity of IgG3 was influenced not only by Fc region (CH2 and CH3 domains) but also by the CH1 domain. Second, functional affinity to the polyvalent antigen (B antigen) was modulated by the CH2 charge. In contrast, oligomerization in solution required only the presence of the CH2 domain of IgG3 and was insensitive to the introduced charge modifications. The results showed that the mechanism behind high functional affinity may depend on more factors than oligomerization in solution does. The observed influence of the CH1 domain on functional affinity is difficult to explain. The CH1 domain of IgG3 has a more positive net charge than the CH1 domain of IgG1 ( 9 ). It is likely that the net charge of the CH1 domain influences the binding of the domain muteins to erythrocytes, which have a strong negative charge. However, the IgG3-derived CH1 domain also enhanced the binding of IgG1_CH1-3 to a surface with the immobilized trisaccharide B-BSA conjugate. Thus, the results support previous observations ( 23 ) that the CH1 domain may influence a variable domain and a paratope of an antibody. According to the general view, the Fab and Fc fragments are independent parts of an antibody ( 24 ). However, our results demonstrate that the Fc, particularly its CH2 domain, may influence Fab-mediated antigen binding. There are two possible mechanisms of this phenomenon—intramolecular signaling ( 25 ) [called by some authors as an intramolecular allostery ( 16 )] or intermolecular cooperativity. There are several examples of intramolecular signaling observed by different authors investigating how the isotype switching changes an antibody affinity to its antigen [reviewed in Ref. ( 16 , 26 )]. The effects of the CH1 domains or Fc fragments on variable regions are well documented, but considered a rather unique phenomenon ( 16 ). It is more likely that the increased affinity of IgG3 to its antigen results from cooperativity of its CH2 domains. Within this domain, a specific site of self-association may be present, which governs oligomerization of an antibody and pre-determines the increased affinity to multivalent antigens. However, we cannot exclude other scenarios—the involvement of both the CH2 and CH3 domains in IgG3 intermolecular interactions or even sole CH3-CH3 interactions, assuming that the CH2 domains influence the whole molecule structure and promote reciprocal interactions of the CH3 domains of neighboring molecules. Other factors, e.g., influence of the CH1 domain on a paratope, properties of an antigen (charge), spatial distribution of epitopes, intermolecular forces between epitope and paratope, or a variable domain framework may further modulate functional affinity of IgG3 upon multivalent antigen binding. Diebolder at al. described recently an interesting example of Fc-dependent antibody oligomerization. Analyses of antibody binding to DNP-labeled liposomes (a multivalent antigen) revealed that human IgG may form hexamers through non-covalent interactions between their constant regions ( 27 ). Several mutations that enhance these interactions and subsequent complement activation were reported ( 27 ). The Fc-interactions promoting antibody hexamerization did not change affinity to the cognate antigen. Thus, this phenomenon seems to be different from IgG3 oligomerization, and it is still an open question whether mouse antibodies are able to form such hexamers. Currently, no structure is available for a full-length mouse IgG3 or its Fc fragment. We performed some analyses using a molecular model of IgG3 obtained by comparative modeling, but its resolution is not sufficient for in-depth studies. IgG3 crystallization might provide a direct insight into the mechanism of its oligomerization, as was in the case of human IgG1 hexamerization described in the cited work ( 27 ). Complement cascade activation, as an effector function of antibodies, constitutes a first-line of defense against microbial infections. As the cascade progresses, components of the complement are deposited on a pathogen surface and act as opsonins for phagocytic cells. Moreover, the complement lyses invading pathogens by forming membrane attacking complex. We confirmed that C1q-binding motifs, known from human IgG1, are functional in the mouse IgG3 framework. On the other hand, we did not observe complement activation by mouse IgG1 equipped with the motifs. The results indicate that the presence of the known C1q-binding motifs is not sufficient for complement activation by an antibody. The motifs must be surrounded by a permissive framework, provided e.g., by human IgG1 or mouse IgG3. Our work suggests that a novel type of monoclonal antibodies may be generated by replacing the CH2 domain of a human antibody with the homologs fragment of mouse IgG3. Human IgG1 subclass is the most feasible target framework for generation of such IgG3-inspired hybrid mouse/human molecule ( 28 ). Our observation indicates that the generated hybrid antibody should preserve the ability to activate complement and may have increased affinity to polyvalent antigens. Since the mouse IgG3 subclass is highly protective against several life-threatening microbial infections, the hybrid molecule may be very useful in preventing or fighting lethal pathogens. However, the hybrid antibody with the mouse CH2 may be immunogenic. To decrease the risk of an unwanted immune response, the mouse component should be reduced to a minimum. Thus, the properties of the CH2 domain derived from mouse IgG3 should be further investigated and efforts should be made especially to identify fragments of this domain that determines its properties. Author Contributions TK conceived and did all experiments. TK and JB analyzed and discussed the results. The manuscript was written by TK and JB. The authors accepted the final version of the manuscript. Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Supplementary Material The Supplementary Material for this article can be found online at https://www.frontiersin.org/articles/10.3389/fimmu.2018.01096/full#supplementary-material . Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7859429/

Development of a Chimeric Vaccine Against Pseudomonas aeruginosa Based on the Th17-Stimulating Epitopes of PcrV and AmpC

Pulmonary infection caused by Pseudomonas aeruginosa (PA) has created an urgent need for an efficient vaccine, but the protection induced by current candidates is limited, partially because of the high variability of the PA genome. Antigens targeting pulmonary Th17 responses are able to provide antibody-independent and broad-spectrum protection; however, little information about Th17-stimulating antigens in PA is available. Herein, we identified two novel PA antigens that effectively induce Th17-dependent protection, namely, PcrV (PA1706) and AmpC (PA4110). Compared to intramuscular immunization, intranasal immunization enhanced the protection of rePcrV due to activation of a Th17 response. The Th17-stimulating epitopes of PcrV and AmpC were identified, and the recombinant protein PVAC was designed and generated by combining these Th17-stimulating epitopes. PVAC was successfully produced in soluble form and elicited broad protective immunity against PA. Our results provide an alternative strategy for the development of Th17-based vaccines against PA and other pathogens. Introduction Respiratory infections caused by Pseudomonas aeruginosa (PA) are a major health problem globally. Recently, with the emergence of multidrug resistant (MDR) PA strains, the clinical treatment of PA pneumonia has faced enormous challenges, especially for patients with cystic fibrosis (CF), non-CF bronchiectasis (nCFB), or chronic obstructive pulmonary disease (COPD), patients who are undergoing mechanical ventilation and patients with other pulmonary disorders ( 1 , 2 ). To date, most PA vaccine candidates have been developed from LPS O antigens ( 3 ), polysaccharide-protein conjugates, outer membrane proteins ( 4 ) and the type III secretion system component PcrV ( 5 ). Despite substantial research efforts over the past fifty years, a vaccine licensed for clinical use has not yet been approved, and several challenges remain to be addressed. One urgent problem to be solved is that the protective effects of most PA vaccines rely mainly on antibody-mediated opsonophagocytic killing and/or inhibition of toxicity. Phase II clinical trials of the PA vaccine IC43, a fused protein (OprF/I)-based vaccine, showed that the vaccine caused a significant increase in antibody titers in volunteers. However, the infection rate was not significantly decreased ( 4 ). Furthermore, although LPS O antigen-based vaccines can mediate high levels of immunity to PA, the protection is limited to strains that have specific LPS serogroups ( 3 ). It is increasingly clear that the activation of strong opsonophagocytic antibodies alone is insufficient for a successful vaccine against PA. The contribution of Th17 immunity to the prevention of infection by pulmonary pathogens has gradually been recognized ( 6 ). Pulmonary Th17 cells participate in the recruitment of neutrophils, the release of antimicrobial peptides, IL-17–driven Th1 immunity and so on ( 7 ). These effectors provide immunity against a wide range of pathogens through the antibody-independent pathway ( 7 ). Therefore, inducing protective Th17 responses is of great importance when designing a vaccine against a pulmonary pathogen. Vaccine candidates that target Th17 responses have been identified and tested to protect against Yersinia pestis ( 8 ), Mycobacterium tuberculosis ( 9 ), Bordetella pertussis ( 10 ), Streptococcus pneumoniae ( 11 ), Candida albicans ( 12 ), and Staphylococcus aureus ( 13 ). The Th17 response is also protective against pulmonary PA infection ( 14 ). To date, the number of Th17 cells and the concentration of IL-17A in the lungs have been shown to increase significantly as soon as 4 h after PA infection. Blocking the Th17 response leads to more severe pathological damage in the lung ( 14 ). A live attenuated PA vaccine was able to elicit serotype-independent protection in the mouse pneumonia model, and this protection relied not on specific antibodies but on Th17 responses ( 15 ). In an in vitro transcription system, a recombinant PcrH-PopB protein was found to stimulate the Th17 response and confer protection in mice ( 16 , 17 ). In a previous study, we generated an optimized Th17-stimulating antigen, OprL, and investigated its serotype-independent protection ( 18 ). However, current Th17-stimulating vaccine candidates are difficult to apply in industrial settings because of potential safety risks, limited protection, and complicated production processes. In this study, we produced ten soluble PA antigens in E. coli and identified two of them, rePcrV and reAmpC, that were able to induce effective protective Th17-dependent protection. Then, we screened the Th17-stimulating epitopes of rePcrV and reAmpC and generated a novel recombinant protein, PVAC, by rationally fusing the Th17-stimulating epitopes. Epitope-based PVAC-elicited immunity and protection were measured in mice. Materials and Methods Mice and Strains Specific pathogen-free female C57BL/6 mice (six- to eight-week-old) were purchased from Beijing HFK Bioscience Limited Company (Beijing, People's Republic of China). IL-17A gene knockout (IL-17A KO) mice (on a C57BL/6 background) were kindly provided by Richard A. Flavell (Yale University School of Medicine, New Haven, CT, USA). The mice were maintained under barrier conditions in a biohazard animal room. PA XN-1 (CCTCC M2015730) was isolated in Southwest Hospital in Chongqing, China, and deposited at the CCTCC (China Center for Type Culture Collection); its serotype is I. PA 464 and PA 451 were also isolated from Southwest Hospital. Their serotypes are F and A, respectively. PA ZNJ004 was isolated from the No.422 Hospital of the Chinese People's Liberation Army, and its serotype is J. Production of Candidate Antigens in E. coli To identify a possible soluble fragment of the target protein, its conserved domains were analyzed using the BLAST search tool ( http://blast.ncbi.nlm.nih.gov/Blast.cgi ). The nucleotide sequence codon-optimized for expression in E. coli was synthesized and cloned into pGEX-6p-1 (GE Healthcare) using the BamH I and Xho I restriction sites for expression in E. coli as a GST-fusion protein. The sequence of the recombinant plasmid was confirmed by DNA sequencing. Information on the candidate antigens is shown in Table S1 . Then, protein expression in E. coli BL21 was induced by adding 0.3 mM IPTG to 1 L of cell culture in LB medium when the OD600 reached ~0.5. The cell culture was then incubated overnight with shaking at 150 rpm at 16°C. The cells were collected by centrifugation at 5,000 rpm for 20 min at 4°C and lysed by sonication in lysis buffer (20 mM phosphate buffer, pH 7.0, 150 mM NaCl) on ice. Then, the cell lysate was centrifuged at 12,000 rpm for 30 min, and the supernatant was collected and passed through a column of glutathione resin (NEB). After extensive washing with five-fold column volumes of washing buffer (20 mM phosphate buffer, pH 7.0, 1 M NaCl), PreScission Protease (GE Healthcare) was added and incubated overnight at 4°C. The recombinant proteins were then eluted from the glutathione resin, and protein impurities were further removed based on the different properties of the target proteins. The eluted proteins were finally concentrated to 4 to 6 mg/ml in buffer (PBS, pH 7.2) and identified by SDS-PAGE. The Keyhole limpet hemocyanin (KLH) was purchased from Genscript Company in Nanjing, China. The Endotoxin Cap Bestarose 4FF resin (Bestchrom) was used to remove the endotoxin in KLH and recombinant proteins in this study. The concentration of endotoxin in each protein was less than 10 EU/ml when determined by Limulus Amebocyte Lysate (LAL) quantitation kit (Pierce). Finally, these proteins were stored at −80°C for future assays. Immunization of Mice and Murine Pneumonia Model In short, to evaluate the immune response, mice were divided equally into different groups and anesthetized with 0.12 ml of pentobarbital sodium (10 mg/ml) in 0.9% saline injected intraperitoneally (i.p.), then immunized intranasally with 20 μl of curdlan (10 mg/ml) and the corresponding antigen(25 μg/mouse) plus curdlan (10 mg/ml). The mouse immunization time points are 0, 14 and 21 days. Challenge or take tissue samples 14 days after the last immunization. Moreover, IL-17A knockout mice and wild-type (WT) mice were immunized to evaluate IL-17–associated protection. In another experiment, mice were injected i.p. with 2 mg of isotype control (RTK2071, Biolegend) or anti–IL-17 (TC11-18H10.1, Biolegend) antibody 2 days prior to challenge. For acute pneumonia challenge, the mice were anesthetized with pentobarbital sodium followed by the intratracheal injection of PA. The lethal doses of PA 464, PA XN-1, PA ZNJ004, and PA 451 were 2.5 × 10 7 , 1.0 × 10 7 , 3.5 × 10 6 , and 1.0 × 10 7 CFU per mouse, respectively. The number of deaths caused by lethal infection was recorded every 12 h during a 7-day observation period post-challenge. In addition, the other mice were infected with a sublethal dose to investigate the protective mechanism. The sublethal doses of PA 464, PA XN-1, PA ZNJ004, and PA 451 were 3.5 × 10 6 , 1.3 × 10 6 , 5.8 × 10 5 , and 1.4 × 10 6 CFU, respectively. The general information on these PA strains were included in Table S2 . Isolation of Lung Mononuclear Cells and Intracellular Flow Cytometry Analysis Lung tissues were chopped 14 days after the last immunization, and then digested with collagenase D (1 mg/ml, Gibco) and DNase I (10 mg/ml, Sigma-Aldrich) for 1 h at 37°C with agitation. Next, the lungs were passed through a 75 μm cell strainer to obtain a single-cell suspension. Then, the erythrocytes were removed by density gradient centrifugation with Percoll. Finally, lung mononuclear cells were obtained. After labeling with CFSE, the lymphocytes were incubated with the corresponding protein (0.5 μM) or Keyhole limpet hemocyanin (KLH), and stimulated with 5 U/ml IL-2 (PeproTech, Rocky Hill, NJ, USA) in complete RPMI-1640 medium. Half of the medium was removed when it turned yellow after incubation for about 2 days and then replaced with fresh medium. Then, the cells were stimulated with leukocyte activation cocktail with GolgiPlug for 4 h. The cell samples were washed and adjusted to a concentration of 5 × 10 5 cells/ml with flow cytometry buffer (0.5% FBS in PBS). The samples were first stained for surface markers, followed by intracellular cytokine staining. For surface marker staining, the cells were incubated with the fluorochrome-conjugated monoclonal antibodies APC/Cyanine7 anti-mouse CD3 (17A2) and PerCP/Cyanine5.5 anti-mouse CD4 (GK1.5). For intracellular cytokine staining, the cells were processed with the Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences) according to the manufacturer's instructions. The samples were then incubated with fluorochrome-conjugated antibodies against APC anti-mouse IL-17A (TC11-18H10.1), PE anti-mouse IFN-γ (XMG1.2), and PE/Cy7 anti-mouse IL-4 (11B11). The above antibodies were all from Biolegend. Next, the samples were analyzed using BD FACSArray software™ on a BD FACSArray flow cytometer (BD Biosciences). ELISA The levels of IgG were measured in serum samples collected 14 days after the last immunization. The 96-well ELISA plates were precoated with 100 μl of recombinant protein at 2 μg/ml overnight at 4°C. A total of 100 μl of serial two-fold dilutions of serum from each group of immunized mice was added to each well and incubated for 1 h at 37°C. The bound antibody was detected with goat anti-mouse IgG-HRP (Life Technology), and 100 μl of tetramethylbenzidine (Beijing ZSGB-BIO) substrate was added to each well to develop the color. The optical density at 450 nm (OD450) was measured. A well was considered positive when its OD450 was at least 2.1-fold higher than that of the negative control. Opsonophagocytic Killing Assay First, HL-60 cells (ATCC, CCL-240) were differentiated into granulocyte-like cells in medium containing 100 mM N'N-dimethyl formamide for 5 days. Serum samples collected 14 days after the last immunization from immunized mice, and then were heat-inactivated (56°C, 30 min) and serially diluted with opsonization buffer (a mixture of 80 ml of sterile water, 10 ml of 10× Hank's balanced salt solution, 10 ml of 1% gelatin, and 5.3 ml of fetal bovine serum). In 96-well plates, each well received 40 μl containing 4 × 10 5 HL-60 cells, 10 3 CFUs of XN-1 in 10 μl of opsonophagocytic buffer, 20 μl of serum, and 10 μl of 1% infant rabbit serum as a complement source. After incubation for 2 h, 10 μl of each sample was plated onto agar medium. The opsonophagocytic killing effect was defined as the reduction in CFUs after overnight incubation. Evaluation of Bacterial Load and Inflammation The lungs of mice 24 h after challenge with a sublethal dose were collected, weighed, and homogenized in 1 ml of sterilized PBS. The homogenates were then plated on PIA plates and cultured at 37°C overnight. Then, the numbers of CFUs per gram of tissue were calculated from each plate. In addition, the concentrations of proinflammatory cytokines, such as IL-1β and TNF-α, in the supernatant were quantified by a Mouse Cytokine Quantification ELISA Test Kit (1210122, 1217202, Dakewei). The protocol was performed following the manufacturer's (Dakewei) instructions. For histologic examination, the lungs were collected and fixed in neutral 10% formalin, embedded in paraffin, sectioned and stained with hematoxylin and eosin. The sections were observed at 100-fold magnification via light microscopy. Each lung section received a pathology score of 0 to 10 (from normal to severe) by a single pathologist according to the hemorrhage, edema, hyperemia, alveolar structure and neutrophil infiltration. Adoptive Transfer Experiments Lung mononuclear cells were pooled from 15 immunized mice or 15 naïve mice and prepared as described above. The CD4 + T cells were sorted by negative selection (STEMCELL Technologies). A total of 5 × 10 6 CD4+ T cells were transferred via intravenous injection. The mice were then challenged with PA XN-1 24 h after T cell transfer. Twenty-four hours after infection, pulmonary bacterial colonization and inflammation were detected as described above. Enzyme-Linked Immunospot (ELISPOT) Assay A mouse IL-17A ELISpot PLUS kit (3521-4HPW-2, MabTech) was applied to measure peptide-specific IL-17-producing cells. In brief, lung mononuclear cells from immunized mice were prepared as described previously. Single-cell suspensions were added to the plate at a starting concentration of 1.5 × 10 6 cells/well. The cells were cultured with 10 μg/ml peptide for 48 h and then tested according to the manufacturer's instructions. The fold increase in the number of spot-forming cells (SFCs) was calculated with the following formula: Fold increase = (N immunuzed − N unimmunized )/N unimmunized . N immunuzed indicates the number of SFCs among lymphocytes from immunized mice, while N unimmunized indicates the number of SFCs among lymphocytes from unimmunized mice. Statistical Analysis The data are presented as the mean ± SE. The significance of differences was determined by an unpaired parametric test (Student's t-test for two groups or one-way ANOVA for more than three groups). Bacterial burden was analyzed by the nonparametric Mann-Whitney test. The survival rate was analyzed by Kaplan-Meier survival curves. For pairwise comparisons among 3 or more groups, p values were adjusted by using Tukey's test. SPSS15.0 and GraphPad Prism 8.0 were used for data analysis. Significance was accepted when P < 0.05. Mice and Strains Specific pathogen-free female C57BL/6 mice (six- to eight-week-old) were purchased from Beijing HFK Bioscience Limited Company (Beijing, People's Republic of China). IL-17A gene knockout (IL-17A KO) mice (on a C57BL/6 background) were kindly provided by Richard A. Flavell (Yale University School of Medicine, New Haven, CT, USA). The mice were maintained under barrier conditions in a biohazard animal room. PA XN-1 (CCTCC M2015730) was isolated in Southwest Hospital in Chongqing, China, and deposited at the CCTCC (China Center for Type Culture Collection); its serotype is I. PA 464 and PA 451 were also isolated from Southwest Hospital. Their serotypes are F and A, respectively. PA ZNJ004 was isolated from the No.422 Hospital of the Chinese People's Liberation Army, and its serotype is J. Production of Candidate Antigens in E. coli To identify a possible soluble fragment of the target protein, its conserved domains were analyzed using the BLAST search tool ( http://blast.ncbi.nlm.nih.gov/Blast.cgi ). The nucleotide sequence codon-optimized for expression in E. coli was synthesized and cloned into pGEX-6p-1 (GE Healthcare) using the BamH I and Xho I restriction sites for expression in E. coli as a GST-fusion protein. The sequence of the recombinant plasmid was confirmed by DNA sequencing. Information on the candidate antigens is shown in Table S1 . Then, protein expression in E. coli BL21 was induced by adding 0.3 mM IPTG to 1 L of cell culture in LB medium when the OD600 reached ~0.5. The cell culture was then incubated overnight with shaking at 150 rpm at 16°C. The cells were collected by centrifugation at 5,000 rpm for 20 min at 4°C and lysed by sonication in lysis buffer (20 mM phosphate buffer, pH 7.0, 150 mM NaCl) on ice. Then, the cell lysate was centrifuged at 12,000 rpm for 30 min, and the supernatant was collected and passed through a column of glutathione resin (NEB). After extensive washing with five-fold column volumes of washing buffer (20 mM phosphate buffer, pH 7.0, 1 M NaCl), PreScission Protease (GE Healthcare) was added and incubated overnight at 4°C. The recombinant proteins were then eluted from the glutathione resin, and protein impurities were further removed based on the different properties of the target proteins. The eluted proteins were finally concentrated to 4 to 6 mg/ml in buffer (PBS, pH 7.2) and identified by SDS-PAGE. The Keyhole limpet hemocyanin (KLH) was purchased from Genscript Company in Nanjing, China. The Endotoxin Cap Bestarose 4FF resin (Bestchrom) was used to remove the endotoxin in KLH and recombinant proteins in this study. The concentration of endotoxin in each protein was less than 10 EU/ml when determined by Limulus Amebocyte Lysate (LAL) quantitation kit (Pierce). Finally, these proteins were stored at −80°C for future assays. Immunization of Mice and Murine Pneumonia Model In short, to evaluate the immune response, mice were divided equally into different groups and anesthetized with 0.12 ml of pentobarbital sodium (10 mg/ml) in 0.9% saline injected intraperitoneally (i.p.), then immunized intranasally with 20 μl of curdlan (10 mg/ml) and the corresponding antigen(25 μg/mouse) plus curdlan (10 mg/ml). The mouse immunization time points are 0, 14 and 21 days. Challenge or take tissue samples 14 days after the last immunization. Moreover, IL-17A knockout mice and wild-type (WT) mice were immunized to evaluate IL-17–associated protection. In another experiment, mice were injected i.p. with 2 mg of isotype control (RTK2071, Biolegend) or anti–IL-17 (TC11-18H10.1, Biolegend) antibody 2 days prior to challenge. For acute pneumonia challenge, the mice were anesthetized with pentobarbital sodium followed by the intratracheal injection of PA. The lethal doses of PA 464, PA XN-1, PA ZNJ004, and PA 451 were 2.5 × 10 7 , 1.0 × 10 7 , 3.5 × 10 6 , and 1.0 × 10 7 CFU per mouse, respectively. The number of deaths caused by lethal infection was recorded every 12 h during a 7-day observation period post-challenge. In addition, the other mice were infected with a sublethal dose to investigate the protective mechanism. The sublethal doses of PA 464, PA XN-1, PA ZNJ004, and PA 451 were 3.5 × 10 6 , 1.3 × 10 6 , 5.8 × 10 5 , and 1.4 × 10 6 CFU, respectively. The general information on these PA strains were included in Table S2 . Isolation of Lung Mononuclear Cells and Intracellular Flow Cytometry Analysis Lung tissues were chopped 14 days after the last immunization, and then digested with collagenase D (1 mg/ml, Gibco) and DNase I (10 mg/ml, Sigma-Aldrich) for 1 h at 37°C with agitation. Next, the lungs were passed through a 75 μm cell strainer to obtain a single-cell suspension. Then, the erythrocytes were removed by density gradient centrifugation with Percoll. Finally, lung mononuclear cells were obtained. After labeling with CFSE, the lymphocytes were incubated with the corresponding protein (0.5 μM) or Keyhole limpet hemocyanin (KLH), and stimulated with 5 U/ml IL-2 (PeproTech, Rocky Hill, NJ, USA) in complete RPMI-1640 medium. Half of the medium was removed when it turned yellow after incubation for about 2 days and then replaced with fresh medium. Then, the cells were stimulated with leukocyte activation cocktail with GolgiPlug for 4 h. The cell samples were washed and adjusted to a concentration of 5 × 10 5 cells/ml with flow cytometry buffer (0.5% FBS in PBS). The samples were first stained for surface markers, followed by intracellular cytokine staining. For surface marker staining, the cells were incubated with the fluorochrome-conjugated monoclonal antibodies APC/Cyanine7 anti-mouse CD3 (17A2) and PerCP/Cyanine5.5 anti-mouse CD4 (GK1.5). For intracellular cytokine staining, the cells were processed with the Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences) according to the manufacturer's instructions. The samples were then incubated with fluorochrome-conjugated antibodies against APC anti-mouse IL-17A (TC11-18H10.1), PE anti-mouse IFN-γ (XMG1.2), and PE/Cy7 anti-mouse IL-4 (11B11). The above antibodies were all from Biolegend. Next, the samples were analyzed using BD FACSArray software™ on a BD FACSArray flow cytometer (BD Biosciences). ELISA The levels of IgG were measured in serum samples collected 14 days after the last immunization. The 96-well ELISA plates were precoated with 100 μl of recombinant protein at 2 μg/ml overnight at 4°C. A total of 100 μl of serial two-fold dilutions of serum from each group of immunized mice was added to each well and incubated for 1 h at 37°C. The bound antibody was detected with goat anti-mouse IgG-HRP (Life Technology), and 100 μl of tetramethylbenzidine (Beijing ZSGB-BIO) substrate was added to each well to develop the color. The optical density at 450 nm (OD450) was measured. A well was considered positive when its OD450 was at least 2.1-fold higher than that of the negative control. Opsonophagocytic Killing Assay First, HL-60 cells (ATCC, CCL-240) were differentiated into granulocyte-like cells in medium containing 100 mM N'N-dimethyl formamide for 5 days. Serum samples collected 14 days after the last immunization from immunized mice, and then were heat-inactivated (56°C, 30 min) and serially diluted with opsonization buffer (a mixture of 80 ml of sterile water, 10 ml of 10× Hank's balanced salt solution, 10 ml of 1% gelatin, and 5.3 ml of fetal bovine serum). In 96-well plates, each well received 40 μl containing 4 × 10 5 HL-60 cells, 10 3 CFUs of XN-1 in 10 μl of opsonophagocytic buffer, 20 μl of serum, and 10 μl of 1% infant rabbit serum as a complement source. After incubation for 2 h, 10 μl of each sample was plated onto agar medium. The opsonophagocytic killing effect was defined as the reduction in CFUs after overnight incubation. Evaluation of Bacterial Load and Inflammation The lungs of mice 24 h after challenge with a sublethal dose were collected, weighed, and homogenized in 1 ml of sterilized PBS. The homogenates were then plated on PIA plates and cultured at 37°C overnight. Then, the numbers of CFUs per gram of tissue were calculated from each plate. In addition, the concentrations of proinflammatory cytokines, such as IL-1β and TNF-α, in the supernatant were quantified by a Mouse Cytokine Quantification ELISA Test Kit (1210122, 1217202, Dakewei). The protocol was performed following the manufacturer's (Dakewei) instructions. For histologic examination, the lungs were collected and fixed in neutral 10% formalin, embedded in paraffin, sectioned and stained with hematoxylin and eosin. The sections were observed at 100-fold magnification via light microscopy. Each lung section received a pathology score of 0 to 10 (from normal to severe) by a single pathologist according to the hemorrhage, edema, hyperemia, alveolar structure and neutrophil infiltration. Adoptive Transfer Experiments Lung mononuclear cells were pooled from 15 immunized mice or 15 naïve mice and prepared as described above. The CD4 + T cells were sorted by negative selection (STEMCELL Technologies). A total of 5 × 10 6 CD4+ T cells were transferred via intravenous injection. The mice were then challenged with PA XN-1 24 h after T cell transfer. Twenty-four hours after infection, pulmonary bacterial colonization and inflammation were detected as described above. Enzyme-Linked Immunospot (ELISPOT) Assay A mouse IL-17A ELISpot PLUS kit (3521-4HPW-2, MabTech) was applied to measure peptide-specific IL-17-producing cells. In brief, lung mononuclear cells from immunized mice were prepared as described previously. Single-cell suspensions were added to the plate at a starting concentration of 1.5 × 10 6 cells/well. The cells were cultured with 10 μg/ml peptide for 48 h and then tested according to the manufacturer's instructions. The fold increase in the number of spot-forming cells (SFCs) was calculated with the following formula: Fold increase = (N immunuzed − N unimmunized )/N unimmunized . N immunuzed indicates the number of SFCs among lymphocytes from immunized mice, while N unimmunized indicates the number of SFCs among lymphocytes from unimmunized mice. Statistical Analysis The data are presented as the mean ± SE. The significance of differences was determined by an unpaired parametric test (Student's t-test for two groups or one-way ANOVA for more than three groups). Bacterial burden was analyzed by the nonparametric Mann-Whitney test. The survival rate was analyzed by Kaplan-Meier survival curves. For pairwise comparisons among 3 or more groups, p values were adjusted by using Tukey's test. SPSS15.0 and GraphPad Prism 8.0 were used for data analysis. Significance was accepted when P < 0.05. Results rePcrV and reAmpC Were Able to Induce a Th17 Response and Confer Protection in Mice After critical review of the reported protective antigens and especially the data from the proteomic analysis of outer membrane vesicles ( 19 ), we tried to produce these 37 proteins in E.coli ( Table S1 ). However only 10 of them were soluble and stable in PBS buffer, and were applied for further analysis ( Figure S1 ). Mice were intranasally immunized with the ten purified recombinant proteins, and lung lymphocytes were isolated and costimulated with the corresponding antigens or KLH. Then, the percentage of CD4 + IL-17A + T cells was measured to identify the antigens that efficiently induced a Th17 response. Immunization with reOprL was able to induce a Th17 response, which was consistent with previous observations ( Figure 1A ) ( 18 ). In addition, immunization with two recombinant proteins, namely, reAmpC and rePcrV, was capable of eliciting a Th17 response. The percentage of CD4 + IL-17A + T cells in the reAmpC and rePcrV groups was significantly higher than that in the reOprL group ( Figure 1B ). KLH stimulation did not significantly increase the percentage of CD4 + IL-17A + T cells. Next, we tested the protection induced by the ten candidates via intratracheal challenge with a lethal dose of PA XN-1. As shown in Figure 1C , the survival rate of the reOprL group was 50%, which was significantly higher than that of the curdlan ( P = 0.0007) group. The survival rate of the reAmpC-immunized mice was also significantly higher than that of the control (P = 0.0019). No mice in the adjuvant control group survived more than 60 h. Surprisingly, the survival rate of the rePcrV vaccination group was as high as 100%, which was also significantly higher than that of the control ( P <0.0001) group. The above results indicate that intranasal vaccination with reAmpC or rePcrV not only induces a robust Th17 response but also provides protective effects. Figure 1 Screening for protective antigens that can induce a Th17 response. (A) Mice (n = 5) were intranasally immunized with the ten purified recombinant proteins, and lung lymphocytes were isolated and costimulated with the corresponding antigens. The representative dot plots show the frequency of CD4 + IL-17A + T cells among lung lymphocytes from mice immunized with reAmpC, rePcrV, reIcmP, and the adjuvant curdlan. (B) The bar represents the percentage of CD4 + IL-17A + T cells among lung lymphocytes from mice (n = 5) of the ten groups. The top three proteins were reAmpC (5.4%), rePcrV (3.4%) and reOprL (3.1%). (C) The survival of mice (n = 10) immunized with reOprL, reAmpC or rePcrV after challenge with a lethal dose of PA XN-1. All the immunized groups showed significantly better survival than did the control. Compared With Intramuscular Vaccination, Intranasal Vaccination With rePcrV Enhanced Protection Intramuscular (i.m.) immunization with the PcrV protein formulated with Al(OH) 3 has been studied extensively ( 5 , 20 ), we determined whether intranasal (i.n.) immunization could improve the efficiency of immunization. As expected, the survival rate of the i.n. group was significantly higher ( P = 0.0041) than that of the i.m. group ( Figure 2A ). Subsequently, the immunized mice were challenged with a sublethal dose of PA XN-1 to explore the protective mechanism. Compared with the rePcrV i.m. group and the curdlan i.n. control group, the rePcrV i.n. group showed significant reductions in pathology score, which was characterized according to the cell infiltration, hemorrhage, alveolar collapse and tissue damage ( Figure 2B ). Interesting, more infiltrate cells but less pathogenic lesions was observed rePcrV i.n. group. This observation indicates the participation of protective resident memory T cells, which was usually found in intranasal delivered vaccines ( 21 ). Figure 2 Compared with intramuscular vaccination, intranasal vaccination with rePcrV enhanced protection. (A) Mice were immunized with rePcrV intranasally (i.n.) or intramuscularly (i.m.). The bar indicates the survival of mice (n = 10) after intratracheal injection with a lethal dose of PA XN-1. The survival rate of the i.n. group was significantly higher than that of the i.m. group. (B) HE staining of lungs from immunized mice 24 h after infection with a sublethal dose of PA XN-1. Images were captured at 100× magnification. (C) Evaluation of the bacterial load in the lung in PA XN-1−infected mice (n = 4). The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue. (D) Quantitative measurement of TNF-α and IL-1β in the lungs (n = 3). The data are shown as the mean ± SE. (E) The bar represents the titer of anti-rePcrV IgG antibodies in the serum of immunized mice (n = 5). (G) The opsonophagocytic killing activity of the serum from immunized mice. Serum samples from immunized mice were diluted and incubated with PA XN-1. The bar represents the percentage of killed bacteria in a series of dilutions. The data are presented as the mean ± SE. (F) Frequency of CD4 + CD17 + T cells in the lungs of mice two weeks after the last immunization. The bar represents the frequency of the cells in each group. The "*" indicates a significant difference at P < 0.05. The "n.s" means "no significant difference". In addition, the bacterial loads of the rePcrV i.n. group were significantly lower than those of the rePcrV i.m. ( P <0.0001) and curdlan i.n. control groups (P = 0.0002) ( Figure 2C ). Furthermore, compared with the rePcrV i.m. and curdlan i.n. groups, the levels of the proinflammatory cytokines TNF-α and IL-1β in the BALF of the rePcrV i.n. group were significantly lower ( Figure 2D ). These results indicated that the improved protection of i.n. immunization with rePcrV might be due to the decrease in the bacterial load, the secretion of proinflammatory cytokines and the increase in inflammation in the lung. In addition, the immune response following rePcrV vaccination was determined. Surprisingly, the titer of anti-PcrV antibodies in the i.n. group was significantly lower ( P <0.0001) than that in the i.m. group ( Figure 2E ). However, no significant difference in the opsonophagocytic activity of anti-PcrV antibodies was observed between the two immunization routes ( Figure 2F ). These findings led us to measure the pulmonary Th17 response induced by rePcrV. As shown in Figure 2G , the percentage of CD4 + IL-17A + T cells in the lung of the rePcrV i.n. group was significantly higher ( P <0.0001) than that of the rePcrV i.m. group. These results suggested that lung Th17 cells could be responsible for the enhanced protection induced via the intranasal route. The Enhanced Protection Induced by Intranasal Vaccination With rePcrV Depends on the Th17 Response To clarify the contribution of the Th17 response to the increase in protection, we first tested the protection induced in IL-17A KO mice after intranasal vaccination with rePcrV. As few as 30% of immunized IL-17A KO mice survived, which was significantly lower ( P = 0.0012) than the survival rate observed in WT (wild-type) mice ( Figure 3A ). Similar results were observed in immunized WT mice after treatment with anti−IL-17A mAbs. The survival was significantly decreased ( P = 0.0013) in mice injected with IL-17A mAbs compared with mice injected with control mouse IgG ( Figure 3B ). Next, we isolated CD4 + T cells from the lungs of mice immunized intranasally with rePcrV and then adoptively transferred them into naïve mice. The mice were challenged with PA XN-1 to further evaluate CD4 + T cell-mediated protection. We found that bacterial colonization in the lungs of the mice given CD4 + T cells was significantly lower ( P = 0.0025) than that of the mice in the control group ( Figure 3C ). Similarly, pathological observation showed that lesions such as pulmonary hemorrhage and inflammatory cell infiltration were significantly reduced after injection of CD4 + T cells ( Figure 3D ). Finally, we tested the concentration of proinflammatory factors (TNF-α and IL-1β) in the BALF. Similarly, adoptive transfer of CD4 + T cells from the immunized group significantly ( P = 0.0390, P = 0.0060) reduced the concentration compared with that of the control group ( Figure 3E ). These results suggest that the Th17 response is indispensable for the enhanced protection induced by intranasal vaccination with rePcrV. Figure 3 The enhanced protection of intranasal vaccination with rePcrV depends on the Th17 response. (A) WT (wild-type) mice or IL-17A knockout mice (n = 10) were immunized with rePcrV and then intratracheally challenged with a lethal dose of PA XN-1. The bar represents the survival of the mice in each group. (B) WT mice were immunized with rePcrV and administered anti-IL-17A antibodies (n = 10) and with isotype control IgG as a control. The bar represents the survival rate of the two groups at each observation point. (C) CD4+ T cells from the lungs of rePcrV-immunized mice were transferred into naïve WT mice. Then, the mice were challenged with a sublethal dose of PA XN-1. The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue (n = 4). (D) HE staining of the lungs of mice after infection with a sublethal dose of PA XN-1. Images were captured at 100× magnification. (E) Quantitative measurement of TNF-α and IL-1β in the BALF of mice (n = 3). The data are shown as the mean ± SE. The "*" indicates a significant difference at P < 0.05. The Protection Mediated by reAmpC Relies on the Th17 Response First, we measured the opsonophagocytic killing activity of anti-reAmpC antibodies. However, no significant difference in bacteria-killing rates was observed between sera from reAmpC-immunized mice and unimmunized mice, which suggest a limited contribution of reAmpC-specific antibodies to preventive effects ( Figure 4A ). Building on these findings, we next aimed to clarify whether Th17 immunity contributes to protection. As expected, the survival rate of immunized wild-type mice was 50%, while that of immunized IL-17A KO mice was only 10%. All WT mice that were not immunized died within 60 h after the challenge ( Figure 4B ). In another setting, the immunized mice were treated with anti-IL-17A mAbs to block the Th17 response before the bacterial challenge. The results showed that the survival decreased significantly ( P = 0.0003) in mice injected with anti−IL-17A mAbs compared with control mice injected with nonspecific mouse IgG ( Figure 4C ). Additionally, the CD4 + T cells of reAmpC-immunized mice were isolated and adoptively transferred to naïve mice to evaluate their protective effects. The number of colonizing bacteria in the lungs of mice that received CD4 + T cells from the lungs of reAmpC-immunized mice was significantly ( P <0.0001) lower than that in the lungs of unimmunized mice ( Figure 4D ). Moreover, a similar trend was observed for the levels of the proinflammatory factors IL-1β and TNF-α in the BALF ( Figure 4E ). Taken together, these findings showed that the protection induced by reAmpC relies on the pulmonary Th17 response. Figure 4 The protection mediated by reAmpC relies on the Th17 response. (A) The opsonophagocytic killing activity of serum from immunized mice. The bar represents the percentage of killed bacteria in a series of dilutions. The data are presented as the mean ± SE. (B) WT (wild-type) mice or IL-17A knockout mice were immunized with reAmpC (n = 10) and then intratracheally challenged with PA XN-1. The bar represents the survival of mice in each group. (C) WT mice or IL-17A knockout mice (n = 10) were immunized with reAmpC and then intratracheally challenged with a lethal dose of PA XN-1. The bar represents the survival of mice in each group (n = 10). (D) CD4+ T cells from the lungs of reAmpC-immunized mice were transferred into naïve WT mice. Then, the mice were challenged with a sublethal dose of PA XN-1. The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue (n = 4). (E) Quantitative measurement of TNF-α and IL-1β in the BALF of mice (n = 3). The data are shown as the mean ± SE. The "*" indicates a significant difference at P < 0.05. The "n.s" means "no significant difference". Mapping of Epitopes in rePcrV and reAmpC That Stimulate the Th17 Response To screen for epitopes that can stimulate Th17 responses, we isolated lung lymphocytes from immunized mice and cultured them with synthesized peptides from the library of rePcrV and reAmpC ( Tables S3 and S4 ). IL-17−secreting cells were then identified by ELISPOT assay. The distribution of Th17-stimulating epitopes in rePcrV and reAmpC is shown in Figure 5A . The highest fold change in the number of spot-forming cells (SFCs) was clearly observed for Pc5 (LSEAQVLKALAWLLAANP), which had approximately 2.5-fold more SFCs than the unimmunized mice. Furthermore, most of the Th17 stimulation epitopes are located at the N-terminus of rePcrV ( Figure 5B ). In addition, the top 10 most potent epitopes of reAmpC for promoting the secretion of IL-17A are shown in red ( Figure 5A , lower panel). P10 (FTATLAGYALTQDKMRLD) and P41 (AGNSTPMALQPHRIARLP) were the most predominant epitopes in reAmpC, with approximately 2.0- and 1.9-fold more SFCs than the unimmunized mice, respectively. Unlike rePcrV, the epitopes in reAmpC did not accumulate in one region but spanned the whole molecule ( Figure 5B ). Figure 5 Mapping of epitopes in rePcrV and reAmpC that stimulate a Th17 response. (A) The lung lymphocytes of mice immunized with rePcrV (upper) and reAmpC (lower) were isolated and cocultured with the corresponding peptide library. The number of IL-17−secreting cells induced by each peptide was detected by ELISpot. The bars represent the mean and SE of the fold increase in SFCs (spot-forming cells). The top ten peptides are shown in red. (B) Cartoon image of the 3D structure of PcrV (left) and AmpC (right). The model of PcrV was built with the I-TASSER Suite, while the model of AmpC was obtained from the Protein Data Bank (PDB: 4GZB). The images were generated with PyMol 2.4 software. The regions corresponding to the top ten peptides are shown in red. The Th17-stimulating epitopes in rePcrV accumulated at its N-terminus, while those of reAmpC spanned the whole protein. Vaccination With PVAC Protects Against Four PA Strains of Different Serotypes Building on these findings, we hypothesized that a combination of Th17-stimulating epitopes could induce more robust protection. After several attempts, we successfully produced the soluble recombinant protein PVAC, which was composed of the N-terminus of rePcrV (27Glu-126Gly) connected to full-length reAmpC with the flexible linker GSGGSG ( Figure S3 ). Intranasal immunization with PVAC could also induce a humoral immune response, as indicated by an elevated titer of PVAC-specific antibodies in the serum ( Figure 6A ). More importantly, the number of CD4 + IL17 + T cells in the lung increased dramatically after immunization, which suggested a potent Th17 response ( Figure 6B ). Next, to determine whether PVAC could provide broad protection, immunized mice were challenged with clinical PA isolates of different serotypes: PA 464, PA XN-1, PA ZNJ004, and PA 451. The serotypes of these four strains were F, I, J and A, respectively. The survival rate of immunized mice was 70%, 60%, 100%, and 50% after challenge with PA 464, PA XN-1, PA ZNJ004, and PA 451, respectively ( Figure 6C ). The survival rates of the four immunized groups were significantly higher than that of the control group, which indicates that broad protection was induced by PVAC. Collectively, these results show that the fusion protein PVAC was able to effectively induce a humoral immune response and a Th17 response, which proved that broad protection was induced in the mouse pneumonia model. Figure 6 Vaccination with PVAC protects against PA strains of four different serotypes. (A) PVAC, which is composed of the N-terminus of rePcrV (27Glu-126Gly) connected to full-length reAmpC, was produced and used to immunize mice. The bar represents the titer of anti-PVAC IgG in the serum of immunized mice (n = 5) (B) The bar represents the frequency of CD4 + CD17 + T cells in the lungs of immunized mice. (C) PVAC-immunized mice were challenged via intratracheal injection of PA-464, PA XN-18 PA ZNJ004, and PA-451. The survival rates of the four immunized groups were significantly higher than that of the control group. The "*" indicates a significant difference at P < 0.05. rePcrV and reAmpC Were Able to Induce a Th17 Response and Confer Protection in Mice After critical review of the reported protective antigens and especially the data from the proteomic analysis of outer membrane vesicles ( 19 ), we tried to produce these 37 proteins in E.coli ( Table S1 ). However only 10 of them were soluble and stable in PBS buffer, and were applied for further analysis ( Figure S1 ). Mice were intranasally immunized with the ten purified recombinant proteins, and lung lymphocytes were isolated and costimulated with the corresponding antigens or KLH. Then, the percentage of CD4 + IL-17A + T cells was measured to identify the antigens that efficiently induced a Th17 response. Immunization with reOprL was able to induce a Th17 response, which was consistent with previous observations ( Figure 1A ) ( 18 ). In addition, immunization with two recombinant proteins, namely, reAmpC and rePcrV, was capable of eliciting a Th17 response. The percentage of CD4 + IL-17A + T cells in the reAmpC and rePcrV groups was significantly higher than that in the reOprL group ( Figure 1B ). KLH stimulation did not significantly increase the percentage of CD4 + IL-17A + T cells. Next, we tested the protection induced by the ten candidates via intratracheal challenge with a lethal dose of PA XN-1. As shown in Figure 1C , the survival rate of the reOprL group was 50%, which was significantly higher than that of the curdlan ( P = 0.0007) group. The survival rate of the reAmpC-immunized mice was also significantly higher than that of the control (P = 0.0019). No mice in the adjuvant control group survived more than 60 h. Surprisingly, the survival rate of the rePcrV vaccination group was as high as 100%, which was also significantly higher than that of the control ( P <0.0001) group. The above results indicate that intranasal vaccination with reAmpC or rePcrV not only induces a robust Th17 response but also provides protective effects. Figure 1 Screening for protective antigens that can induce a Th17 response. (A) Mice (n = 5) were intranasally immunized with the ten purified recombinant proteins, and lung lymphocytes were isolated and costimulated with the corresponding antigens. The representative dot plots show the frequency of CD4 + IL-17A + T cells among lung lymphocytes from mice immunized with reAmpC, rePcrV, reIcmP, and the adjuvant curdlan. (B) The bar represents the percentage of CD4 + IL-17A + T cells among lung lymphocytes from mice (n = 5) of the ten groups. The top three proteins were reAmpC (5.4%), rePcrV (3.4%) and reOprL (3.1%). (C) The survival of mice (n = 10) immunized with reOprL, reAmpC or rePcrV after challenge with a lethal dose of PA XN-1. All the immunized groups showed significantly better survival than did the control. Compared With Intramuscular Vaccination, Intranasal Vaccination With rePcrV Enhanced Protection Intramuscular (i.m.) immunization with the PcrV protein formulated with Al(OH) 3 has been studied extensively ( 5 , 20 ), we determined whether intranasal (i.n.) immunization could improve the efficiency of immunization. As expected, the survival rate of the i.n. group was significantly higher ( P = 0.0041) than that of the i.m. group ( Figure 2A ). Subsequently, the immunized mice were challenged with a sublethal dose of PA XN-1 to explore the protective mechanism. Compared with the rePcrV i.m. group and the curdlan i.n. control group, the rePcrV i.n. group showed significant reductions in pathology score, which was characterized according to the cell infiltration, hemorrhage, alveolar collapse and tissue damage ( Figure 2B ). Interesting, more infiltrate cells but less pathogenic lesions was observed rePcrV i.n. group. This observation indicates the participation of protective resident memory T cells, which was usually found in intranasal delivered vaccines ( 21 ). Figure 2 Compared with intramuscular vaccination, intranasal vaccination with rePcrV enhanced protection. (A) Mice were immunized with rePcrV intranasally (i.n.) or intramuscularly (i.m.). The bar indicates the survival of mice (n = 10) after intratracheal injection with a lethal dose of PA XN-1. The survival rate of the i.n. group was significantly higher than that of the i.m. group. (B) HE staining of lungs from immunized mice 24 h after infection with a sublethal dose of PA XN-1. Images were captured at 100× magnification. (C) Evaluation of the bacterial load in the lung in PA XN-1−infected mice (n = 4). The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue. (D) Quantitative measurement of TNF-α and IL-1β in the lungs (n = 3). The data are shown as the mean ± SE. (E) The bar represents the titer of anti-rePcrV IgG antibodies in the serum of immunized mice (n = 5). (G) The opsonophagocytic killing activity of the serum from immunized mice. Serum samples from immunized mice were diluted and incubated with PA XN-1. The bar represents the percentage of killed bacteria in a series of dilutions. The data are presented as the mean ± SE. (F) Frequency of CD4 + CD17 + T cells in the lungs of mice two weeks after the last immunization. The bar represents the frequency of the cells in each group. The "*" indicates a significant difference at P < 0.05. The "n.s" means "no significant difference". In addition, the bacterial loads of the rePcrV i.n. group were significantly lower than those of the rePcrV i.m. ( P <0.0001) and curdlan i.n. control groups (P = 0.0002) ( Figure 2C ). Furthermore, compared with the rePcrV i.m. and curdlan i.n. groups, the levels of the proinflammatory cytokines TNF-α and IL-1β in the BALF of the rePcrV i.n. group were significantly lower ( Figure 2D ). These results indicated that the improved protection of i.n. immunization with rePcrV might be due to the decrease in the bacterial load, the secretion of proinflammatory cytokines and the increase in inflammation in the lung. In addition, the immune response following rePcrV vaccination was determined. Surprisingly, the titer of anti-PcrV antibodies in the i.n. group was significantly lower ( P <0.0001) than that in the i.m. group ( Figure 2E ). However, no significant difference in the opsonophagocytic activity of anti-PcrV antibodies was observed between the two immunization routes ( Figure 2F ). These findings led us to measure the pulmonary Th17 response induced by rePcrV. As shown in Figure 2G , the percentage of CD4 + IL-17A + T cells in the lung of the rePcrV i.n. group was significantly higher ( P <0.0001) than that of the rePcrV i.m. group. These results suggested that lung Th17 cells could be responsible for the enhanced protection induced via the intranasal route. The Enhanced Protection Induced by Intranasal Vaccination With rePcrV Depends on the Th17 Response To clarify the contribution of the Th17 response to the increase in protection, we first tested the protection induced in IL-17A KO mice after intranasal vaccination with rePcrV. As few as 30% of immunized IL-17A KO mice survived, which was significantly lower ( P = 0.0012) than the survival rate observed in WT (wild-type) mice ( Figure 3A ). Similar results were observed in immunized WT mice after treatment with anti−IL-17A mAbs. The survival was significantly decreased ( P = 0.0013) in mice injected with IL-17A mAbs compared with mice injected with control mouse IgG ( Figure 3B ). Next, we isolated CD4 + T cells from the lungs of mice immunized intranasally with rePcrV and then adoptively transferred them into naïve mice. The mice were challenged with PA XN-1 to further evaluate CD4 + T cell-mediated protection. We found that bacterial colonization in the lungs of the mice given CD4 + T cells was significantly lower ( P = 0.0025) than that of the mice in the control group ( Figure 3C ). Similarly, pathological observation showed that lesions such as pulmonary hemorrhage and inflammatory cell infiltration were significantly reduced after injection of CD4 + T cells ( Figure 3D ). Finally, we tested the concentration of proinflammatory factors (TNF-α and IL-1β) in the BALF. Similarly, adoptive transfer of CD4 + T cells from the immunized group significantly ( P = 0.0390, P = 0.0060) reduced the concentration compared with that of the control group ( Figure 3E ). These results suggest that the Th17 response is indispensable for the enhanced protection induced by intranasal vaccination with rePcrV. Figure 3 The enhanced protection of intranasal vaccination with rePcrV depends on the Th17 response. (A) WT (wild-type) mice or IL-17A knockout mice (n = 10) were immunized with rePcrV and then intratracheally challenged with a lethal dose of PA XN-1. The bar represents the survival of the mice in each group. (B) WT mice were immunized with rePcrV and administered anti-IL-17A antibodies (n = 10) and with isotype control IgG as a control. The bar represents the survival rate of the two groups at each observation point. (C) CD4+ T cells from the lungs of rePcrV-immunized mice were transferred into naïve WT mice. Then, the mice were challenged with a sublethal dose of PA XN-1. The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue (n = 4). (D) HE staining of the lungs of mice after infection with a sublethal dose of PA XN-1. Images were captured at 100× magnification. (E) Quantitative measurement of TNF-α and IL-1β in the BALF of mice (n = 3). The data are shown as the mean ± SE. The "*" indicates a significant difference at P < 0.05. The Protection Mediated by reAmpC Relies on the Th17 Response First, we measured the opsonophagocytic killing activity of anti-reAmpC antibodies. However, no significant difference in bacteria-killing rates was observed between sera from reAmpC-immunized mice and unimmunized mice, which suggest a limited contribution of reAmpC-specific antibodies to preventive effects ( Figure 4A ). Building on these findings, we next aimed to clarify whether Th17 immunity contributes to protection. As expected, the survival rate of immunized wild-type mice was 50%, while that of immunized IL-17A KO mice was only 10%. All WT mice that were not immunized died within 60 h after the challenge ( Figure 4B ). In another setting, the immunized mice were treated with anti-IL-17A mAbs to block the Th17 response before the bacterial challenge. The results showed that the survival decreased significantly ( P = 0.0003) in mice injected with anti−IL-17A mAbs compared with control mice injected with nonspecific mouse IgG ( Figure 4C ). Additionally, the CD4 + T cells of reAmpC-immunized mice were isolated and adoptively transferred to naïve mice to evaluate their protective effects. The number of colonizing bacteria in the lungs of mice that received CD4 + T cells from the lungs of reAmpC-immunized mice was significantly ( P <0.0001) lower than that in the lungs of unimmunized mice ( Figure 4D ). Moreover, a similar trend was observed for the levels of the proinflammatory factors IL-1β and TNF-α in the BALF ( Figure 4E ). Taken together, these findings showed that the protection induced by reAmpC relies on the pulmonary Th17 response. Figure 4 The protection mediated by reAmpC relies on the Th17 response. (A) The opsonophagocytic killing activity of serum from immunized mice. The bar represents the percentage of killed bacteria in a series of dilutions. The data are presented as the mean ± SE. (B) WT (wild-type) mice or IL-17A knockout mice were immunized with reAmpC (n = 10) and then intratracheally challenged with PA XN-1. The bar represents the survival of mice in each group. (C) WT mice or IL-17A knockout mice (n = 10) were immunized with reAmpC and then intratracheally challenged with a lethal dose of PA XN-1. The bar represents the survival of mice in each group (n = 10). (D) CD4+ T cells from the lungs of reAmpC-immunized mice were transferred into naïve WT mice. Then, the mice were challenged with a sublethal dose of PA XN-1. The bar represents the average and SE of the log CFU of PA XN-1 per mg of lung tissue (n = 4). (E) Quantitative measurement of TNF-α and IL-1β in the BALF of mice (n = 3). The data are shown as the mean ± SE. The "*" indicates a significant difference at P < 0.05. The "n.s" means "no significant difference". Mapping of Epitopes in rePcrV and reAmpC That Stimulate the Th17 Response To screen for epitopes that can stimulate Th17 responses, we isolated lung lymphocytes from immunized mice and cultured them with synthesized peptides from the library of rePcrV and reAmpC ( Tables S3 and S4 ). IL-17−secreting cells were then identified by ELISPOT assay. The distribution of Th17-stimulating epitopes in rePcrV and reAmpC is shown in Figure 5A . The highest fold change in the number of spot-forming cells (SFCs) was clearly observed for Pc5 (LSEAQVLKALAWLLAANP), which had approximately 2.5-fold more SFCs than the unimmunized mice. Furthermore, most of the Th17 stimulation epitopes are located at the N-terminus of rePcrV ( Figure 5B ). In addition, the top 10 most potent epitopes of reAmpC for promoting the secretion of IL-17A are shown in red ( Figure 5A , lower panel). P10 (FTATLAGYALTQDKMRLD) and P41 (AGNSTPMALQPHRIARLP) were the most predominant epitopes in reAmpC, with approximately 2.0- and 1.9-fold more SFCs than the unimmunized mice, respectively. Unlike rePcrV, the epitopes in reAmpC did not accumulate in one region but spanned the whole molecule ( Figure 5B ). Figure 5 Mapping of epitopes in rePcrV and reAmpC that stimulate a Th17 response. (A) The lung lymphocytes of mice immunized with rePcrV (upper) and reAmpC (lower) were isolated and cocultured with the corresponding peptide library. The number of IL-17−secreting cells induced by each peptide was detected by ELISpot. The bars represent the mean and SE of the fold increase in SFCs (spot-forming cells). The top ten peptides are shown in red. (B) Cartoon image of the 3D structure of PcrV (left) and AmpC (right). The model of PcrV was built with the I-TASSER Suite, while the model of AmpC was obtained from the Protein Data Bank (PDB: 4GZB). The images were generated with PyMol 2.4 software. The regions corresponding to the top ten peptides are shown in red. The Th17-stimulating epitopes in rePcrV accumulated at its N-terminus, while those of reAmpC spanned the whole protein. Vaccination With PVAC Protects Against Four PA Strains of Different Serotypes Building on these findings, we hypothesized that a combination of Th17-stimulating epitopes could induce more robust protection. After several attempts, we successfully produced the soluble recombinant protein PVAC, which was composed of the N-terminus of rePcrV (27Glu-126Gly) connected to full-length reAmpC with the flexible linker GSGGSG ( Figure S3 ). Intranasal immunization with PVAC could also induce a humoral immune response, as indicated by an elevated titer of PVAC-specific antibodies in the serum ( Figure 6A ). More importantly, the number of CD4 + IL17 + T cells in the lung increased dramatically after immunization, which suggested a potent Th17 response ( Figure 6B ). Next, to determine whether PVAC could provide broad protection, immunized mice were challenged with clinical PA isolates of different serotypes: PA 464, PA XN-1, PA ZNJ004, and PA 451. The serotypes of these four strains were F, I, J and A, respectively. The survival rate of immunized mice was 70%, 60%, 100%, and 50% after challenge with PA 464, PA XN-1, PA ZNJ004, and PA 451, respectively ( Figure 6C ). The survival rates of the four immunized groups were significantly higher than that of the control group, which indicates that broad protection was induced by PVAC. Collectively, these results show that the fusion protein PVAC was able to effectively induce a humoral immune response and a Th17 response, which proved that broad protection was induced in the mouse pneumonia model. Figure 6 Vaccination with PVAC protects against PA strains of four different serotypes. (A) PVAC, which is composed of the N-terminus of rePcrV (27Glu-126Gly) connected to full-length reAmpC, was produced and used to immunize mice. The bar represents the titer of anti-PVAC IgG in the serum of immunized mice (n = 5) (B) The bar represents the frequency of CD4 + CD17 + T cells in the lungs of immunized mice. (C) PVAC-immunized mice were challenged via intratracheal injection of PA-464, PA XN-18 PA ZNJ004, and PA-451. The survival rates of the four immunized groups were significantly higher than that of the control group. The "*" indicates a significant difference at P < 0.05. Discussion One major challenge for developing a PA vaccine is the hypervariation of its genome, which requires a "universal vaccine" to ensure broad protection ( 22 ). By targeting the Th17 response, serotype-independent protection was elicited by attenuated PA ( 15 ), recombinant reOprL ( 18 ) or PopB coexpressed with PcrH ( 16 ). Despite these efforts, a Th17-based universal PA vaccine has yet to be attained. In fact, PopB-PcrH is too difficult to produce, and the protection induced by reOprL is limited. In this study, we generated recombinant PVAC according to the Th17-stimulating epitopes of PcrV and AmpC. The results showed that recombinant PVAC was efficiently produced in E. coli and was able to induce broad protective immunity against PA. Our results provide additional data for a Th17-based vaccine and will benefit the future development of PA vaccines. One interesting finding is that the serum anti-PcrV antibodies induced by intranasal immunization were opsonophagocytic and protective, which is very different from the antibodies induced by other Th17-stimulating antigens, such as reOprL, reAmpC and PopB ( 17 , 18 ). One possible explanation is that PcrV is the critical component of the type three secretion system, which is located in the outer membrane of bacteria ( 23 ). Thus, the antibodies may be able to bind to PcrV and block the effects of toxins secreted by the secretion apparatus. In addition, anti-PcrV antibodies promote the internalization of PA and the localization of ingested PA into acidified vacuoles, which ultimately stimulates antibody-dependent cellular cytotoxicity ( 20 , 24 ). To date, two therapeutic antibodies targeting PcrV, KB001 and MEDI3902 have been proven to be protective in animal models and are being tested in clinical trials ( 24 , 25 ). As a result, it is not surprising that elevated protection is mediated by intranasal immunization with PcrV because Th17 cells and antibodies induce simultaneous effects. In this study we found that the PVAC, which contains the predominant Th17-stimulating epitopes, could provide wide-spectrum protection against PA. However, the protective efficacy of PVAC was lower than that of rePcrV in the case of PA XN-1. The decrease of efficacy is possibly due to the change of humoral response because PVAC contains the most Th17-stimulating epitopes of PcrV. Since PcrV is a key component of the T3SS, one mechanism for anti-PcrV immunity is based on the inhibition of the cytotoxic effects of T3SS toxins ( 5 , 26 ). The epitopes recognized by the protective antibodies are predominantly located in the 144–257 region ( 27 , 28 ). However, these amino acids are not present in the PVAC. Consequently, antibodies against this region are missing in mice immunized with PVAC, resulting in decreased protection. These data suggest that protective epitopes recognized by B-cells should be integrated to improve the protective effect of the vaccine in future studies. The variability of vaccine targets is closely correlated to vaccine efficacy. The pcrV gene was found to be conserved among diverse PA isolates ( 29 ). AmpC is a class C lactamase that provides resistance against antipseudomonal cephalosporins, and it is present in most PA strains ( 30 ). Epidemiological studies showed that mutations in AmpC commonly increase the resistance to antipseudomonal cephalosporins, which potentially impairs the efficacy of PVAC-mediated protection ( 31 , 32 ). However, AmpC mutations are usually individual amino acid substitutions, which only affect a small minority of epitopes. In addition, these mutations were only detected in a small proportion (about 1%) of clinical PA isolates, which is different from the extensive diversity of PA habitats ( 31 ). As a result, we concluded that AmpC mutations are likely to have a very limited effect on the protection mediated by PVAC. One bottleneck for the development of Th17-based vaccines is the lack of effective adjuvants. Traditional aluminum adjuvants preferentially induce Th2 and antibody responses, although it is possible for these adjuvants to promote a Th17 response by activating the inflammasome via the release of uric acid ( 33 ). Several cytokines, such as IL-1β, IL-6, and IL-23, were found to be able to induce or enhance a strong Th17 response ( 34 , 35 ). Additionally, recombinant E. coli heat-labile enterotoxin and cholera toxin were able to activate dendritic cells and enhance the secretion of cytokines to induce a Th17 response ( 36 , 37 ). Another type of Th17 adjuvant is TLR receptor agonists. For example, a TLR4 agonist, monophospholipid A-trehalose dimycolate (TDM), is capable of enhancing the differentiation of Th17 cells. In this study, the TLR4 agonist curdlan, which has previously been proven to be able to induce a Th17 response, was applied as an adjuvant. However, no adjuvant that preferentially generates a Th17 response is currently available due to potential toxicity or insolubility, which highlights an urgent need for novel adjuvants for the development of Th17-based vaccines. One major concern related to Th17-based vaccines is safety. Despite their protective effects against pathogens at mucosal sites, Th17 responses have been proven to be pathogenic in many diseases ( 38 ). For example, Th17 cells can infiltrate tissues and secrete proinflammatory TNF-α and GM-CSF, which exacerbate the pathogenesis of rheumatoid arthritis (RA), multiple sclerosis (MS) and other autoimmune diseases ( 39 , 40 ). In addition, long-lived Th17 cells may be the reason for the persistence of HIV infection ( 41 , 42 ). Importantly, accumulating data indicates that Th17 cells are essential for the pathogenesis of chronic obstructive pulmonary disease (COPD) and cystic fibrosis (CF) ( 43 ). However, patients with COPD or CF are a major target population of the PA vaccine due to their susceptibility to this infection. Thus, careful assessment of the safety of the Th17-based PA vaccine is required before immunization of patients with COPD, CF and autoimmune diseases. The immunogenicity of multi-epitope vaccines may be restricted when applied to humans because of the difference of major histocompatibility complex (MHC) between humans and mice. To bridge this gap, transgenic HLA mice were applied as reliable animal models in pre-clinical investigation of vaccines ( 44 ). For example, a vaccine against Coccidioides posadasii was generated using human Th17-driven epitopes ( 45 ), and its protective effect was evaluated in humanized HLA-DR4 transgenic mice ( 46 ). Adequate investments are needed to evaluate the efficacy of PVAC in HLA class I and class II humanized transgenic mice in order to develop a human vaccine against PA in the future. Data Availability Statement The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors. Ethics Statement The animal study was reviewed and approved by the Animal Ethical and Experimental Committee of the Third Military Medical University. Author Contributions YW, XC, CW, JW, and CG performed the experiments. YZ, LP, PL, and DL supervised the experiments. YW and JG wrote and revised the manuscript. HZ and QZ designed the project and supervised the experiments. All authors contributed to the article and approved the submitted version. Funding This research was supported by the National Natural Science Foundation of China (81772155 and 81571621) and Natural Science Foundation of Chongqing (CSTC2020JCYJ-MSXM2301). Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Supplementary Material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2020.601601/full#supplementary-material Click here for additional data file. Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3249468/

Ultra-thin Layer MALDI Mass Spectrometry of Membrane Proteins in Nanodiscs

Nanodiscs have become a leading technology to solubilize membrane proteins for biophysical, enzymatic, and structural investigations. Nanodiscs are nanoscale, discoidal lipid bilayers surrounded by an amphipathic membrane scaffold protein (MSP) belt. A variety of analytical tools has been applied to membrane proteins in Nanodiscs, including several recent mass spectrometry studies. Mass spectrometry of full-length proteins is an important technique for analyzing protein modifications, for structural studies, and for identification of proteins present in binding assays. However, traditional MALDI-TOF mass spectrometry methods for analyzing full-length membrane proteins solubilized in Nanodiscs are limited by strong signal from the MSP belt and weak signal from the membrane protein inside the Nanodisc. Herein we show that an optimized ultra-thin layer MALDI sample preparation technique dramatically enhances the membrane protein signal and nearly completely eliminates the MSP signal. First shot MALDI and MALDI imaging are used to characterize the spots formed by the ultra-thin layer method. Furthermore, the membrane protein enhancement and MSP suppression is shown to be independent of the type of membrane protein and is applicable to mixtures of membrane proteins in Nanodiscs.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4467858/

Wnt directs the endosomal flux of LDL-derived cholesterol and lipid droplet homeostasis

The Wnt pathway, which controls crucial steps of the development and differentiation programs, has been proposed to influence lipid storage and homeostasis. In this paper, using an unbiased strategy based on high-content genome-wide RNAi screens that monitored lipid distribution and amounts, we find that Wnt3a regulates cellular cholesterol. We show that Wnt3a stimulates the production of lipid droplets and that this stimulation strictly depends on endocytosed, LDL-derived cholesterol and on functional early and late endosomes. We also show that Wnt signaling itself controls cholesterol endocytosis and flux along the endosomal pathway, which in turn modulates cellular lipid homeostasis. These results underscore the importance of endosome functions for LD formation and reveal a previously unknown regulatory mechanism of the cellular programs controlling lipid storage and endosome transport under the control of Wnt signaling.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3557791/

High immunogenicity of nicotine vaccines obtained by intradermal delivery with safe adjuvants

Immunotherapy for tobacco addiction may offer a safe, alternative treatment if the immunogenicity of the current nicotine vaccines can be improved. We show here that intradermal (ID) immunization induces the production of antibody directed against nicotine (NicAb) at a much higher level than conventional intramuscular (IM) immunization. The magnitude and duration of NicAb production was further increased robustly by non-inflammatory laser vaccine adjuvant (LVA), slightly inflammatory monophosphoryl lipid A (MPL) or a combination of MPL and CpG adjuvants. Consequently, significantly fewer vaccination doses were required to attain a high level of NicAb production for an extended period of time and reduce nicotine entry into the brain in the presence of LVA, MPL or MPL/CpG adjuvant, respectively. Yet, the potency of these adjuvants to augment ID nicotine vaccine immunogenicity came at the expense of local skin reactogenicity, with LVA causing little skin reaction and MPL/CpG stimulating overt skin irritation. These observations underscore a necessity of a balance between optimal adjuvant potency and undesired local reactogenicity. In summary, our study presents a novel approach to significantly improve nicotine vaccine immunogenicity by a combination of safe cutaneous vaccine adjuvants with ID immunization.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7089334/

Nanotechnology for the Food and Bioprocessing Industries

Several complex set of engineering and scientific challenges in the food and bioprocessing industries for manufacturing high quality and safe food through efficient and sustainable means can be solved through nanotechnology. Bacteria identification and food quality monitoring using biosensors; intelligent, active, and smart food packaging systems; and nanoencapsulation of bioactive food compounds are few examples of emerging applications of nanotechnology for the food industry. We review the background about the potential of nanotechnology, provide an overview of the current and future applications of nanotechnology relevant to food and bioprocessing industry, and identify the societal implications for successful implementation of nanotechnology. Introduction Nanotechnology is generally defined as the design, production, and application of structures, devices, and systems through control of the size and shape of the material at the 10 −9 of a meter scale. The National Nanotechnology Initiative (Arlington, VA, USA) defines nanotechnology as 'the understanding and control of matter at dimensions of roughly 1-100 nm, where unique phenomena enable novel applications'. Nanotechnology is truly an interdisciplinary field that stretches across a whole spectrum of science including physics, chemistry, and biology as well as engineering including micro-fabrication techniques. The physical, chemical, and biological properties of structures and systems at nanoscale are substantially different than the macro-scale counterparts due to the interactions of individual atoms and molecules thereby offering unique and novel functional applications. As the size of the particles gets reduced to nanoscale range, there is an immense increase in the surface to volume ratio which increases reactivity and changes the mechanical, electrical, and optical properties of the particles. The food and bioprocessing industry is facing enormous challenges for developing and implementing systems that can produce high quality, safe foods as well as feeds while also being efficient, environmentally acceptable, and sustainable (Manufuture 2006 ). To answer these complex set of engineering and scientific challenges, innovation is needed for new processes, products, and tools in the food industry. Nanotechnology is gaining momentum and becoming a worldwide important tool for the food and bioprocessing industry in meeting the foreseeable increasing world demand that will result from population growth and increasing incomes in developing countries. Nanotechnology can possibly improve production processes to provide products with better characteristics and new functionalities in the food and bioprocessing industry (Roco 2002 ). Total global investment in nanotechnologies in the year 2004 was US $7 billion (European Commission 2004 ). The annual value of nanotechnology related products for the years 2011–2015 has been estimated to be $1 trillion (Roco and Bainbridge 2001 ). The nanofood market is expected to surge from US $7 billion in 2006 to US $20.4 billion in 2010 (Helmut Kaiser Consultancy 2004 ). In the year 2006, there were about 400 agricultural and food companies around the world actively pursuing nanotechnology research and development and this number is expected to increase to more than 1,000 by 2015 (Joseph and Morrison 2006 ). The term 'nanofood' describes food which has been cultivated, produced, processed, or packaged using nanotechnology techniques or tools, or to which manufactured nanomaterials have been added (Joseph and Morrison 2006 ). To communicate the merits of nanotechnology in food applications and to avoid misunderstanding and confusion, a new definition of nanotechnology for food applications is essential (Kampers 2007 ). Nanotechnology has potential applications in all aspects of food chain including storage, quality monitoring, food processing, and food packaging. Nanotechnology applications in the food industry range from intelligent packaging to creation of on-demand interactive food that allows consumers to modify food, depending on the nutritional needs and tastes. The objective of this review is to provide a background on agri-food nanotechnology and an up-to-date account of known and possible futuristic applications of nanotechnology in the food and bioprocessing industry. The brand and the company names mentioned in the manuscript here are for the purpose of information only and not intended as an act of promotion or endorsement. Food Quality Monitoring Quality assurance in food and bioprocessing industry is of utmost importance because consumers demand safe and wholesome food as well as governments impose stringent regulations to ensure food safety and feed hygiene. Sensors or detection systems for rapid detection of spoilage of product components, for quality control, and for abuse detection at source and during production chain is possible through nanotechnology. Nanosensors Nanosensors can provide quality assurance by tracking microbes, toxins, and contaminants throughout food processing chain through data capture for automatic control functions and documentation. Nanotechnology also enables to implement low cost nanosensors in food packaging to monitor the quality of food during various stages of the logistic process to guarantee product quality up until consumption. Grain quality monitoring nanosensors (Fig. 1 ), that are being developed by researchers at the Canadian Wheat Board Centre for Grain Storage Research, University of Manitoba, Canada; use conducting polymer nanoparticles (Neethirajan et al. 2009a ), which respond to analytes and volatiles in food storage environment and thereby detect the source and the type of spoilage. The advantage of this sensor system is that thousands of nanoparticles can be placed on a single sensor to accurately detect the presence of insects or fungus inside stored grain bulk in bins. Because of the miniaturization and low power requirement, the nanosensors can be fabricated small and light weight (Neethirajan and Jayas 2007 ) and can be deployed and distributed into the crevices of grain bulk, where the stored product pests often hide. Fig. 1 Example of a futuristic wireless nanosensor network for grain quality monitoring. 1 Control panel, 2 grain auger, 3 air plenum, 4 fan, 5 auger to transfer grain, if needed, 6 wireless transmitter Ruengruglikit et al. ( 2004 ) have developed an electronic tongue for inclusion in food packaging that consists of an array of nanosensors that are extremely sensitive to gases released by food as it spoils, causing the sensor strip to change color as a result, giving a clear visible signal of whether the food is fresh or not. Bacteria Identification Horner et al. ( 2006 ) have developed an analytical technology called reflective interferometry, using nanotechnology which provides specific, rapid, and label-free optical detection of biomolecules in complex mixtures. This new platform technology has provided food quality assurance by detecting Escherichia coli ( E. coli ) bacteria in a food sample by measuring and detecting light scattering by cell mitochondria. This sensor works on the principle that a protein of a known and characterized bacterium set on a silicon chip can bind with any other E. coli bacteria present in the food sample. This binding will result in a nanosized light scattering detectable by analysis of digital images. A biosensor developed by Fu et al. ( 2008 ) uses fluorescent dye particles attached to anti-salmonella antibodies on a silicon/gold nanorod array. When the salmonella bacteria present in the food is being tested, the nanosized dye particles on the sensor become visible. Unlike the time-consuming conventional lab tests that are based on bacterial cultures, this biosensor can detect the salmonella in food instantly. Campylobacter jejuni are bacteria which cause abdominal cramps and diarrhea in humans. The campylobacter infections can be traced to poultry meat products which have been contaminated with intestinal contents during processing. To address this food safety problem, Stutzenberger et al. ( 2007 ) have developed a novel strategy that employs bioactive nanoparticles in the chicken feed specifically designed to bind to the biomolecular structures on the surfaces of campylobacters. The feed enriched by antibiotic-functioning nanocarbohydrate particles binds with the bacterium's surface to remove it through the bird's feces. Agromicron Ltd, Hong Kong has developed a low cost Nano Bioluminescent Spray (Plexus Institute 2006 ), which can react with the pathogen strain on food and produce a visual glow for easy detection. The spray is made of nanoparticles and would work based on its reactivity with the bacteria. The higher the number of connections between bacteria and molecules, the more intense the glow produced by the particles. This spray can identify a broad range of food-related pathogens, such as Salmonella and E. coli . Cheng et al. ( 2009 ) demonstrated rapid detection of E. coli in food using biofunctional magnetic nanoparticles (about 20 nm in diameter) in combination with adenosine triphosphate bioluminescence. Zhao et al. ( 2004 ) developed an ultrasensitive immunoassay for in situ pathogen quantification in spiked ground beef samples using antibody-conjugated silica fluorescent nanoparticles (about 60 nm in diameter). Nanosensors Nanosensors can provide quality assurance by tracking microbes, toxins, and contaminants throughout food processing chain through data capture for automatic control functions and documentation. Nanotechnology also enables to implement low cost nanosensors in food packaging to monitor the quality of food during various stages of the logistic process to guarantee product quality up until consumption. Grain quality monitoring nanosensors (Fig. 1 ), that are being developed by researchers at the Canadian Wheat Board Centre for Grain Storage Research, University of Manitoba, Canada; use conducting polymer nanoparticles (Neethirajan et al. 2009a ), which respond to analytes and volatiles in food storage environment and thereby detect the source and the type of spoilage. The advantage of this sensor system is that thousands of nanoparticles can be placed on a single sensor to accurately detect the presence of insects or fungus inside stored grain bulk in bins. Because of the miniaturization and low power requirement, the nanosensors can be fabricated small and light weight (Neethirajan and Jayas 2007 ) and can be deployed and distributed into the crevices of grain bulk, where the stored product pests often hide. Fig. 1 Example of a futuristic wireless nanosensor network for grain quality monitoring. 1 Control panel, 2 grain auger, 3 air plenum, 4 fan, 5 auger to transfer grain, if needed, 6 wireless transmitter Ruengruglikit et al. ( 2004 ) have developed an electronic tongue for inclusion in food packaging that consists of an array of nanosensors that are extremely sensitive to gases released by food as it spoils, causing the sensor strip to change color as a result, giving a clear visible signal of whether the food is fresh or not. Bacteria Identification Horner et al. ( 2006 ) have developed an analytical technology called reflective interferometry, using nanotechnology which provides specific, rapid, and label-free optical detection of biomolecules in complex mixtures. This new platform technology has provided food quality assurance by detecting Escherichia coli ( E. coli ) bacteria in a food sample by measuring and detecting light scattering by cell mitochondria. This sensor works on the principle that a protein of a known and characterized bacterium set on a silicon chip can bind with any other E. coli bacteria present in the food sample. This binding will result in a nanosized light scattering detectable by analysis of digital images. A biosensor developed by Fu et al. ( 2008 ) uses fluorescent dye particles attached to anti-salmonella antibodies on a silicon/gold nanorod array. When the salmonella bacteria present in the food is being tested, the nanosized dye particles on the sensor become visible. Unlike the time-consuming conventional lab tests that are based on bacterial cultures, this biosensor can detect the salmonella in food instantly. Campylobacter jejuni are bacteria which cause abdominal cramps and diarrhea in humans. The campylobacter infections can be traced to poultry meat products which have been contaminated with intestinal contents during processing. To address this food safety problem, Stutzenberger et al. ( 2007 ) have developed a novel strategy that employs bioactive nanoparticles in the chicken feed specifically designed to bind to the biomolecular structures on the surfaces of campylobacters. The feed enriched by antibiotic-functioning nanocarbohydrate particles binds with the bacterium's surface to remove it through the bird's feces. Agromicron Ltd, Hong Kong has developed a low cost Nano Bioluminescent Spray (Plexus Institute 2006 ), which can react with the pathogen strain on food and produce a visual glow for easy detection. The spray is made of nanoparticles and would work based on its reactivity with the bacteria. The higher the number of connections between bacteria and molecules, the more intense the glow produced by the particles. This spray can identify a broad range of food-related pathogens, such as Salmonella and E. coli . Cheng et al. ( 2009 ) demonstrated rapid detection of E. coli in food using biofunctional magnetic nanoparticles (about 20 nm in diameter) in combination with adenosine triphosphate bioluminescence. Zhao et al. ( 2004 ) developed an ultrasensitive immunoassay for in situ pathogen quantification in spiked ground beef samples using antibody-conjugated silica fluorescent nanoparticles (about 60 nm in diameter). Food Packaging The purpose of food packaging is to increase food shelf life by avoiding spoilage, bacteria, or the loss of food nutrient. Nanotechnology offers higher hopes in food packaging by promising longer shelf life, safer packaging, better traceability of food products, and healthier food. Polymer nanocomposite technology holds the key to future advances in flexible, intelligent, and active packaging. Intelligent, smart, and active packaging systems produced by nanotechnology would be able to repair the tears and leakages (self healing property), and respond to environmental conditions (e.g., change in temperature and moisture). Intelligent food packaging can sense when its contents are spoiling, and alert the consumer, while active packaging will release a preservative such as antimicrobials, flavors, colors, or nutritional supplements into the food when it begins to spoil. Nanotechnology can provide solutions for food packaging by modifying the permeation behavior of foils, increasing barrier properties (mechanical, chemical, and microbial), providing antimicrobial properties, and by improving heat-resistance properties (Brody 2003 ; Chaudhry et al. 2008 ). Antimicrobial Packaging Antimicrobial packaging systems are important for the food industry and the consumers because these systems can extend the product shelf life and maintain food safety by reducing the growth rate of microorganisms. Anti-microbial nanoparticle coatings in the matrix of the packaging material can reduce the development of bacteria on or near the food product, inhibiting the microbial growth on non-sterilized foods and maintain the sterility of pasteurized foods by preventing the post-contamination. Antimicrobial packaging systems include adding an antimicrobial nanoparticle sachet into the package, dispersing bioactive agents in the packaging; coating bioactive agents on the surface of the packaging material, or utilizing antimicrobial macromolecules with film forming properties or edible matrices (Coma 2008 ). The schematic of a typical antimicrobial coating nanopackaging film (120 µm thickness) based on research by Buonocore et al. ( 2005 ) is shown in Fig. 2 . Fig. 2 a Schematic showing exploded view of a typical antimicrobial coating nanopackaging film; b Antimicrobial active packaging microorganisms hydrolyses starch based particles causing release of the antimicrobial lysozyme resulting in inhibitors of microbial growth (based on Buonocore et al. 2005) Foods such as cheese, sliced meat, and bakery that are prone to spoiling on the surface can be protected by contact packaging imbued with antimicrobial nanoparticles. Antifungal active paper packaging developed by Rodriguez et al. ( 2008 ) incorporating cinnamon oil with solid wax paraffin using nanotechnology as an active coating was shown to be used as an effective packaging material for bakery products. Working with oregano oil and apple puree, Rojas-Grau et al. ( 2006 ) have created edible food films that are able to kill E. coli bacteria. Antimicrobial nanoparticles that have been synthesized and tested for applications in antimicrobial packaging and food storage boxes include silver oxide nanoparticles (Sondi and Salopek-Sondi 2004 ), zinc oxide, and magnesium oxide nanoparticles (Jones et al. 2008 ) and nisin particles produced from the fermenation of a bacteria (Gadang et al. 2008 ). CTC Nanotechnology GmbH, Merzig, Germany has manufactured and is selling a nanoscale dirt-repellent coating (CTC Nanotechnology 2009 ) to create self-cleaning surfaces for use in food packages and meat-processing plants. The technology concept is based on the sol-gel process where the nanoparticles are suspended in a fluid medium and by the action of nanohydrophobisation; the absorbency of the surfaces to be treated is eliminated so that they remain resistant to the environmental factors after cleaning. The added advantages of this product are that they are biodegradable and approved and certified for use with food. Improved Food Storage Oxygen inside food packaging is the main cause for food deterioration due to oxidation of fats and oils and growth of microorganisms. Also, oxygen accelerates the processes inside food packaging leading to discoloration, changes in texture, rancidity and off-odor, and flavor problems. Nanotechnology can effectively produce oxygen scavengers for sliced processed meat, beer, beverages, cooked pastas, and ready-to-eat snacks; moisture absorber sheets for fresh meat, poultry, and fish; and ethylene-scavenging bags for packaging of fruit and vegetables. Active packaging films for selective control of oxygen transmission and aroma affecting enzymes has been developed based on the nanotechnology approach (Rivett and Speer 2009 ). The modification of the surface of nanosized materials by dispersing agents can act as substrates for the oxidoreductase enzymes. Oxygen absorbing sachets (Bioka Ltd, Finland; Sealed Air Corporation, USA; Constar International Inc, USA; Actipak, India) based on reactions catalyzed by food grade enzymes are also commercially available in the market. Packaging film enriched with silicate nanoparticles produced by Bayer Polymers, Germany reduces the entrance of oxygen and other gasses, and the exit of moisture and can prevent the food from spoilage. Nanocor Inc, Chicago, IL, USA has developed a nanocomposite containing clay nanoparticles (Advantage Magazine 2004 ) for manufacturing polyethylene terephthalate bottles to ship beer, fruit juice, and soft drinks. The clay nanoparticles embedded in the plastic bottles stiffen the packaging, reducing gas permeability, and minimizes the loss of carbon dioxide from the beer and the ingress of oxygen to the bottle, keeping the beer fresher and increases the shelf life to more than six months. Green Packaging Natural biopolymer bio-nanocomposites-based packaging materials have great potential for enhancing food quality, safety, and stability as an innovative packaging and processing technology. Plantic Technologies Ltd, Altona, Australia has manufactured and is selling biodegradable and fully compostable bioplastics packaging (CSIRO 2006 ), made from organic corn starch using nanotechnology. Bio degradable bio-nanocomposites prepared from natural biopolymers such as starch and protein exhibited advantages as a food packaging material by providing enhanced organoleptic characteristics such as appearance, odor, and flavor (Zhao et al. 2008 ). The unique advantages of the natural biopolymer packaging are that these can handle particulate foods, can act as carriers for functionally active substances, and provide nutritional supplements (Rhim and Ng 2007 ). Kriegel et al. ( 2009 ) have developed a methodology using electrospinning technique for making biodegradable green food packaging from chitin. Chitin is a natural polymer and a main component of lobster shells. The electrospinning technique involves dissolving chitin in a solvent and drawing it through a tiny hole with applied electricity to produce nanoslim fiber spins. These strong and naturally antimicrobial nanofibers were used for developing the green food packaging. BASF, Ludwigshafen, Germany; New Ice, Durango, USA; Archer Daniels Midland CO, Decatur, USA; Sharp Interpack, Aylesham, UK and RPC Group, Northamptonshire, UK (BASF 2009 ; Bordes et al. 2009 ; Coating & Converting Magazine 2008 ) have produced food packaging bags and sachets from biodegradable polylactic acid and polycaprolactone obtained from polymer nanocomposites of corn plant. Tracking, Tracing, and Brand Protection Nanotechnology can help food industries in providing authentication, and track and trace features of a food product for avoiding counterfeiting; preventing adulteration and diversion of products destined for a specific market. To help in the tracking and tracing, nanotechnology provides complex invisible nanobarcodes with batch information which can be encrypted directly onto the food products and packaging. This nanobarcode technology offers food safety by allowing the brand owners to monitor their supply chains without having to share company information to distributors and wholesalers. Oxonica, Oxford, UK offers solutions for food product identification using a biological fingerprint combined with recorded quality characteristics in the form of nanobarcodes. The technology involves nanoparticles (Oxonica 2007 ) made up of gold, silver, and platinum varying in width, length, and amount to create stripes of different reflectivity. By altering the stripe orders, different codes can be created and be assigned for every food item providing brand and authenticity in tracing food batches. NanoInk, Skokie, USA has developed a patterning technique called Dip Pen Nanolithography (Zhang et al. 2009 ) to encrypt information directly onto food products or pharmaceutical pills and on packaging. The technique involves using a scanning probe molecule-coated tip to deposit a chemically engineered ink material to create nanolithographic pattern onto the food surface. Authentix, Addison, USA has developed and is marketing nanoscale markers that can be incorporated into product packaging. Nam et al. ( 2003 ) have made nanodisks of gold and nickel to encrypt information to be used as biological labels in applications such as DNA detection and as tags for tracking food products. The nanodisks were functionalized with dye molecules called chromophores that emit a unique light spectrum when illuminated with a laser beam. Li et al. ( 2005 ) have created a nanobarcode detection system that fluoresces under ultraviolet light in a combination of color that can be read by a computer scanner. Food and biological samples containing various combinations of E. coli , anthrax, and tularemia bacteria, and Ebola and SARS viruses has been tested using this system and several pathogens were clearly distinguished simultaneously by different color codes. Antimicrobial Packaging Antimicrobial packaging systems are important for the food industry and the consumers because these systems can extend the product shelf life and maintain food safety by reducing the growth rate of microorganisms. Anti-microbial nanoparticle coatings in the matrix of the packaging material can reduce the development of bacteria on or near the food product, inhibiting the microbial growth on non-sterilized foods and maintain the sterility of pasteurized foods by preventing the post-contamination. Antimicrobial packaging systems include adding an antimicrobial nanoparticle sachet into the package, dispersing bioactive agents in the packaging; coating bioactive agents on the surface of the packaging material, or utilizing antimicrobial macromolecules with film forming properties or edible matrices (Coma 2008 ). The schematic of a typical antimicrobial coating nanopackaging film (120 µm thickness) based on research by Buonocore et al. ( 2005 ) is shown in Fig. 2 . Fig. 2 a Schematic showing exploded view of a typical antimicrobial coating nanopackaging film; b Antimicrobial active packaging microorganisms hydrolyses starch based particles causing release of the antimicrobial lysozyme resulting in inhibitors of microbial growth (based on Buonocore et al. 2005) Foods such as cheese, sliced meat, and bakery that are prone to spoiling on the surface can be protected by contact packaging imbued with antimicrobial nanoparticles. Antifungal active paper packaging developed by Rodriguez et al. ( 2008 ) incorporating cinnamon oil with solid wax paraffin using nanotechnology as an active coating was shown to be used as an effective packaging material for bakery products. Working with oregano oil and apple puree, Rojas-Grau et al. ( 2006 ) have created edible food films that are able to kill E. coli bacteria. Antimicrobial nanoparticles that have been synthesized and tested for applications in antimicrobial packaging and food storage boxes include silver oxide nanoparticles (Sondi and Salopek-Sondi 2004 ), zinc oxide, and magnesium oxide nanoparticles (Jones et al. 2008 ) and nisin particles produced from the fermenation of a bacteria (Gadang et al. 2008 ). CTC Nanotechnology GmbH, Merzig, Germany has manufactured and is selling a nanoscale dirt-repellent coating (CTC Nanotechnology 2009 ) to create self-cleaning surfaces for use in food packages and meat-processing plants. The technology concept is based on the sol-gel process where the nanoparticles are suspended in a fluid medium and by the action of nanohydrophobisation; the absorbency of the surfaces to be treated is eliminated so that they remain resistant to the environmental factors after cleaning. The added advantages of this product are that they are biodegradable and approved and certified for use with food. Improved Food Storage Oxygen inside food packaging is the main cause for food deterioration due to oxidation of fats and oils and growth of microorganisms. Also, oxygen accelerates the processes inside food packaging leading to discoloration, changes in texture, rancidity and off-odor, and flavor problems. Nanotechnology can effectively produce oxygen scavengers for sliced processed meat, beer, beverages, cooked pastas, and ready-to-eat snacks; moisture absorber sheets for fresh meat, poultry, and fish; and ethylene-scavenging bags for packaging of fruit and vegetables. Active packaging films for selective control of oxygen transmission and aroma affecting enzymes has been developed based on the nanotechnology approach (Rivett and Speer 2009 ). The modification of the surface of nanosized materials by dispersing agents can act as substrates for the oxidoreductase enzymes. Oxygen absorbing sachets (Bioka Ltd, Finland; Sealed Air Corporation, USA; Constar International Inc, USA; Actipak, India) based on reactions catalyzed by food grade enzymes are also commercially available in the market. Packaging film enriched with silicate nanoparticles produced by Bayer Polymers, Germany reduces the entrance of oxygen and other gasses, and the exit of moisture and can prevent the food from spoilage. Nanocor Inc, Chicago, IL, USA has developed a nanocomposite containing clay nanoparticles (Advantage Magazine 2004 ) for manufacturing polyethylene terephthalate bottles to ship beer, fruit juice, and soft drinks. The clay nanoparticles embedded in the plastic bottles stiffen the packaging, reducing gas permeability, and minimizes the loss of carbon dioxide from the beer and the ingress of oxygen to the bottle, keeping the beer fresher and increases the shelf life to more than six months. Green Packaging Natural biopolymer bio-nanocomposites-based packaging materials have great potential for enhancing food quality, safety, and stability as an innovative packaging and processing technology. Plantic Technologies Ltd, Altona, Australia has manufactured and is selling biodegradable and fully compostable bioplastics packaging (CSIRO 2006 ), made from organic corn starch using nanotechnology. Bio degradable bio-nanocomposites prepared from natural biopolymers such as starch and protein exhibited advantages as a food packaging material by providing enhanced organoleptic characteristics such as appearance, odor, and flavor (Zhao et al. 2008 ). The unique advantages of the natural biopolymer packaging are that these can handle particulate foods, can act as carriers for functionally active substances, and provide nutritional supplements (Rhim and Ng 2007 ). Kriegel et al. ( 2009 ) have developed a methodology using electrospinning technique for making biodegradable green food packaging from chitin. Chitin is a natural polymer and a main component of lobster shells. The electrospinning technique involves dissolving chitin in a solvent and drawing it through a tiny hole with applied electricity to produce nanoslim fiber spins. These strong and naturally antimicrobial nanofibers were used for developing the green food packaging. BASF, Ludwigshafen, Germany; New Ice, Durango, USA; Archer Daniels Midland CO, Decatur, USA; Sharp Interpack, Aylesham, UK and RPC Group, Northamptonshire, UK (BASF 2009 ; Bordes et al. 2009 ; Coating & Converting Magazine 2008 ) have produced food packaging bags and sachets from biodegradable polylactic acid and polycaprolactone obtained from polymer nanocomposites of corn plant. Tracking, Tracing, and Brand Protection Nanotechnology can help food industries in providing authentication, and track and trace features of a food product for avoiding counterfeiting; preventing adulteration and diversion of products destined for a specific market. To help in the tracking and tracing, nanotechnology provides complex invisible nanobarcodes with batch information which can be encrypted directly onto the food products and packaging. This nanobarcode technology offers food safety by allowing the brand owners to monitor their supply chains without having to share company information to distributors and wholesalers. Oxonica, Oxford, UK offers solutions for food product identification using a biological fingerprint combined with recorded quality characteristics in the form of nanobarcodes. The technology involves nanoparticles (Oxonica 2007 ) made up of gold, silver, and platinum varying in width, length, and amount to create stripes of different reflectivity. By altering the stripe orders, different codes can be created and be assigned for every food item providing brand and authenticity in tracing food batches. NanoInk, Skokie, USA has developed a patterning technique called Dip Pen Nanolithography (Zhang et al. 2009 ) to encrypt information directly onto food products or pharmaceutical pills and on packaging. The technique involves using a scanning probe molecule-coated tip to deposit a chemically engineered ink material to create nanolithographic pattern onto the food surface. Authentix, Addison, USA has developed and is marketing nanoscale markers that can be incorporated into product packaging. Nam et al. ( 2003 ) have made nanodisks of gold and nickel to encrypt information to be used as biological labels in applications such as DNA detection and as tags for tracking food products. The nanodisks were functionalized with dye molecules called chromophores that emit a unique light spectrum when illuminated with a laser beam. Li et al. ( 2005 ) have created a nanobarcode detection system that fluoresces under ultraviolet light in a combination of color that can be read by a computer scanner. Food and biological samples containing various combinations of E. coli , anthrax, and tularemia bacteria, and Ebola and SARS viruses has been tested using this system and several pathogens were clearly distinguished simultaneously by different color codes. Encapsulation and Delivery The nanoencapsulation system offers numerous benefits (Shefer 2008 ) including ease of handling, enhanced stability, protection against oxidation, retention of volatile ingredients, taste masking, moisture-triggered controlled release, pH-triggered controlled release, consecutive delivery of multiple active ingredients, change in flavor character, long lasting organoleptic perception, and enhanced bioavailability and efficacy. Nanomaterials for food and bioprocessing applications can be produced from engineered plants or microbes and through the processing of waste materials such as stalks and other cellulosic materials (Robinson and Morrison 2009 ). Nanosilicas produced from plants can be used for encapsulating enzymes that in turn can be used for in vivo drug or nutrient release systems (Neethirajan et al. 2009 ). Bioactive Compounds Bioactive compounds are extra nutritional constituents that typically occur in small quantities in foods. Examples include beta-carotene from carrots, lycopene from tomato, beta-glucan from oats, omega-3 acid from salmon oil, conjugated linoleic acid from cheese, lactobacillus from yogurt, and isoflavones from soybeans. Nanotechnology has shown greater potential in improving the efficiency of delivery of nutraceuticals and bioactive compounds in functional foods to improve human health. Nanotechnology can enhance solubility, improve bioavailability, and protect the stability of micronutrients and bioactive compounds during processing, storage and distribution (Chen et al. 2006 ). Nanocapsules have been used by George Weston Foods, Australia to mask the taste and odor of tuna fish oil (source of omega-3 fatty acids) which is integrated into bread. The nanocapsules break open only when they reach the stomach and hence the unpleasant fish oil taste can be avoided. Nanoencapsules has been used for the protection and controlled release of beneficial live probiotic species to promote healthy gut function. The viability of probiotic organisms including Lactobacillus acidophilus , Lactobacillus casei , Lactobacillus rhamnosus , and Bifidobacterium spp. within freeze dried yogurt can be improved by nanoencapsulation with calciumalginate (Kailasapathy and Rybka 1997 ). Nanoencapsulated Bifidobacteria with starch by spray coating exhibited an affordable and industrially convenient encapsulation process (O'Riordan et al. 2001 ). The bioavailability of lycopene (antioxidant from tomato), can be increased by fortifying nanoparticles of lycopene in tomato juice, pasta sauce, and jam (Auweter et al. 1999 ). Milk protein, casein, was used to make nanosized micelles and has been employed as a vehicle for delivering sensitive health-promoting ingredients including vitamin D2 (Semo et al. 2007 ). Biopolymer zein (maize protein) nanofibers prepared by electrospinning technique (Fernandez et al. 2009 ) for encapsulating beta-carotene demonstrates the potential of nanotechnology in food and nutraceutical formulation and coatings, bioactive food packaging, and food processing industries. Self-assembled nanotubes, developed from hydrolysed milk protein α-lactalbumin, offers a new naturally derived carrier for nanoencapsulation of nutrients, supplements, and pharmaceuticals (Graveland-Bikker and de Kruif 2006 ). Crystalline nanocochleates of about 50 nm in size derived from soya bean can protect micronutrients and antioxidants from degradation during manufacture and storage (BioDelivery Sciences International, Raleigh, NC, USA). Interactive Foods Nanotechnology helps to make interactive foods which can allow consumers to modify the food depending on their own nutritional needs or tastes. The concept of on-demand food states that thousands of nanocapsules containing flavor or color enhancers or added nutritional elements would remain dormant in the food and will only be released when triggered by the consumer (Dunn 2004 ). Kraft, the leader in food industry, has established a consortium called 'Nanotek' to collaborate with universities and research laboratories in USA for developing interactive foods and nanoscale sensors (Forbes 2005 ). Development of foods capable of changing their color, flavor, or nutritional properties according to a person's dietary needs, allergies, or taste preferences is on the research agenda of Nestle and Kraft. Nanotechnology can enable methods to make foods such as soft drinks, ice cream, chocolate, or chips to be marketed as 'health' foods by reducing fat, carbohydrate or calorie content or by increasing protein, fiber or vitamin content. Also, nanotechnology can aid in the production of stronger flavorings, colorings, and nutritional additives, and processing aids to increase the pace of manufacturing and to lower costs of ingredients and processing (Burdo 2005 ). Nanofilters and membranes can screen out or pass through certain molecules based on the shape and/or size to remove toxins or adjust flavors. Nestle and Unilever are reported to be developing a nanoemulsion based ice cream with a lower fat content that retains a fatty texture and flavor (Renton 2006 ). Texture The size and the structure of food influence the functionality of foods by providing the taste, texture, and stability properties that consumers want. Nanotechnology can play a vital role in controlling the size and structure of food to a greater extent. These include healthier foods (lower fat, lower salt) with desirable sensory properties; ingredients with improved properties; and the potential for removal of certain additives without loss of stability, for example in emulsions, and in smart-aids for processing foods to remove allergens such as peanut protein. Scaling down the size of food molecules to nanosized crystals creates more particles for an overall greater surface area. Smaller particles improve food's spreadability and stability, and can aid in developing healthier low-fat food products. Multiple emulsions such as water-in-oil-in-water can distribute the lipids more evenly to reduce extra stabilizers and thickeners to achieve a desirable food texture (Garti and Benichou 2004 ). Bitter blockers (Senomyx, San Diego, USA) prepared from nanoscale assays (Wenner 2008 ) can activate the taste receptors of human tongue and can reduce the bitterness naturally inherent in some foods. A photocatalytic process using nanogold particles (80-120 nm size) was developed by Lin et al. ( 2008 ) for shortening the aging period and enhancing the sensory quality of sorghum spirits. Contreras et al. ( 2009 ) showed that nanozinc can be potentially used to optimize conditions for surface enhancement of infrared absorption of food components. They were able to demonstrate that butter treated with nanozinc particles provided trans fat spectral information along with the degree and the unsaturation of the acyl groups. These results indicate the potential of nanomaterials in imaging to reveal useful information concerning food allergens, bioactive compounds, and microbial pathogens. Bioactive Compounds Bioactive compounds are extra nutritional constituents that typically occur in small quantities in foods. Examples include beta-carotene from carrots, lycopene from tomato, beta-glucan from oats, omega-3 acid from salmon oil, conjugated linoleic acid from cheese, lactobacillus from yogurt, and isoflavones from soybeans. Nanotechnology has shown greater potential in improving the efficiency of delivery of nutraceuticals and bioactive compounds in functional foods to improve human health. Nanotechnology can enhance solubility, improve bioavailability, and protect the stability of micronutrients and bioactive compounds during processing, storage and distribution (Chen et al. 2006 ). Nanocapsules have been used by George Weston Foods, Australia to mask the taste and odor of tuna fish oil (source of omega-3 fatty acids) which is integrated into bread. The nanocapsules break open only when they reach the stomach and hence the unpleasant fish oil taste can be avoided. Nanoencapsules has been used for the protection and controlled release of beneficial live probiotic species to promote healthy gut function. The viability of probiotic organisms including Lactobacillus acidophilus , Lactobacillus casei , Lactobacillus rhamnosus , and Bifidobacterium spp. within freeze dried yogurt can be improved by nanoencapsulation with calciumalginate (Kailasapathy and Rybka 1997 ). Nanoencapsulated Bifidobacteria with starch by spray coating exhibited an affordable and industrially convenient encapsulation process (O'Riordan et al. 2001 ). The bioavailability of lycopene (antioxidant from tomato), can be increased by fortifying nanoparticles of lycopene in tomato juice, pasta sauce, and jam (Auweter et al. 1999 ). Milk protein, casein, was used to make nanosized micelles and has been employed as a vehicle for delivering sensitive health-promoting ingredients including vitamin D2 (Semo et al. 2007 ). Biopolymer zein (maize protein) nanofibers prepared by electrospinning technique (Fernandez et al. 2009 ) for encapsulating beta-carotene demonstrates the potential of nanotechnology in food and nutraceutical formulation and coatings, bioactive food packaging, and food processing industries. Self-assembled nanotubes, developed from hydrolysed milk protein α-lactalbumin, offers a new naturally derived carrier for nanoencapsulation of nutrients, supplements, and pharmaceuticals (Graveland-Bikker and de Kruif 2006 ). Crystalline nanocochleates of about 50 nm in size derived from soya bean can protect micronutrients and antioxidants from degradation during manufacture and storage (BioDelivery Sciences International, Raleigh, NC, USA). Interactive Foods Nanotechnology helps to make interactive foods which can allow consumers to modify the food depending on their own nutritional needs or tastes. The concept of on-demand food states that thousands of nanocapsules containing flavor or color enhancers or added nutritional elements would remain dormant in the food and will only be released when triggered by the consumer (Dunn 2004 ). Kraft, the leader in food industry, has established a consortium called 'Nanotek' to collaborate with universities and research laboratories in USA for developing interactive foods and nanoscale sensors (Forbes 2005 ). Development of foods capable of changing their color, flavor, or nutritional properties according to a person's dietary needs, allergies, or taste preferences is on the research agenda of Nestle and Kraft. Nanotechnology can enable methods to make foods such as soft drinks, ice cream, chocolate, or chips to be marketed as 'health' foods by reducing fat, carbohydrate or calorie content or by increasing protein, fiber or vitamin content. Also, nanotechnology can aid in the production of stronger flavorings, colorings, and nutritional additives, and processing aids to increase the pace of manufacturing and to lower costs of ingredients and processing (Burdo 2005 ). Nanofilters and membranes can screen out or pass through certain molecules based on the shape and/or size to remove toxins or adjust flavors. Nestle and Unilever are reported to be developing a nanoemulsion based ice cream with a lower fat content that retains a fatty texture and flavor (Renton 2006 ). Texture The size and the structure of food influence the functionality of foods by providing the taste, texture, and stability properties that consumers want. Nanotechnology can play a vital role in controlling the size and structure of food to a greater extent. These include healthier foods (lower fat, lower salt) with desirable sensory properties; ingredients with improved properties; and the potential for removal of certain additives without loss of stability, for example in emulsions, and in smart-aids for processing foods to remove allergens such as peanut protein. Scaling down the size of food molecules to nanosized crystals creates more particles for an overall greater surface area. Smaller particles improve food's spreadability and stability, and can aid in developing healthier low-fat food products. Multiple emulsions such as water-in-oil-in-water can distribute the lipids more evenly to reduce extra stabilizers and thickeners to achieve a desirable food texture (Garti and Benichou 2004 ). Bitter blockers (Senomyx, San Diego, USA) prepared from nanoscale assays (Wenner 2008 ) can activate the taste receptors of human tongue and can reduce the bitterness naturally inherent in some foods. A photocatalytic process using nanogold particles (80-120 nm size) was developed by Lin et al. ( 2008 ) for shortening the aging period and enhancing the sensory quality of sorghum spirits. Contreras et al. ( 2009 ) showed that nanozinc can be potentially used to optimize conditions for surface enhancement of infrared absorption of food components. They were able to demonstrate that butter treated with nanozinc particles provided trans fat spectral information along with the degree and the unsaturation of the acyl groups. These results indicate the potential of nanomaterials in imaging to reveal useful information concerning food allergens, bioactive compounds, and microbial pathogens. Safety and Societal Implications The existing safety laws, safety testing methods, and the workplace health procedures are inadequate to measure the exposure and assess the risks posed by nanofoods, nanofood packaging, and nanobased chemicals, as summarized in Table 1 . The nanomaterials used for manufacturing nanofoods and nanopackaging materials are not assessed as new chemicals and currently, the industries follow established guidelines in the safety assessments. Interaction of nanoparticles with living cells and the implications for industry and consumers is not yet understood completely. Regulations governing nanomaterial developments, verification of their safety, fate, and how to dispose them through remediation treatments need to be understood. Experimental studies and research tests should be performed to generate hazard and exposure data leading to risk assessments and to answer concerns about the possible toxicological effects of exposure to nanoparticles in the air pollution. Table 1 Promising nanotechnology applications for food and bioprocessing industries Technology Description Benefits Nanostructures of food ingredients Nanosized ingredients, additives Improved texture, flavor, taste; Reduction in the amount of salt and sugar; enhanced bioavailability Nanoencapsulaton of supplements based on micelles and liposomes Delivery systems for supplements Taste masking; protection from degradation during processing Nanoparticle form of additives and supplements Nano-engineered particulate additives Antimicrobial; health benefits; enhanced bioavailability of nutrients Improved and active nano-composites, intelligent and smart packaging Food packaging Improve flexibility, durability, temperature/moisture stability, barrier properties Nutrient delivery Enzymatic structure, modification, emulsion and foams Targeted delivery of nutrients, increased bioavailability of nutrients Membrane filtration Effective separation of target material from food Higher quality food products and fluids Surface disinfectant Engineering nanoparticles Non-contaminated foods, protection from pathogens Nanoparticle-based intelligent inks; reactive nanolayers Nanolithography depositions Traceability, authentication, prevention of adulteration Acquiring evidence in the nanosafety area is fundamental for the development of proportionate controls and associated legislation. Being a very new technology, lack of data introduces potentially high uncertainty into assessments of environmental risk. The key areas of uncertainty must first be identified, and the strategic approach for addressing and managing these areas of uncertainty needs to be developed. A study from the USA (Monteiller et al. 2007 ) shows that toxicities of nanoparticles and large particles were similar when the dose was expressed in surface area. Hence, the complexity of the nanomaterials behavior in natural systems and the uncertainty introduced for hazard and exposure assessment can be answered by building on experiences with chemicals risk assessment. Toxicological assessment of nanomaterials in food applications by high content screening technique and Zebrafish model can provide valuable developmental toxicity information in terms of endpoint identification and mechanism elucidation (Donofrio 2006 ). Further research into human exposure to nanomaterials and their toxicology and biokinetics is needed. A large number of initiatives have been established, 421 over the last few years, to address potential health and environmental safety issues associated with nanomaterials and nanotechnologies including European and American research projects and networks, International Risk Governance Council projects on nanotechnologies, and European and international standardization activities. Investigations into the health effects of inhaled nanotubes and the surface reactivity and free radical generating potential of nanomaterials are being carried out at National Nanotoxiciology Research Centre, UK. An international approach to regulate the risks from nanomaterials through Organization for Economic Co-operation and Development (OECD) has been developed. This has drawn research work into occupational exposure undertaken by the Health and Safety laboratories as part of multinational programs in the G8 countries (Canada, France, Germany, Italy, Japan, Russia, UK, and USA). The toxicokinetic properties of engineered nanomaterials after oral exposure into the human body should be correlated with their physicochemical properties to determine whether these nanomaterials can be categorized based on appropriate dose metrics (European Food Safety Authority Report 2009). Nanotechnology related terminology and nomenclature and validated measurement and characterization protocols in addressing the nanosafety issues to consider social and ethical concerns and demands are being undertaken by the British Standards Institution, International and European Committee for Standardization, and OECD. Conclusions Influenced by nanotechnology, the food and bioprocessing industry will see great advances with intelligent innovations in the upcoming years and will lead to improved food quality and safety. Nanotechnology has provided sensors and diagnostic devices with improved sensitivity and selectivity to monitor food processes and assure food quality measurements along the production lines. The brand protection and track and trace applications using nanotechnology is mostly confined to the research laboratories and is expected to grow exponentially. Nanotechnology will open up new possibilities in controlling structural changes in the food product and might permit decrease in the power consumption for food production and processing. Nanotechnology offers intriguing opportunities for research in food nanoscience and provides new chances for innovation with tremendous possibilities in bringing solutions for the food and bioprocessing industry. Knowledge gap in addressing and framing the regulations of nanotechnology usage for foods, food additives, and food packaging materials is underway through various regional and international agencies. The success of nanotechnology in the food and bioprocessing industry depends on the perception of consumers and societal acceptance.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2689355/

Immune Modulation by Adjuvants Combined with Diphtheria Toxoid Administered Topically in BALB/c Mice After Microneedle Array Pretreatment

Purpose In this study, modulation of the immune response against diphtheria toxoid (DT) by various adjuvants in transcutaneous immunization (TCI) with microneedle array pretreatment was investigated. Methods TCI was performed on BALB/c mice with or without microneedle array pretreatment using DT as a model antigen co-administrated with lipopolysaccharide (LPS), Quil A, CpG oligo deoxynucleotide (CpG) or cholera toxin (CT) as adjuvant. The immunogenicity was evaluated by measuring serum IgG subtype titers and neutralizing antibody titers. Results TCI with microneedle array pretreatment resulted in a 1,000-fold increase of DT-specific serum IgG levels as compared to TCI. The immune response was further improved by co-administration of adjuvants, showing a progressive increase in serum IgG titers when adjuvanted with LPS, Quil A, CpG and CT. IgG titers of the CT-adjuvanted group reached levels comparable to those obtained after DT-alum subcutaneous injection. The IgG1/IgG2a ratio of DT-specific antibodies decreased in the following sequence: plain DT, Quil A, CT and CpG, suggesting that the immune response was skewed towards the Th1 direction. Conclusions The potency and the quality of the immune response against DT administered by microneedle array mediated TCI can be modulated by co-administration of adjuvants. Purpose In this study, modulation of the immune response against diphtheria toxoid (DT) by various adjuvants in transcutaneous immunization (TCI) with microneedle array pretreatment was investigated. Methods TCI was performed on BALB/c mice with or without microneedle array pretreatment using DT as a model antigen co-administrated with lipopolysaccharide (LPS), Quil A, CpG oligo deoxynucleotide (CpG) or cholera toxin (CT) as adjuvant. The immunogenicity was evaluated by measuring serum IgG subtype titers and neutralizing antibody titers. Results TCI with microneedle array pretreatment resulted in a 1,000-fold increase of DT-specific serum IgG levels as compared to TCI. The immune response was further improved by co-administration of adjuvants, showing a progressive increase in serum IgG titers when adjuvanted with LPS, Quil A, CpG and CT. IgG titers of the CT-adjuvanted group reached levels comparable to those obtained after DT-alum subcutaneous injection. The IgG1/IgG2a ratio of DT-specific antibodies decreased in the following sequence: plain DT, Quil A, CT and CpG, suggesting that the immune response was skewed towards the Th1 direction. Conclusions The potency and the quality of the immune response against DT administered by microneedle array mediated TCI can be modulated by co-administration of adjuvants. INTRODUCTION Currently, most vaccines are administrated by injection. Costs, risks and discomforts associated with the use and abuse of needles have boosted research on needle-free vaccinations ( 1 ). Around 10 years ago, Glenn, G.M et al. reported for the first time data on transcutaneous immunization (TCI) and showed that strong immune responses could be induced by topically applied cholera toxin ( 2 ). TCI is particularly attractive because of the high accessibility of the skin and the presence of antigen-presenting cells (APCs) in the epidermis and dermis, in particular the Langerhans cells (LCs) and the dermal dendritic cells (DCs) ( 3 ). However, the upper layer of the skin, the stratum corneum, acts as a barrier for diffusion of macromolecules and therefore is a major obstacle to dermal vaccine delivery. To overcome this barrier and achieve effective TCI, physical methods such as intradermal injection ( 4 , 5 ), thermal ablation ( 6 ), microdermabrasion ( 7 ) electroporation ( 8 ) and cavitational ultrasound ( 9 ) have been used. Physical disruption of the skin barrier increases the percutanous penetration of the antigen and makes the antigen more readily available for sampling by APCs ( 10 ). Moreover, disruption of the skin barrier may induce a chain of molecular events that lead to the secretion of pro-inflammatory cytokines and facilitate APC activation. A relatively novel approach to disrupt the skin barrier in a controlled manner with little pain sensation is the use of microneedle arrays. It was proposed first by Gerstel and Place already in the 1970s ( 11 ). Ten years ago, when the technology for fabrication in micron dimensions became readily available, Prausnitz resumed the study using microneedle arrays in transdermal drug delivery ( 12 ). When used as a pretreatment, microneedle arrays enable antigens to diffuse along the transiently formed tiny conduits through the stratum corneum. Thereby antigens may be able to approach the LCs in the epidermis and the DCs in the dermis ( 13 ). Using ovalbumin-coated microneedle array, Matriano et al. evaluated the uniformity of skin piercing, and studied the dose of the vaccine used and the kinetics and magnitude of antibody titers induced in hairless guinea pigs ( 14 ). Widera et al . investigated the influence of important fabrication parameters, e.g. length and density of the microneedles, area and coating of the microneedle arrays, on vaccination efficiency of TCI ( 15 ). Hooper et al. reported that smallpox DNA vaccine-coated microneedle arrays applied topically in combination with electroporation protected mice against lethal challenge ( 16 ). Recently, Van Damme et al. tested the injectable microneedle array in human volunteers using influenza vaccine, resulting in a comparable seroprotection rate as compared to i.m. injection with 5 fold dose sparing ( 17 ). These interesting results show that TCI using microneedle array is promising. However, after more than 10 years of extensive research ( 18 , 19 ), there continues to be a need for further improvement of microneedle array mediated TCI, e.g. by using potent adjuvants or novel ways of applying the microneedle arrays. Recently a new electric applicator was developed in our lab. It is designed to insert microneedle arrays into the skin with a predetermined velocity and thereby counteracts the elasticity of the skin. This applicator enables us to reproducibly pierce human and mouse skin in vivo with microneedles with a length of 300 μm or less ( 20 ). DT was recruited as a model antigen to evaluate the potential of microneedle array pretreatment in TCI. In a previous study the immunogenicity of topically applied DT was dramatically improved by microneedle array pretreatment as compared to untreated skin ( 21 ). The objective of the present study was to determine the effect of adjuvants on the quantity and quality of the immune response against DT after TCI with microneedle array pretreatment (M-TCI). The Th1/Th2 balance of the immune response depends on several factors including the nature of the antigen and the adjuvant, the delivery route and the targeted APCs, as suggested by the ratio of IgG1/IgG2a antibody titers ( 22 ). The adjuvants included in this study, cholera toxin (CT), lipopolysaccharide (LPS), synthetic oligo deoxynucleotide containing a CpG motif (CpG), immunostimulatory fractions extracted from the bark of the tree Quillaja saponica (Quil A) and aluminum phosphate (alum), differ in their adjuvant mechanism and ability to modulate the immune response (see Table I ) ( 22 ). Immune modulation by these adjuvants was evaluated in (M-)TCI and compared with conventional subcutaneous (s.c.) injection of DT or DT-alum, by measuring serum IgG (subtype) titers and neutralizing antibody titers. Table I Properties of the Adjuvants Employed in the Current Study (Adapted from ( 22 , 47 , 48 )) Adjuvant Type Cell-mediated immunity (Th1) Humoral immunity (Th2) Receptor Cholera toxin Exotoxin + a +++ GM1 ganglioside lpxL1 LPS Endotoxin ++ ++ TLR4 Quil A Saponin based +++ ++ Not identified CpG Bacterial DNA ++++ + TLR9 Alum Inorganic salt + +++ Not identified a Humoral and cellular immunity in arbitrary units represent the ability of adjuvants to enhance Th2 response or CTL or Th1, respectively, to foreign antigens MATERIALS AND METHODS Materials Diphtheria toxin (batch 79/1), diphtheria toxoid (batch 98/40, protein content 12.6 mg/ml by BCA assay, 1 μg equal to approximately 0.3 Lf) and the lpxL1 LPS were provided by the Netherlands Vaccine Institute (NVI, Bilthoven, The Netherlands). Horseradish peroxidase-conjugated goat anti-mouse (HRP-GAM) IgG (γ-chain specific), IgG1 (γ1-chain specific) and IgG2a (γ2a-chain specific) were purchased from Southern Biotech (Birmingham, USA). Quil A and Adju-Phos® (alum) were obtained from Brenntag Biosector (Copenhagen, Denmark). CpG oligo deoxynucleotide 1826 (5′-tcc atg acg ttc ctg acg tt-3′, phosphorothioated) was synthesized by Isogen Biosolutions (IJsselstein, The Netherlands). Chromogen 3, 3′, 5, 5′-tetramethylbenzidine (TMB) and the substrate buffer were purchased from Biosource B.V. (Nivelles, Belgium). Cholera toxin was ordered from Sigma-Aldrich (Zwijndrecht, The Netherlands). Nimatek® (100 mg/ml Ketamine, Euovet Animal Health B.V., Bladel, The Netherlands), Rompun® (20 mg/ml Xylasine, Bayer B.V., Mijdrecht, The Netherlands) and the injection fluid (0.9% NaCl) were obtained from a local pharmacy. All other chemicals used were of analytical grade and all solutions were prepared with distilled water. Animals Female BALB/c mice (H2d), 8-week old at the start of the experiment were purchased from Charles River (Maastricht, The Netherlands), and maintained under standardized conditions in the animal facility of the Leiden/Amsterdam Center for Drug Research, Leiden University. The study was conducted in conformity with the Public Health Service Policy on use of laboratory animals and had been approved by the Research Ethical Committee of Leiden University (UDEC, nr. 07016). Methods Microneedle Array and Applicator Assembled microneedle arrays were manufactured from commercially available 30G hypodermic needles (BD, Alphen a/d Rijn, The Netherlands) as described previously ( 13 ). The needles were assembled as a 300 μm-long, 4 × 4 array on a polymer back plate with a surface area of about 0.5 cm 2 . The electric applicator was developed and optimized as reported previously ( 21 ). The microneedles were pierced into mouse skin using a velocity of 3 m/s. Immunization Study The DT-alum formulations were prepared as previously described and the adsorption of DT to alum was between 70% and 80% ( 23 ). As control groups, 5 μg of DT (∼1.5 Lf) with and without alum in 100 μl solution/suspension was administered per mouse by s.c injection. The other vaccine-adjuvant formulations were freshly prepared by mixing DT and the adjuvants in buffer solution in appropriate amounts as indicated in Table II . One hundred micrograms DT and adjuvants per mouse were applied on intact skin (TCI) or on microneedle array pretreated skin (M-TCI) as previously described ( 21 , 24 ). A mutant of LPS, lpxL1 LPS, with reduced toxicity but retained adjuvanticity, was employed in this study ( 25 ). The dose of alum and lpxL1 LPS used in M-TCI were based on the w / w ratio of antigen/adjuvant used in injection control group and previous immunization study ( 25 ), respectively. Table II Formulations Prepared for In Vivo (M-)TCI Study DT dose (μg) Adjuvants Adjuvant dose (μg) Solvent (pH) Volume (μl) for s.c. injection 5 AlPO 4 150 0.9% saline (7.0) 100 5 – – 0.9% saline (7.0) 100 (M)-TCI 100 – – PBS a (7.4) 70 100 lpxL1 LPS 25 PBS/Tris b (7.4) 70 100 Quil A 100 PBS (7.4) 70 100 CT 100 PBS (7.4) 70 100 CpG 100 PBS (7.4) 70 100 AlPO 4 3,000 0.9% saline (7.0) 70 a PBS: 2.67 mM KCl, 1.47 mM KH 2 PO 4 , 137.93 mM NaCl, 8.06 mM Na 2 HPO 4 ·7H 2 O, without Ca 2 + and Mg 2+ b PBS/Tris: PBS mixed with 1 mM Tris–HCl ( v / v = 5:3). During vaccination, mice were anaesthetized by intraperitoneal injection of 150 mg/kg ketamine and 10 mg/kg xylazine. For all groups receiving transcutaneous vaccination, the abdominal skin of the mice, shaved 24 h prior to vaccination, was first wiped with 70% ethanol. For M-TCI group, a skin fold was supported by styrofoam and pierced using the microneedle array and the electric applicator. Then 70 μl DT-adjuvant formulation was carefully spread to wet the entire skin area of application (microneedle array-pretreated or untreated, ∼2 cm 2 area restricted by a metal ring). After 1 h of occlusive incubation, the skin area was extensively washed with lukewarm tap water and patted dry twice. All mice were immunized three times on day 1, 21 and 42 (at approximately the same skin region for all the TCI groups) and sacrificed on day 56. Blood was sampled from the tail vein one day before each immunization and the whole blood was collected from the femoral artery during sacrifice. Cell free sera were obtained using MiniCollect ® tubes (Greiner bio-one, Alphen a/d Rijn, The Netherlands) by centrifugation after clot formation and stored at −80°C. Serum Antibody Assay Serum IgG, IgG1 and IgG2a titers were determined by sandwich ELISA as previously described ( 26 ). Briefly, ELISA plate (Microlon®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands) wells were coated with DT at 4°C overnight. Two-fold serial dilutions of serum samples were applied in the plates and the containing DT-specific antibodies were detected by HRP-GAM IgG, IgG1 or IgG2a using TMB as substrate. Antibody titers are expressed as the reciprocal of the calculated sample dilution corresponding to half of the maximum absorbance at 450 nm of a complete s-shaped absorbance-log dilution curve. If samples were not diluted in the optimal range, additional measurements were performed to generate an s-shaped curve. Subsequently, the titers were calculated using a four-parameter fitting of the curve. Samples that did not reach the half-saturated absorbance value at the lowest (ten fold) dilution were considered as non-responders. Neutralizing Antibody Assay Immunity against diphtheria depends on the presence of circulating toxin-neutralizing antibodies. These antibodies were evaluated using Vero cell test, the WHO standard method to assess the success of diphtheria vaccination, which relies on the inhibition of a cytotoxic dose of diphtheria toxin ( 26 ). In brief, after complement inactivation, two-fold serial dilutions of serum samples were prepared with complete medium 199 (CM199, Gibco, Breda, The Netherlands) and applied to microtiter plates (CELLSTAR®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands). Subsequently, 2.5 × 10 −5 Lf diphtheria toxin was added to the wells. After 2 h incubation at 37°C for neutralization, Vero cells suspension in CM199 was added to each well. Covered with a plate sealer, Vero cells were incubated at 37°C in 5% CO 2 for 6 days. The end point was taken as the highest serum dilution protecting the Vero cells. Statistical Analysis ELISA titers were logarithmically transformed for better normality before statistical analysis. Two way ANOVA with Bonferroni posttest, one way ANOVA with Tukey posttest, or Kruskal–Wallis nonparametric test with Dunns posttest were performed as indicated. Statistical analysis was carried out using Prism Graphpad and a p value less than 0.05 was considered to be significant. Materials Diphtheria toxin (batch 79/1), diphtheria toxoid (batch 98/40, protein content 12.6 mg/ml by BCA assay, 1 μg equal to approximately 0.3 Lf) and the lpxL1 LPS were provided by the Netherlands Vaccine Institute (NVI, Bilthoven, The Netherlands). Horseradish peroxidase-conjugated goat anti-mouse (HRP-GAM) IgG (γ-chain specific), IgG1 (γ1-chain specific) and IgG2a (γ2a-chain specific) were purchased from Southern Biotech (Birmingham, USA). Quil A and Adju-Phos® (alum) were obtained from Brenntag Biosector (Copenhagen, Denmark). CpG oligo deoxynucleotide 1826 (5′-tcc atg acg ttc ctg acg tt-3′, phosphorothioated) was synthesized by Isogen Biosolutions (IJsselstein, The Netherlands). Chromogen 3, 3′, 5, 5′-tetramethylbenzidine (TMB) and the substrate buffer were purchased from Biosource B.V. (Nivelles, Belgium). Cholera toxin was ordered from Sigma-Aldrich (Zwijndrecht, The Netherlands). Nimatek® (100 mg/ml Ketamine, Euovet Animal Health B.V., Bladel, The Netherlands), Rompun® (20 mg/ml Xylasine, Bayer B.V., Mijdrecht, The Netherlands) and the injection fluid (0.9% NaCl) were obtained from a local pharmacy. All other chemicals used were of analytical grade and all solutions were prepared with distilled water. Animals Female BALB/c mice (H2d), 8-week old at the start of the experiment were purchased from Charles River (Maastricht, The Netherlands), and maintained under standardized conditions in the animal facility of the Leiden/Amsterdam Center for Drug Research, Leiden University. The study was conducted in conformity with the Public Health Service Policy on use of laboratory animals and had been approved by the Research Ethical Committee of Leiden University (UDEC, nr. 07016). Animals Female BALB/c mice (H2d), 8-week old at the start of the experiment were purchased from Charles River (Maastricht, The Netherlands), and maintained under standardized conditions in the animal facility of the Leiden/Amsterdam Center for Drug Research, Leiden University. The study was conducted in conformity with the Public Health Service Policy on use of laboratory animals and had been approved by the Research Ethical Committee of Leiden University (UDEC, nr. 07016). Methods Microneedle Array and Applicator Assembled microneedle arrays were manufactured from commercially available 30G hypodermic needles (BD, Alphen a/d Rijn, The Netherlands) as described previously ( 13 ). The needles were assembled as a 300 μm-long, 4 × 4 array on a polymer back plate with a surface area of about 0.5 cm 2 . The electric applicator was developed and optimized as reported previously ( 21 ). The microneedles were pierced into mouse skin using a velocity of 3 m/s. Immunization Study The DT-alum formulations were prepared as previously described and the adsorption of DT to alum was between 70% and 80% ( 23 ). As control groups, 5 μg of DT (∼1.5 Lf) with and without alum in 100 μl solution/suspension was administered per mouse by s.c injection. The other vaccine-adjuvant formulations were freshly prepared by mixing DT and the adjuvants in buffer solution in appropriate amounts as indicated in Table II . One hundred micrograms DT and adjuvants per mouse were applied on intact skin (TCI) or on microneedle array pretreated skin (M-TCI) as previously described ( 21 , 24 ). A mutant of LPS, lpxL1 LPS, with reduced toxicity but retained adjuvanticity, was employed in this study ( 25 ). The dose of alum and lpxL1 LPS used in M-TCI were based on the w / w ratio of antigen/adjuvant used in injection control group and previous immunization study ( 25 ), respectively. Table II Formulations Prepared for In Vivo (M-)TCI Study DT dose (μg) Adjuvants Adjuvant dose (μg) Solvent (pH) Volume (μl) for s.c. injection 5 AlPO 4 150 0.9% saline (7.0) 100 5 – – 0.9% saline (7.0) 100 (M)-TCI 100 – – PBS a (7.4) 70 100 lpxL1 LPS 25 PBS/Tris b (7.4) 70 100 Quil A 100 PBS (7.4) 70 100 CT 100 PBS (7.4) 70 100 CpG 100 PBS (7.4) 70 100 AlPO 4 3,000 0.9% saline (7.0) 70 a PBS: 2.67 mM KCl, 1.47 mM KH 2 PO 4 , 137.93 mM NaCl, 8.06 mM Na 2 HPO 4 ·7H 2 O, without Ca 2 + and Mg 2+ b PBS/Tris: PBS mixed with 1 mM Tris–HCl ( v / v = 5:3). During vaccination, mice were anaesthetized by intraperitoneal injection of 150 mg/kg ketamine and 10 mg/kg xylazine. For all groups receiving transcutaneous vaccination, the abdominal skin of the mice, shaved 24 h prior to vaccination, was first wiped with 70% ethanol. For M-TCI group, a skin fold was supported by styrofoam and pierced using the microneedle array and the electric applicator. Then 70 μl DT-adjuvant formulation was carefully spread to wet the entire skin area of application (microneedle array-pretreated or untreated, ∼2 cm 2 area restricted by a metal ring). After 1 h of occlusive incubation, the skin area was extensively washed with lukewarm tap water and patted dry twice. All mice were immunized three times on day 1, 21 and 42 (at approximately the same skin region for all the TCI groups) and sacrificed on day 56. Blood was sampled from the tail vein one day before each immunization and the whole blood was collected from the femoral artery during sacrifice. Cell free sera were obtained using MiniCollect ® tubes (Greiner bio-one, Alphen a/d Rijn, The Netherlands) by centrifugation after clot formation and stored at −80°C. Serum Antibody Assay Serum IgG, IgG1 and IgG2a titers were determined by sandwich ELISA as previously described ( 26 ). Briefly, ELISA plate (Microlon®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands) wells were coated with DT at 4°C overnight. Two-fold serial dilutions of serum samples were applied in the plates and the containing DT-specific antibodies were detected by HRP-GAM IgG, IgG1 or IgG2a using TMB as substrate. Antibody titers are expressed as the reciprocal of the calculated sample dilution corresponding to half of the maximum absorbance at 450 nm of a complete s-shaped absorbance-log dilution curve. If samples were not diluted in the optimal range, additional measurements were performed to generate an s-shaped curve. Subsequently, the titers were calculated using a four-parameter fitting of the curve. Samples that did not reach the half-saturated absorbance value at the lowest (ten fold) dilution were considered as non-responders. Neutralizing Antibody Assay Immunity against diphtheria depends on the presence of circulating toxin-neutralizing antibodies. These antibodies were evaluated using Vero cell test, the WHO standard method to assess the success of diphtheria vaccination, which relies on the inhibition of a cytotoxic dose of diphtheria toxin ( 26 ). In brief, after complement inactivation, two-fold serial dilutions of serum samples were prepared with complete medium 199 (CM199, Gibco, Breda, The Netherlands) and applied to microtiter plates (CELLSTAR®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands). Subsequently, 2.5 × 10 −5 Lf diphtheria toxin was added to the wells. After 2 h incubation at 37°C for neutralization, Vero cells suspension in CM199 was added to each well. Covered with a plate sealer, Vero cells were incubated at 37°C in 5% CO 2 for 6 days. The end point was taken as the highest serum dilution protecting the Vero cells. Statistical Analysis ELISA titers were logarithmically transformed for better normality before statistical analysis. Two way ANOVA with Bonferroni posttest, one way ANOVA with Tukey posttest, or Kruskal–Wallis nonparametric test with Dunns posttest were performed as indicated. Statistical analysis was carried out using Prism Graphpad and a p value less than 0.05 was considered to be significant. Microneedle Array and Applicator Assembled microneedle arrays were manufactured from commercially available 30G hypodermic needles (BD, Alphen a/d Rijn, The Netherlands) as described previously ( 13 ). The needles were assembled as a 300 μm-long, 4 × 4 array on a polymer back plate with a surface area of about 0.5 cm 2 . The electric applicator was developed and optimized as reported previously ( 21 ). The microneedles were pierced into mouse skin using a velocity of 3 m/s. Immunization Study The DT-alum formulations were prepared as previously described and the adsorption of DT to alum was between 70% and 80% ( 23 ). As control groups, 5 μg of DT (∼1.5 Lf) with and without alum in 100 μl solution/suspension was administered per mouse by s.c injection. The other vaccine-adjuvant formulations were freshly prepared by mixing DT and the adjuvants in buffer solution in appropriate amounts as indicated in Table II . One hundred micrograms DT and adjuvants per mouse were applied on intact skin (TCI) or on microneedle array pretreated skin (M-TCI) as previously described ( 21 , 24 ). A mutant of LPS, lpxL1 LPS, with reduced toxicity but retained adjuvanticity, was employed in this study ( 25 ). The dose of alum and lpxL1 LPS used in M-TCI were based on the w / w ratio of antigen/adjuvant used in injection control group and previous immunization study ( 25 ), respectively. Table II Formulations Prepared for In Vivo (M-)TCI Study DT dose (μg) Adjuvants Adjuvant dose (μg) Solvent (pH) Volume (μl) for s.c. injection 5 AlPO 4 150 0.9% saline (7.0) 100 5 – – 0.9% saline (7.0) 100 (M)-TCI 100 – – PBS a (7.4) 70 100 lpxL1 LPS 25 PBS/Tris b (7.4) 70 100 Quil A 100 PBS (7.4) 70 100 CT 100 PBS (7.4) 70 100 CpG 100 PBS (7.4) 70 100 AlPO 4 3,000 0.9% saline (7.0) 70 a PBS: 2.67 mM KCl, 1.47 mM KH 2 PO 4 , 137.93 mM NaCl, 8.06 mM Na 2 HPO 4 ·7H 2 O, without Ca 2 + and Mg 2+ b PBS/Tris: PBS mixed with 1 mM Tris–HCl ( v / v = 5:3). During vaccination, mice were anaesthetized by intraperitoneal injection of 150 mg/kg ketamine and 10 mg/kg xylazine. For all groups receiving transcutaneous vaccination, the abdominal skin of the mice, shaved 24 h prior to vaccination, was first wiped with 70% ethanol. For M-TCI group, a skin fold was supported by styrofoam and pierced using the microneedle array and the electric applicator. Then 70 μl DT-adjuvant formulation was carefully spread to wet the entire skin area of application (microneedle array-pretreated or untreated, ∼2 cm 2 area restricted by a metal ring). After 1 h of occlusive incubation, the skin area was extensively washed with lukewarm tap water and patted dry twice. All mice were immunized three times on day 1, 21 and 42 (at approximately the same skin region for all the TCI groups) and sacrificed on day 56. Blood was sampled from the tail vein one day before each immunization and the whole blood was collected from the femoral artery during sacrifice. Cell free sera were obtained using MiniCollect ® tubes (Greiner bio-one, Alphen a/d Rijn, The Netherlands) by centrifugation after clot formation and stored at −80°C. Serum Antibody Assay Serum IgG, IgG1 and IgG2a titers were determined by sandwich ELISA as previously described ( 26 ). Briefly, ELISA plate (Microlon®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands) wells were coated with DT at 4°C overnight. Two-fold serial dilutions of serum samples were applied in the plates and the containing DT-specific antibodies were detected by HRP-GAM IgG, IgG1 or IgG2a using TMB as substrate. Antibody titers are expressed as the reciprocal of the calculated sample dilution corresponding to half of the maximum absorbance at 450 nm of a complete s-shaped absorbance-log dilution curve. If samples were not diluted in the optimal range, additional measurements were performed to generate an s-shaped curve. Subsequently, the titers were calculated using a four-parameter fitting of the curve. Samples that did not reach the half-saturated absorbance value at the lowest (ten fold) dilution were considered as non-responders. Neutralizing Antibody Assay Immunity against diphtheria depends on the presence of circulating toxin-neutralizing antibodies. These antibodies were evaluated using Vero cell test, the WHO standard method to assess the success of diphtheria vaccination, which relies on the inhibition of a cytotoxic dose of diphtheria toxin ( 26 ). In brief, after complement inactivation, two-fold serial dilutions of serum samples were prepared with complete medium 199 (CM199, Gibco, Breda, The Netherlands) and applied to microtiter plates (CELLSTAR®, Greiner Bio-one, Alphen a/d Rijn, The Netherlands). Subsequently, 2.5 × 10 −5 Lf diphtheria toxin was added to the wells. After 2 h incubation at 37°C for neutralization, Vero cells suspension in CM199 was added to each well. Covered with a plate sealer, Vero cells were incubated at 37°C in 5% CO 2 for 6 days. The end point was taken as the highest serum dilution protecting the Vero cells. Statistical Analysis ELISA titers were logarithmically transformed for better normality before statistical analysis. Two way ANOVA with Bonferroni posttest, one way ANOVA with Tukey posttest, or Kruskal–Wallis nonparametric test with Dunns posttest were performed as indicated. Statistical analysis was carried out using Prism Graphpad and a p value less than 0.05 was considered to be significant. RESULTS During the immunization study, there was no adverse effect from the shaving, anesthesia, piercing, immunization, or washing procedure observed. Neither erythema nor induration was seen at the immunization site after exposure of antigen and adjuvants. Immune Response Improved by Microneedle Array Pretreatment The IgG titers of mice from all groups after prime, first boost and second boost (day 20, 41 and 56) are shown in Fig. 1 a–e and an overview of the DT-IgG titers after the second boost is shown in Fig. 1 f. It is clear that non-adjuvanted DT did not result in a substantial IgG response when applied on intact skin (TCI). Microneedle array pretreatment provided major improvement to the immunogenicity of DT. After prime, more mice responded with higher mean IgG titers in the M-TCI group than in the TCI group. After the first boost, all mice responded and the IgG titers were 100 fold higher in the M-TCI group ( p  0.05). For the adjuvants in the absence of microneedle array pretreatment, their effects on the DT immunogenicity were more variable and less pronounced. Only CT moderately improved the IgG titers after the second boost (Fig. 1 a, e, p  0.05). As expected, the groups that had received formulations on untreated skin (TCI) did not show detectable neutralizing antibody titers. Fig. 4. Diphtheria toxin-neutralizing antibody titers after M-TCI. Serum samples were collected after the second boost (day 56) and determined by Vero cell test. Data are expressed as logarithm of the highest dilution that was still capable of protecting the Vero cells from the challenge of diphtheria toxin ( filled circle the M-TCI groups, empty circle the injection groups, * p  0.05). For the adjuvants in the absence of microneedle array pretreatment, their effects on the DT immunogenicity were more variable and less pronounced. Only CT moderately improved the IgG titers after the second boost (Fig. 1 a, e, p  0.05). As expected, the groups that had received formulations on untreated skin (TCI) did not show detectable neutralizing antibody titers. Fig. 4. Diphtheria toxin-neutralizing antibody titers after M-TCI. Serum samples were collected after the second boost (day 56) and determined by Vero cell test. Data are expressed as logarithm of the highest dilution that was still capable of protecting the Vero cells from the challenge of diphtheria toxin ( filled circle the M-TCI groups, empty circle the injection groups, * p  0.05 after the second boost). The ratios of anti-DT IgG1/IgG2a titers were significantly lowered as compared to M-TCI of plain DT ( p  0.05). This may be due to the following factors: (a) LCs do not express TLR4 and do not respond to bacteria or LPS ( 39 ). TLR4 expression by mouse keratinocytes also seems limited. Besides few data available from mice, reports on TLR4 expression by human keratinocytes are conflicting ( 40 , 41 ); (b) Although little is known about the actual diffusion of these dispersions across the conduits induced by the microneedles, previous studies of our group showed that fluorescently labeled particles of ca. 200 nm did pass through the conduits created by the same microneedle arrays as used in the present study ( 20 ). However, as no quantitative information is available, it is possible that only a limited amount of the DT- lpxL1 LPS dispersion reached the DCs in the dermis; (c) the LPS mutant may work in a less efficient way as the original LPS in the epidermal/dermal microenvironment. Further studies are required to elucidate the exact mechanism. The different reactivity of skin immune system to CpG and LPS may be involved in a strategic control of host defense to the bacterial commensal skin flora ( 39 ). Alum did not show any adjuvanticity in M-TCI. It induced lower IgG and IgG1 titers than those of non-adjuvanted DT group after each vaccination. Furthermore, alum in M-TCI did not induce detectable IgG2a titers and provided no protection in the Vero cell test after the second boost (data not shown). This may partially be due to the size of DT-alum particles, several microns in diameter, which prevents their diffusion through the conduits in sufficient amount to exert a 'depot' effect ( 42 ). The adjuvanticity of alum could also be dependent on the epithelial microenvironment; the danger signal it induces through injection, the uric acid production ( 43 ), may not be provoked in the dermis and epidermis. Interestingly, when comparing IgG subtype titers and neutralizing antibody titers between CpG and Quil A groups (Figs. 1 and 4 ), the CpG group, with similar IgG titers, but much higher IgG2a titers than the other, induced similar to slightly lower neutralizing antibody titers, which indicates that IgG2a might not contribute to toxin-neutralization. This was confirmed by the comparison between CpG and CT groups, also showing that higher IgG2a titers did not result in higher neutralizing antibody titers. IgG1 seems to be the main neutralizing antibody for protection, as IgG1 titers significantly correlated with neutralizing titers. However, IgG1 titers could not be used to predict individual neutralizing antibody titers with great accuracy, which is in line with a diphtheria vaccination study done in human infants ( 44 ). Therefore, from an application perspective, CpG in M-TCI is more suitable for anti-viral immune responses such as influenza vaccination where IgG2a provides the main immune protection and a Th1 biased response is more desired ( 45 ). In this study, a high dose of DT, 20 times of that used in the s.c. injection group, was applied in M-TCI. This dose is similar to the dose of DT used by Glenn et al. in previous studies ( 24 , 46 ). High doses were used as apparently only a fraction of DT applied actually enters epidermis/dermis. A limiting factor for diffusion of the antigen into the skin is that the mice used for this study can only be kept under anesthesia for about 1 h. This is not a limitation of the transcutaneous route, but a limitation of the animal model used. The high doses used until now also indicate that formulation improvement to increase the efficiency of the immune response is very important in transcutaneous vaccination. Conjugation and encapsulation of adjuvants into vaccine-containing particles and specific DC and LC targeting approaches could be attractive strategies to further improve the potency of vaccine formulations for M-TCI. CONCLUSION We have shown that the type of adjuvant used has a significant effect on the immune response and protective immunity against DT in M-TCI. The epithelial microenvironment and DC heterogeneity also play an important role in the regulation of the immune response. This delivery method is applicable to many other vaccines by formulating with proper adjuvants and holds a lot of promise for future use.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5970133/

Isolation and Identification of an Anthracimycin Analogue from Nocardiopsis kunsanensis , a Halophile from a Saltern, by Genomic Mining Strategy

Modern medicine is unthinkable without antibiotics; yet, growing issues with microbial drug resistance require intensified search for new active compounds. Natural products generated by Actinobacteria have been a rich source of candidate antibiotics, for example anthracimycin that, so far, is only known to be produced by Streptomyces species. Based on sequence similarity with the respective biosynthetic cluster, we sifted through available microbial genome data with the goal to find alternative anthracimycin-producing organisms. In this work, we report about the prediction and experimental verification of the production of anthracimycin derivatives by Nocardiopsis kunsanensis , a non- Streptomyces actinobacterial microorganism. We discovered N. kunsanensis to predominantly produce a new anthracimycin derivative with methyl group at C-8 and none at C-2, labeled anthracimycin BII-2619, besides a minor amount of anthracimycin. It displays activity against Gram-positive bacteria with similar low level of mammalian cytotoxicity as that of anthracimycin. Introduction The biosynthetic gene cluster of the macrolide anthracimycin ( atc ) has been identified in less than a handful of bacterial strains belonging to Streptomyces sp. It was not long ago that it was first reported for the marine strain Streptomyces sp. T676 isolated off St. John's Island in Singapore and described as a type I trans -AT PKS. 1 . In contrast to cis -AT PKSs, trans -AT PKS enzymes show an amount of architectural diversity that makes the cis -AT co-linearity rule not directly applicable. For instance, they can have a cross-module domain mode of action, unique domains, split modules between two proteins, despite their trans -acting functionality and other additional unusual features 2 . In this regard, evolution by horizontal gene transfer and extensive recombination of PKS gene fragments seem to be a preferred mode of action by trans -AT PKSs 2 , 3 . Historically, the atc compound ( 1 in Figure 1 ) was discovered by MerLion Pharmaceuticals Pte Ltd as a result of a high-throughput screening for new antibacterials (unpublished data) in 2001. However, it was only in 2013 that its chemical structure was reported, when the compound was isolated from Streptomyces sp. CNH365 4 , a marine strain found in sediments near-shore off Santa Barbara, CA. In 2014, the year before the atc cluster was reported, the genome of Streptomyces sp. NRRL F-5065 became available (accession no. JOHV00000000) 5 . In 2015, the draft genome of another anthracimycin producing strain, Streptomyces sp. TP-A0875, isolated from a compost sample in Ishikawa, Japan, was published 6 , 7 . In 2008, the structure of chlorotonil A ( 3 in Figure 1 ), a compound isolated from the myxobacterium Sorangium cellulosum has been reported 8 . Both anthracimycin and chlorotonil A are tricyclic metabolites with a similar carbon skeleton; chlorotonil A seems to be almost the optical isomer of anthracimycin 6 . Its difference from anthracimycin includes its unique gem-dichloro-1,3-dione moiety and a methyl group in position C-8. The biosynthetic gene cluster for chlorotonil A has been published at the same time as the one for anthracimycin and was shown to also produce new chlorotonil congeners. Its cluster organization, as well as its proposed pathway, showed some striking differences to that of anthracimycin 9 . The recent increase in attention to anthracimycin is due to its strong activity against Gram-positive bacteria, including some drug-resistant strains such as Bacillus anthracis , methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE), its rapid bacterial killing kinetics and in vivo efficacy in the murine peritonitis model 4 , 10 . Chlorotonil A has most notably shown potent bioactivity against Plasmodium falciparum , one of the malaria pathogens 11 . Thus, the search for different organisms producing the same compound or its analogues is of great importance for future advances, as different organisms might lead to greater compound production yields or generate other biologically effective derivatives. We searched for microbial genomes that contain sequence segments similar to the known atc biosynthetic gene cluster in Streptomyces . In this work, we report the prediction of a potential anthracimycin derivative-producing gene cluster in Nocardiopsis kunsanensis , an actinobacterial microorganism belonging to Streptosporangiales, a different order compared to the reported anthracimycin-producing Streptomyces . Most interestingly, experimental verification of this finding revealed that N. kunsanensis predominantly synthesized a new anthracimycin derivative and only a very small amount of anthracimycin. After the elucidation of its structure, we named this new analogue compound anthracimycin BII-2619 (see 2 , Figure 1 ). We also compared its biological activity with anthracimycin. Methods Bioinformatics search methods: prediction of the anthracimycin ( atc ) gene cluster in Streptosporangiales In order to identify other potential anthracimycin producing microorganisms, we first focused on the sequences of the functional domains as defined for Streptomyces sp. T676 in the public database (accession no. LN871452). For this entry, there are nine polyketide synthase (PKS) gene products (atcA, atcB, atcC, atcD, atcE, atcF, atcG, atcH, atcI), where four of them (atcC-atcF) have been proposed to be functionally involved in the generation of anthracimycin 1 . These four products were further split into modules according to the definition of functional modules commonly present in biosynthetic megaenzymes, where independently folding protein domains are joined together via so-called 'linker' regions 12 . The boundaries of these domains were annotated in the entry as misc_feature . For the atcD gene product, an extra feature was annotated as " repeated sequence; additional in PacBio sequence ". This region of 62 residues is rich in low complexity, as seen in the ANNOTATOR platform 13 using a consensus disorder prediction 14 , and so was the only region not included in this analysis. All other 41 domains were blasted and psi-blasted 15 against nr (Dec 2015) 16 , so as to identify species that shared the highest number of domains significantly similar to the ones in the atc cluster. 1292 and 2000 organisms were found by blast and psi-blast (E-value cut off 0.001), respectively, to produce at least 1 significant domain hit. In this work, we limited ourselves to display the top strains that displayed 39 or more hits to all domains present in the atc cluster (Supplementary Table S1 ). In order to identify if the protein domains found in N. kunsanensis would have the architecture of a biosynthetic gene cluster, tblastn was used with the annotated T676 protein domains as queries, to map them in the N. kunsanensis nucleotide sequence. All the protein domains had a significant hit (E-value cut off 0.001) and could be mapped to N. kunsanensis DSM 44524 contig 32 (accession no. NZ_ANAY01000032) (Figure 2 ). All the regions plotted in Figure 2 refer to the first tblastn hit result, with the exception of domains DH5 (in AtcE) and KR8 (in AtcF) that were the 2 nd and 3 rd best hits, respectively. Sequence alignment and phylogenetic analysis Each one of the four functional proteins in the cluster (atcC, atcD, atcE and atcF) had their amino acid sequences aligned among the five strains with the MAFFT E-INS-i algorithm 17 . Individual phylogenetic trees were calculated with MEGA6 18 for each of the four aligned proteins. These included Neighbor-Joining and Maximum Likelihood with and without bootstrap methods (data not shown). Given that there was no observed difference among the generated trees for each of these four proteins, regardless of the method, these aligned sequences were intentionally merged so as to produce a general picture of the evolutionary relationship of these four functional proteins among the five strains that reflects their individual protein trees (Figure 3 ). Bacterial strain acquisition and fermentation conditions Nocardiopsis kunsanensis DSM 44524 was purchased from Leibniz-Institut DSMZ - Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) 19 . Nocardiopsis kunsanensis DSM 44524 was grown on Complex Medium (CM) agar for 7 days at 28°C and was used to prepare 50 mL seed culture in CM (7.5 g casamino acids, 10 g yeast extract, 3 g sodium citrate, 10 g magnesium sulfate, 2 g potassium chloride, 1 mL 4.98% iron sulfate, 50 g NaCl, 1 L distilled water at pH 7.4 20 ). After 7 days cultivation at 28°C on a rotary shaker at 200 rpm, 0.5 mL of the seed culture was used to inoculate 50 mL of CM, CA02LB (20 g mannitol, 20 g soybean meal, 1 L distilled water at pH 7.5) or CA07LB (15 g Glycerol, 30 g oatmeal, 5 g yeast extract, 5 g KH 2 PO 4 , 5 g Na 2 HPO 4 .12H 2 O, 1 g MgCl 2 .6H 2 O, 1 L distilled water at its natural pH) in 250 mL Erlenmeyer flasks. All media was supplemented with 5-10% NaCl (w/v) when necessary. The fermentation was carried out for 7 days at 28 °C at 200 rpm. Chemicals and reagents for media preparation were purchased from Sigma. Extraction and isolation For the initial chemical analysis the entire 50 ml cultures of N. kunsanensis grown in various media were freeze-dried before extracting with methanol for LCMS analysis. For the scale up fermentation for compound isolation, the cultures (80 × 50 mL, total 4 L) of N. kunsanensis from all flasks were first combined and centrifuged to separate the culture supernatant and the mycelia. The combined mycelia were freeze-dried, shaken overnight 2 times with MeOH/ CH 2 Cl 2 (4 L), filtered and added to 1 L of water. The organic layer from the solvent partition was evaporated to dryness using rotary evaporation. The dried dichloromethane extract was dissolved in DMSO and separated by C 18 reversed-phase preparative HPLC (solvent A: H 2 O + 0.1% HCOOH, solvent B: ACN + 0.1% HCOOH; flow rate: 18 mL/min, gradient conditions:70:30 isocratic for 5 minutes; 30% to 40% of solvent B over 5 minutes, followed by 40% to 70% of solvent B over 40 minutes, and an increase from 70% to 100% of solvent B over 50 minutes, and to 100% solvent B in 15 minutes) to give 0.2 mg of anthracimycin ( 1 ) with impurities and 0.9 mg of anthracimycin BII-2619 ( 2 ). Chemical Structural Data The NMR spectra of the compounds are provided in Supplementary Figures S1 (S1a-S1d) to S2 (S2a-S2f). Below, we list the chemical structural data including the measured molar circular dichroism (Δε). Anthracimycin standard Amorphous white powder; [α] D -7 ( c 2.93, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.21) and 286 (3.76) nm; ECD (MeOH) λ (Δε) 202 (+40.47), 214 (+12.96), 236 (+34.84), 283 (-28.84); HRESIMS m/ z 397.2374 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Anthracimycin BII-2619 (2) Amorphous white powder; [α] D -47 ( c 0.16, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.17) and 286 (3.62) nm; ECD (MeOH) λ (Δε) 201 (+30.64), 213 (+7.53), 236 (+18.14), 283 (-15.98); HRESIMS m/ z 397.2371 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Analytical Chemistry Procedures Optical rotations were recorded on a JASCO P-2000 digital polarimeter. Electronic circular dichroism (ECD) was measured on a JASCO J815 - circular dichroism (CD) spectrometer. The CD spectra of all samples were recorded in methanol at molar concentration of 0.025 mM in 195-400 nm region. UV spectra were obtained on a General Electric Ultrospec 9000 spectrophotometer. NMR spectra were collected on a Bruker DRX-400 NMR spectrometer with Cryoprobe, using 5-mm BBI ( 1 H, G-COSY, multiplicity-edited G-HSQC, and G-HMBC spectra) or BBO ( 13 C spectra) probe heads equipped with z-gradients. Spectra were calibrated to residual protonated solvent signals (C H Cl 3 δ H 7.24 and C DCl 3 δ C 77.23). The HRESIMS and MS/MS spectra were acquired on Agilent UHPLC 1290 Infinity coupled to Agilent 6540 accurate-mass quadrupole time-of-flight (QTOF) mass spectrometer equipped with a splitter and an ESI source. The analysis was performed with a C18 4.6 x 75 mm, 2.7 µm column at flowrate of 2 mL/min, under standard gradient condition of 100% water with 0.1% formic acid to 100% acetonitrile with 0.1% formic acid over 15 minutes. The chiral analysis was performed on a chiral column Lux® 3 µm amylose-1 at flowrate of 0.4 mL/min, under gradient condition of 75% to 100% acetonitrile/water with 0.1% formic acid over 6.5 minutes. The typical QTOF operating parameters were as follows: positive ionization mode; sheath gas nitrogen, 12 L/min at 295 °C; drying gas nitrogen flow, 8L/min at 275 °C; nebulizer pressure, 30 psig; nozzle voltage, 1.5 kV; capillary voltage, 4 kV. Lock masses in positive ion mode: purine ion at m/z 121.0509 and HP-921 ion at m/z 922.0098. MS/MS data of precursor ions were acquired using collision energies between 10-40 eV and acquisition rate at 2 spectra/s. Preparative HPLC was performed on the Agilent 1260 Infinity Preparative-Scale LCMS Purification System, completed with Agilent OpenLAB CDS ChemStation Edition for LC and LCMS Systems. Anthracimycin and analogues were separated on an Agilent Prep C18 column (100 × 30 mm) by gradient elution with a mixture of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). Bacterial whole cell assays MIC was determined according to guidelines of the Clinical Laboratory Standards Institute (CLSI) 19 with slight modifications using the broth dilution method in a 384-well microplate format. Inoculum of each test organism was prepared by resuspending isolated colonies in sterile saline to a 0.5 McFarland standard. An inoculum of 5 x 10 5 CFU/mL was incubated with the compounds for 24 hours at 37°C. The effect of the compounds on bacterial growth was evaluated by measuring the optical density at 600 nm using a microplate reader. Mammalian cell cytotoxicity A549 human lung carcinoma cells seeded at 1500 cells/well in a 384-well microplate were treated with the compounds for 72 hours at 37°C in the presence of 5% CO2. PrestoBlue™ cell viability reagent (Life Technologies) was used to assess the cytotoxic effect of the compounds. Microplates were incubated with this dye for 2 hours before fluorescence reading at excitation 560 nm and emission 590 nm. Bioinformatics search methods: prediction of the anthracimycin ( atc ) gene cluster in Streptosporangiales In order to identify other potential anthracimycin producing microorganisms, we first focused on the sequences of the functional domains as defined for Streptomyces sp. T676 in the public database (accession no. LN871452). For this entry, there are nine polyketide synthase (PKS) gene products (atcA, atcB, atcC, atcD, atcE, atcF, atcG, atcH, atcI), where four of them (atcC-atcF) have been proposed to be functionally involved in the generation of anthracimycin 1 . These four products were further split into modules according to the definition of functional modules commonly present in biosynthetic megaenzymes, where independently folding protein domains are joined together via so-called 'linker' regions 12 . The boundaries of these domains were annotated in the entry as misc_feature . For the atcD gene product, an extra feature was annotated as " repeated sequence; additional in PacBio sequence ". This region of 62 residues is rich in low complexity, as seen in the ANNOTATOR platform 13 using a consensus disorder prediction 14 , and so was the only region not included in this analysis. All other 41 domains were blasted and psi-blasted 15 against nr (Dec 2015) 16 , so as to identify species that shared the highest number of domains significantly similar to the ones in the atc cluster. 1292 and 2000 organisms were found by blast and psi-blast (E-value cut off 0.001), respectively, to produce at least 1 significant domain hit. In this work, we limited ourselves to display the top strains that displayed 39 or more hits to all domains present in the atc cluster (Supplementary Table S1 ). In order to identify if the protein domains found in N. kunsanensis would have the architecture of a biosynthetic gene cluster, tblastn was used with the annotated T676 protein domains as queries, to map them in the N. kunsanensis nucleotide sequence. All the protein domains had a significant hit (E-value cut off 0.001) and could be mapped to N. kunsanensis DSM 44524 contig 32 (accession no. NZ_ANAY01000032) (Figure 2 ). All the regions plotted in Figure 2 refer to the first tblastn hit result, with the exception of domains DH5 (in AtcE) and KR8 (in AtcF) that were the 2 nd and 3 rd best hits, respectively. Sequence alignment and phylogenetic analysis Each one of the four functional proteins in the cluster (atcC, atcD, atcE and atcF) had their amino acid sequences aligned among the five strains with the MAFFT E-INS-i algorithm 17 . Individual phylogenetic trees were calculated with MEGA6 18 for each of the four aligned proteins. These included Neighbor-Joining and Maximum Likelihood with and without bootstrap methods (data not shown). Given that there was no observed difference among the generated trees for each of these four proteins, regardless of the method, these aligned sequences were intentionally merged so as to produce a general picture of the evolutionary relationship of these four functional proteins among the five strains that reflects their individual protein trees (Figure 3 ). Bacterial strain acquisition and fermentation conditions Nocardiopsis kunsanensis DSM 44524 was purchased from Leibniz-Institut DSMZ - Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) 19 . Nocardiopsis kunsanensis DSM 44524 was grown on Complex Medium (CM) agar for 7 days at 28°C and was used to prepare 50 mL seed culture in CM (7.5 g casamino acids, 10 g yeast extract, 3 g sodium citrate, 10 g magnesium sulfate, 2 g potassium chloride, 1 mL 4.98% iron sulfate, 50 g NaCl, 1 L distilled water at pH 7.4 20 ). After 7 days cultivation at 28°C on a rotary shaker at 200 rpm, 0.5 mL of the seed culture was used to inoculate 50 mL of CM, CA02LB (20 g mannitol, 20 g soybean meal, 1 L distilled water at pH 7.5) or CA07LB (15 g Glycerol, 30 g oatmeal, 5 g yeast extract, 5 g KH 2 PO 4 , 5 g Na 2 HPO 4 .12H 2 O, 1 g MgCl 2 .6H 2 O, 1 L distilled water at its natural pH) in 250 mL Erlenmeyer flasks. All media was supplemented with 5-10% NaCl (w/v) when necessary. The fermentation was carried out for 7 days at 28 °C at 200 rpm. Chemicals and reagents for media preparation were purchased from Sigma. Extraction and isolation For the initial chemical analysis the entire 50 ml cultures of N. kunsanensis grown in various media were freeze-dried before extracting with methanol for LCMS analysis. For the scale up fermentation for compound isolation, the cultures (80 × 50 mL, total 4 L) of N. kunsanensis from all flasks were first combined and centrifuged to separate the culture supernatant and the mycelia. The combined mycelia were freeze-dried, shaken overnight 2 times with MeOH/ CH 2 Cl 2 (4 L), filtered and added to 1 L of water. The organic layer from the solvent partition was evaporated to dryness using rotary evaporation. The dried dichloromethane extract was dissolved in DMSO and separated by C 18 reversed-phase preparative HPLC (solvent A: H 2 O + 0.1% HCOOH, solvent B: ACN + 0.1% HCOOH; flow rate: 18 mL/min, gradient conditions:70:30 isocratic for 5 minutes; 30% to 40% of solvent B over 5 minutes, followed by 40% to 70% of solvent B over 40 minutes, and an increase from 70% to 100% of solvent B over 50 minutes, and to 100% solvent B in 15 minutes) to give 0.2 mg of anthracimycin ( 1 ) with impurities and 0.9 mg of anthracimycin BII-2619 ( 2 ). Chemical Structural Data The NMR spectra of the compounds are provided in Supplementary Figures S1 (S1a-S1d) to S2 (S2a-S2f). Below, we list the chemical structural data including the measured molar circular dichroism (Δε). Anthracimycin standard Amorphous white powder; [α] D -7 ( c 2.93, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.21) and 286 (3.76) nm; ECD (MeOH) λ (Δε) 202 (+40.47), 214 (+12.96), 236 (+34.84), 283 (-28.84); HRESIMS m/ z 397.2374 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Anthracimycin BII-2619 (2) Amorphous white powder; [α] D -47 ( c 0.16, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.17) and 286 (3.62) nm; ECD (MeOH) λ (Δε) 201 (+30.64), 213 (+7.53), 236 (+18.14), 283 (-15.98); HRESIMS m/ z 397.2371 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Anthracimycin standard Amorphous white powder; [α] D -7 ( c 2.93, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.21) and 286 (3.76) nm; ECD (MeOH) λ (Δε) 202 (+40.47), 214 (+12.96), 236 (+34.84), 283 (-28.84); HRESIMS m/ z 397.2374 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Anthracimycin BII-2619 (2) Amorphous white powder; [α] D -47 ( c 0.16, CH 2 Cl 2 ); UV (MeOH) λ max (log ε) 235 (4.17) and 286 (3.62) nm; ECD (MeOH) λ (Δε) 201 (+30.64), 213 (+7.53), 236 (+18.14), 283 (-15.98); HRESIMS m/ z 397.2371 [M+H] + (calculated for C 25 H 32 O 4 , 397.2379). Analytical Chemistry Procedures Optical rotations were recorded on a JASCO P-2000 digital polarimeter. Electronic circular dichroism (ECD) was measured on a JASCO J815 - circular dichroism (CD) spectrometer. The CD spectra of all samples were recorded in methanol at molar concentration of 0.025 mM in 195-400 nm region. UV spectra were obtained on a General Electric Ultrospec 9000 spectrophotometer. NMR spectra were collected on a Bruker DRX-400 NMR spectrometer with Cryoprobe, using 5-mm BBI ( 1 H, G-COSY, multiplicity-edited G-HSQC, and G-HMBC spectra) or BBO ( 13 C spectra) probe heads equipped with z-gradients. Spectra were calibrated to residual protonated solvent signals (C H Cl 3 δ H 7.24 and C DCl 3 δ C 77.23). The HRESIMS and MS/MS spectra were acquired on Agilent UHPLC 1290 Infinity coupled to Agilent 6540 accurate-mass quadrupole time-of-flight (QTOF) mass spectrometer equipped with a splitter and an ESI source. The analysis was performed with a C18 4.6 x 75 mm, 2.7 µm column at flowrate of 2 mL/min, under standard gradient condition of 100% water with 0.1% formic acid to 100% acetonitrile with 0.1% formic acid over 15 minutes. The chiral analysis was performed on a chiral column Lux® 3 µm amylose-1 at flowrate of 0.4 mL/min, under gradient condition of 75% to 100% acetonitrile/water with 0.1% formic acid over 6.5 minutes. The typical QTOF operating parameters were as follows: positive ionization mode; sheath gas nitrogen, 12 L/min at 295 °C; drying gas nitrogen flow, 8L/min at 275 °C; nebulizer pressure, 30 psig; nozzle voltage, 1.5 kV; capillary voltage, 4 kV. Lock masses in positive ion mode: purine ion at m/z 121.0509 and HP-921 ion at m/z 922.0098. MS/MS data of precursor ions were acquired using collision energies between 10-40 eV and acquisition rate at 2 spectra/s. Preparative HPLC was performed on the Agilent 1260 Infinity Preparative-Scale LCMS Purification System, completed with Agilent OpenLAB CDS ChemStation Edition for LC and LCMS Systems. Anthracimycin and analogues were separated on an Agilent Prep C18 column (100 × 30 mm) by gradient elution with a mixture of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). Bacterial whole cell assays MIC was determined according to guidelines of the Clinical Laboratory Standards Institute (CLSI) 19 with slight modifications using the broth dilution method in a 384-well microplate format. Inoculum of each test organism was prepared by resuspending isolated colonies in sterile saline to a 0.5 McFarland standard. An inoculum of 5 x 10 5 CFU/mL was incubated with the compounds for 24 hours at 37°C. The effect of the compounds on bacterial growth was evaluated by measuring the optical density at 600 nm using a microplate reader. Mammalian cell cytotoxicity A549 human lung carcinoma cells seeded at 1500 cells/well in a 384-well microplate were treated with the compounds for 72 hours at 37°C in the presence of 5% CO2. PrestoBlue™ cell viability reagent (Life Technologies) was used to assess the cytotoxic effect of the compounds. Microplates were incubated with this dye for 2 hours before fluorescence reading at excitation 560 nm and emission 590 nm. Results Predicting the atc biosynthetic gene cluster in a different organism The published atc biosynthetic gene cluster in Streptomyces sp. T676 comprises nine genes/proteins (AtcA-AtcI, accession no. LN871452), where four of them (AtcC-AtcF) have been proposed to be functionally involved in the generation of anthracimycin. These four proteins can be split into 10 Polyketide Synthetase (PKS) modules, and each module harbours specific functional domains required for their molecular function 1 . As a result of a blast/psi-blast search 15 with all functional domains described in the Streptomyces sp. T676 atc cluster against nr 16 , we generated a list of organisms harbouring all or almost all of these domains (E-value cut-off 0.001, see Methods). As expected, we retrieved the four Streptomyces strains discussed above first but, in addition, we found sequence equivalents for all 41 functional domains for AtcC-AtcF in the genome of Nocardiopsis kunsanensis , a species belonging to the same taxonomic class (Actinobacteria) but in a different order (Streptosporangiales) compared to Streptomyces (Supplementary Table S1 ). The next best hit is the species Sorangium cellulosum missing one domain from module 9. Interestingly, Sorangium cellulosum is known to produce chlorotonil A, a compound with a scaffold very similar to that of anthracimycin. As a next step, we tried to understand whether the detected functional domains in N. kunsanensis are sequentially organized in a cluster and, if so, whether this cluster resembles the atc biosynthetic cluster. Tblastn searches with the annotated T676 functional protein domains as queries against N. kunsanensis contigs were carried out. All functional domains had a significant hit and could be mapped to N. kunsanensis DSM 44524 contig 32 (accession no. NZ_ANAY01000032) (Figure 2 ). Furthermore, blast searches with the four Streptomyces sp. T676 atc (AtcC-AtcF) proteins (accession no. CTQ34880, CTQ34881, CTQ34882, CTQ34883, respectively) against the N. kunsanensis proteome produced significant hits (E-value = 0.0 for all of them) to four N. kunsanensis proteins: WP_017577639 (60% sequence identity to AtcC), WP_017577638 (50% sequence identity to AtcD), WP_017577637 (54% sequence identity to AtcE), and WP_020480252 (55% sequence identity to AtcF). For all results, the sequence alignment coverage was ≥ 98% (Supplementary Table S2 ). Interestingly, WP_017577639 is annotated as [acyl-carrier-protein] S-malonyltransferase; the other three proteins are annotated as hypothetical proteins. Finally, we tested the detected N. kunsanensis contig 32 (NZ_ANAY01000032) with antiSMASH 21 . The result clearly supports the existence of an anthracimycin producing trans -AT PKS cluster in N. kunsanensis (Supplementary Figure S3 ). Notably, the N. kunsanensis NZ_ ANAY01000032 cluster was not part of the antiSMASH gene cluster database when this study was carried out in December 2015. Based on the alignment of the four AtcC, AtcD, AtcE and AtcF proteins of the four Streptomyces strains discussed above and N. kunsanensis , a phylogenetic tree was created (Figure 3 ). It shows that the four Streptomyces strains cluster together, and the N. kunsanensis cluster sequence is the outlier, being the most distant one. This is also true for each single PKS module, with the Nocardiopsis sequences showing the lowest sequence identity to Streptomyces sp. T676 compared to all other Streptomyces studied (Supplementary Figure S4 ). But the inspection of core regions of the critical functional domains, keto-synthase (KS), dehydratase (DH), ketoreductase (KR), methyltransferase (MT), acyl carrier protein (ACP) and thioesterase (TE) (see Figure 5 in 1 ), shows that all critical amino acid residues are strongly conserved among all four Streptomyces strains and N. kunsanensis (Supplementary Figure S5 a). To conclude, in December 2015 we suggested that the genome of N. kunsanensis carries an atc trans -AT PKS biosynthetic cluster. It appears to contain all functional domains compared to the known Streptomyces atc cluster, and the critical residues in the core regions of all ACP domains, DH domains, KR domains, KS domains, MT domains, and TE domains are present. The sequence of the N. kunsanensis atc cluster is the most distant one compared to the four available anthracimycin producing Streptomyces strains, nevertheless, we speculated that N. kunsanensis is probably able to produce anthracimycin or its derivatives despite the evolutionary divergence from the Streptomyces genus. Experimental validation of anthracimycin derivative production by N. kunsanensis In order to experimentally validate the production of anthracimycin or its derivative by N. kunsanensis , following the identification of the presence of the respective biosynthetic gene cluster in its genome, we analysed the methanol extracts derived from the a small scale 50 mL fermentation of N. kunsanensis in 3 different media, CM, CA02LB and CA07LB. For this initial analysis, we freeze-dried the entire 50 mL cultures and extracted the metabolites with methanol. The methanol extracts were then analysed with ultra-performance liquid chromatography (UPLC), combined with quadrupole-time of flight high resolution mass spectrometry under positive ionization (ESI) mode. The extracts were screened for the precursor ion of anthracimycin ( m/z 397.2379) and matched against an authentic anthracimycin standard ( 1 ) isolated by our group from the Streptomyces sp. T676 1 (retention time of 13.4 min, Supplementary Figure S6 a). The extracted ion chromatogram (EIC) of m/z 397.2379 showed that N. kunsanensis produced the highest amount of the metabolite in CM (Supplementary Figure S6 e); however, the retention time for the peak was observed at 13.2 min. The observed difference of 0.2 min is one order of magnitude higher than the retention drift in our LC/QTOF system suggesting the production of a compound that is different from anthracimycin, but most likely its derivative. Furthermore, close examination of its MS/MS spectra showed identical fragment ions to those of anthracimycin standard ( 1 ) at higher mass range ( m/z > 311) but with a difference of ~14.02 Da at lower mass range (Supplementary Figures S7 a-c and S8a-c). The fragmentation patterns suggested that the major compound from N. kunsanensis is a new analogue of anthracimycin with structural difference in a methyl position. A 4 liter fermentation of N. kunsanensis in complex medium (CM) was conducted with the aim of isolating sufficient amount of the new analogue to determine its structure. After seven days of cultivation a small aliquot of the cell biomass was separated from the broth and both were extracted with methanol for analysis by LCMS. The extracted ion chromatogram of the biomass extract showed a major peak at 13.2 min for the potential new analogue of anthracimycin as well as a minor peak at 13.4 min for anthracmycin ( 1 ) (Supplementary Figure S6 a). To note, the expected error rate for HPLC device is clearly below 0.05 min 22 . Only trace amount of anthracimycins were detected in the broth (data not shown). The biomass collected from the 4 liter fermentation was then freeze-dried before extracting with dichloromethane/methanol (1:1) and partitioned with water. The dichloromethane fraction was dried, dissolved in DMSO and fractionated by reverse phase preparative HPLC to afford a suite of compounds. A new anthracimycin analogue, anthracimycin BII-2619 ( 2 ) was isolated together with a minor amount of the known compound anthracimycin ( 1 ). The structure of anthracimycin was determined by 1 H NMR and LCMS/MS experiments (Supplementary Figure S7 ), whereas that of anthracimycin BII-2619 was elucidated by 1D and 2D NMR as well as LCMS/MS experiments (Supplementary Figure S8 ). In order to further ascertain that 0.2 min difference in retention time is not a measurement error between anthracimycin ( 1 ) and anthracimycin BII-2619 ( 2 ), the three NMR samples (anthracimycin ( 1 ), anthracimycin standard and anthracimycin BII-2619 ( 2 )) were re-analyzed using an isocratic gradient. Extracted ion chromatogram of m/z 397.2379 (Supplementary Figure S6 b) showed anthracimycin ( 1 ) and anthracimycin standard at retention time 3.6 min, whereas anthracimycin BII-2619 ( 2 ) is indeed clearly different at 3.2 min. Anthracimycin ( 1 ) was obtained from the 4 liter fermentation of N . kunsanensis with impurities, but was found to have a comparable 1 H NMR spectra (Table 1 , Supplementary Figure S1 a and b) as our anthracimycin standard that we purified from Streptomyces sp. T676. The retention time, UV absorption spectrum and MS/MS fragmentation patterns of these two sources of anthracimycin on a chiral column (Lux® 3 µm amylose-1) were identical (Supplementary Figure S9 ). The electronic circular dichroism (ECD) results for our anthracimycin standard (Supplementary Figure S10 ) were in good agreement with those of the recently reported anthracimycin (Supplementary page S20 in ref. 4 ). This suggests that the anthracimycin produced by N. kunsanensis and Streptomyces sp. T676 are of the same chirality as that produced by Streptomyces sp. CNH365. The full structure and absolute configuration at all asymmetric carbon centers of anthracimycin ( 1 ) have been assigned by Jang et al. 4 on the basis of X-ray and NMR analysis. Hence, the structure of anthracimycin ( 1 ) was defined as depicted in Figure 1 . Anthracimycin BII-2619 ( 2 ) was purified as a white powder and was found to have the same molecular formula C 25 H 32 O 4 and UV as anthracimycin ( 1 ). As we have much less than 10mg of the compound, X-ray crystallographic structural determination, the most reliable approach, is not possible. Our structural proposal ( 2 ) resides on the following arguments: The MS/MS spectra showed characteristic fragment ions of anthracimycin ( 1 ) at m/z 379.2272 [M-H 2 O+H] + , 351.2317 [M-HCOOH+H] + , 329.1751 [M-C 5 H 7 +H] + and 311.1645 [M-C 5 H 9 O+H] + . However, the lower mass range at m/ z 269.1540, 205.0861 and 107.0858 showed a difference of 14.0159 Da to that of anthracimycin ( 1 ) ( m/ z 255.1383, 219.1020 and 93.0699 in Supplementary Figures S7 a-c and S8a-c). Thus, the new compound has the same upper MS/MS spectrum as anthracimycin but a difference by one methyl-group mass in lower end. Thus, the most likely reason for a structural difference is the changed position of a methyl group. The fragmentation patterns suggested a change in the methyl position, possibly from C-2 to C-8 in ( 2 ) as compared to ( 1 ). The 1D 1 H and 13 C NMR data and the 2D COSY, HSQC and HMBC data for the new compound are very similar to that of anthracimycin except for that of a differently placed methyl group. Therefore, we conclude that the structural frame of the new compound repeats the structure of anthracimycin which is known from X-ray crystallography. The 1 H and 13 C NMR data of ( 1 ) and ( 2 ) were very similar except for the presence of an isolated methylene group at δ H 3.21 (d, J 11.5 Hz), 3.47 (d, J 11.5 Hz) and a methyl signal at δ H 0.73 (3H, d, J 7 Hz) in the 1 H NMR spectrum of ( 2 ). The isolated methylene was positioned at C-2, in between an enolised β-diketone and a lactone moiety, and was supported by the 2 J CH and 3 J CH HMBC correlations (Figure 4 ) from H-2 to C-1, H-2 to C-3 and H-4 to C-2. The HMBC data indicates that the methyl group shifted from C-2 to C-8: The observed methyl signal at δ H 0.73 was assigned to C-24 attached to C-8, based on the evidence provided by the 2 J CH and 3 J CH HMBC correlations from H-24 to C-7, H-24 to C-8, and H-24 to C-9. The deduced structure and the relative configurations of ( 2 ) are entirely consistent with the rest of the HMBC and the NOESY correlations (Figure 4 , Supplementary Figures S2 e and S2f). The absolute configuration of anthracimycin BII-2619 ( 2 ) was assumed to be the same as anthracimycin ( 1 ), except C-24-methyl, based on the similarities in their 1 H and 13 C chemical shifts as well as (a) due the essentially identical ECD spectra compared with that of the anthracimycin standard (they really pinpoint the chirality adjacent to the chromophore β-diketone, see Supplementary Figure S10 ) 23 and (b) the absence of an enantiodivergence step in their biosynthesis. Thus, the absolute configuration of anthracimycin BII-2619 ( 2 ) at positions 6, 7, 12, 15, 16, 21 were assumed to be the same as anthracimycin ( 1 ) 4 , based on the similarities in their 1 H and 13 C chemical shifts, coupling constants and no enantiondivergence step in their biosynthesis. Therefore, we can interpret the NOESY data for deriving the relative spatial position of C-24 methyl attached to C-8. The α orientation of the C-24 methyl group was deduced by the unusual shielding effect observed for the C-24 methyl group at δ H 0.73 (3H, d, J 7 Hz), which is due to the anisotropy effect from the proximate C-5 ketone function, and is only possible if the C-24 methyl group is in the α position. This was further supported by the observed H-24/H-6 and H-24/H-12 NOESY correlations (Supplementary Figure S2 f). In fact, the spectral data which contains the C-24 methyl group of anthracimycin BII-2619 ( 2 ) resemble those of chlorotonil A ( 3 ) 8 , 9 . Hence, the structure of anthracimycin BII-2619 ( 2 ) was established as shown in Figure 1 . To emphasize, although anthracimycin BII-2619 ( 2 ) was the major product purified, N. kunansensis does produce also minor amounts of classic anthracimycin ( 1 ) as well as even lesser amounts of other analogues (data not shown). Both the 1 H NMR (Supplementary Figure S1 a) and EIC (Supplementary Figure 6 a) showed that classic anthracimycin ( 1 ) isolated from N. kunsanensis was not pure, due to its low abundance. To make the 1 H NMR data derived from the anthracimycin standard as well as from anthracimycin ( 1 ) and anthracimycin BII-2619 ( 2 ) purified from N. kunsanensis more comparable, we show the data fused into one figure with x -axis (Figure S1 d). Indeed, the 1 H data of ( 1 ) and ( 2 ) are very similar except for the presence of an isolated methylene group at δ 3.21 (d, J 11.5 Hz), 3.47 (d, J 11.5 Hz) (see left box in Figure S1 d) and a methyl signal at δ 0.73 (3H, d, J 7 Hz) in the 1 H NMR spectrum of ( 2 ) (see right box in Figure S1 d, lane (c)). Thus, the quartet signal is seen only in lanes (a) and (b) of Figure S1 d (anthracimycin standard and anthracimycin from N. kunansensis ) but not in lane (c) (anthracimycin BII-2619). In summary, Figure S1 d shows the 8-methyl signal (position C-24) at δ 0.73 ppm (right box, lane (c)) for anthracimycin BII-2619 and the 2-methyl signal (position C-23) at δ 1.39 ppm for anthracimycin ( 1 ) and anthracimycin standard (lanes (a) and (b), respectively). These results suggest that the methyl signal is absent in anthramycin ( 1 ) and anthracimycin standard 1 H NMR spectrum. Thus, the LCMS and NMR data have unambiguously shown that N. kunsanensis produces both anthracimycin and anthracimycin BII-2619. Biological activities Antibacterial activity of anthracimycin BII-2619 ( 2 ) was assessed against methicillin-sensitive and -resistant Staphylococcus aureus (MSSA and MRSA) and compared with that of anthracimycin standard. Both compounds showed minimum inhibitory concentration (MIC) values of ≤1μM, and were thus further profiled against other bacterial cultures and a human laboratory cancer cell line (Table 2 ). Anthracimycin and anthracimycin BII-2619 ( 2 ) displayed activity against Bacillus subtilis and Enterococcus faecalis but were inactive against the tested Gram-negative bacteria. Anthracimycin was, however, still more potent against Enterococcus faecalis (~60-fold) and Bacillus subtilis (~100-fold) when compared to anthracimycin BII-2619 ( 2 ). Both compounds demonstrated moderate, very similar low cytotoxic activity. Predicting the atc biosynthetic gene cluster in a different organism The published atc biosynthetic gene cluster in Streptomyces sp. T676 comprises nine genes/proteins (AtcA-AtcI, accession no. LN871452), where four of them (AtcC-AtcF) have been proposed to be functionally involved in the generation of anthracimycin. These four proteins can be split into 10 Polyketide Synthetase (PKS) modules, and each module harbours specific functional domains required for their molecular function 1 . As a result of a blast/psi-blast search 15 with all functional domains described in the Streptomyces sp. T676 atc cluster against nr 16 , we generated a list of organisms harbouring all or almost all of these domains (E-value cut-off 0.001, see Methods). As expected, we retrieved the four Streptomyces strains discussed above first but, in addition, we found sequence equivalents for all 41 functional domains for AtcC-AtcF in the genome of Nocardiopsis kunsanensis , a species belonging to the same taxonomic class (Actinobacteria) but in a different order (Streptosporangiales) compared to Streptomyces (Supplementary Table S1 ). The next best hit is the species Sorangium cellulosum missing one domain from module 9. Interestingly, Sorangium cellulosum is known to produce chlorotonil A, a compound with a scaffold very similar to that of anthracimycin. As a next step, we tried to understand whether the detected functional domains in N. kunsanensis are sequentially organized in a cluster and, if so, whether this cluster resembles the atc biosynthetic cluster. Tblastn searches with the annotated T676 functional protein domains as queries against N. kunsanensis contigs were carried out. All functional domains had a significant hit and could be mapped to N. kunsanensis DSM 44524 contig 32 (accession no. NZ_ANAY01000032) (Figure 2 ). Furthermore, blast searches with the four Streptomyces sp. T676 atc (AtcC-AtcF) proteins (accession no. CTQ34880, CTQ34881, CTQ34882, CTQ34883, respectively) against the N. kunsanensis proteome produced significant hits (E-value = 0.0 for all of them) to four N. kunsanensis proteins: WP_017577639 (60% sequence identity to AtcC), WP_017577638 (50% sequence identity to AtcD), WP_017577637 (54% sequence identity to AtcE), and WP_020480252 (55% sequence identity to AtcF). For all results, the sequence alignment coverage was ≥ 98% (Supplementary Table S2 ). Interestingly, WP_017577639 is annotated as [acyl-carrier-protein] S-malonyltransferase; the other three proteins are annotated as hypothetical proteins. Finally, we tested the detected N. kunsanensis contig 32 (NZ_ANAY01000032) with antiSMASH 21 . The result clearly supports the existence of an anthracimycin producing trans -AT PKS cluster in N. kunsanensis (Supplementary Figure S3 ). Notably, the N. kunsanensis NZ_ ANAY01000032 cluster was not part of the antiSMASH gene cluster database when this study was carried out in December 2015. Based on the alignment of the four AtcC, AtcD, AtcE and AtcF proteins of the four Streptomyces strains discussed above and N. kunsanensis , a phylogenetic tree was created (Figure 3 ). It shows that the four Streptomyces strains cluster together, and the N. kunsanensis cluster sequence is the outlier, being the most distant one. This is also true for each single PKS module, with the Nocardiopsis sequences showing the lowest sequence identity to Streptomyces sp. T676 compared to all other Streptomyces studied (Supplementary Figure S4 ). But the inspection of core regions of the critical functional domains, keto-synthase (KS), dehydratase (DH), ketoreductase (KR), methyltransferase (MT), acyl carrier protein (ACP) and thioesterase (TE) (see Figure 5 in 1 ), shows that all critical amino acid residues are strongly conserved among all four Streptomyces strains and N. kunsanensis (Supplementary Figure S5 a). To conclude, in December 2015 we suggested that the genome of N. kunsanensis carries an atc trans -AT PKS biosynthetic cluster. It appears to contain all functional domains compared to the known Streptomyces atc cluster, and the critical residues in the core regions of all ACP domains, DH domains, KR domains, KS domains, MT domains, and TE domains are present. The sequence of the N. kunsanensis atc cluster is the most distant one compared to the four available anthracimycin producing Streptomyces strains, nevertheless, we speculated that N. kunsanensis is probably able to produce anthracimycin or its derivatives despite the evolutionary divergence from the Streptomyces genus. Experimental validation of anthracimycin derivative production by N. kunsanensis In order to experimentally validate the production of anthracimycin or its derivative by N. kunsanensis , following the identification of the presence of the respective biosynthetic gene cluster in its genome, we analysed the methanol extracts derived from the a small scale 50 mL fermentation of N. kunsanensis in 3 different media, CM, CA02LB and CA07LB. For this initial analysis, we freeze-dried the entire 50 mL cultures and extracted the metabolites with methanol. The methanol extracts were then analysed with ultra-performance liquid chromatography (UPLC), combined with quadrupole-time of flight high resolution mass spectrometry under positive ionization (ESI) mode. The extracts were screened for the precursor ion of anthracimycin ( m/z 397.2379) and matched against an authentic anthracimycin standard ( 1 ) isolated by our group from the Streptomyces sp. T676 1 (retention time of 13.4 min, Supplementary Figure S6 a). The extracted ion chromatogram (EIC) of m/z 397.2379 showed that N. kunsanensis produced the highest amount of the metabolite in CM (Supplementary Figure S6 e); however, the retention time for the peak was observed at 13.2 min. The observed difference of 0.2 min is one order of magnitude higher than the retention drift in our LC/QTOF system suggesting the production of a compound that is different from anthracimycin, but most likely its derivative. Furthermore, close examination of its MS/MS spectra showed identical fragment ions to those of anthracimycin standard ( 1 ) at higher mass range ( m/z > 311) but with a difference of ~14.02 Da at lower mass range (Supplementary Figures S7 a-c and S8a-c). The fragmentation patterns suggested that the major compound from N. kunsanensis is a new analogue of anthracimycin with structural difference in a methyl position. A 4 liter fermentation of N. kunsanensis in complex medium (CM) was conducted with the aim of isolating sufficient amount of the new analogue to determine its structure. After seven days of cultivation a small aliquot of the cell biomass was separated from the broth and both were extracted with methanol for analysis by LCMS. The extracted ion chromatogram of the biomass extract showed a major peak at 13.2 min for the potential new analogue of anthracimycin as well as a minor peak at 13.4 min for anthracmycin ( 1 ) (Supplementary Figure S6 a). To note, the expected error rate for HPLC device is clearly below 0.05 min 22 . Only trace amount of anthracimycins were detected in the broth (data not shown). The biomass collected from the 4 liter fermentation was then freeze-dried before extracting with dichloromethane/methanol (1:1) and partitioned with water. The dichloromethane fraction was dried, dissolved in DMSO and fractionated by reverse phase preparative HPLC to afford a suite of compounds. A new anthracimycin analogue, anthracimycin BII-2619 ( 2 ) was isolated together with a minor amount of the known compound anthracimycin ( 1 ). The structure of anthracimycin was determined by 1 H NMR and LCMS/MS experiments (Supplementary Figure S7 ), whereas that of anthracimycin BII-2619 was elucidated by 1D and 2D NMR as well as LCMS/MS experiments (Supplementary Figure S8 ). In order to further ascertain that 0.2 min difference in retention time is not a measurement error between anthracimycin ( 1 ) and anthracimycin BII-2619 ( 2 ), the three NMR samples (anthracimycin ( 1 ), anthracimycin standard and anthracimycin BII-2619 ( 2 )) were re-analyzed using an isocratic gradient. Extracted ion chromatogram of m/z 397.2379 (Supplementary Figure S6 b) showed anthracimycin ( 1 ) and anthracimycin standard at retention time 3.6 min, whereas anthracimycin BII-2619 ( 2 ) is indeed clearly different at 3.2 min. Anthracimycin ( 1 ) was obtained from the 4 liter fermentation of N . kunsanensis with impurities, but was found to have a comparable 1 H NMR spectra (Table 1 , Supplementary Figure S1 a and b) as our anthracimycin standard that we purified from Streptomyces sp. T676. The retention time, UV absorption spectrum and MS/MS fragmentation patterns of these two sources of anthracimycin on a chiral column (Lux® 3 µm amylose-1) were identical (Supplementary Figure S9 ). The electronic circular dichroism (ECD) results for our anthracimycin standard (Supplementary Figure S10 ) were in good agreement with those of the recently reported anthracimycin (Supplementary page S20 in ref. 4 ). This suggests that the anthracimycin produced by N. kunsanensis and Streptomyces sp. T676 are of the same chirality as that produced by Streptomyces sp. CNH365. The full structure and absolute configuration at all asymmetric carbon centers of anthracimycin ( 1 ) have been assigned by Jang et al. 4 on the basis of X-ray and NMR analysis. Hence, the structure of anthracimycin ( 1 ) was defined as depicted in Figure 1 . Anthracimycin BII-2619 ( 2 ) was purified as a white powder and was found to have the same molecular formula C 25 H 32 O 4 and UV as anthracimycin ( 1 ). As we have much less than 10mg of the compound, X-ray crystallographic structural determination, the most reliable approach, is not possible. Our structural proposal ( 2 ) resides on the following arguments: The MS/MS spectra showed characteristic fragment ions of anthracimycin ( 1 ) at m/z 379.2272 [M-H 2 O+H] + , 351.2317 [M-HCOOH+H] + , 329.1751 [M-C 5 H 7 +H] + and 311.1645 [M-C 5 H 9 O+H] + . However, the lower mass range at m/ z 269.1540, 205.0861 and 107.0858 showed a difference of 14.0159 Da to that of anthracimycin ( 1 ) ( m/ z 255.1383, 219.1020 and 93.0699 in Supplementary Figures S7 a-c and S8a-c). Thus, the new compound has the same upper MS/MS spectrum as anthracimycin but a difference by one methyl-group mass in lower end. Thus, the most likely reason for a structural difference is the changed position of a methyl group. The fragmentation patterns suggested a change in the methyl position, possibly from C-2 to C-8 in ( 2 ) as compared to ( 1 ). The 1D 1 H and 13 C NMR data and the 2D COSY, HSQC and HMBC data for the new compound are very similar to that of anthracimycin except for that of a differently placed methyl group. Therefore, we conclude that the structural frame of the new compound repeats the structure of anthracimycin which is known from X-ray crystallography. The 1 H and 13 C NMR data of ( 1 ) and ( 2 ) were very similar except for the presence of an isolated methylene group at δ H 3.21 (d, J 11.5 Hz), 3.47 (d, J 11.5 Hz) and a methyl signal at δ H 0.73 (3H, d, J 7 Hz) in the 1 H NMR spectrum of ( 2 ). The isolated methylene was positioned at C-2, in between an enolised β-diketone and a lactone moiety, and was supported by the 2 J CH and 3 J CH HMBC correlations (Figure 4 ) from H-2 to C-1, H-2 to C-3 and H-4 to C-2. The HMBC data indicates that the methyl group shifted from C-2 to C-8: The observed methyl signal at δ H 0.73 was assigned to C-24 attached to C-8, based on the evidence provided by the 2 J CH and 3 J CH HMBC correlations from H-24 to C-7, H-24 to C-8, and H-24 to C-9. The deduced structure and the relative configurations of ( 2 ) are entirely consistent with the rest of the HMBC and the NOESY correlations (Figure 4 , Supplementary Figures S2 e and S2f). The absolute configuration of anthracimycin BII-2619 ( 2 ) was assumed to be the same as anthracimycin ( 1 ), except C-24-methyl, based on the similarities in their 1 H and 13 C chemical shifts as well as (a) due the essentially identical ECD spectra compared with that of the anthracimycin standard (they really pinpoint the chirality adjacent to the chromophore β-diketone, see Supplementary Figure S10 ) 23 and (b) the absence of an enantiodivergence step in their biosynthesis. Thus, the absolute configuration of anthracimycin BII-2619 ( 2 ) at positions 6, 7, 12, 15, 16, 21 were assumed to be the same as anthracimycin ( 1 ) 4 , based on the similarities in their 1 H and 13 C chemical shifts, coupling constants and no enantiondivergence step in their biosynthesis. Therefore, we can interpret the NOESY data for deriving the relative spatial position of C-24 methyl attached to C-8. The α orientation of the C-24 methyl group was deduced by the unusual shielding effect observed for the C-24 methyl group at δ H 0.73 (3H, d, J 7 Hz), which is due to the anisotropy effect from the proximate C-5 ketone function, and is only possible if the C-24 methyl group is in the α position. This was further supported by the observed H-24/H-6 and H-24/H-12 NOESY correlations (Supplementary Figure S2 f). In fact, the spectral data which contains the C-24 methyl group of anthracimycin BII-2619 ( 2 ) resemble those of chlorotonil A ( 3 ) 8 , 9 . Hence, the structure of anthracimycin BII-2619 ( 2 ) was established as shown in Figure 1 . To emphasize, although anthracimycin BII-2619 ( 2 ) was the major product purified, N. kunansensis does produce also minor amounts of classic anthracimycin ( 1 ) as well as even lesser amounts of other analogues (data not shown). Both the 1 H NMR (Supplementary Figure S1 a) and EIC (Supplementary Figure 6 a) showed that classic anthracimycin ( 1 ) isolated from N. kunsanensis was not pure, due to its low abundance. To make the 1 H NMR data derived from the anthracimycin standard as well as from anthracimycin ( 1 ) and anthracimycin BII-2619 ( 2 ) purified from N. kunsanensis more comparable, we show the data fused into one figure with x -axis (Figure S1 d). Indeed, the 1 H data of ( 1 ) and ( 2 ) are very similar except for the presence of an isolated methylene group at δ 3.21 (d, J 11.5 Hz), 3.47 (d, J 11.5 Hz) (see left box in Figure S1 d) and a methyl signal at δ 0.73 (3H, d, J 7 Hz) in the 1 H NMR spectrum of ( 2 ) (see right box in Figure S1 d, lane (c)). Thus, the quartet signal is seen only in lanes (a) and (b) of Figure S1 d (anthracimycin standard and anthracimycin from N. kunansensis ) but not in lane (c) (anthracimycin BII-2619). In summary, Figure S1 d shows the 8-methyl signal (position C-24) at δ 0.73 ppm (right box, lane (c)) for anthracimycin BII-2619 and the 2-methyl signal (position C-23) at δ 1.39 ppm for anthracimycin ( 1 ) and anthracimycin standard (lanes (a) and (b), respectively). These results suggest that the methyl signal is absent in anthramycin ( 1 ) and anthracimycin standard 1 H NMR spectrum. Thus, the LCMS and NMR data have unambiguously shown that N. kunsanensis produces both anthracimycin and anthracimycin BII-2619. Biological activities Antibacterial activity of anthracimycin BII-2619 ( 2 ) was assessed against methicillin-sensitive and -resistant Staphylococcus aureus (MSSA and MRSA) and compared with that of anthracimycin standard. Both compounds showed minimum inhibitory concentration (MIC) values of ≤1μM, and were thus further profiled against other bacterial cultures and a human laboratory cancer cell line (Table 2 ). Anthracimycin and anthracimycin BII-2619 ( 2 ) displayed activity against Bacillus subtilis and Enterococcus faecalis but were inactive against the tested Gram-negative bacteria. Anthracimycin was, however, still more potent against Enterococcus faecalis (~60-fold) and Bacillus subtilis (~100-fold) when compared to anthracimycin BII-2619 ( 2 ). Both compounds demonstrated moderate, very similar low cytotoxic activity. Discussion The need for the identification of novel antibiotics in order to deal with the increasing problem of drug-resistance development has led to a continuous increase in the search for natural products 24 . As most antibiotics are derived from natural products 25 , this research area will greatly benefit from -omics technologies since mining of genome sequences in silico will efficiently generate lists of organisms with the potential ability to synthesize compounds of desired structure classes 24 , 26 . The respective shortcuts will then direct experimental validation efforts and, thus, save valuable time and cost. This approach was applied in the present work, as we screened the existing genomic data to find organisms that carry sequences similar to the atc biosynthetic cluster of Streptomyces . For example, we found in our analysis that Streptomyces sp. NRRL F-5065 carries the atc biosynthetic gene cluster; yet, it is the only reported strain not experimentally tested to produce anthracimycin. We were surprised to also hit Nocardiopsis kunsanensis , a distant relative of Streptomyces from another taxonomic order. Occasionally observing similar or identical genes or even gene clusters in diverse genera as a consequence of horizontal gene transfer is to be expected in the context of large scale comparative genomics efforts 27 - 29 . Nevertheless, it is interesting to see the atc biosynthetic gene cluster outside the Streptomycetales phylogenetic branch as an additional example of a trans -AT PKS cluster that could potentially evolve by this mode of action. The identification of alternative organisms producing any known compound is of importance on different levels. For instance, they can lead to the discovery of new major compounds that are analogues of the existing ones. This has been the case here with the identification of compound anthracimycin BII-2619 produced by N. kunsanensis . This new analogue automatically becomes a candidate compound for further testing. Our experimental results prove that N. kunsanensis produces the derivative anthracimycin BII-2619 as the major compound and anthracimycin in minor quantities. This finding further exemplifies the exceptional importance of Actinobacteria for natural product discovery 26 , serving as potential source of chemical agents against different drug-resistant infections 25 , for refactoring synthetic pathways 30 , as well as for finding unique chemical structures that may form the scaffold of novel drugs 31 . In addition, knowing more than one organism that produce the same compound can be of crucial help to evolutionary studies on the origin of their biosynthetic gene clusters. The only chemical difference observed between anthracimycin ( 1 ) and anthracimycin BII-2619 ( 2 ) is the presence of methyl groups at the C-2 and C-8 positions, respectively (Figure 1 ). In total, anthracimycin and its derivative carry up to four methyl groups at C-2, C-8, C-10 and C-16. Since the atc cluster contains only three methyltransferases (MT3, MT7, and MT10; Figure 2 ) and these methyltransferases are implied to catalyze the methylation at C-16, C-10 and C-2 respectively 1 , we have to consider that methylation at C-8 is possibly carried out by either a post-PKS tailoring reaction catalyzed by a methyltransferase that is external to the atc PKS cluster, by MT7 if it is to assume its significant deviation from colinearity and possible cross-modular function as described by Alt and Wilkinson 1 , or by MT10, due to its C-terminal localization that could help it to function as a post-PKS tailoring enzyme. There are currently >20 proteins from N. kunsanensis labelled as S-adenosylmethionine- (SAM-) dependent methyltransferases out of >70 annotated methyltransferases. Attempts to identify the actual enzyme for this reaction would require an extensive effort that is out of the scope of this work. Yet, at this point, alternative scenarios cannot be excluded such as some PKS cluster methyltransferases being less than 100% efficient or with methylation site promiscuity. To note, a similar hypothesis was already proposed for the chlorotonil A biosynthesis. The PKS domain organization of chlorotonil A producing strain S. cellulosum 1525 is different from the anthracimycin producing Streptomyces and N. kunsanensis . In this case, the methyltransferases MT3, MT6, and MT7 of the PKS cluster were implied to be responsible for the methylation at C-16, C-8, and C-10, respectively, whereas facultative methylation at C-2 was accredited to the free-standing SAM-dependent methyltransferase CtoF (accession no. ALD83694.1), which is not part of the PKS cluster 9 . The final answer to this question will require a dedicated analytical experiment. We wish to emphasize that the structure of anthracinycin BII-2619 ( 2 ) as well as those of anthracimycin ( 1 ) and chlorotonil A also cannot be strictly explained based on the type I PKS organization, that several significant deviation from collinearity of enzymes have been described and, generally, that the co-linearity rule is only well applied to type I PKS with cis -AT organization 1 , 2 , 9 . The biosynthetic gene cluster reported here falls into the trans -AT PKS classification and so do the anthracimycin ( 1 ) and chlorotonil A clusters as well. Importantly, the biosynthetic gene cluster for chlorotonil A was shown to also produce new chlorotonil congeners that additionally exhibit carbon skeleton variations, which illustrates the partial promiscuity of these megaenzymes and the difficulty to ascertain all reaction steps to individual modules and the assignments of functions in the previous well cited work as well as in our new one remains tentative. One of their scenarios to explain such differences is the inefficient methylation functionality in module 7. Although it is true that we used the genome of N. kunsanensis DSM 44524, which by definition is a draft genome, one should be aware that current sequencing and assembly technology produces draft genomes that cover 95-99% of the full genome in the case of bacteria, as benchmarked in GAGE 32 and in cases of lower coverage, such as for Rhodobacter sphaeroides , the missing regions came from short low complexity repeat regions that are difficult to assemble but do not harbor biosynthetic gene clusters. Still, we ran antiSMASH against the complete draft genome of N. kunsanensis to find all its predicted biosynthetic gene clusters. We were able to identify the cluster here reported as the only trans -AT PKS. To compare this cluster to other predicted PKS types of clusters, we extracted all the predicted KS domains, aligned and calculated their phylogenetic distances as in the Figure S5 b. This tree is in perfect agreement with the Figure S16 in suppl. material of reference 9 allowing to identify clades of KS domains based on their substrate specificity. This tree shows that all other PKS clusters found in the genome are not closely related to the one producing anthracimycin BII-2619, which itself remains closely related to the one producing chlorotonil A. We also compared the biological activity of anthracimycin standard and anthracimycin BII-2619 ( 2 ). We found that both compounds were effective against Gram-positive bacteria, but anthracimycin BII-2619 is 50- to 100-fold less active compared to anthracimycin. Both compounds are inefficient for suppressing the growth of Gram-negative microbes and have similar level of toxicity against the lung cancer cell line A549. Low cytotoxicity was also observed for HeLa cells 10 ; however it was recently reported that anthracimycin was approximately 10-fold more toxic against three human hepatocellular carcinoma cell lines 33 . Thus, the toxicity of anthracimycin and its analogs needs to be further assessed against more cell types. As stated above, the only structural difference between these two compounds is the presence of methyl groups at the C-2 in anthracimycin ( 1 ) and C-8 position for anthracimycin BII-2619 ( 2 ). This shift of the methyl group from the C-2 to C-8 position appeared to have no effect on the cytotoxicity of this compound on A549 cells, but significantly decreased its antibacterial potency. To conclude, we report the anthracimycin derivative producing biosynthetic gene cluster in N. kunsanensis that was identified through genome sequence screening. We show that this computational approach is able to identify different organisms that produce the same or similar compounds. It is the first time that the atc biosynthetic cluster has been reported for another order rather than Streptomycetales. We were able to prove that this strain also produces anthracimycin in small quantities under specific fermentation conditions. The identification of its main product anthracimycin BII-2619 ( 2 ) together with its bioactivity tests contribute to the understanding of anthracimycin as a potential candidate antibiotic. Supplementary Material Supplementary figures and tables. Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3491559/

Acquisition of Maternal Antibodies both from the Placenta and by Lactation Protects Mouse Offspring from Yersinia pestis Challenge

Artificially passive immunization has been demonstrated to be effective against Yersinia pestis infection in animals. However, maternal antibodies' protective efficacy against plague has not yet been demonstrated. Here, we evaluated the kinetics, protective efficacy, and transmission modes of maternal antibodies, using mice immunized with plague subunit vaccine SV1 (20 μg of F1 and 10 μg of rV270). The results showed that the rV270- and F1-specific antibodies could be detected in the sera of newborn mice (NM) until 10 and 14 weeks of age, respectively. There was no antibody titer difference between the parturient mice immunized with SV1 (PM-S) and the caesarean-section newborns (CSN) from the PM-S or between the lactating mice immunized by SV1 (LM-S) and the cross-fostered mice (CFM) during 3 weeks of lactation. The NM had a 72% protection against 4,800 CFU Y. pestis strain 141 challenge at 6 weeks of age, whereas at 14 weeks of age, NM all succumbed to 5,700 CFU of Y. pestis challenge. After 7 weeks of age, CFM had an 84% protection against 5,000 CFU of Y. pestis challenge. These results indicated that maternal antibodies induced by the plague subunit vaccine in mother mice can be transferred to NM by both placenta and lactation. Passive antibodies from the immunized mothers could persist for 3 months and provide early protection for NM. The degree of early protection is dependent on levels of the passively acquired antibody. The results indicate that passive immunization should be an effective countermeasure against plague during its epidemics.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3583336/

Acute and Potentially Life-Threatening Tropical Diseases in Western Travelers—A GeoSentinel Multicenter Study, 1996–2011

We performed a descriptive analysis of acute and potentially life-threatening tropical diseases among 82,825 ill western travelers reported to GeoSentinel from June of 1996 to August of 2011. We identified 3,655 patients (4.4%) with a total of 3,666 diagnoses representing 13 diseases, including falciparum malaria (76.9%), enteric fever (18.1%), and leptospirosis (2.4%). Ninety-one percent of the patients had fever; the median time from travel to presentation was 16 days. Thirteen (0.4%) patients died: 10 with falciparum malaria, 2 with melioidosis, and 1 with severe dengue. Falciparum malaria was mainly acquired in West Africa, and enteric fever was largely contracted on the Indian subcontinent; leptospirosis, scrub typhus, and murine typhus were principally acquired in Southeast Asia. Western physicians seeing febrile and recently returned travelers from the tropics need to consider a wide profile of potentially life-threatening tropical illnesses, with a specific focus on the most likely diseases described in our large case series.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9885324/

Poria cocos polysaccharide—functionalized graphene oxide nanosheet induces efficient cancer immunotherapy in mice

Introduction: Tumor vaccines that induce robust humoral and cellular immune responses have attracted tremendous interest for cancer immunotherapy. Despite the tremendous potential of tumor vaccines as an effective approach for cancer treatment and prevention, a major challenge in achieving sustained antitumor immunity is inefficient antigen delivery to secondary lymphoid organs, even with adjuvant aid. Methods: Herein, we present antigen/adjuvant integrated nanocomplexes termed nsGO/PCP/OVA by employing graphene oxide nanosheet (nsGO) as antigen nanocarriers loaded with model antigen ovalbumin (OVA) and adjuvant, Poria cocos polysaccharides (PCP). We evaluated the efficacy of nsGO/PCP/OVA in activating antigen-specific humoral as well as cellular immune responses and consequent tumor prevention and rejection in vivo . Results: The optimally formed nsGO/PCP/OVA was approximately 120–150 nm in diameter with a uniform size distribution. Nanoparticles can be effectively engulfed by dendritic cells (DCs) through receptor-mediated endocytosis, induced the maturation of DCs and improved the delivery efficiency both in vitro and in vivo. The nsGO/PCP/OVA nanoparticles also induced a significant enhancement of OVA antigen-specific Th1 and Th2 immune responses in vivo . In addition, vaccination with nsGO/PCP/OVA not only significantly suppressed tumor growth in prophylactic treatments, but also achieved a therapeutic effect in inhibiting the growth of already-established tumors. Conclusion: Therefore, this potent nanovaccine platform with nanocarrier nsGO and PCP as adjuvants provides a promising strategy for boosting anti-tumor immunity for cancer immunotherapy. Introduction Recently, cancer immunotherapy has become one of the most promising techniques of both cancer prevention and intervention by virtue of its effective avoidance of off-target effects which can better improve anti-tumor immune responses ( Arya et al., 2018 ). Among all kinds of cancer immunotherapy, the introduction of vaccines is rapidly becoming a growing trend in cancer treatment ( Bachmann and Jennings, 2010 ). Recent studies have focused on different types of cancer vaccines, including tumor cell lysates, dendritic cells (DCs), nucleic acids (such as mRNA), and neoantigens ( Bao et al., 2011 ). However, antigens alone are poor activators of adaptive immune responses. In the absence of adjuvants, antigens targeting immature DCs without inflammation or any microbial stimulation induce tolerance instead of effective immune responses. Recently, a number of adjuvants, including MF59, CpG ODN, AS04, and AS01, have been approved for use in human vaccines; however, the systemic immune toxicity induced by adjuvants continues to be an obstacle for widespread applications ( Didierlaurent et al., 2017 ; Wilkins et al., 2017 ). Thus, innovative adjuvants with fewer adverse effects and greater modifiability are urgently required. Poria cocos is an edible mushroom and has been used for medicine for a long history owing to its specific characterization and biological activities ( Zhu et al., 2019 ; Liang et al., 2022 ; Zheng et al., 2022 ). Poria cocos is composed of multiple chemical composition, including triterpenes, polysaccharides, steroids, amino acids, choline, and histidine. Previous studies have revealed that Poria cocos polysaccharides (PCPs) and their derivatives have various biological activities, such as anti-tumor, anti-inflammatory, anti-viral, and anti-oxidant ( Chen et al., 2009 ; Ke et al., 2010 ; Xu et al., 2022 ). The capability of PCP to enhance cellular immunity and humoral immunity and its outstanding safety profile make it a promising candidate for innovative adjuvant development ( Zhang et al., 2019 ). However, even with adjuvants, inefficient delivery of antigens and adjuvants to secondary lymphoid organs often results in poor immune responses ( Arya et al., 2018 ). Thus, an adjuvant can be co-formulated with an antigen within the same delivery carrier to induce effective immune responses. Nanoparticles are minute particles that are typically <200 nm in diameter ( Chrastina et al., 2011 ) and exhibit unique properties, such as large surface-to-volume ratio, drug-loading compacity and tunable surface chemistry, confer many advantages in terms of vaccine delivery ( Xu et al., 2015 ). Recently, graphene oxide (GO) have been widely utilized as vehicles for drug delivery ( Zheng et al., 2018 ), biological sensors ( Xu et al., 2022 ) and have been applied successfully in photodynamic therapy ( Wei et al., 2016 ), cancer treatment ( Arya et al., 2018 ) and antibacterial therapy ( Wang et al., 2022 ). GO is derived from graphite via a variety of oxidation processes, the most common of which is the enhanced Hummers method ( Zhang et al., 2022 ). Because of its oxygen-containing functional groups, aromatic lattice, and large interfacial surface area, GO has the flexibility and capacity to load a wide range of compounds including drugs, nucleic acids, and peptides by non-covalent interactions (hydrophobic interactions, hydrogen bonding, and π–π stacking). In recent decades, there has been an increasing amount of research regarding the capacity of GO for the encapsulation and delivery of antigens. According to Dudek et al. (2016) antigen-loaded GO stimulated the immune system by up-regulating inflammatory cytokines, inducing lymphocyte proliferation and differentiation, and thus aided in the elimination of intracellular pathogens and infected cells after immunization. GO nanosheets (nsGOs) are derived from GO and have a small size (−100 nm) and tight size distribution. nsGO exhibits effective cell membrane permeability and low cytotoxicity ( Wang et al., 2014 ). Thus, nsGO can be used as a potential carrier platform for antigen and adjuvant co-delivery. In view of these considerations, we generated a safe and effective nsGO nanovaccine that could co-deliver the model antigen ovalbumin (OVA) protein and adjuvant PCP to induce robust immune responses and antitumor effects in a tumor-bearing mouse model. The formed nsGO/PCP/OVA nanoparticles induced DC maturation without detectable cytotoxicity and promoted antigen uptake both in vitro and in vivo . In addition, the nsGO/PCP/OVA nanoparticles provoked strong Th1 and Th2 type immune responses in vaccinated mice and functioned as a prophylactic vaccine to protect mice from E.G7-OVA tumor challenge, as well as a therapeutic vaccine to achieve better anti-tumor effects. These data demonstrate that nsGO/PCP/OVA may be an effective approach for enhancing antigen-specific adaptive immune responses against cancer cells. Methods Materials CpG ODN 1668 (TCCATGACGTTC CTGATGCT) with a single-stranded phosphonothioate was obtained from Sangon Biotech (Sangon Biotech, China). PCP was obtained from the YuanYe Company (YuanYe inc, China). OVA protein was purchased from Sigma (Sigma, MO, United States), OVA peptide 257–264 (SIINFEKL) was from the Chinese Peptide Company (Chinese Peptide Company, China), and Alexa Fluor™ 647 conjugated OVA was purchased from Thermo (Thermo Fisher Scientific, MA, United States). All materials used in vaccines were purified using Pierce™ High-Capacity Endotoxin Removal Spin Columns (Thermo Fisher Scientific, MA, United States). Afterwards, the endotoxin levels were measured to be constantly below 5 Endotoxin Unit (EU)/mL using the ToxinSensor™ Endpoint Chromosome Endotoxin Detection Kit (Genscript, NJ, United States). Cell culture Mouse lymphoma cell line E.G7-OVA cell was obtained from the American Type Culture Collection (ATCC, CRL-2113, MD, United States) and cultured in RPMI 1640 medium (ATCC, Cat# 30-2001, MD, USA) at 37°C in a humidified atmosphere containing 5% CO 2 . Mouse dendritic cell line DC2.4 cells were purchased from Bena Culture Collection (BNCC Inc, China) and incubated in RPMI 1640 medium (Gibco, NY, United States) with 2 mM glutamine, streptomycin-penicillin solution, 50 μM 2-mercaptoethanol, and 10% heat-inactivated fetal calf serum (FBS, Gibco, NY, United States). Bone marrow-derived DCs (BMDCs) were obtained from female C57BL/6 mice. Briefly, fresh BMDCs were isolated from the femur and tibia of C57BL/6 mice. Cells were then cultured in RPMI 1640 medium (Gibco, NY, United States) with 20 ng/mL GM-CSF (PeproTech, NJ, United States), penicillin-streptomycin solution (Solarbio, China), 50 μM 2-mercaptoethanol (Invitrogen, CA, United States), and 10% FBS (Gibco, NY, United States). Non-adherent and loosely adherent cells were harvested on day 5 or day 6 as immature BMDCs. Animals 57BL/6 (6–8 weeks old) female mice were obtained from Vital River Laboratory Animal Technology Co., Ltd. (Vital River Laboratory Animal Technology Co., Ltd. China) and housed in a specific-pathogen-free (SPF) animal laboratory. All animals are free for sterile food and water. After the injection procedure, the animals were closely monitored for symptoms of food intake, pain or distress, and motility. Mice were euthanized by cervical dislocation at humane endpoints or at the end of the experiments. All animal experiments were reviewed and approved by the Institutional Animal Treatment and Use Committee of the China Academy of Chinese Medica Sciences (code:2021B218). Preparation of nsGO GO was generously provided by Dr. Dongtang Zhang (Beijing University of Technology, China) and nsGO was prepared according to Ying's protocol ( Wang et al., 2014 ). Briefly, GO was dissolved in water at a concentration of 0.2 mg/mL and sonicated in water bath for 2 h. After sonication in an ice bath with a sonification power of 40 W, NaOH solution was added to reach a final concentration of 5 M NaOH. Then, the resultant solution was sonicated in water bath for 2 h, and the pH of the solution was adjusted to neutral. The solution was centrifuged at 1,6128 × g for 10 min, and the supernatant was designated as nsGO. Preparation and characterization of nsGO/PCP/OVA Epichlorohydrin was added to the nsGO solution at 40°C for 4 h under N 2 protection. Once the unreacted epichlorohydrin was removed by ultrafiltration, endotoxin-free PCP solution (10 mg/mL) was added to epichlorohydrin-GO solution in water bath at 42°C for 3 h. After the pH of the nsGO/PCP solution was adjusted to weak acidity, a 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydro/N-hydroxy succinimide (EDC/NHS) mixture was added, followed by incubation for 30 min. Subsequently, the OVA protein solution (1 mg/mL) was added to the prepared solution for 1 h at room temperature. Then, the unreacted protein was removed by ultrafiltration and the lyophilized solid was collected as nsGO/PCP/OVA nanoparticles. The concentration of nsGO/PCP/OVA was quantified by the OVA level using the Bicinchoninic Acid Assay kit (BCA, Biomed Inc, CA, United States). The encapsulation efficiency was measured using the following formula: [(Total OVA–Free OVA)/Total OVA] ⋠100%. The nanoparticles were characterized using transmission electron microscopy (TEM, Hitachi, Japan), Fourier Transform infrared spectroscopy (FT-IR spectroscopy, Spectrum 100 FTIR, PerkinElmer, MA, USA), and dynamic light scattering (DLS, Malvern, Germany). TEM images were analyzed for the nanoparticle size distribution using Nano measurer 1.2 software (Fudan University, China). Cell viability To evaluate the cytotoxic effects of nanoparticles, DC2.4 cells were incubated with nsGO/PCP/OVA ranging from 0 to 100 μg/mL OVA. The absorbance at 450 nm was measured using the Cell Counting Kit-8 (Dojindo, Japan) 24 h after treatment. The percentage of cell viability of DC2.4 cells was calculated. Flow cytometry analysis For cell surface marker detecting, the cells were washed with PBS and blocked with anti-mouse CD16/32 antibody (BioLegend, CA, United States). After treatment for 15 min at room temperature, the cells were incubated with different antibodies in PBS with 1% FBS (Gibco, NY, United States) for 1 h. After washed 3 times with PBS, cells were and resuspended in PBS with 1% FBS. Flow cytometry analysis was performed using a BD FACS Calibur™ flow cytometer (BD Bioscience, CA, United States), and data were analyzed using FlowJo software V10 (Tree star, OR, United States). Bone marrow-derived DCs activation Immature BMDCs were collected and treated with 20 μg/mL OVA, 5 μg/mL nsGO, 100 μg/mL PCP, nsGO/PCP, and nsGO/PCP/OVA, according to Dong's research ( Dong et al., 2021 ). After 24 h, the supernatants were collected and measured for IL-6 and IL12 production (ELISA MAX Deluxe kits, BioLegend, CA, United States). The expression of CD80, CD86, and MHC-II on BMDC cell surface was analyzed using PE anti-mouse MHC-II (BioLegend, CA, United States), APC anti-mouse CD80 (BioLegend, CA, United States), and PerCP/Cyanine5.5 anti-mouse CD86 (BioLegend, CA, United States) by FACS analysis. Uptake assays For in vitro uptake assays, BMDCs were incubated with PBS, 5 μg/mL OVA-FITC, and 2.5, 5, and 10 μg/mL nsGO/PCP/OVA-FITC. Cells incubated with nsGO/PCP/OVA-FITC (5 μg/mL) at 4°C were used as controls. After 30 min of incubation, the BMDC cells were collected and stained with APC anti-mouse CD11c (BioLegend, CA, United States) antibody before FACS analysis. For uptake competition assays, cells were pre-treated with PCP, OVA, or 200 μg/mL mannans (Solarbio, China) for 30 min. BMDCs were then treated with nsGO/PCP/OVA-FITC (10 μg/mL) for 45 min. BMDCs were also incubated with nsGO/PCP/OVA (10 μg/mL) for 30 min, followed by co-culture with 100 μg/mL Lucifer yellow VS. dilithium salt (LY, Sigma, MO, United StatesA) for 45 min before FACS analysis. To monitor the in vivo uptake of nanoparticles, mice were injected with OVA-Alexa 647 (20 μg) or nsGO/PCP/OVA-Alexa 647 (containing 20 μg OVA) in both hind footpads. After 6 h, the popliteal lymph nodes were dissected and incubated with collagen D (Sigma, MO, United States) and DNase (Sigma-Aldrich, MO, United States) to prepare single-cell suspensions. Cells were incubated with anti-mouse CD16/32 antibody (BioLegend, CA, United States) for 15 min and then stained with FITC-anti-mouse CD11c antibody (BioLegend, CA, United States) before FACS analysis. Immunization of mice Female C57BL/6 mice were randomly allocated to six groups. On days 0, 7, and 14, mice were immunized subcutaneously with PBS (control), OVA (20 μg per mouse), PCP + OVA (250 μg PCP and 20 μg OVA per mouse), CpG ODN + OVA (10 μg CpG and 20 μg OVA per mouse), nsGO/PCP + OVA (20 μg OVA per mouse), and nsGO/PCP/OVA (containing 20 μg OVA per mouse). On day 21, the mice were anesthetized with isoflurane and sacrificed following the collection of serum and spleen for subsequent analysis. Measurement of Th1 and Th2 immune responses Splenocytes were harvested from mice on day 21 after the three vaccinations. For ELISPOT, cells were cultured at 2 × 10 5 cells/well in ELISPOT plates (Dakewe, China), and stimulated with 1 μg/mL OVA peptide 257-264 (SIINFEKL, OVA â ) or 300 μg/mL OVA protein for 36 h. The number of spots was counted, and the results were expressed as spot-forming cells (SFCs) per 10 5 splenocytes. For the cell proliferation assay, splenocytes were cultured at 2 × 10 6 cells/mL with 300 μg/mL OVA 300 μg/mL or 1 μg/mL OVA â 1 μg/mL. After 48 h, cells were collected, incubated with anti-mouse CD16/32 antibody, and stained with anti-mouse CD69, anti-mouse CD3 (Biolegend, CA, United States), anti-mouse CD4 (Biolegend, CA, United States), and anti-mouse CD8 (Biolegend, CA, United States) antibodies at room temperature for 30 min. To measure cytokine release, splenocytes were cultured with OVA (300 μg/mL) or OVA I (1 μg/mL) for 72 h, supernatants were collected and measured for IL-4 and IFN-γ concentrations using ELISA kits (BioLegend, CA, United States). Prophylactic tumor challenge assays For prophylactic vaccination, C57BL/6 mice were first primed with different formulations on days −21 and boosted on days −14 and −7 as above. On day 0, mice were subcutaneously injected with 2 × 10 5 E.G7-OVA cells in the right flank. On day 21, all mice were euthanized when a humane endpoint was reached, tumor masses were measured, and tumor volume was calculated as length × width 2 × 0.5. Blood samples were collected for antibody detection. Therapeutic tumor vaccination C57BL/6 mice were subcutaneously injected with 2 × 10 5 E.G7-OVA cells into the right flank. When the mean tumor size reached approximately 100 mm 3 , mice were inoculated with different formulations. All mice were euthanized on day 18, when the control mice reached the humane endpoint. Blood and spleen tissues were collected for further analysis. ELISA For antibody detection in prophylactic tumor assays, serum was collected from mice on day 21. The antibody titers of serum OVA-specific IgG1 and IgG2a were measured using an IgG Mouse ELISA kit (Thermo Fisher Scientific, MA, United States). In therapeutic tumor assay, to detect anti-OVA specific antibodies, serum was collected on day 18 and diluted (1:1,000 to detect IgG, 1:500 to detect IgG1, and 1:3,000 for IgG2a) for serum OVA-specific IgG, IgG1, and IgG2a measurements using ELISA kit (Thermo Fisher Scientific, MA, United States). For cytokine measurements in tumor-bearing mice, splenocytes were collected and cultured at 2 × 10 6 cells/mL with 300 μg/mL OVA. After 72 h, the supernatant of cell culture was collected and IL-2 and IL-4 concentrations were measured using ELISA kits (BioLegend, CA, United States). Statistical analyses Statistical analyses were performed using Prism 7.0 (GraphPad Software Inc, CA, United States). All data analysis results in this study are expressed as means ± SD and were analyzed by one-way analysis of variance (ANOVA). p < 0.05 was defined as statistical significance. Materials CpG ODN 1668 (TCCATGACGTTC CTGATGCT) with a single-stranded phosphonothioate was obtained from Sangon Biotech (Sangon Biotech, China). PCP was obtained from the YuanYe Company (YuanYe inc, China). OVA protein was purchased from Sigma (Sigma, MO, United States), OVA peptide 257–264 (SIINFEKL) was from the Chinese Peptide Company (Chinese Peptide Company, China), and Alexa Fluor™ 647 conjugated OVA was purchased from Thermo (Thermo Fisher Scientific, MA, United States). All materials used in vaccines were purified using Pierce™ High-Capacity Endotoxin Removal Spin Columns (Thermo Fisher Scientific, MA, United States). Afterwards, the endotoxin levels were measured to be constantly below 5 Endotoxin Unit (EU)/mL using the ToxinSensor™ Endpoint Chromosome Endotoxin Detection Kit (Genscript, NJ, United States). Cell culture Mouse lymphoma cell line E.G7-OVA cell was obtained from the American Type Culture Collection (ATCC, CRL-2113, MD, United States) and cultured in RPMI 1640 medium (ATCC, Cat# 30-2001, MD, USA) at 37°C in a humidified atmosphere containing 5% CO 2 . Mouse dendritic cell line DC2.4 cells were purchased from Bena Culture Collection (BNCC Inc, China) and incubated in RPMI 1640 medium (Gibco, NY, United States) with 2 mM glutamine, streptomycin-penicillin solution, 50 μM 2-mercaptoethanol, and 10% heat-inactivated fetal calf serum (FBS, Gibco, NY, United States). Bone marrow-derived DCs (BMDCs) were obtained from female C57BL/6 mice. Briefly, fresh BMDCs were isolated from the femur and tibia of C57BL/6 mice. Cells were then cultured in RPMI 1640 medium (Gibco, NY, United States) with 20 ng/mL GM-CSF (PeproTech, NJ, United States), penicillin-streptomycin solution (Solarbio, China), 50 μM 2-mercaptoethanol (Invitrogen, CA, United States), and 10% FBS (Gibco, NY, United States). Non-adherent and loosely adherent cells were harvested on day 5 or day 6 as immature BMDCs. Animals 57BL/6 (6–8 weeks old) female mice were obtained from Vital River Laboratory Animal Technology Co., Ltd. (Vital River Laboratory Animal Technology Co., Ltd. China) and housed in a specific-pathogen-free (SPF) animal laboratory. All animals are free for sterile food and water. After the injection procedure, the animals were closely monitored for symptoms of food intake, pain or distress, and motility. Mice were euthanized by cervical dislocation at humane endpoints or at the end of the experiments. All animal experiments were reviewed and approved by the Institutional Animal Treatment and Use Committee of the China Academy of Chinese Medica Sciences (code:2021B218). Preparation of nsGO GO was generously provided by Dr. Dongtang Zhang (Beijing University of Technology, China) and nsGO was prepared according to Ying's protocol ( Wang et al., 2014 ). Briefly, GO was dissolved in water at a concentration of 0.2 mg/mL and sonicated in water bath for 2 h. After sonication in an ice bath with a sonification power of 40 W, NaOH solution was added to reach a final concentration of 5 M NaOH. Then, the resultant solution was sonicated in water bath for 2 h, and the pH of the solution was adjusted to neutral. The solution was centrifuged at 1,6128 × g for 10 min, and the supernatant was designated as nsGO. Preparation and characterization of nsGO/PCP/OVA Epichlorohydrin was added to the nsGO solution at 40°C for 4 h under N 2 protection. Once the unreacted epichlorohydrin was removed by ultrafiltration, endotoxin-free PCP solution (10 mg/mL) was added to epichlorohydrin-GO solution in water bath at 42°C for 3 h. After the pH of the nsGO/PCP solution was adjusted to weak acidity, a 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydro/N-hydroxy succinimide (EDC/NHS) mixture was added, followed by incubation for 30 min. Subsequently, the OVA protein solution (1 mg/mL) was added to the prepared solution for 1 h at room temperature. Then, the unreacted protein was removed by ultrafiltration and the lyophilized solid was collected as nsGO/PCP/OVA nanoparticles. The concentration of nsGO/PCP/OVA was quantified by the OVA level using the Bicinchoninic Acid Assay kit (BCA, Biomed Inc, CA, United States). The encapsulation efficiency was measured using the following formula: [(Total OVA–Free OVA)/Total OVA] ⋠100%. The nanoparticles were characterized using transmission electron microscopy (TEM, Hitachi, Japan), Fourier Transform infrared spectroscopy (FT-IR spectroscopy, Spectrum 100 FTIR, PerkinElmer, MA, USA), and dynamic light scattering (DLS, Malvern, Germany). TEM images were analyzed for the nanoparticle size distribution using Nano measurer 1.2 software (Fudan University, China). Cell viability To evaluate the cytotoxic effects of nanoparticles, DC2.4 cells were incubated with nsGO/PCP/OVA ranging from 0 to 100 μg/mL OVA. The absorbance at 450 nm was measured using the Cell Counting Kit-8 (Dojindo, Japan) 24 h after treatment. The percentage of cell viability of DC2.4 cells was calculated. Flow cytometry analysis For cell surface marker detecting, the cells were washed with PBS and blocked with anti-mouse CD16/32 antibody (BioLegend, CA, United States). After treatment for 15 min at room temperature, the cells were incubated with different antibodies in PBS with 1% FBS (Gibco, NY, United States) for 1 h. After washed 3 times with PBS, cells were and resuspended in PBS with 1% FBS. Flow cytometry analysis was performed using a BD FACS Calibur™ flow cytometer (BD Bioscience, CA, United States), and data were analyzed using FlowJo software V10 (Tree star, OR, United States). Bone marrow-derived DCs activation Immature BMDCs were collected and treated with 20 μg/mL OVA, 5 μg/mL nsGO, 100 μg/mL PCP, nsGO/PCP, and nsGO/PCP/OVA, according to Dong's research ( Dong et al., 2021 ). After 24 h, the supernatants were collected and measured for IL-6 and IL12 production (ELISA MAX Deluxe kits, BioLegend, CA, United States). The expression of CD80, CD86, and MHC-II on BMDC cell surface was analyzed using PE anti-mouse MHC-II (BioLegend, CA, United States), APC anti-mouse CD80 (BioLegend, CA, United States), and PerCP/Cyanine5.5 anti-mouse CD86 (BioLegend, CA, United States) by FACS analysis. Uptake assays For in vitro uptake assays, BMDCs were incubated with PBS, 5 μg/mL OVA-FITC, and 2.5, 5, and 10 μg/mL nsGO/PCP/OVA-FITC. Cells incubated with nsGO/PCP/OVA-FITC (5 μg/mL) at 4°C were used as controls. After 30 min of incubation, the BMDC cells were collected and stained with APC anti-mouse CD11c (BioLegend, CA, United States) antibody before FACS analysis. For uptake competition assays, cells were pre-treated with PCP, OVA, or 200 μg/mL mannans (Solarbio, China) for 30 min. BMDCs were then treated with nsGO/PCP/OVA-FITC (10 μg/mL) for 45 min. BMDCs were also incubated with nsGO/PCP/OVA (10 μg/mL) for 30 min, followed by co-culture with 100 μg/mL Lucifer yellow VS. dilithium salt (LY, Sigma, MO, United StatesA) for 45 min before FACS analysis. To monitor the in vivo uptake of nanoparticles, mice were injected with OVA-Alexa 647 (20 μg) or nsGO/PCP/OVA-Alexa 647 (containing 20 μg OVA) in both hind footpads. After 6 h, the popliteal lymph nodes were dissected and incubated with collagen D (Sigma, MO, United States) and DNase (Sigma-Aldrich, MO, United States) to prepare single-cell suspensions. Cells were incubated with anti-mouse CD16/32 antibody (BioLegend, CA, United States) for 15 min and then stained with FITC-anti-mouse CD11c antibody (BioLegend, CA, United States) before FACS analysis. Immunization of mice Female C57BL/6 mice were randomly allocated to six groups. On days 0, 7, and 14, mice were immunized subcutaneously with PBS (control), OVA (20 μg per mouse), PCP + OVA (250 μg PCP and 20 μg OVA per mouse), CpG ODN + OVA (10 μg CpG and 20 μg OVA per mouse), nsGO/PCP + OVA (20 μg OVA per mouse), and nsGO/PCP/OVA (containing 20 μg OVA per mouse). On day 21, the mice were anesthetized with isoflurane and sacrificed following the collection of serum and spleen for subsequent analysis. Measurement of Th1 and Th2 immune responses Splenocytes were harvested from mice on day 21 after the three vaccinations. For ELISPOT, cells were cultured at 2 × 10 5 cells/well in ELISPOT plates (Dakewe, China), and stimulated with 1 μg/mL OVA peptide 257-264 (SIINFEKL, OVA â ) or 300 μg/mL OVA protein for 36 h. The number of spots was counted, and the results were expressed as spot-forming cells (SFCs) per 10 5 splenocytes. For the cell proliferation assay, splenocytes were cultured at 2 × 10 6 cells/mL with 300 μg/mL OVA 300 μg/mL or 1 μg/mL OVA â 1 μg/mL. After 48 h, cells were collected, incubated with anti-mouse CD16/32 antibody, and stained with anti-mouse CD69, anti-mouse CD3 (Biolegend, CA, United States), anti-mouse CD4 (Biolegend, CA, United States), and anti-mouse CD8 (Biolegend, CA, United States) antibodies at room temperature for 30 min. To measure cytokine release, splenocytes were cultured with OVA (300 μg/mL) or OVA I (1 μg/mL) for 72 h, supernatants were collected and measured for IL-4 and IFN-γ concentrations using ELISA kits (BioLegend, CA, United States). Prophylactic tumor challenge assays For prophylactic vaccination, C57BL/6 mice were first primed with different formulations on days −21 and boosted on days −14 and −7 as above. On day 0, mice were subcutaneously injected with 2 × 10 5 E.G7-OVA cells in the right flank. On day 21, all mice were euthanized when a humane endpoint was reached, tumor masses were measured, and tumor volume was calculated as length × width 2 × 0.5. Blood samples were collected for antibody detection. Therapeutic tumor vaccination C57BL/6 mice were subcutaneously injected with 2 × 10 5 E.G7-OVA cells into the right flank. When the mean tumor size reached approximately 100 mm 3 , mice were inoculated with different formulations. All mice were euthanized on day 18, when the control mice reached the humane endpoint. Blood and spleen tissues were collected for further analysis. ELISA For antibody detection in prophylactic tumor assays, serum was collected from mice on day 21. The antibody titers of serum OVA-specific IgG1 and IgG2a were measured using an IgG Mouse ELISA kit (Thermo Fisher Scientific, MA, United States). In therapeutic tumor assay, to detect anti-OVA specific antibodies, serum was collected on day 18 and diluted (1:1,000 to detect IgG, 1:500 to detect IgG1, and 1:3,000 for IgG2a) for serum OVA-specific IgG, IgG1, and IgG2a measurements using ELISA kit (Thermo Fisher Scientific, MA, United States). For cytokine measurements in tumor-bearing mice, splenocytes were collected and cultured at 2 × 10 6 cells/mL with 300 μg/mL OVA. After 72 h, the supernatant of cell culture was collected and IL-2 and IL-4 concentrations were measured using ELISA kits (BioLegend, CA, United States). Statistical analyses Statistical analyses were performed using Prism 7.0 (GraphPad Software Inc, CA, United States). All data analysis results in this study are expressed as means ± SD and were analyzed by one-way analysis of variance (ANOVA). p < 0.05 was defined as statistical significance. Result Preparation and characterization of nsGO/PCP/OVA The synthetic route to nsGO/PCP/OVA is illustrated in Figure 1A . nsGO was first conjugated with PCP and then with OVA, and the OVA antigen encapsulation efficiency was calculated as 43.5 ± 4.5%. TEM revealed that nsGO/PCP/OVA was well-dispersed and exhibited a size of approximately 120–150 nm ( Figure 1B ). The combination of PCP and OVA with nsGO was also confirmed by FT-IR spectroscopy ( Figure 1C ), and a broad band at 3,406 cm −1 , attributed to OH groups, and bands at 1,729 cm −1 , typical of carbonyl groups, were observed in GO nanosheets. The presence of a C-O-C stretching peak at 1,075 cm −1 indicated that the polysaccharide was inserted into the nsGO. Two distinct amide I and II peaks for the protein were observed at 1,650 cm −1 and 1,545 cm −1 . The ζ potentials and size distribution of nsGO/PCP/OVA were monitored by DLS analysis for 7 days, and the constructed nanoparticles exhibited sustained storage stability ( Figure 1D ). These data suggested that PCP and OVA successfully bonded to the GO nanosheets. We also evaluated the cytotoxicity of nsGO/PCP/OVA on DC2.4 cells ( Figure 1E ) as well as nsGO ( Supplementary Figure S1 ) and nsGO/PCP/OVA on BMDCs ( Supplementary Figure S2 ). As revealed by CCK-8 assays, DC2.4 cells treated with 0–100 μg/mL nsGO/PCP/OVA showed no obvious cytotoxicity, indicating no detectable cytotoxic effect of nsGO/PCP/OVA in vitro at a concentration of 100 μg/mL. FIGURE 1 Preparation and characterization of nsGO/PCP/OVA. (A) Schematic representation for the preparation of nsGO/PCP/OVA. (B) TEM image and size distribution of nsGO/PCP/OVA. The inset in (B) is the size of nsGO/PCP/OVA nanoparticle. (C) FT-IR spectra of nsGO, nsGO/PCP and nsGO/PCP/OVA. (D) ζ potentials and size distribution of nsGO/PCP/OVA for a week. (E) Viability of DC2.4 cells after nsGO/PCP/OVA treatment was evaluated by CCK-8 assay. Data shown are representative of 3 replicate experiments. nsGO/PCP/OVA induced maturation of BMDCs and enhanced OVA uptake nsGO/PCP/OVA-induced maturation of DCs was assessed by measuring co-stimulatory factor expression and cytokine release by BMDCs. Compared to the control, nsGO/PCP and nsGO/PCP/OVA induced a 2-3-fold upregulation in the surface expression of CD80 ( Figure 2A ), CD86 ( Figure 2B ), and MHC II (referred to as I-A/I-E, Figure 2C ). In addition, nsGO/PCP/OVA-treated DCs exhibited significantly higher production of interleukin 6 (IL-6) and interleukin 12 (IL-12) as determined by ELISA ( Figures 2D,E ). These data suggested that nsGO/PCP/OVA significantly induced BMDC maturation. FIGURE 2 In vitro BMDC maturation induced by nsGO/PCP/OVA. Immature BMDCs were isolated from C57BL/6 female mice and stimulated with 20 μg/mL OVA, 5 μg/mL nsGO, 100 μg/mL PCP, nsGO/PCP, and nsGO/PCP/OVA for 24 h. Expression of BMDC surface markers CD80 (A) , CD86 (B) , and I-A/I-E (C) were analyzed by flow cytometry. Supernatant of BMDC culture was collected, IL-6 (D) and IL-12 (E) secretion by BMDCs was measured by ELISA. Values presented are the means ± SD of three replicates, **, p < 0.01; ****, p < 0.0001 vs. PBS control. The cellular uptake of FITC-labeled nsGO/PCP/OVA by DCs was measured using flow cytometry. According to Figure 3A , there is a dose-dependent increase in uptake of nsGO/PCP/OVA nanoparticles by DCs compared to free OVA and nsGO/PCP/OVA at 4°C. To get a better understanding of the uptake route of nsGO/PCP/OVA, BMDCs were pre-incubated with free PCP, free OVA, and mannan, which was used to block the mannose receptor, a well-recognized endocytic receptor responsible for OVA uptake by BMDCs. The results showed that free PCP, OVA, and mannan caused a decrease in the percentage of OVA-FITC positive cells, indicating that PCP and OVA could both mediate the uptake of nsGO/PCP/OVA, and multiple receptors, including the mannose receptor, were involved in the receptor-mediated endocytosis of nsGO/PCP/OVA ( Figure 3B ). In addition, we incubated BMDCs with Lucifer Yellow, a well-known marker for pinocytosis. As depicted in Figure 3C , there is no reduction in the uptake of Lucifer Yellow, suggesting that the nanoparticles interfered with receptor-mediated endocytosis, but not pinocytosis. Moreover, the efficiency of nanovaccines in co-delivering antigens and adjuvants to lymph nodes was assessed in vivo . After injection of Alexa 647-labeled nsGO/PCP/OVA, the draining lymph nodes of mice were collected, and CD11c + DCs were prepared for analyzing OVA-Alexa 647 positive cells. As shown in Figure 3D , nsGO/PCP/OVA significantly enhanced the uptake of OVA by DCs in draining lymph nodes, compared to OVA alone. These data suggested that nsGO/PCP/OVA promoted the uptake of antigens both in vitro and in vivo . FIGURE 3 Uptake of nsGO/PCP/OVA in vitro and in vivo . (A) BMDCs were pre-incubated with PBS, OVA-FITC at 5 μg/mL, nsGO/PCP/OVA-FITC at 2.5, 5, 10 μg/mL at 37°C or nsGO/PCP/OVA-FITC at 5 μg/mL at 4°C for 30 min, OVA-FITC-positive BMDCs were detected by FACS analysis. (B) BMDCs were pre-incubated with PCP, OVA or mannan for 30 min, cells were cultured with nsGO/PCP/OVA-FITC (10 μg/mL) for 45 min in 37°C before FACS analysis. (C) BMDCs were first incubated with nsGO/PCP/OVA-FITC (10 μg/mL) for 30 min, then BMDCs were co-cultured with Lucifer Yellow for 45 min at 37°C before analysis of antigen uptake by CD11c + BMDCs. The uptake of Lucifer Yellow is shown in histograms (Grey area: cells without pre-incubation with nsGO/PCP/OVA, red line: cells incubated with nsGO/PCP/OVA). (D) Mice were injected with PBS, OVA-Alexa 647 (20 μg per mouse) and nsGO/PCP/OVA-Alexa 647 (20 μg OVA per mouse) in both footpads. After 6 h, popliteal lymph nodes were isolated and prepare for single-cell suspension. Cells were then stained with anti-mouse CD11c antibody to analyze Alexa 647-positive cells using FACS analysis. Values presented are the means ± SD of three replicates (n = 3), *, p < 0.05; ***, p < 0.001 vs. PBS control. nsGO/PCP/OVA induced Th1 and Th2 immune responses in vivo After three vaccinations with nsGO/PCP/OVA at 1 week intervals, mouse spleens were collected and assayed for the number of IFN-γ-producing cells, CD69 expression, as well as cytokine production. After stimulation with OVA or OVA I for 36 h, IFN-γ-secreting cells were quantitated by ELISPOT ( Figure 4A ), which showed that both OVA-specific ( Figure 4B ) and OVA I-specific ( Figure 4C ) IFN-γ-secreting CD8 + T cell numbers significantly increased after nsGO/PCP/OVA vaccination, suggesting that nsGO/PCP/OVA activated OVA-specific CD4 + and CD8 + T cells. The proliferation of OVA-specific CD8 + T cells after nsGO/PCP/OVA treatment was further supported by upregulation of CD69 expression in CD8 + T cells ( Figures 4D,E ). Moreover, immunization with nsGO/PCP/OVA elevated the secretion of IFN-γ ( Figure 4F ) and IL-4 ( Figure 4G ) compared to the control group. Together, these results demonstrated the robust adjuvant effect of nsGO/PCP/OVA with enhanced Th1 and Th2 immune responses. FIGURE 4 Stimulation of T cells in immunized mice by nsGO/PCP/OVA. Mice were immunized 3 times (s.c.) with PBS, OVA, PCP + OVA, CpG ODN + OVA, nsGO/PCP + OVA or nsGO/PCP/OVA at 1 week intervals. Blood and spleen tissues were collected on day 21. (A) Splenocytes were stimulated with 300 μg/mL OVA protein or 1 μg/mL OVA peptide 257–264 (OVAI) per well. After 36 h, IFN-γ secreting cells were visualized and analyzed by ELISPOT. The number of OVA-stimulated (B) and OVA â - (C) stimulated IFN-γ secreting cells was measured by ELISPOT. Splenocytes were co-cultured with OVA (300 μg/mL) or OVA â (1 μg/mL) for 48 h, CD69 + CD4 + T (D) and CD69 + CD8 + T (E) cells were analyzed using FACS. Supernatants of splenocytes were collected at 72 h and measured for IFN-γ (F) and IL-4 (G) levels using ELISA kits. Data are presented as means ± SD (n = 5), *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 vs. PBS control. Prophylactic and therapeutic effects of nsGO/PCP/OVA on E.G7-OVA tumor-bearing mice The prophylactic anti-tumor effect against E.G7-OVA tumor cells, a lymphoma cell line that stably expresses OVA, was evaluated to assess the potential of nsGO/PCP/OVA as an antineoplastic vaccine. C57BL/6 female mice were subcutaneously immunized with nsGO/PCP/OVA or control formulations 3 times at 1week intervals. As depicted in Figure 5A , all mice were challenged with E.G7-OVA cells 1 week after the final vaccination. As shown in Figure 5B , tumor growth in the nsGO/PCP/OVA group was significantly lower than that in the control group. In addition, data concerning serum OVA-specific antibody production showed that nsGO/PCP/OVA upregulated anti-OVA IgG1 ( Figure 5C ) and IgG2a ( Figure 5D ) antibodies, indicating that nsGO/PCP/OVA induced a potent immune response in tumor prophylaxis. FIGURE 5 Prophylactic anti-tumor effects of nsGO/PCP/OVA on mice. (A) Immunization scheme for tumor prevention. Mice were primed subcutaneously with either PBS, OVA, CpG + OVA, PCP + OVA, nsGO/PCP + OVA, or nsGO/PCP/OVA and boosted with the same antigens twice at 1 week intervals. On day 0, 2 × 10 5 (E) G7-OVA cells were injected subcutaneously into the right flank of mice. (B) On day 21, mice were sacrificed and tumor masses were measured. Blood was collected on day 21. The concentrations of OVA-specific IgG1 (C) and IgG2a (D) secretions in mouse serum were measured by ELISA. Data are presented as means ± SD (n = 5), **, p < 0.01; ****, p < 0.0001 vs. PBS control. Next, we evaluated the elimination of established tumors by inoculation with nsGO/PCP/OVA. One week after inoculation with E.G7-OVA tumor cells, mice received vaccination of nsGO/PCP/OVA and control formulas, respectively ( Figure 6A ). All mice were euthanized on day 18, when the control mice reached the humane endpoint, and a decline in tumor volume ( Figure 6B ) and weight ( Figure 6C ) was observed in the nsGO/PCP/OVA-treated mice. In addition, we analyzed OVA-specific anti-OVA IgG ( Figure 6D ), IgG1 ( Figure 6E ), IgG2c ( Figure 6F ) in serum and cytokine IL-4 ( Figure 6G ), IL-2 ( Figure 6H ) release from splenocytes. We found that nsGO/PCP/OVA significantly upregulated serum anti-OVA IgG ( Figure 6D ), IgG1 ( Figure 6E ) antibodies, and IL-4 release ( Figure 6G ), indicating that nsGO/PCP/OVA induced a potent humoral immune response in mice with tumors. FIGURE 6 Therapeutic anti-tumor effect of nsGO/PCP/OVA on tumor-bearing mice. (A) The scheme of vaccines for tumor treatment. Mice were subcutaneously injected with 2 × 10 5 (E) G7-OVA cells into the right flank. Seven days later, the mice were administered either PBS, OVA, CpG + OVA, PCP + OVA, nsGO/PCP + OVA, or nsGO/PCP/OVA. Tumor volumes (B) and weight (C) were measured after sacrifice on day 18 because the control mice reached a humane endpoint. Blood was collected on day 18 from tumor-bearing mice. The concentrations of OVA-specific IgG (D) , IgG1 (E) and IgG2a (F) secretions in mouse serum were measured by ELISA. Splenocytes were collected on day 18 and cultured with OVA (300 μg/mL) for 72 h, IL-4 (G) and IL-2 (H) concentrations in the supernatants of splenocytes were measured using ELISA kits. Data are presented as means ± SD (n = 5), *, p < 0.05; **, p < 0.01; ****, p < 0.0001 vs. PBS control. Preparation and characterization of nsGO/PCP/OVA The synthetic route to nsGO/PCP/OVA is illustrated in Figure 1A . nsGO was first conjugated with PCP and then with OVA, and the OVA antigen encapsulation efficiency was calculated as 43.5 ± 4.5%. TEM revealed that nsGO/PCP/OVA was well-dispersed and exhibited a size of approximately 120–150 nm ( Figure 1B ). The combination of PCP and OVA with nsGO was also confirmed by FT-IR spectroscopy ( Figure 1C ), and a broad band at 3,406 cm −1 , attributed to OH groups, and bands at 1,729 cm −1 , typical of carbonyl groups, were observed in GO nanosheets. The presence of a C-O-C stretching peak at 1,075 cm −1 indicated that the polysaccharide was inserted into the nsGO. Two distinct amide I and II peaks for the protein were observed at 1,650 cm −1 and 1,545 cm −1 . The ζ potentials and size distribution of nsGO/PCP/OVA were monitored by DLS analysis for 7 days, and the constructed nanoparticles exhibited sustained storage stability ( Figure 1D ). These data suggested that PCP and OVA successfully bonded to the GO nanosheets. We also evaluated the cytotoxicity of nsGO/PCP/OVA on DC2.4 cells ( Figure 1E ) as well as nsGO ( Supplementary Figure S1 ) and nsGO/PCP/OVA on BMDCs ( Supplementary Figure S2 ). As revealed by CCK-8 assays, DC2.4 cells treated with 0–100 μg/mL nsGO/PCP/OVA showed no obvious cytotoxicity, indicating no detectable cytotoxic effect of nsGO/PCP/OVA in vitro at a concentration of 100 μg/mL. FIGURE 1 Preparation and characterization of nsGO/PCP/OVA. (A) Schematic representation for the preparation of nsGO/PCP/OVA. (B) TEM image and size distribution of nsGO/PCP/OVA. The inset in (B) is the size of nsGO/PCP/OVA nanoparticle. (C) FT-IR spectra of nsGO, nsGO/PCP and nsGO/PCP/OVA. (D) ζ potentials and size distribution of nsGO/PCP/OVA for a week. (E) Viability of DC2.4 cells after nsGO/PCP/OVA treatment was evaluated by CCK-8 assay. Data shown are representative of 3 replicate experiments. nsGO/PCP/OVA induced maturation of BMDCs and enhanced OVA uptake nsGO/PCP/OVA-induced maturation of DCs was assessed by measuring co-stimulatory factor expression and cytokine release by BMDCs. Compared to the control, nsGO/PCP and nsGO/PCP/OVA induced a 2-3-fold upregulation in the surface expression of CD80 ( Figure 2A ), CD86 ( Figure 2B ), and MHC II (referred to as I-A/I-E, Figure 2C ). In addition, nsGO/PCP/OVA-treated DCs exhibited significantly higher production of interleukin 6 (IL-6) and interleukin 12 (IL-12) as determined by ELISA ( Figures 2D,E ). These data suggested that nsGO/PCP/OVA significantly induced BMDC maturation. FIGURE 2 In vitro BMDC maturation induced by nsGO/PCP/OVA. Immature BMDCs were isolated from C57BL/6 female mice and stimulated with 20 μg/mL OVA, 5 μg/mL nsGO, 100 μg/mL PCP, nsGO/PCP, and nsGO/PCP/OVA for 24 h. Expression of BMDC surface markers CD80 (A) , CD86 (B) , and I-A/I-E (C) were analyzed by flow cytometry. Supernatant of BMDC culture was collected, IL-6 (D) and IL-12 (E) secretion by BMDCs was measured by ELISA. Values presented are the means ± SD of three replicates, **, p < 0.01; ****, p < 0.0001 vs. PBS control. The cellular uptake of FITC-labeled nsGO/PCP/OVA by DCs was measured using flow cytometry. According to Figure 3A , there is a dose-dependent increase in uptake of nsGO/PCP/OVA nanoparticles by DCs compared to free OVA and nsGO/PCP/OVA at 4°C. To get a better understanding of the uptake route of nsGO/PCP/OVA, BMDCs were pre-incubated with free PCP, free OVA, and mannan, which was used to block the mannose receptor, a well-recognized endocytic receptor responsible for OVA uptake by BMDCs. The results showed that free PCP, OVA, and mannan caused a decrease in the percentage of OVA-FITC positive cells, indicating that PCP and OVA could both mediate the uptake of nsGO/PCP/OVA, and multiple receptors, including the mannose receptor, were involved in the receptor-mediated endocytosis of nsGO/PCP/OVA ( Figure 3B ). In addition, we incubated BMDCs with Lucifer Yellow, a well-known marker for pinocytosis. As depicted in Figure 3C , there is no reduction in the uptake of Lucifer Yellow, suggesting that the nanoparticles interfered with receptor-mediated endocytosis, but not pinocytosis. Moreover, the efficiency of nanovaccines in co-delivering antigens and adjuvants to lymph nodes was assessed in vivo . After injection of Alexa 647-labeled nsGO/PCP/OVA, the draining lymph nodes of mice were collected, and CD11c + DCs were prepared for analyzing OVA-Alexa 647 positive cells. As shown in Figure 3D , nsGO/PCP/OVA significantly enhanced the uptake of OVA by DCs in draining lymph nodes, compared to OVA alone. These data suggested that nsGO/PCP/OVA promoted the uptake of antigens both in vitro and in vivo . FIGURE 3 Uptake of nsGO/PCP/OVA in vitro and in vivo . (A) BMDCs were pre-incubated with PBS, OVA-FITC at 5 μg/mL, nsGO/PCP/OVA-FITC at 2.5, 5, 10 μg/mL at 37°C or nsGO/PCP/OVA-FITC at 5 μg/mL at 4°C for 30 min, OVA-FITC-positive BMDCs were detected by FACS analysis. (B) BMDCs were pre-incubated with PCP, OVA or mannan for 30 min, cells were cultured with nsGO/PCP/OVA-FITC (10 μg/mL) for 45 min in 37°C before FACS analysis. (C) BMDCs were first incubated with nsGO/PCP/OVA-FITC (10 μg/mL) for 30 min, then BMDCs were co-cultured with Lucifer Yellow for 45 min at 37°C before analysis of antigen uptake by CD11c + BMDCs. The uptake of Lucifer Yellow is shown in histograms (Grey area: cells without pre-incubation with nsGO/PCP/OVA, red line: cells incubated with nsGO/PCP/OVA). (D) Mice were injected with PBS, OVA-Alexa 647 (20 μg per mouse) and nsGO/PCP/OVA-Alexa 647 (20 μg OVA per mouse) in both footpads. After 6 h, popliteal lymph nodes were isolated and prepare for single-cell suspension. Cells were then stained with anti-mouse CD11c antibody to analyze Alexa 647-positive cells using FACS analysis. Values presented are the means ± SD of three replicates (n = 3), *, p < 0.05; ***, p < 0.001 vs. PBS control. nsGO/PCP/OVA induced Th1 and Th2 immune responses in vivo After three vaccinations with nsGO/PCP/OVA at 1 week intervals, mouse spleens were collected and assayed for the number of IFN-γ-producing cells, CD69 expression, as well as cytokine production. After stimulation with OVA or OVA I for 36 h, IFN-γ-secreting cells were quantitated by ELISPOT ( Figure 4A ), which showed that both OVA-specific ( Figure 4B ) and OVA I-specific ( Figure 4C ) IFN-γ-secreting CD8 + T cell numbers significantly increased after nsGO/PCP/OVA vaccination, suggesting that nsGO/PCP/OVA activated OVA-specific CD4 + and CD8 + T cells. The proliferation of OVA-specific CD8 + T cells after nsGO/PCP/OVA treatment was further supported by upregulation of CD69 expression in CD8 + T cells ( Figures 4D,E ). Moreover, immunization with nsGO/PCP/OVA elevated the secretion of IFN-γ ( Figure 4F ) and IL-4 ( Figure 4G ) compared to the control group. Together, these results demonstrated the robust adjuvant effect of nsGO/PCP/OVA with enhanced Th1 and Th2 immune responses. FIGURE 4 Stimulation of T cells in immunized mice by nsGO/PCP/OVA. Mice were immunized 3 times (s.c.) with PBS, OVA, PCP + OVA, CpG ODN + OVA, nsGO/PCP + OVA or nsGO/PCP/OVA at 1 week intervals. Blood and spleen tissues were collected on day 21. (A) Splenocytes were stimulated with 300 μg/mL OVA protein or 1 μg/mL OVA peptide 257–264 (OVAI) per well. After 36 h, IFN-γ secreting cells were visualized and analyzed by ELISPOT. The number of OVA-stimulated (B) and OVA â - (C) stimulated IFN-γ secreting cells was measured by ELISPOT. Splenocytes were co-cultured with OVA (300 μg/mL) or OVA â (1 μg/mL) for 48 h, CD69 + CD4 + T (D) and CD69 + CD8 + T (E) cells were analyzed using FACS. Supernatants of splenocytes were collected at 72 h and measured for IFN-γ (F) and IL-4 (G) levels using ELISA kits. Data are presented as means ± SD (n = 5), *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 vs. PBS control. Prophylactic and therapeutic effects of nsGO/PCP/OVA on E.G7-OVA tumor-bearing mice The prophylactic anti-tumor effect against E.G7-OVA tumor cells, a lymphoma cell line that stably expresses OVA, was evaluated to assess the potential of nsGO/PCP/OVA as an antineoplastic vaccine. C57BL/6 female mice were subcutaneously immunized with nsGO/PCP/OVA or control formulations 3 times at 1week intervals. As depicted in Figure 5A , all mice were challenged with E.G7-OVA cells 1 week after the final vaccination. As shown in Figure 5B , tumor growth in the nsGO/PCP/OVA group was significantly lower than that in the control group. In addition, data concerning serum OVA-specific antibody production showed that nsGO/PCP/OVA upregulated anti-OVA IgG1 ( Figure 5C ) and IgG2a ( Figure 5D ) antibodies, indicating that nsGO/PCP/OVA induced a potent immune response in tumor prophylaxis. FIGURE 5 Prophylactic anti-tumor effects of nsGO/PCP/OVA on mice. (A) Immunization scheme for tumor prevention. Mice were primed subcutaneously with either PBS, OVA, CpG + OVA, PCP + OVA, nsGO/PCP + OVA, or nsGO/PCP/OVA and boosted with the same antigens twice at 1 week intervals. On day 0, 2 × 10 5 (E) G7-OVA cells were injected subcutaneously into the right flank of mice. (B) On day 21, mice were sacrificed and tumor masses were measured. Blood was collected on day 21. The concentrations of OVA-specific IgG1 (C) and IgG2a (D) secretions in mouse serum were measured by ELISA. Data are presented as means ± SD (n = 5), **, p < 0.01; ****, p < 0.0001 vs. PBS control. Next, we evaluated the elimination of established tumors by inoculation with nsGO/PCP/OVA. One week after inoculation with E.G7-OVA tumor cells, mice received vaccination of nsGO/PCP/OVA and control formulas, respectively ( Figure 6A ). All mice were euthanized on day 18, when the control mice reached the humane endpoint, and a decline in tumor volume ( Figure 6B ) and weight ( Figure 6C ) was observed in the nsGO/PCP/OVA-treated mice. In addition, we analyzed OVA-specific anti-OVA IgG ( Figure 6D ), IgG1 ( Figure 6E ), IgG2c ( Figure 6F ) in serum and cytokine IL-4 ( Figure 6G ), IL-2 ( Figure 6H ) release from splenocytes. We found that nsGO/PCP/OVA significantly upregulated serum anti-OVA IgG ( Figure 6D ), IgG1 ( Figure 6E ) antibodies, and IL-4 release ( Figure 6G ), indicating that nsGO/PCP/OVA induced a potent humoral immune response in mice with tumors. FIGURE 6 Therapeutic anti-tumor effect of nsGO/PCP/OVA on tumor-bearing mice. (A) The scheme of vaccines for tumor treatment. Mice were subcutaneously injected with 2 × 10 5 (E) G7-OVA cells into the right flank. Seven days later, the mice were administered either PBS, OVA, CpG + OVA, PCP + OVA, nsGO/PCP + OVA, or nsGO/PCP/OVA. Tumor volumes (B) and weight (C) were measured after sacrifice on day 18 because the control mice reached a humane endpoint. Blood was collected on day 18 from tumor-bearing mice. The concentrations of OVA-specific IgG (D) , IgG1 (E) and IgG2a (F) secretions in mouse serum were measured by ELISA. Splenocytes were collected on day 18 and cultured with OVA (300 μg/mL) for 72 h, IL-4 (G) and IL-2 (H) concentrations in the supernatants of splenocytes were measured using ELISA kits. Data are presented as means ± SD (n = 5), *, p < 0.05; **, p < 0.01; ****, p < 0.0001 vs. PBS control. Discussion As an adjuvant, PCP is known for its potential to improve immunogenicity by triggering antigen-specific immune responses against cancer metastasis ( Wu et al., 2016 ; Gai et al., 2017 ; Liu et al., 2020 ). However, owing to its relatively poor stability and untargeted features ( Zhao et al., 2022 ), PCP is limited in clinical applications. Recently, the construction of nanoparticles has been shown to reduce drug loss during delivery, enhance the solubility of hydrophobic drugs, improve drug targeting, and extend drug release ( Dudek et al., 2016 ; Wen et al., 2019 ). In the present study, we generated nanoparticles assembled from PCP adjuvant, OVA antigen, and nsGO nanosheets, and investigated their potential to enhance humoral and cellular immune responses and their therapeutic and prophylactic antitumor effects in E.G7-OVA-bearing mice. Our data demonstrated that nsGO/PCP/OVA induced robust activation of BMDCs and enhanced antigen uptake both in vitro and in vivo . Compared to free PCP, nsGO/PCP/OVA elicited stronger Th1 and Th2 responses in mice, as shown by the significant upregulation of IFN-γ-secreting CD8 + and CD4 + T cells, as well as the production of IFN-γ and IL-4. Furthermore, nsGO/PCP/OVA treatment exhibited antitumor effects against E.G7-OVA in both prophylactic and therapeutic mouse models. The application of nanotechnology in drug delivery marks an unparalleled opportunity to change the foreseeable future of the pharmaceutical and biotechnological industries. GO offers excellent opportunities for; vaccination, such as enhancing antigen uptake by DCs and stimulating antigen-specific humoral and cellular immunity, thereby achieving robust cancer immunotherapy ( Wang et al., 2022 ). GO has a typical two-dimensional crystal structure with a single atomic layer and oxygen functional groups. The basic skeleton of this 2D atomic planar structure is composed of crumpled sheets of sp 2– and sp 3– hybridized carbon atoms arranged in a hexagonal grid, which provides GO with hydrophobic nature and a large surface area. Hydrophilic groups including epoxy, carboxyl, and hydroxyl are dispersed over the basal planes and edges of the skeleton ( Bao et al., 2011 ; Sharma et al., 2018 ). GO exhibit a large surface area that is almost 10 times the size of other nanomaterials ( Bao et al., 2011 ), which endows GO with superiority on delivery over other materials. Owing to its unique physicochemical properties, GO has the potential to boost the immune system and thus could be employed to deliver antigens into DCs ( Zhang et al., 2022 ). However, the toxic effects of GO on living cells and organs limit its application in the medical field. Researchers have demonstrated that Graphene-Family Nanomaterials (GFNs) can be toxic to cells; in particular, GFNs cause dose-dependent oxidative stress in cells owing to their inherent chemical properties (oxidation state and lateral size), and it is speculated that the biological response can be different over the material family depending on the number of layers, stiffness, lateral size, surface functionalization, hydrophobicity, and dose ( Zhang et al., 2010 ; Chang et al., 2011 ; Wang et al., 2011 ; Sanchez et al., 2012 ; Jia et al., 2019 ). Previous reports have noted that reduced GO (rGO) and carboxylated graphene exhibited lower toxicity than GO or native graphene, indicating that the surface modification of graphene can affect its toxicity ( Yang et al., 2010 ; Sasidharan et al., 2011 ). The functional derivatives of GO possess distinctive features that make it more effective for biomedical applications. Functionalized GO is distinguished by its ability to disperse in numerous solvents, which facilitates its use and lowers toxicity ( Paredes et al., 2008 ). Therefore, in our study, we utilized nsGO with functional modifications to co-deliver antigens and adjuvants effectively with fewer side effects compared to GO, and no cytotoxicity of nsGO/PCP/OVA was detected in cell viability assays at a concentration of 100 μg/mL, further suggesting the safety of the formed nanoparticles. Among all types of immunotherapies, the efficiency of delivering the antigen peptide and adjuvant to lymph nodes and engulfment by APCs is key factors affecting immunotherapeutic effects. Prior studies have noted the importance of particle size in determining vaccine access to lymph nodes ( Zhang et al., 2017 ). Generally, nanoparticles with a size of 10–200 nm can enter lymphatic vessels and be effectively engulfed by APCs in the lymph nodes ( Bachmann and Jennings, 2010 ). In this study, the size of the formed nsGO/PCP/OVA was stable at approximately 120–150 nm, which suggests its potential lymph node targeting capability. In addition, results from in vivo uptake assays further demonstrated the superiority of nanovaccines in delivering antigens and adjuvants to draining lymph nodes. The results from the in vitro study also confirmed that nsGO/PCP/OVA could effectively co-deliver PCP and antigenic peptide-OVA into DCs, as shown by the maturation and activation of DCs. Sufficient expression of antigen and co-stimulatory molecules as well as the secretion of IL-6 and IL-12 are all necessary for effective DC function ( Trinchieri, 2003 ; Kaka et al., 2008 ). According to our findings, nsGO/PCP/OVA induced DC maturation by enhancing the expression of MHC class II and co-stimulatory molecule, as well as cytokine secretion. To explore how nsGO/PCP/OVA entered DCs, we conducted a compete uptake assay by adding sufficient free PCP, free OVA, and mannan. Previous reports have noted that mannan is broadly utilized in the blockage of mannose receptors on DCs and the mannose receptor-mediated pathway has been reported to be involved in the uptake of OVA by DCs ( Becker et al., 2006 ; Burgdorf et al., 2007 ). The data showed that mannan, PCP, and OVA significantly inhibited the uptake of nsGO/PCP/OVA, indicating that PCP and OVA mediated the uptake of nanoparticles, and that the mannose receptor was involved in receptor-mediated endocytosis. Furthermore, no significant reduction was found in Lucifer Yellow uptake, indicating that nsGO/PCP/OVA primarily affected receptor-mediated endocytosis rather than pinocytosis. To further elucidate the capability of nsGO/PCP/OVA to enhance cellular immune responses, OVA was utilized as a model antigen, whose MHC I epitope (OVA 257–264) and MHC II epitope (OVA 323–339) have been well studied ( Rötzschke et al., 1991 ; Lipford et al., 1993 ). We also used CpG ODN 1668 as a positive control to obtain a better understanding of the cellular immune response induced by nsGO/PCP/OVA. The TLR9 ligand CpG ODN, a well-documented Th1-related adjuvant, has been authorized for the use in the Heplisav-B (HepB-CpG) vaccine to aids cross-presentation of MHC I-restricted antigens since 2018 ( Schillie et al., 2018 ; Hyer and Janssen, 2019 ). After three vaccinations, splenocytes of immunized mice were harvested and re-stimulated with OVA or OVA I. Successful antitumor vaccines likely require both CD 4 + and CD8 + T cell responses, as reported ( Joffre et al., 2012 ). Effective cross-presentation of antigens by DCs plays crucial role of initiating optimal CD8 + T cell responses, especially in vaccines against intracellular antigens (intracellular microbes, viruses) and cancer. If nsGO/PCP/OVA could promote the cross-presentation of exogenous OVA protein in DCs, CD8 + T lymphocytes that particularly identify the SIINFEKL (MHC class I-restricted) epitope of OVA would proliferate and be capable of increasing IFN-γ secretion. This hypothesis was confirmed by FACS analysis, ELISPOT assays, and ELISA results in that nsGO/PCP/OVA were potent at cross-priming, activating specific cytotoxic CD8 + T cells, increasing the production of Th1 type cytokines as well, and triggering Th2 immune responses. The results of the prophylactic tumor assay showed that nsGO/PCP/OVA significantly inhibited tumor growth, suggesting the superior capability of nsGO/PCP/OVA to prevent tumor occurrence. The production of IgG antibodies in the nanoparticle groups further confirmed the ability of nsGO/PCP/OVA to induce strong humoral immune responses. As nsGO/PCP/OVA vaccination could effectively inhibit tumor growth after prophylactic immunization, we sought to investigate nanoparticle-mediated adaptive immune responses in therapeutic vaccinations. Similar to the prophylactic vaccination, tumors generated in nsGO/PCP/OVA-immunized mice exhibited a tendency of suppression compared to that in the PBS- or OVA + PCP-immunized groups. In addition, the production of IL-4 and IgG antibodies in nanoparticle groups were significantly increased compared to other groups. A recent study showed that GO can induce the differentiation of Th0 cells into Th2 cells, resulting in the promotion of humoral immunity ( Lategan et al., 2018 ). Thus, our results further confirmed the hypothesis that tumor antigen-specific CD4 + T cells are functionally activated by nsGO/PCP/OVA nanoparticle and exhibit strong humoral immune responses. In addition, our research concerning nanoparticle vaccines leaves some room for improvement. In subsequent research, it will be necessary to investigate immunological memory induced by nanoparticle vaccines in surviving mice as a supplement ( Chen et al., 2016 ; Yang et al., 2017 ). Moreover, several reports have shown that cancer vaccines can be combined with immune checkpoint blockade (ICB) therapies, such as α-PD-1, to improve therapeutic outcomes ( Egen et al., 2002 ; Topalian et al., 2012 ; Crittenden et al., 2015 ; Melero et al., 2015 ). Hence, further research should be undertaken to compare the therapeutic efficacy of nsGO/PCP/OVA nanovaccines and nanovaccines in combination with α-PD-1 therapy against established tumors. Although nsGO vaccine technology has shown robust immune effects, its relatively poor stability and low solubility limit its application in the human body ( Zheng et al., 2018 ). Therefore, further modifications of nsGO are required to achieve improved biocompatibility. Conclusion In this study, nsGO/PCP/OVA nanoparticles were constructed for tumor immunotherapy by encapsulating the antigen protein OVA and adjuvant PCP with nsGO nanosheets. Nanoparticles enhance the cellular uptake of antigens, promote the maturation of BMDCs in vitro , and induce both Th1 and Th2 responses in vivo . Importantly, nsGO/PCP/OVA could be utilized not only as a prophylactic nanovaccine against E.G7-OVA tumor challenge, but also as a therapeutic nanovaccine to inhibit the growth of existing tumors. Given the urgent need for safe vaccination platforms that trigger humoral and cellular immunity in the treatment of malignancies, the nanoparticle vaccine described here is of potential value for future applications in cancer treatment. Data availability statement The original contributions presented in the study are included in the article/ Supplementary Material , further inquiries can be directed to the corresponding authors. Ethics statement The animal study was reviewed and approved by the Institutional Animal Treatment and Use Committee, China Academy of Chinese Medica Sciences. Author contributions JY, XD, and QH conceived of and designed the study. JY, XD, and QH designed experiments. JY, XD, BL, TC, BY, and QH performed or assisted with the experiments. JY, XD, and QH analyzed the data. JY and QH wrote the manuscript, with contributions from XW, XD, and BP. All authors have approved the final manuscript. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fbioe.2022.1050077/full#supplementary-material Click here for additional data file. Click here for additional data file. Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6504709/

MicroRNA Post-transcriptional Regulation of the NLRP3 Inflammasome in Immunopathologies

Inflammation has a crucial role in protection against various pathogens. The inflammasome is an intracellular multiprotein signaling complex that is linked to pathogen sensing and initiation of the inflammatory response in physiological and pathological conditions. The most characterized inflammasome is the NLRP3 inflammasome, which is a known sensor of cell stress and is tightly regulated in resting cells. However, altered regulation of the NLRP3 inflammasome is found in several pathological conditions, including autoimmune disease and cancer. NLRP3 expression was shown to be post-transcriptionally regulated and multiple miRNA have been implicated in post-transcriptional regulation of the inflammasome. Therefore, in recent years, miRNA based post-transcriptional control of NLRP3 has become a focus of much research, especially as a potential therapeutic approach. In this review, we provide a summary of the recent investigations on the role of miRNA in the post-transcriptional control of the NLRP3 inflammasome, a key regulator of pro-inflammatory IL-1β and IL-18 cytokine production. Current approaches to targeting the inflammasome product were shown to be an effective treatment for diseases linked to NLRP3 overexpression. Although utilizing NLRP3 targeting miRNAs was shown to be a successful therapeutic approach in several animal models, their therapeutic application in patients remains to be determined. Inflammasome Structure In 2002, the ground breaking work published by Martinon et al. (2002) has demonstrated the role of the inflammasome, a multi-protein complex, in the activation of pro-inflammatory caspases. The authors described the multistep process of the inflammasome assembly which is initiated by the detection of pathogen-associated molecular patterns (PAMPs) or danger signals released by damaged cells ( Duncan et al., 2009 ; Ichinohe et al., 2010 ; Costa et al., 2012 ). Several inflammasome sensors were later identified including the nucleotide-binding oligomerization domain (NOD) like receptors (NLRs), the absent in melanoma-2 like receptors (ALRs) and pyrin ( Ting et al., 2008 ). In the past decade our understanding of NLR containing inflammasomes structure and assembly mechanisms has advanced considerably, largely due to their potential involvement in pathogenesis of several diseases ( Hoffman et al., 2001 ; Alexander So and Borbála Pazár, 2010 ; Song et al., 2017 ). NLRs contain three domains, an N-terminal domain, a NOD, and a C-terminal leucine-rich repeat (LRR) ( Inohara and Nunez, 2003 ). The N-terminal domain contains a caspase recruitment domain (CARD) or pyrin domain (PYD), which function to interact with downstream molecules, such as apoptosis-associated speck-like protein containing (ASC) ( Inohara and Nunez, 2003 ; Schroder and Tschopp, 2010 ). The NOD domain is linked to LRR detecting PAMPs ( Boekhout et al., 2011 ). Upon sensing PAMPs, the NLRs polymerize followed by the interaction between the PYD or CARD domains of LLR and ASC ( Stutz et al., 2013 ). Once activated the inflammasome adopts a wheel-like structure ( Hu et al., 2015 ), where CARD–CARD interactions are essential for recruiting pro-caspase 1 (PC1) into close proximity with the complex ( Faustin et al., 2007 ). PC1 becomes proteolytically cleaved by the CARD domain releasing an active caspase 1 (AC1) p10/p20 tetramer ( Martinon et al., 2002 ; Kanneganti et al., 2006 ; Boucher et al., 2018 ). NLR Inflammasomes This family of inflammasomes includes two subgroups based on the presence of CARD or pyrin in the N terminus. Only nucleotide-binding domain leucine-rich repeats proteins (NLRP)1, NLRP3, and NLRC4 were shown to form inflammasomes that produce AC1 ( Mao et al., 2014 ). In contrast, NLRP6, NLRP9b, and NLRP12 are believed to form inflammasomes, but their roles as inflammasome sensors are less recognized ( Anand et al., 2012 ; Vladimer et al., 2012 ; Zhu et al., 2017 ). NLRP1 NLRP1 was the first identified cytosolic receptor capable of forming active inflammasomes ( Martinon et al., 2002 ). PYD, NBD, and LRR domains, a 'function-to find' domain (FIIND) and a C-terminal CARD are the structural components of NLRP1 ( Jin et al., 2013b ). Our knowledge of NLRP1 function comes largely from studying animal models. It appears that NLRP1 senses and protects against microbial pathogens, as was shown using a mouse model of Bacillus anthracis and Shigella flexneri infection ( Boyden and Dietrich, 2006 ; Sandstrom et al., 2019 ). Additionally, NLRP1 inflammasomes facilitate parasite clearance and protection as demonstrated in Toxoplasma gondii infection in mouse and rat models ( Cirelli et al., 2014 ; Gorfu et al., 2014 ). The clinical relevance of NLRP1 inflammasomes against Toxoplasma gondii is also evident in individuals with specific single-nucleotide polymorphisms in the NLRP1 gene, which are linked to congenital toxoplasmosis ( Witola et al., 2011 ). Aberrant activation of NLRP1 is linked to a pathogenesis of inflammatory diseases. Polymorphisms in the NLRP1 gene are linked to Crohn's disease, rheumatoid arthritis (RA) and systemic sclerosis ( Finger et al., 2012 ). Although the mechanism of NLRP1 activation remains largely unknown, recently, the failure of inflammasome inhibition by dipeptidyl dipeptidase 9 (DDP9), linked to antigen processing ( Zhong et al., 2018 ), was demonstrated to play role in pathogenesis of an autoimmune diseases ( Zhong et al., 2018 ). The authors identified that a single mutation in the FIIND domain of NLRP1 abrogates binding to DPP9, triggering over activation of the inflammasome in autoinflammatory disease AIADK. NLRC4 Similar to NLRP1, NLRC4 establishes protection against infectious pathogens ( Mariathasan et al., 2004 ; Franchi et al., 2006 ; Zhao et al., 2011 ). In the absence of stimulus, NLRC4 remains inactive, where its NBD domain retains a closed conformation by binding to the winged helix domain ( Tenthorey et al., 2014 ). NLRC4 activation is indirect, and it requires NLR family apoptosis inhibitory proteins (NAIPs) for the initial sensing of the microbial ligand ( Rayamajhi et al., 2013 ; Yang et al., 2013 ; Kortmann et al., 2015 ). NAIPs trigger NLRC4 oligomerization, which is essential for inflammasome activation ( Hu et al., 2015 ). Loss of the control over NLRC4 expression and subsequent production of AC1 and release of IL-1β by macrophages was suggested to play role in the pathogenesis of inflammasome linked autoinflammation ( von Moltke et al., 2012 ; Canna et al., 2014 ). Also, a missense mutation in the NLRC4 gene was found in familial cold autoinflammatory syndrome ( Kitamura et al., 2014 ). Multiple mutations in NLRC4 were identified in several autoinflammatory diseases including atopic dermatitis, periodic fever, and fatal or near-fatal episodes of autoinflammation ( Nakamura et al., 2010 ; Canna et al., 2014 ; Bonora et al., 2015 ). These data suggest that NLRC4 plays role in protection against microbial pathogens and autoinflammation. NLRP6 NLRP6 is an inflammasome which plays a role in gut health and maintaining mucosal response to pathogens ( Elinav et al., 2011 ; Anand et al., 2012 ). A microbial metabolite, taurine, was identified as an NLRP6 activator ( Levy et al., 2015 ). The NLRP6-taurite axis appears to be essential for the health of the gut mucosa and microbiome. Taurite produced by the normal microbiota activates NLRP6 which prevents dysbacteriosis by promoting production of antimicrobial peptides ( Levy et al., 2015 ). NLRP12 NLRP12 is intracellular protein expressed in cells of myeloid lineages ( Arthur et al., 2010 ). NLRP12 inflammasome expression can be downregulated by microbial ligands ( Williams et al., 2005 ; Lich et al., 2007 ) via canonical and non-canonical inhibition of NF-κB ( Zaki et al., 2011 ; Allen et al., 2012 ). Several ligands were identified as NLRP12 activators including microbes ( Allen et al., 2012 ; Vladimer et al., 2012 ). ALR Family Inflammasomes ALR family inflammasomes contain an N-terminal PYD and a C-terminal hematopoietic interferon-inducible nuclear protein with 200-amino acid repeat (HIN200) domain ( Cridland et al., 2012 ). ALR inflammasomes sense cytosolic double stranded DNA (dsDNA) ( Burckstummer et al., 2009 ; Ferreri et al., 2010 ). Absent in melanoma 2 (AIM2) is the best characterized member of ALR inflammasomes. Similar to other ALR family members, AIM2 senses dsDNA; however, it appears that dsDNA recognition is independent of nucleic acid sequence as it could bind to both, microbial and host genomic material ( Jin et al., 2012 ). dsDNA binding to HIN200 causes its dissociation from the PYD domain ( Jin et al., 2012 ), allowing the freed PYD domain to interact with ASC, and inflammasome assembly ( Jin et al., 2013c ). AIM2 was implicated in the recognition of microbial, host and tumor derived dsDNA ( Davis B.K. et al., 2011 ; Choubey, 2012 ; Dihlmann et al., 2014 ). Pyrin Pyrin is an inflammasome sensor complex, which contains a N-terminal PYD, central B-box and coiled-coil domain, and a C-terminal B30.2/SPRY domain ( Heilig and Broz, 2018 ). Pyrin was proposed to sense the changes in actin cytoskeletal dynamics as it was found co-localized with stress actin filaments ( Xu et al., 2014 ). Microtubules promote ASC recruitment and the oligomerization ( Gao et al., 2016 ); however, the physiological relevance of this interaction remains largely unknown. Also, microbial toxins which cause impairment of Rho GTPase activity were identified as strong activators of the pyrin inflammasome ( Dumas et al., 2014 ; Xu et al., 2014 ). Several monogenic autoinflammatory syndromes were linked to pyrin inflammasome dysregulation including familial Mediterranean fever (FMF), pyrin-associated autoinflammation with neutrophilic dermatosis, pyogenic arthritis, pyoderma gangrenosum, acne, etc. ( Jamilloux et al., 2018 ). FMF is the most investigated pyrin inflammasome disease, characterized by repeating, self-limited, episodes of fever and polyserositis ( Bernot et al., 1998 ). FMF is linked to a mutation in the Mediterranean Fever (MEFV) gene in a region encoding the B30.2 domain of pyrin ( Omenetti et al., 2014 ). Also, the high prevalence of FMF within certain populations could indicate a selective pressure to preserve this mutation ( Schaner et al., 2001 ). Pyroptosis Pyroptosis is an inflammatory form of programmed cell death linked exclusively to PC1 activation ( Hilbi et al., 1998 ). AC1 is a product of several inflammasomes: NLRP1, NLRP3, NLRC4, and AIM2. Therefore, pyroptosis is often associated with inflammasome activation. Pyroptosis differs from apoptosis in many respects including lack of DNA fragmentation ( Watson et al., 2000 ) and sustained structural integrity of the nucleus ( Zychlinsky et al., 1992 ). Also, pyroptosis is characterized by cell membrane pore formation, which causes cell swelling in contrast to apoptosis, where cells shrink ( Fink and Cookson, 2006 ). Additionally, an increased intracellular osmotic pressure generates large spherical protrusions of the membrane in pyroptotic cells, which coalescence and rupture ( Ona et al., 1999 ). Multiple studies revealed the role of pyroptosis in clearance of microbial pathogens ( Sansonetti et al., 2000 ; Tsuji et al., 2004 ; Lara-Tejero et al., 2006 ). However, over activation of AC1 could lead to pyroptosis associated tissue damage and autoimmunity ( Ona et al., 1999 ; Siegmund et al., 2001 ; Frantz et al., 2003 ). Structure In 2002, the ground breaking work published by Martinon et al. (2002) has demonstrated the role of the inflammasome, a multi-protein complex, in the activation of pro-inflammatory caspases. The authors described the multistep process of the inflammasome assembly which is initiated by the detection of pathogen-associated molecular patterns (PAMPs) or danger signals released by damaged cells ( Duncan et al., 2009 ; Ichinohe et al., 2010 ; Costa et al., 2012 ). Several inflammasome sensors were later identified including the nucleotide-binding oligomerization domain (NOD) like receptors (NLRs), the absent in melanoma-2 like receptors (ALRs) and pyrin ( Ting et al., 2008 ). In the past decade our understanding of NLR containing inflammasomes structure and assembly mechanisms has advanced considerably, largely due to their potential involvement in pathogenesis of several diseases ( Hoffman et al., 2001 ; Alexander So and Borbála Pazár, 2010 ; Song et al., 2017 ). NLRs contain three domains, an N-terminal domain, a NOD, and a C-terminal leucine-rich repeat (LRR) ( Inohara and Nunez, 2003 ). The N-terminal domain contains a caspase recruitment domain (CARD) or pyrin domain (PYD), which function to interact with downstream molecules, such as apoptosis-associated speck-like protein containing (ASC) ( Inohara and Nunez, 2003 ; Schroder and Tschopp, 2010 ). The NOD domain is linked to LRR detecting PAMPs ( Boekhout et al., 2011 ). Upon sensing PAMPs, the NLRs polymerize followed by the interaction between the PYD or CARD domains of LLR and ASC ( Stutz et al., 2013 ). Once activated the inflammasome adopts a wheel-like structure ( Hu et al., 2015 ), where CARD–CARD interactions are essential for recruiting pro-caspase 1 (PC1) into close proximity with the complex ( Faustin et al., 2007 ). PC1 becomes proteolytically cleaved by the CARD domain releasing an active caspase 1 (AC1) p10/p20 tetramer ( Martinon et al., 2002 ; Kanneganti et al., 2006 ; Boucher et al., 2018 ). NLR Inflammasomes This family of inflammasomes includes two subgroups based on the presence of CARD or pyrin in the N terminus. Only nucleotide-binding domain leucine-rich repeats proteins (NLRP)1, NLRP3, and NLRC4 were shown to form inflammasomes that produce AC1 ( Mao et al., 2014 ). In contrast, NLRP6, NLRP9b, and NLRP12 are believed to form inflammasomes, but their roles as inflammasome sensors are less recognized ( Anand et al., 2012 ; Vladimer et al., 2012 ; Zhu et al., 2017 ). NLRP1 NLRP1 was the first identified cytosolic receptor capable of forming active inflammasomes ( Martinon et al., 2002 ). PYD, NBD, and LRR domains, a 'function-to find' domain (FIIND) and a C-terminal CARD are the structural components of NLRP1 ( Jin et al., 2013b ). Our knowledge of NLRP1 function comes largely from studying animal models. It appears that NLRP1 senses and protects against microbial pathogens, as was shown using a mouse model of Bacillus anthracis and Shigella flexneri infection ( Boyden and Dietrich, 2006 ; Sandstrom et al., 2019 ). Additionally, NLRP1 inflammasomes facilitate parasite clearance and protection as demonstrated in Toxoplasma gondii infection in mouse and rat models ( Cirelli et al., 2014 ; Gorfu et al., 2014 ). The clinical relevance of NLRP1 inflammasomes against Toxoplasma gondii is also evident in individuals with specific single-nucleotide polymorphisms in the NLRP1 gene, which are linked to congenital toxoplasmosis ( Witola et al., 2011 ). Aberrant activation of NLRP1 is linked to a pathogenesis of inflammatory diseases. Polymorphisms in the NLRP1 gene are linked to Crohn's disease, rheumatoid arthritis (RA) and systemic sclerosis ( Finger et al., 2012 ). Although the mechanism of NLRP1 activation remains largely unknown, recently, the failure of inflammasome inhibition by dipeptidyl dipeptidase 9 (DDP9), linked to antigen processing ( Zhong et al., 2018 ), was demonstrated to play role in pathogenesis of an autoimmune diseases ( Zhong et al., 2018 ). The authors identified that a single mutation in the FIIND domain of NLRP1 abrogates binding to DPP9, triggering over activation of the inflammasome in autoinflammatory disease AIADK. NLRC4 Similar to NLRP1, NLRC4 establishes protection against infectious pathogens ( Mariathasan et al., 2004 ; Franchi et al., 2006 ; Zhao et al., 2011 ). In the absence of stimulus, NLRC4 remains inactive, where its NBD domain retains a closed conformation by binding to the winged helix domain ( Tenthorey et al., 2014 ). NLRC4 activation is indirect, and it requires NLR family apoptosis inhibitory proteins (NAIPs) for the initial sensing of the microbial ligand ( Rayamajhi et al., 2013 ; Yang et al., 2013 ; Kortmann et al., 2015 ). NAIPs trigger NLRC4 oligomerization, which is essential for inflammasome activation ( Hu et al., 2015 ). Loss of the control over NLRC4 expression and subsequent production of AC1 and release of IL-1β by macrophages was suggested to play role in the pathogenesis of inflammasome linked autoinflammation ( von Moltke et al., 2012 ; Canna et al., 2014 ). Also, a missense mutation in the NLRC4 gene was found in familial cold autoinflammatory syndrome ( Kitamura et al., 2014 ). Multiple mutations in NLRC4 were identified in several autoinflammatory diseases including atopic dermatitis, periodic fever, and fatal or near-fatal episodes of autoinflammation ( Nakamura et al., 2010 ; Canna et al., 2014 ; Bonora et al., 2015 ). These data suggest that NLRC4 plays role in protection against microbial pathogens and autoinflammation. NLRP6 NLRP6 is an inflammasome which plays a role in gut health and maintaining mucosal response to pathogens ( Elinav et al., 2011 ; Anand et al., 2012 ). A microbial metabolite, taurine, was identified as an NLRP6 activator ( Levy et al., 2015 ). The NLRP6-taurite axis appears to be essential for the health of the gut mucosa and microbiome. Taurite produced by the normal microbiota activates NLRP6 which prevents dysbacteriosis by promoting production of antimicrobial peptides ( Levy et al., 2015 ). NLRP12 NLRP12 is intracellular protein expressed in cells of myeloid lineages ( Arthur et al., 2010 ). NLRP12 inflammasome expression can be downregulated by microbial ligands ( Williams et al., 2005 ; Lich et al., 2007 ) via canonical and non-canonical inhibition of NF-κB ( Zaki et al., 2011 ; Allen et al., 2012 ). Several ligands were identified as NLRP12 activators including microbes ( Allen et al., 2012 ; Vladimer et al., 2012 ). NLRP1 NLRP1 was the first identified cytosolic receptor capable of forming active inflammasomes ( Martinon et al., 2002 ). PYD, NBD, and LRR domains, a 'function-to find' domain (FIIND) and a C-terminal CARD are the structural components of NLRP1 ( Jin et al., 2013b ). Our knowledge of NLRP1 function comes largely from studying animal models. It appears that NLRP1 senses and protects against microbial pathogens, as was shown using a mouse model of Bacillus anthracis and Shigella flexneri infection ( Boyden and Dietrich, 2006 ; Sandstrom et al., 2019 ). Additionally, NLRP1 inflammasomes facilitate parasite clearance and protection as demonstrated in Toxoplasma gondii infection in mouse and rat models ( Cirelli et al., 2014 ; Gorfu et al., 2014 ). The clinical relevance of NLRP1 inflammasomes against Toxoplasma gondii is also evident in individuals with specific single-nucleotide polymorphisms in the NLRP1 gene, which are linked to congenital toxoplasmosis ( Witola et al., 2011 ). Aberrant activation of NLRP1 is linked to a pathogenesis of inflammatory diseases. Polymorphisms in the NLRP1 gene are linked to Crohn's disease, rheumatoid arthritis (RA) and systemic sclerosis ( Finger et al., 2012 ). Although the mechanism of NLRP1 activation remains largely unknown, recently, the failure of inflammasome inhibition by dipeptidyl dipeptidase 9 (DDP9), linked to antigen processing ( Zhong et al., 2018 ), was demonstrated to play role in pathogenesis of an autoimmune diseases ( Zhong et al., 2018 ). The authors identified that a single mutation in the FIIND domain of NLRP1 abrogates binding to DPP9, triggering over activation of the inflammasome in autoinflammatory disease AIADK. NLRC4 Similar to NLRP1, NLRC4 establishes protection against infectious pathogens ( Mariathasan et al., 2004 ; Franchi et al., 2006 ; Zhao et al., 2011 ). In the absence of stimulus, NLRC4 remains inactive, where its NBD domain retains a closed conformation by binding to the winged helix domain ( Tenthorey et al., 2014 ). NLRC4 activation is indirect, and it requires NLR family apoptosis inhibitory proteins (NAIPs) for the initial sensing of the microbial ligand ( Rayamajhi et al., 2013 ; Yang et al., 2013 ; Kortmann et al., 2015 ). NAIPs trigger NLRC4 oligomerization, which is essential for inflammasome activation ( Hu et al., 2015 ). Loss of the control over NLRC4 expression and subsequent production of AC1 and release of IL-1β by macrophages was suggested to play role in the pathogenesis of inflammasome linked autoinflammation ( von Moltke et al., 2012 ; Canna et al., 2014 ). Also, a missense mutation in the NLRC4 gene was found in familial cold autoinflammatory syndrome ( Kitamura et al., 2014 ). Multiple mutations in NLRC4 were identified in several autoinflammatory diseases including atopic dermatitis, periodic fever, and fatal or near-fatal episodes of autoinflammation ( Nakamura et al., 2010 ; Canna et al., 2014 ; Bonora et al., 2015 ). These data suggest that NLRC4 plays role in protection against microbial pathogens and autoinflammation. NLRP6 NLRP6 is an inflammasome which plays a role in gut health and maintaining mucosal response to pathogens ( Elinav et al., 2011 ; Anand et al., 2012 ). A microbial metabolite, taurine, was identified as an NLRP6 activator ( Levy et al., 2015 ). The NLRP6-taurite axis appears to be essential for the health of the gut mucosa and microbiome. Taurite produced by the normal microbiota activates NLRP6 which prevents dysbacteriosis by promoting production of antimicrobial peptides ( Levy et al., 2015 ). NLRP12 NLRP12 is intracellular protein expressed in cells of myeloid lineages ( Arthur et al., 2010 ). NLRP12 inflammasome expression can be downregulated by microbial ligands ( Williams et al., 2005 ; Lich et al., 2007 ) via canonical and non-canonical inhibition of NF-κB ( Zaki et al., 2011 ; Allen et al., 2012 ). Several ligands were identified as NLRP12 activators including microbes ( Allen et al., 2012 ; Vladimer et al., 2012 ). ALR Family Inflammasomes ALR family inflammasomes contain an N-terminal PYD and a C-terminal hematopoietic interferon-inducible nuclear protein with 200-amino acid repeat (HIN200) domain ( Cridland et al., 2012 ). ALR inflammasomes sense cytosolic double stranded DNA (dsDNA) ( Burckstummer et al., 2009 ; Ferreri et al., 2010 ). Absent in melanoma 2 (AIM2) is the best characterized member of ALR inflammasomes. Similar to other ALR family members, AIM2 senses dsDNA; however, it appears that dsDNA recognition is independent of nucleic acid sequence as it could bind to both, microbial and host genomic material ( Jin et al., 2012 ). dsDNA binding to HIN200 causes its dissociation from the PYD domain ( Jin et al., 2012 ), allowing the freed PYD domain to interact with ASC, and inflammasome assembly ( Jin et al., 2013c ). AIM2 was implicated in the recognition of microbial, host and tumor derived dsDNA ( Davis B.K. et al., 2011 ; Choubey, 2012 ; Dihlmann et al., 2014 ). Pyrin Pyrin is an inflammasome sensor complex, which contains a N-terminal PYD, central B-box and coiled-coil domain, and a C-terminal B30.2/SPRY domain ( Heilig and Broz, 2018 ). Pyrin was proposed to sense the changes in actin cytoskeletal dynamics as it was found co-localized with stress actin filaments ( Xu et al., 2014 ). Microtubules promote ASC recruitment and the oligomerization ( Gao et al., 2016 ); however, the physiological relevance of this interaction remains largely unknown. Also, microbial toxins which cause impairment of Rho GTPase activity were identified as strong activators of the pyrin inflammasome ( Dumas et al., 2014 ; Xu et al., 2014 ). Several monogenic autoinflammatory syndromes were linked to pyrin inflammasome dysregulation including familial Mediterranean fever (FMF), pyrin-associated autoinflammation with neutrophilic dermatosis, pyogenic arthritis, pyoderma gangrenosum, acne, etc. ( Jamilloux et al., 2018 ). FMF is the most investigated pyrin inflammasome disease, characterized by repeating, self-limited, episodes of fever and polyserositis ( Bernot et al., 1998 ). FMF is linked to a mutation in the Mediterranean Fever (MEFV) gene in a region encoding the B30.2 domain of pyrin ( Omenetti et al., 2014 ). Also, the high prevalence of FMF within certain populations could indicate a selective pressure to preserve this mutation ( Schaner et al., 2001 ). Pyroptosis Pyroptosis is an inflammatory form of programmed cell death linked exclusively to PC1 activation ( Hilbi et al., 1998 ). AC1 is a product of several inflammasomes: NLRP1, NLRP3, NLRC4, and AIM2. Therefore, pyroptosis is often associated with inflammasome activation. Pyroptosis differs from apoptosis in many respects including lack of DNA fragmentation ( Watson et al., 2000 ) and sustained structural integrity of the nucleus ( Zychlinsky et al., 1992 ). Also, pyroptosis is characterized by cell membrane pore formation, which causes cell swelling in contrast to apoptosis, where cells shrink ( Fink and Cookson, 2006 ). Additionally, an increased intracellular osmotic pressure generates large spherical protrusions of the membrane in pyroptotic cells, which coalescence and rupture ( Ona et al., 1999 ). Multiple studies revealed the role of pyroptosis in clearance of microbial pathogens ( Sansonetti et al., 2000 ; Tsuji et al., 2004 ; Lara-Tejero et al., 2006 ). However, over activation of AC1 could lead to pyroptosis associated tissue damage and autoimmunity ( Ona et al., 1999 ; Siegmund et al., 2001 ; Frantz et al., 2003 ). NLRP3 Inflammasomes Molecular Mechanism of Activation NLRP3 is the most characterized inflammasome, and its expression is tightly regulated in resting cells ( Bauernfeind et al., 2009 ). While NLRP3 levels in unstimulated cells are insufficient to trigger assembly of an active inflammasome complex, sensing of pathogen ligands or danger signals, triggers complex formation and pro-inflammatory cytokine production. There are multiple stimuli shown to activate NLRP3 including ATP, toxins, K + efflux, reactive oxygen species and mitochondrial dysfunction ( Dostert et al., 2008 ; Piccini et al., 2008 ). Upon sensing the stimulus, the nucleotide binding domain (NBD) polymerizes initiating PYD–PYD oligomerization with ASC ( Lu A. et al., 2014 ). The CARD of ASC recruits PC1, which becomes cleaved liberating AC1 ( Boucher et al., 2018 ). It appears that within the large family of inflammasomes, NLRP3 is the main PC1 activator ( Agostini et al., 2004 ; Davis E.E. et al., 2011 ). Inflammatory AC1 liberates functional IL-1β and IL-18 ( Afonina et al., 2015 ), pleotropic cytokines regulating inflammation and innate immune response ( Garlanda et al., 2013 ). The classic pathway of NLRP3 activation requires two steps: priming and activation ( Figure 1 ). Toll-like receptor (TLR), FAS-associated death domain protein and IL-1R ligands were identified as NLRP3 priming stimuli ( Allam et al., 2014 ; Gurung et al., 2014 ; He Y. et al., 2016 ). The priming step includes transcriptional activation of NLRP3 via NF-κB signaling ( Bauernfeind et al., 2009 ; Costa et al., 2012 ); however, it fails to initiate functional inflammasome formation, which requires a second stimulus ( Jo et al., 2016 ). The second signal can be provided by multiple pathogen and danger associated ligands ( Franchi et al., 2012 ; Koizumi et al., 2012 ), promoting the assembly of an adaptor (ASC) and PC1. The formed complex cleaves the PC1, which subsequently processes and releases functional IL-1β and IL-18 ( Alnemri et al., 1996 ). FIGURE 1 NLRP3 inflammasome activation. There are two signals required for NLRP3 inflammasome activation. Signal 1 is a priming trigger (microbial ligands, cytokines, etc.) required for the upregulation of NLRP3 and pro-IL-1β transcription and protein synthesis. Signal 2 is an activation trigger (ATP, toxins, viral RNA, etc.) which is essential for formation of an active NLRP3 inflammasome. The second stimulus promotes NLRP3, PC1, pro-IL-1β, and pro-IL-18 protein synthesis. The N-terminal NBD of NLRP3 polymerizes initiating PYD–PYD oligomerization with ASC. The CARD of ASC recruits PC1, which become cleaved liberating AC1. Inflammatory AC1 liberates functional IL-1β and IL-18, pleotropic cytokines regulating inflammation and innate immune response. Molecular Mechanism of Activation NLRP3 is the most characterized inflammasome, and its expression is tightly regulated in resting cells ( Bauernfeind et al., 2009 ). While NLRP3 levels in unstimulated cells are insufficient to trigger assembly of an active inflammasome complex, sensing of pathogen ligands or danger signals, triggers complex formation and pro-inflammatory cytokine production. There are multiple stimuli shown to activate NLRP3 including ATP, toxins, K + efflux, reactive oxygen species and mitochondrial dysfunction ( Dostert et al., 2008 ; Piccini et al., 2008 ). Upon sensing the stimulus, the nucleotide binding domain (NBD) polymerizes initiating PYD–PYD oligomerization with ASC ( Lu A. et al., 2014 ). The CARD of ASC recruits PC1, which becomes cleaved liberating AC1 ( Boucher et al., 2018 ). It appears that within the large family of inflammasomes, NLRP3 is the main PC1 activator ( Agostini et al., 2004 ; Davis E.E. et al., 2011 ). Inflammatory AC1 liberates functional IL-1β and IL-18 ( Afonina et al., 2015 ), pleotropic cytokines regulating inflammation and innate immune response ( Garlanda et al., 2013 ). The classic pathway of NLRP3 activation requires two steps: priming and activation ( Figure 1 ). Toll-like receptor (TLR), FAS-associated death domain protein and IL-1R ligands were identified as NLRP3 priming stimuli ( Allam et al., 2014 ; Gurung et al., 2014 ; He Y. et al., 2016 ). The priming step includes transcriptional activation of NLRP3 via NF-κB signaling ( Bauernfeind et al., 2009 ; Costa et al., 2012 ); however, it fails to initiate functional inflammasome formation, which requires a second stimulus ( Jo et al., 2016 ). The second signal can be provided by multiple pathogen and danger associated ligands ( Franchi et al., 2012 ; Koizumi et al., 2012 ), promoting the assembly of an adaptor (ASC) and PC1. The formed complex cleaves the PC1, which subsequently processes and releases functional IL-1β and IL-18 ( Alnemri et al., 1996 ). FIGURE 1 NLRP3 inflammasome activation. There are two signals required for NLRP3 inflammasome activation. Signal 1 is a priming trigger (microbial ligands, cytokines, etc.) required for the upregulation of NLRP3 and pro-IL-1β transcription and protein synthesis. Signal 2 is an activation trigger (ATP, toxins, viral RNA, etc.) which is essential for formation of an active NLRP3 inflammasome. The second stimulus promotes NLRP3, PC1, pro-IL-1β, and pro-IL-18 protein synthesis. The N-terminal NBD of NLRP3 polymerizes initiating PYD–PYD oligomerization with ASC. The CARD of ASC recruits PC1, which become cleaved liberating AC1. Inflammatory AC1 liberates functional IL-1β and IL-18, pleotropic cytokines regulating inflammation and innate immune response. Epigenetic Factors and Post-Transcriptional Mechanisms Regulating NLRP3 Inflammasome Activation The term "epigenetic" was originally presented by Waddington (1956) to describe regulation of gene expression during the embryogenesis. Since then, definition of "epigenetic" has changed, and now refers to a stably heritable modulation of gene expression without altering DNA sequence ( Berger et al., 2009 ). Epigenetic factors include DNA methylation at cytosine followed by guanine (CpGs) nucleotide and histone posttranslational modifications ( Peschansky and Wahlestedt, 2014 ). Initially, epigenetic control was demonstrated in normal development and differentiation; however, its role in pathogenesis of acute and chronic inflammation has become increasingly recognized ( Bayarsaihan, 2011 ). DNA Methylation DNA methylation is dynamic and changes during the embryonic development and differentiation ( Berger, 2007 ). It was shown that DNA methylation silences genes to ensure monoallelic expression, prevent endogenous retrovirus expression and transposon actions ( Walsh et al., 1998 ; Bourc'his et al., 2001 ; Bourc'his and Bestor, 2004 ). DNA methylation is essential for normal cell function; however, its role in the pathogenesis of several diseases has also been confirmed ( Wei et al., 2016 ; Vento-Tormo et al., 2017 ). DNA demethylation is often detected near promoters, suggesting that gene overexpression could play role in pathogenesis of many pathologies ( Ryan et al., 2010 ; Bierne et al., 2012 ). NLRP3 inflammasome expression can also be regulated by changes in gene methylation status. For example, NLRP3 gene expression is silenced in health which appears to be essential for inhibiting inflammation ( Ryan et al., 2010 ; Bierne et al., 2012 ; Wei et al., 2016 ). However, demethylation and, subsequent, overexpression of NLRP3 was linked to pathogenesis of cryopyrin-associated periodic syndromes (CAPS) ( Vento-Tormo et al., 2017 ) and Mycobacterium tuberculosis infection ( Wei et al., 2016 ). Histone Modifications The effect of epigenetic modification of histones was studied using several inflammatory models ( Bayarsaihan, 2011 ). Histone acetylation is essential for initiation of an activation phase of inflammation, which is characterized by the release of pro-inflammatory cytokines via CREB, mitogen-activated protein kinases (MAPKs), nuclear factor-κB (NF-κB) and signal transducer and activator of transcription (STAT) factors ( Escobar et al., 2012 ). In contrast, histone deacetylations regulate the late, an attenuation phase of inflammation ( Villagra et al., 2010 ). It appears that inflammasome activation can also be regulated by affecting the acetylation status of histones, as it was recently shown by Liu C.C. et al. (2018) . The authors demonstrated upregulation of NLRP3 in patients diagnosed with painful neuropathy, which could be prevented by inhibition of histone acetylation. Non-coding RNAs In addition to epigenetic modulation non-coding RNAs are also involved in NLRP3 regulation ( Bayarsaihan, 2011 ), as was demonstrated in the setting of inflammation caused by microbial and viral infection ( Li et al., 2010 ; Ryan et al., 2010 ; Bierne et al., 2012 ; Jin et al., 2013a ; Chen and Ichinohe, 2015 ). This inflammation is post-transcriptionally regulated via non-coding RNAs targeting inflammasome components, where mRNA stability and inhibition of translation were most commonly affected ( Bayarsaihan, 2011 ). Post-transcriptional Regulation of NLRP3 Inflammasomes: MicroRNA (miRNA) MicroRNAs are endogenous conservative, single-stranded non-coding RNAs which are 19–24 nucleotides long. Usually, miRNAs are derived from transcripts with a hairpin structure and are loaded into the Argonaute protein within a silencing complex ( Hutvagner and Zamore, 2002 ; Mourelatos et al., 2002 ; Bartel, 2004 ). The inhibitory effect of miRNAs is explained by their binding to the untranslated regions (UTRs) of transcripts which modulates the stability and translation of the target mRNA ( Figure 2 ) ( Ruvkun, 2001 ; Filipowicz et al., 2008 ; Bartel, 2009 ; Coll and O'Neill, 2010 ). miRNAs can modulate the expression of histone modifies including histone deacetylases and DNA methyltransferases resulting in modulation of histone modifications and DNA methylation ( Tuddenham et al., 2006 ; Fabbri et al., 2007 ). FIGURE 2 miRNA regulation of NLRP3 inflammasome expression. (A) Priming signal triggers NLRP3, PC1, IL-1β, and IL-18 transcription and protein synthesis. Activation signal initiates inflammasome formation and release of AC1. AC1 proteolytically cleaves pro-IL-1β and pro-IL-18, liberating active cytokines. (B) Suppression of NLRP3 protein translation and inflammasome formation by miRNA. Priming stimulus triggers NLRP3 transcription; however, miR-223 , miR-22 , miR-30e , and miR-7 bind to the UTR region of NLRP3 mRNA and interrupt protein translation. Absence of NLRP3 protein leads failure of the inflammasome protein complex formation, when the second stimulus present. NLRP3 activation is tightly regulated where two signals are required to initiate functional inflammasome formation. The first signal includes cell priming with TLR ligands ( Bauernfeind et al., 2009 ; Franchi et al., 2009 ). Therefore, it could be suggested that targeting TLR expression will impact the inflammasome activity. Indirect regulation of TLR expression includes modulation of the downstream pathways molecules, which has been shown in injuries, inflammation and cancer ( Coll and O'Neill, 2010 ; Sheedy et al., 2010 ; Nahid et al., 2011 ; Anzola et al., 2018 ; Tan et al., 2018 ; Zhi et al., 2018 ). TLR4 ligands are the most studied priming signals of NLRP3 activation ( Groslambert and Py, 2018 ). It was shown that the TLR ligand binding increases the level of several miRNAs, including miR-155 , miR-146a , miR-21 , and miR-132 , which were linked to inhibition of TLR4/MyD88/NF-κB signaling ( Coll and O'Neill, 2010 ; Sheedy et al., 2010 ; Nahid et al., 2011 ; Anzola et al., 2018 ; Tan et al., 2018 ; Zhi et al., 2018 ). It appears that upregulation of miRNAs is a component of a negative feedback mechanism designed to down-modulate inflammatory cytokine production after response to microbial stimuli ( Ceppi et al., 2009 ). A direct inhibitory effect of let-7 family miRNAs on TLR4 mRNA has been demonstrated ( Chen et al., 2007 ). Let-7 miRNA regulation of TLR4 was shown to occur via post-transcriptional suppression ( Androulidaki et al., 2009 ). It was suggested that let-7 miRNA downregulation of TLR4 could have detrimental effect on host defense against microbes, promoting microbial survival and propagation ( Chen et al., 2005 ; Muxel et al., 2018 ). Post-transcriptional regulation of TLR signaling and its impact on diseases are reviewed by Nahid et al. (2011) . Active inflammasome complex formation requires a second signal, initiating substantial NLRP3 transcription ( Dostert et al., 2008 ; Piccini et al., 2008 ). During this transcriptionally active phase, NLRP3 mRNA could be regulated by miRNA, as was shown by miR-223 ( Bauernfeind et al., 2012 ). According to an in silico analysis, miR-223 can bind to a highly conserved region of the 3′UTR of NLRP3 mRNA and subsequently interfere with protein translation ( Lewis et al., 2005 ). Interestingly, miR-223 appears to be an important NLRP3 regulator in leukocytes ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ), where the miRNA levels have been shown to vary in different leukocyte subsets. For example, this miRNA was found absent in T and B lymphocytes ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ). In contrast, the miR-223 was demonstrated in myeloid cells, where it was highest in neutrophils, followed by macrophages and dendritic cells ( Bauernfeind et al., 2012 ). It has been suggested that this miRNA plays role in granulocyte production and regulation of inflammation ( Johnnidis et al., 2008 ; Neudecker et al., 2017 ). Decreased production of pro-inflammatory cytokines such as IL-1β and IL-18 was demonstrated in cells treated with miR-223 or its mimics ( Neudecker et al., 2017 ; Ding Q. et al., 2018 ). These data suggest that miR-223 could be a potential target for regulation of NLRP3 expression, where increased miRNA could reduce inflammasome activation and, subsequently, abrogate the inflammation ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ). Since several miRNAs could regulate expression of a single transcript ( Krek et al., 2005 ), it is likely that in addition to miR-223 , other miRNAs can alter NLRP3 transcription ( Figure 3 ). FIGURE 3 UTR binding sites of NLRP3 for miRNAs responsible for the regulation of inflammation. Numerous studies have identified that pathogens, trauma and cancer can cause abnormal expression of miRNAs which impair NLPR3 inflammasome function disrupt the functional complex formation and its signaling ( Table 1 ). Table 1 Aberrant miRNA expressions linked to inflammasome related diseases. Disease miRNA Regulation of miRNA Target cell Target gene References Inflammatory bowel diseases miR-223 ↑ Intestinal biopsies NLRP3 Neudecker et al., 2017 ↑ Circulating monocytes, neutrophils Johnnidis et al., 2008 ; Bauernfeind et al., 2012 ; Neudecker et al., 2017 ↓ Macrophages Rheumatoid arthritis miR-33 ↑ Macrophages PGC1-α Karunakaran et al., 2015 ; Xie Q. et al., 2018 Type 1 diabetes miR-146a ↓ Macrophages TLR2, TLR4 Bhatt et al., 2016 ; Xie Z. et al., 2018 Type 2 diabetes miR-146a ↓ Macrophages TLR2, TLR4 Balasubramanyam et al., 2011 Systemic lupus erythematosus miR-23b ↓ Inflammatory lesions TAB2, TAB3, IKK-α Zhu et al., 2012 Parkinson's disease miR-7 ↓ Microglia NLRP3 Zhou Y. et al., 2016 miR-30e ↓ NLRP3 Li D. et al., 2018 Atherosclerosis miR-22 ↓ Monocytes, macrophages NLRP3 Huang W.Q. et al., 2017 miR-9 ↓ JAK1 Wang F. et al., 2017 miR-30e-3p ↓ FOXO3 Li P. et al., 2018 Acute lung injury/acute respiratory distress syndrome miR-223 ↑ Ly6G+ neutrophils NLRP3 Feng et al., 2017 Hepatocellular carcinoma miR-223 ↑ Tumor cell line NLRP3, EPB41L3, FOXO1 Li X. et al., 2011 ; Kim et al., 2017 miR-223 ↓ Patient's sera NLRP3 Bhattacharya et al., 2016 miR-30e ↓ NLRP3 Bhattacharya et al., 2016 Colorectal cancer miR-223 Tumor type specific Tumor tissue, tumor cell line NLRP3, FoxO3a Ju et al., 2018 miR-22 ↓ SP-1 Xia et al., 2017 Gastric cancer miR-223 ↑ Tumor tissue NLRP3 Haneklaus et al., 2012 miR-22 ↓ Macrophages NLRP3 Li S. et al., 2018 Oral squamous cell carcinoma miR-223 ↑ Tumor tissue RHOB Manikandan et al., 2016 miR-22 ↓ NLRP3 Feng et al., 2018 Cervical cancer miR-223 ↓ Tumor tissue, tumor cell line FOXO1 Wu et al., 2012 miR-22 ↓ HDAC6 Wongjampa et al., 2018 Glioblastoma miR-223 Controversial Tumor tissue, tumor cell line NFIA, PAX6 Fazi et al., 2005 ; Glasgow et al., 2013 ; Cheng et al., 2017 ; Ding Q. et al., 2018 miR-22 ↓ SIRT1 Li W.B. et al., 2013 miRNA in Regulation of Inflammasome in Infections Inflammasome activation is an important component of infectious pathogens surveillance and antimicrobial immune and inflammatory responses. This inflammasome was shown to be activated by several bacterial pathogens including Staphylococcus aureus , Salmonella typhimurium , Listeria monocytogenes , Mycobacterium , Streptococcus pyogenes , Neisseria gonorrhoeae as well as fungi such as Candida albicans and Aspergillus fumigatus ( Franchi et al., 2006 ; Mariathasan et al., 2006 ; Miao et al., 2006 ; Craven et al., 2009 ; Duncan et al., 2009 ; Harder et al., 2009 ; Hise et al., 2009 ; Joly et al., 2009 ; Munoz-Planillo et al., 2009 ; Broz et al., 2010 ; Carlsson et al., 2010 ; McElvania Tekippe et al., 2010 ; Said-Sadier et al., 2010 ). NAIP/NLRC4 inflammasome can protect against Salmonella Typhimurium and C. rodentium invasion by bacteria expulsion from intestinal epithelial cells together with IL-18 and eicosanoid lipid mediators release ( Nordlander et al., 2014 ; Sellin et al., 2014 ; Rauch et al., 2017 ). It appears that NLRP3 activation is essential for establishing the inflammatory milieu in the target tissue and augmenting the phagocytic capacity of the local macrophages ( Master et al., 2008 ; Melehani and Duncan, 2016 ; Cohen et al., 2018 ). Enhanced macrophage bactericidal activity is the most commonly identified mechanism of inflammasome antimicrobial effect ( Master et al., 2008 ; Cohen et al., 2018 ). Additionally, NLRP3 activation induced death of macrophages was described as an effort to prevent microbial propagation and spread ( Miao et al., 2010 ; Sagulenko et al., 2013 ). However, there is a growing body of evidence suggesting that there is a threshold of NLRP3 activity, which acts as a safeguard mechanism to prevent inflammasome over-activation. It appears that aberrant NLRP3 activation could have a detrimental effect on tissues homeostasis and compromise barrier integrity ( Bortolotti et al., 2018 ; McKenzie et al., 2018 ). It is this detrimental effect of the inflammasome over-activation that is often employed by microbes to ensure spread and propagation ( Duncan et al., 2009 ; Harder et al., 2009 ; Carlsson et al., 2010 ). Microbial virulence factors often act as NLRP3 activators. For example, it was shown that the detrimental (to the host) role of Esx1, a membrane lysis factor of Mycobacterium ( Stanley et al., 2003 ), is linked to inflammasome activation ( Carlsson et al., 2010 ). Two virulence factors of group A Streptococcus (GAS), M protein and streptolysin O, were also identified as contributing into NLRP3 activation and IL-1β production ( Harder et al., 2009 ; Valderrama et al., 2017 ). Both virulence factors are commonly detected in association with invasive GAS infections, including necrotizing fasciitis and toxic shock syndrome. Therefore, NLRP3 activation by virulent factors could promote microbe propagation and aid their escape from immune clearance. Restoring the NLRP3 activation threshold could be a novel therapeutic approach for treatment of invasive infections. In this respect, miRNA may be a tool to regain control over NLRP3. It has been shown that miR-223 expression is consistently high in NLRP3 responsive cells, suggesting the high efficacy of this miRNA in prevention of inflammasome over-activation ( Bauernfeind et al., 2012 ). Dorhoi et al. (2013) demonstrated that miR-223 is upregulated in the blood and lung parenchyma of patients diagnosed with tuberculosis. Also, data collected using animal models confirmed the link between deletion of miR-223 and increased susceptibility to Mycobacterium tuberculosum infection ( Dorhoi et al., 2013 ). Similarly, a protective role of miR-223 in Staphylococcus aureus infection was demonstrated by Fang et al. (2016) . Additionally, the effect of targeting TLR4 for NLRP3 regulation in Listeria monocytogenes infection was demonstrated by Schnitger et al. (2011) . The authors identified that, miR-146a can directly inhibit TLR4 receptor expression, which can downregulate inflammasome activity ( Schnitger et al., 2011 ). Many viruses can activate inflammasomes, including Influenza virus, Hepatitis C virus, Herpes simplex virus-1, etc. ( Delaloye et al., 2009 ; Ichinohe et al., 2010 ; Ito et al., 2012 ; Kaushik et al., 2012 ; Negash et al., 2013 ; Triantafilou et al., 2013a , b ; Wu et al., 2013 ; Ermler et al., 2014 ; Chen and Ichinohe, 2015 ). Inflammasome activation appears to be essential for anti-viral protection, serving as viral genome sensors and triggering innate immune response ( Muruve et al., 2008 ; Lupfer et al., 2015 ). The protective role of inflammasomes was shown in influenza virus infection as an increased viral clearance was NLRP3 dependent ( Allen et al., 2009 ). Also, inflammasome activation improved the survival rate in an animal model of influenza ( Ichinohe et al., 2009 ). Thomas et al. (2009) demonstrated that, the innate immune response activation by NLRP3 inflammasomes is essential for animal protection. However, our understanding of the mechanisms of inflammasome antiviral defense remains limited ( Anand et al., 2011 ). Some viruses were shown to post-transcriptionally regulate inflammasome expression to benefit self-replication and propagation ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). For example, miRNA suppression of inflammasomes was shown in Epstein Barr Virus (EBV) infected cells ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). It appears that, EBV can avert NLRP3 inflammasome activation by expressing miRNAs encoded by three BHRF1 -regions and 40 BART -regions of the viral genome ( Albanese et al., 2016 ; Tagawa et al., 2016 ; Farrell, 2018 ). Additionally, two miRNAs encoded by EBV, miR-BART11-5p and miR-BART15 , were identified by Haneklaus et al. (2012) , which could bind to the 3′-UTR of NLRP3 , the same site targeted by miR-223 , and inhibit the inflammasome. It remains to be determined whether these viral miRNA could be used as therapeutic targets. miRNA Regulation of Inflammasome in Autoimmune Diseases Autoimmune diseases are often the result of a dysregulated immune response, characterized by inflammation and organ damage ( Chang, 2013 ; Yang and Chiang, 2015 ). Chronic inflammation is frequently identified as a predisposing factor for an autoimmune reaction ( Yang and Chiang, 2015 ). Multiple mechanisms were suggested to explain prolonged inflammation leading to autoimmunity; where failure to control inflammasome activation was recently identified in some autoimmune conditions ( Yang and Chiang, 2015 ). It has been established that in addition to inflammation, an increased secretion of IL-1β and IL-18, can stimulate proliferation and organ distribution of the effector T cells, which can cause tissue damage ( Oyanguren-Desez et al., 2011 ; Celhar et al., 2012 ). Therefore, targeting the inflammasome could be suggested to restore control over the inflammatory and immune response. Therapeutic potentials of several NLRP3 targeting miRNAs were investigated in autoimmune diseases such as inflammatory bowel diseases (IBDs) ( Neudecker et al., 2017 ), RA ( Xie Z. et al., 2018 ), type 1 diabetes (T1D) ( Yang and Chiang, 2015 ), type 2 diabetes (T2D) ( Yang and Chiang, 2015 ), and systemic lupus erythematosus (SLE) ( Zhu et al., 2012 ). Inflammatory bowel diseases (IBDs) Inflammatory bowel diseases are characterized by chronic inflammation of the intestine and comprise two disorders Crohn's disease and ulcerative colitis. It is believed that the pathogenesis of IBDs is associated with dysregulation of innate and adaptive immune responses, triggered by microbial antigens. This could result in chronic inflammation of the digestive tract and damage to the intestinal mucosa ( Fiocchi, 1998 ). The role of the inflammasome in intestinal inflammation is controversial. Zaki et al. (2010) reported that, NLRP3 induced production of IL-18 in intestinal epithelial cells can be protective, and contributes to epithelium integrity in experimental colitis. In contrast, Seo et al. (2015) have demonstrated the role of inflammasome in exacerbation of an intestinal pathology. The damaging effect of the inflammasome was also confirmed by Shouval et al. (2016) , who identified that IL-1β inhibition improves the course of IBDs. It appears that increased IL-1β levels and tissue damage in IBDs are linked to NLRP3 activation in myeloid leukocytes infiltrating the gut tissue ( Neudecker et al., 2017 ). The role of the inflammasome in IBDs pathogenesis was also confirmed by using a miR-223 deficient animal model of colitis ( Neudecker et al., 2017 ). miR-223 deficient mice develop experimental colitis manifesting with colonic ulceration, inflammatory leukocyte infiltration and tissue injury which resembles closely IBDs ( Neudecker et al., 2017 ). Tissue injury in these mice was linked to an enhanced NLRP3 expression and elevated IL-1β ( Neudecker et al., 2017 ). Treatment of animals with miR-223 mimetics alleviated symptoms of the colitis which coincided with reduced NLRP3 RNA and IL-1β levels ( Neudecker et al., 2017 ). This data presents miR-223 as a novel biomarker and therapeutic target in subsets of IBDs and colitis ( Polytarchou et al., 2015 ). Rheumatoid arthritis (RA) Rheumatoid arthritis is a chronic, systemic inflammatory disease affecting joints as well as skin, eyes, lungs, heart, and blood vessels ( Scott et al., 2010 ). It was suggested that RA pathogenesis is related to activation of the NLRP3/IL-1β axis, where inflammasome activation was linked to worsening symptoms of the disease ( Xie Q. et al., 2018 ). It was shown that activation of NLRP3 leads to an abundant expression of IL-1β ( Guo et al., 2018 ), which can trigger T helper type 17 (Th17) cell differentiations and osteoclasts activation in RA ( Dayer, 2003 ; McInnes and Schett, 2011 ; Zhang et al., 2015b ). Th17 cells play a central role in RA pathogenesis, by maintaining chronic inflammation, recruiting neutrophils and promoting joint degradation ( Cai et al., 2001 ; Shahrara et al., 2009 ; Leipe et al., 2010 ). Recently, an indirect effect of miR-33 on NLRP3 activation was demonstrated in RA ( Xie Q. et al., 2018 ), which could be explained by miRNA controlled dysregulation of mitochondrial function ( Schroder et al., 2010 ; Zhou et al., 2011 ; Miao et al., 2014 ; Ouimet et al., 2015 ). Xie Q. et al. (2018) suggested that miR-33 increases mitochondrial oxygen consumption and accumulation of reactive oxygen species which upregulates expression of NLRP3 and PCA1 in RA. Also, both miR - 33 expression and NLRP3 inflammasome activity were found to be higher in RA monocytes as compared to controls ( Xie Q. et al., 2018 ). These findings indicate that miR - 33 could play an indirect role in pathogenesis of RA through NLRP3 inflammasome activation. Additional studies will provide more insight into the miRNA regulation of NLRP3 in RA and its therapeutic and prognostic implications. Type 1 diabetes (T1D) Type 1 diabetes is caused by autoimmune targeted elimination of pancreatic β cells islet ( Kloppel et al., 1985 ). It was shown that TLRs play an essential role in the pathogenesis of T1D ( Xie Z. et al., 2018 ). Upregulated expression of TLR4 as well as increased activity of the downstream targets was demonstrated in monocytes from T1D ( Devaraj et al., 2008 ). Increased expression of activated TLRs was explained as a reaction to a high levels of circulating ligands in TID ( Devaraj et al., 2009 ). Also, epigenetic regulation was associated with an aberrant TLR signaling and an increased IL-1β expression in T1D ( Grishman et al., 2012 ). Several miRNAs were found altered in pre-TID patients, where levels of nine miRNAs ( miR-146a , miR-561 , and miR-548a-3p , miR-184 , and miR-200a ) were decreased, and two miRNAs ( miR-30c and miR-487a ) were increased ( Grieco et al., 2018 ). Supporting these results was data published by Wang G. et al. (2018) demonstrating lower levels of miR-150 , miR-146a , and miR-424 compared to controls. One of the most consistent findings was the decreased miR-146a levels in T1D. It appears that miR-146a deficiency could play role in T1D exacerbation and increased IL-1β and IL-18 expression ( Bhatt et al., 2016 ). Increased IL-1β levels could indicate inflammasome activation in T1D, although the role of inflammasome in the disease pathogenesis remains largely unknown. Type 2 diabetes (T2D) Circulating autoantibodies to β cells, self-reactive T cells and the glucose-lowering efficacy of some immunomodulatory therapies are suggestive of the autoimmune nature of the T2D ( Itariu and Stulnig, 2014 ). Interestingly, a role for miRNA regulation of gene expression was demonstrated in T2D, where Balasubramanyam et al. (2011) have shown reduced miR-146a which was associated with increased NF-κB , TNF-α and IL-6 mRNA levels. It is the same miRNA, which was found implicated to pathogenesis of T1D ( Xie Z. et al., 2018 ), indicating potential similarities in the pathogenesis of both diseases. Recently in vivo studies demonstrated that miR-146a deficiency could increase expression of M1 and suppress expression of M2 markers in macrophages collected from patients with diabetes ( Bhatt et al., 2016 ). Macrophage polarization occurs in the presence of IFNγ (M1) or IL-4 (M2) ( Nathan et al., 1983 ; Stein et al., 1992 ) and is linked to pro-inflammatory and anti-inflammatory activities, respectively. M1 macrophages were shown to support inflammation by producing pro-inflammatory cytokines, including the inflammasome product IL-1β ( Bhatt et al., 2016 ). Therefore, a link could be suggested between low miR-146a levels and inflammasome activation in M1 cells. More investigation is required to identify the connection between miR-146a and inflammasome activation and the role of this in T2D pathogenesis. Systemic lupus erythematosus (SLE) Systemic lupus erythematosus is an autoimmune disease caused by the loss of immune tolerance to ubiquitous autoantigens ( Tsokos, 2011 ). Inflammation plays essential role in SLE pathogenesis ( Yang et al., 2014 ; Rose and Dorner, 2017 ), where high levels of circulating proinflammatory cytokines are commonly detected ( Yao et al., 2016 ; Mende et al., 2018 ). Inflammasome activation is proposed as one of the mechanisms underlying increased proinflammatory cytokine level in SLE ( Kahlenberg and Kaplan, 2014 ). This assumption is supported by a report where IL-1β deficient mice were found to be resistant to experimental SLE ( Voronov et al., 2006 ). Also, an increased expression of NLRP3 and AC1 have been reported in SLE nephritis biopsies ( Kahlenberg et al., 2011 ). Kahlenberg and Kaplan (2014) have shown that SLE macrophages are highly reactive to innate immune stimuli, often leading to inflammasome activation. Therefore, targeting inflammasome activity could be a novel approach for SLE treatment. The expression of several miRNAs targeting the inflammasome and its products were found differentially expressed in SLE. For example, Wang et al. (2012) have demonstrated high levels of circulating miR-223 , which was shown to inhibit NLRP3 , in SLE. Also, reduced levels of circulating miR-146a , which regulates priming of TLRs, was found in SLE plasma ( Wang et al., 2012 ). Interestingly, expression of miR-23b , which indirectly inhibits IL-1β responses, was shown to be downregulated in inflammatory lesions of SLE patients and animal model ( Zhu et al., 2012 ). More studies are required to determine the role of miRNAs in pathogenesis of SLE and their therapeutic potential. miRNA Regulation of Inflammasome in Neurodegenerative Disorders Inflammasome products, IL-1β and IL-18, were shown to be essential for the health and functional competence of the nervous system ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ). NLRP3 expression was demonstrated in microglia and astrocytes, which could explain the constitutive level of these cytokines in the brain ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ; Savage et al., 2012 ; Minkiewicz et al., 2013 ; Cho et al., 2014 ; Lu M. et al., 2014 ). Interestingly, higher than normal levels of IL-1β and IL-18 were found in several neurodegenerative disorders, suggesting that over-activation of inflammasomes may play a role in pathogenesis of these diseases ( Cho et al., 2014 ; Lu M. et al., 2014 ; Denes et al., 2015 ; Mamik and Power, 2017 ; Song et al., 2017 ). The significance of miRNA in the regulation of inflammasome activation in the pathogenesis of neurodegenerative diseases remains largely unknown. However, the role of an aberrant miRNA in regulation of NLRP3 expression was previously demonstrated in Parkinson's disease (PD). Parkinson's disease is a neurodegenerative disease which is characterized by progressive loss of dopaminergic neurons in substantia nigra compacta ( Gasser, 2009 ). It is believed that accumulation of α-Syn fibrillary aggregates in the brain, most notably in the nigral dopaminergic neurons, induces the neuroinflammation ( Eriksen et al., 2003 ). According to Zhou Y. et al. (2016) , α-Syn can activate NLRP3 inflammasomes in microglia leading to an increased production of IL-1β. The authors also demonstrated that, miR-7 and miR-30e analogs can inhibit NLRP3 inflammasome mediated neuroinflammation in the brain and protect dopaminergic neurons ( Zhou Y. et al., 2016 ). It appears that the anti-inflammatory effects of miR-7 and miR-30e are associated with their targeting of NLRP3 mRNA in microglial cells. Interestingly, decreased miR-7 and miR-30e expression was demonstrated in PD, which could lead to the loss of the regulatory control of α-Syn induced NLRP3 activation ( Li D. et al., 2018 ). miRNA Regulation of the Inflammasome in Cardiovascular Diseases (CVDs) The physiological significance of inflammation is confirmed as it facilitates elimination of destructive stimuli and pathogens. However, aberrant inflammatory responses could cause tissue damage, tissue fibrosis and chronic diseases ( Liu D. et al., 2018 ). Inflammation is recognized as a major risk factor for CVDs ( Zhou et al., 2018 ), where chronic inflammasome activation was shown to contribute to the pathogenesis of atherosclerosis, ischemic and non-ischemic heart diseases ( Zhou et al., 2018 ). Therefore, regulation of inflammasome activity using miRNA could be used for treatment and prevention of CVDs. Currently, strong evidence for the role of NLRP3 activation has been demonstrated in pathogenesis atherosclerosis. Atherosclerosis is a form of CVD characterized by narrowing of the blood vessel lumen due to plaque formation, continuous dyslipidemia and inflammation ( Ross, 1993 ). Chronic inflammation is commonly found in and around the atherosclerotic plaques which has an adverse effect on the arterial wall structure and function ( Bernhagen et al., 2007 ). It is believed that atherogenic lipid mediators, involved in the formation of chronic inflammation in atherosclerotic plaque ( Chen et al., 2006 ), can trigger peripheral blood monocytes migration and differentiation into macrophages within the intima of the arterial wall ( Chen et al., 2006 ). T cells were also detected in atherosclerotic lesions ( Kleemann et al., 2008 ), where, together with activated macrophages, they were shown to secrete proinflammatory mediators such as interferons, interleukins, and proteases ( Østerud and Bjørklid, 2003 ; Shashkin et al., 2005 ; Tabas, 2005 ; Chen et al., 2006 ). IL-1β expression was identified in the early phase of atherosclerotic plaque formation and this stimulates secretion of additional cytokines and chemokines ( Kleemann et al., 2008 ). Therefore, inflammasome activation in macrophages and T cell within the atherosclerotic lesion contributes to the pathogenesis of chronic inflammation. miR-22 , a miRNA inhibiting NLRP3 , is decreased in peripheral blood mononuclear cells from coronary atherosclerosis ( Chen B. et al., 2016 ), suggesting that upregulation of this miRNA could have therapeutic potential in CVD. Supporting this assumption, Huang W.Q. et al. (2017) investigated the effect of miR-22 on the NLRP3 inflammasome and endothelial cell damage in an in vivo model of coronary heart disease. The authors demonstrated that miR-22 mimics could decrease the release of inflammatory cytokines such as IL-1β and IL-18 by suppressing NLRP3 expression in monocytes and macrophages ( Huang W.Q. et al., 2017 ). Two additional miRNAs, miR-9 and mir-30e-5p were found to indirectly affect inflammasome activation in atherosclerosis ( Wang Y. et al., 2017 ; Li P. et al., 2018 ). It appears that miR-9 could indirectly suppress inflammasome activation by targeting an atherogenic lipid mediator, oxidized low-density lipoprotein (oxLDL), in atherosclerosis ( Liu W. et al., 2014 ). In another report, Wang Y. et al. (2017) reported that miR-9 inhibits NLRP3 inflammasome activation induced by oxLDL in human THP-1 derived macrophages and peripheral blood monocytes in an in vitro atherosclerosis model. miR-9 targets Janus kinase 1 ( JAK1 ) pathway ( Wang Y. et al., 2017 ) inhibiting expression of NF-κB p65 which is required for the first step of NLRP3 inflammasome activation ( Wang Y. et al., 2017 ). In addition, miR-30c-5p was linked to an indirect regulation of NLRP3 expression in atherosclerosis ( Li P. et al., 2018 ). Li P. et al. (2018) reported that miR-30c-5p protects human aortic endothelial cells (HAECs) from the oxLDL insult by targeting FOXO3 . The authors showed that miR-30c-5p can suppress FOXO3 expression and, consequently, decrease levels of NLRP3, AC1, IL-18 and IL-1β in HAECs ( Li P. et al., 2018 ). As evidence emerges supporting the role of NLRP3 in the pathogenesis of atherosclerosis, targeting the inflammasome becomes an attractive therapeutic approach, where miRNAs could be suitable novel tools. miRNA in Regulation of Inflammasome in Cancer The role of the inflammasome in tumorigenesis remains controversial. Some reports indicate that NLRP3 inflammasome activation and IL-18 signaling protect against colorectal cancer ( Karki et al., 2017 ), whereas progression of breast cancer, fibrosarcoma, gastric carcinoma, and lung metastasis were shown to be supported by the inflammasome ( Okamoto et al., 2010 ; Kolb et al., 2014 ). Inflammasome regulation is complex, where multiple factors are implicated, making identification of the key regulatory elements challenging. As the inflammasome involvement in pathogenesis of some malignancies becomes more evident, understanding the regulatory mechanisms could lead to the discovery of novel therapeutic targets for cancer treatment. Hepatocellular carcinoma (HCC) Hepatocellular carcinoma (HCC) is a frequent sequelae of hepatitis B and hepatitis C viral infection ( Perz et al., 2006 ). It is understood that these viruses activate NLRP3 inflammasomes causing hepatocyte pyroptosis, apoptosis and fibrosis ( Kofahi et al., 2016 ). However, HCC tissue analysis failed to detect inflammasome activation; in fact, it was found to be significantly down-regulated when compared to the adjacent normal tissue ( Zhu et al., 2011 ; Wei et al., 2014 ). To explain this inconsistency, Wei et al. (2014) suggested that NLRP3 expression is dynamic changing during the progression of HCC. It appears that NLRP3 expression was increased in liver cells at the early stages of transformation, while inflammasome levels were decreased in malignant cells when compared to adjacent normal tissue ( Wei et al., 2014 ). Interestingly, levels of miR-223 , a negative regulator of NLRP3 , were found to be increased in Hep3B cells derived from HCC ( Wan et al., 2018 ). Increased miR-223 was shown to coincide with tumor growth, suggesting a role in post-transcriptional mechanisms in malignant progression. In addition to NLRP3 , miR-223 was shown to target erythrocyte membrane protein band 4.1 like 3 ( EPB41L3 ) and FOXO1 ( Li and Rana, 2014 ; Kim et al., 2017 ). FOXO1 transcription factor binds to the thioredoxin-interacting protein (TXNIP) and regulates genes involved in cell death as well as the oxidative stress responses ( Kim et al., 2017 ). TXNIP interacts with the NLRP3 inflammasome and activates AC1 in murine β-cells ( Zhou et al., 2010 ). In addition, miR-223 appears to be released systemically, where the level of this miRNA in the plasma was significantly lower in HCC cases ( Giray et al., 2014 ). In addition to miR-223 , decreased circulating miR-30e , which also targets NLRP3 , was found in HCC cases ( Bhattacharya et al., 2016 ). Therefore, it could be suggested that analysis of serum levels of miR-223 and miR-30e could be used for diagnosis of HCC as well as an indicator of the efficacy of anticancer therapeutics. Colorectal cancer (CRC) Data on the role of NLRP3 in colorectal cancer (CRC) pathogenesis is inconsistent, where some evidence suggests a pro-tumorigenic role for the inflammasome, while others identified that the inflammasomes protects against tumor ( Allen et al., 2010 ; Huber et al., 2012 ; Guo et al., 2014 ; Wang et al., 2016 ). Inflammasome expression analysis also demonstrated contradicting results where Wang et al. (2016) reported high NLRP3 in mesenchymal-like colon cancer cells, while Allen et al. (2010) demonstrated decreased inflammasome expression in colitis-associated cancer. Inflammasome contribution to tumorigenesis varies depending on the target cell type in the intestinal tissue ( Lissner and Siegmund, 2011 ). According to Lissner and Siegmund (2011) , inflammasome activation is required to maintain integrity of the epithelium. However, aggravated activation of the inflammasome stimulates intestinal inflammation, which could have a detrimental effect on epithelium permeability and increase its leakage ( Lissner and Siegmund, 2011 ). It was identified that damage to the intestinal epithelium could trigger NLRP3 activation and secretion of IL-18, a proinflammatory cytokine ( Huber et al., 2012 ). Subsequently, it was shown that IL-18 could reduce the expression of IL-22 binding protein (IL-22BP) and increase levels of IL-22 ( Huber et al., 2012 ). Although IL-22 is protective against malignancies, aberrant over expression of IL-22 could trigger gut epithelial cell transformation and CRC development ( Huber et al., 2012 ). Therefore, it is believed that IL-18, a NLRP3 product, has a promoting role in CRC development ( Huber et al., 2012 ). Targeting the inflammasome was suggested as a potential approach for treatment of CRC ( Guo et al., 2014 ). NLRP3 expression was shown to be regulated by multiple miRNAs in various diseases ( Haneklaus et al., 2012 ; Feng et al., 2018 ; Wan et al., 2018 ; Xie Q. et al., 2018 ). However, the role of miRNAs in cancer pathogenesis is not straight forward. There are inconsistent results regarding the expression status of miR-223 , a known regulator of NLRP3 expression, in CRC cell lines and primary tumors. In a clinical study, the expression of miR-223 was found to be significantly higher in stage III/IV patients ( Ding J. et al., 2018 ). However, levels of miR-223 vary significantly in colon tumor derived cell lines ( Ding J. et al., 2018 ). Wu et al. (2012) reported reduced expression of miR-223 in a HCT116, a CRC cell line. In contrast, several research groups demonstrated up-regulation of miR-223 in CRC cell lines and primary tissues ( Wang F. et al., 2017 ; Ju et al., 2018 ; Wei et al., 2018 ). Similar to these results, Ju et al. (2018) demonstrated up-regulation of miR-223 in SW620, a CRC cell line. It was identified that high expression of miR-223 suppresses FoxO3a and enhances cancer cell proliferation ( Ju et al., 2018 ). It appears that the protumorigenic effect of Foxo3a is via NF-κB activation, which is essential for upregulation of the inflammasome linked proinflammatory signaling pathways ( Thompson et al., 2015 ). Unlike miR-223 , data on miR-22 expression status in CRC consistently demonstrates that miR-22 expression is significantly lower in CRC tissues and cell lines ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Also, absence of miR-22 was shown to positively correlate with increased cancer cell proliferation, migration, invasion, and metastasis ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Multiple genes were identified as targets for miR-22 including TIAM1 ( Li B. et al., 2013 ), BTG1 ( Zhang et al., 2015a ), HuR ( Liu Y. et al., 2018 ), and SP-1 ( Xia et al., 2017 ). Among these genes, only SP-1 gene expression was linked to inflammasome regulation ( Hofmann et al., 2015 ). According to Hofmann et al. (2015) , Sp-1 protein could contribute to NLRP3 inflammasome activation in monocytes in chronic recurrent multifocal osteomyelitis. However, the role of Sp-1 in activation of the NLRP3 inflammasome in CRC tumor tissues and monocytes remains largely unknown. Recent finding revealed that, in addition to miR-22 , another negative regulator of NLRP3, miR-30e , is absent in CRC tumors as compared to normal colon tissues ( Laudato et al., 2017 ). However, the role of miR-30e in CRC pathogenesis remains unknown. Gastric cancer (GC) It was shown that NLRP3 inflammasome activation promotes gastric cancer (GC) cells proliferation ( Li S. et al., 2018 ). Over expression of miR-223 supports GC invasion and metastasis in primary GC tumors ( Haneklaus et al., 2012 ). Additionally, Li S. et al. (2018) reported that increased NLRP3 expression in GC tumors and macrophages negatively correlates with miR-22 expression. The authors also demonstrated that constitutive expression of miR-22 dramatically decreases NLRP3 mRNA expression and IL-1β secretion in macrophages ( Li S. et al., 2018 ). Therefore, the effect of targeting NLRP3 expression with miRNAs in tumors and immune cells may vary depending on tumor and/or cell type. Oral squamous cell carcinoma (OSCC) High NLRP3 expression was found in oral squamous cell carcinoma (OSCC) cells and tissues ( Wang H. et al., 2018 ). A role for NLRP3 supporting OSCC proliferation and growth was demonstrated in several reports. Wang G. et al. (2018) demonstrated a positive correlation between NLRP3 expression and tumor size, lymph node status and IL-1β expression in OSCC tissue specimens and in vivo models of OSCC. Also, the authors showed that, silencing of NLRP3 in OSCC cell lines reduced cell proliferation, migration, and invasion in vitro ( Wang H. et al., 2018 ). Additionally, high expression of the NLRP3 inflammasome mediates chemoresistance in OSCC ( Feng et al., 2018 ). Therefore, downregulation of NLRP3 could have a therapeutic potential in OSCC. Surprisingly, high expression of miR-223 , which targets NLRP3 , was found in primary OSCC tissue ( Manikandan et al., 2016 ). In silico analysis identified a Ras Homolog Family Member B ( RHOB ) as a potential target for miR-223 in OSCC ( Manikandan et al., 2016 ). It appears that miR-223 could indirectly suppress NLRP3 and TLR4/NF-κB signaling via RHOB ( Yan et al., 2019 ). These data provide a novel potential target for OSCC treatment, where miR-223 inhibition of NLRP3 could be attained through RHOB. Overexpression of miR-22 in OSCC was shown to reduce NLRP3 activation and decrease OSCC malignancy ( Feng et al., 2018 ). miR-22 levels were shown to be inversely correlated with NLRP3 expression and miR-22 levels were significantly lower in OSCC compared to adjacent non-cancerous tissue ( Feng et al., 2018 ). The inhibitory effect of miR-22 on OSCC migration was confirmed using a lentiviral expression system. As expected an inhibitor of miR-22 promoted OSCC spread ( Feng et al., 2018 ). The 3′-UTR of the NLRP3 gene was identified as a miR-22 target site ( Feng et al., 2018 ). It appears that NLRP3 promotes OSCC growth and tumor spread, which makes miR-22 a potential therapeutic target for cancer treatment. Two miRNAs, miR-223 and miR-22 , were identified as inhibiting the inflammasome and, subsequently, suppressing tumor growth. Therefore, the anti-tumor effect of these molecules in OSCC warrants further investigation. Cervical cancer (CC) Human papillomavirus (HPV) infection and persistent chronic inflammation were identified as fundamental for the pathogenesis of cervical cancer (CC) ( de Castro-Sobrinho et al., 2016 ; Kriek et al., 2016 ). HPV can cause chronic inflammation by inducing TLR4 expression and impairing the TLR4-NF-κB pathway ( Wang et al., 2014 ; He A. et al., 2016 ). Wu et al. (2012) reported reduced expression of miR-223 , which targets NLRP3 , in the CC cell line HeLa. The authors also demonstrated that over-expression of miR-223 inhibits tumor cell proliferation by targeting FOXO1 ( Wu et al., 2012 ). In addition, another direct post-transcriptional regulator of NLRP3 , miR-22, was found to be down-regulated in CC cell lines and tissues ( Xin et al., 2016 ; Wongjampa et al., 2018 ). Furthermore, Wongjampa et al. (2018) reported an inverse correlation between histone deacetylase 6 (HDAC6) and miR-22 . It was previously shown that HDAC6 directly binds to NLRP3 via its ubiquitin-binding domain to regulate NLRP3 inflammasome expression ( Hwang et al., 2015 ). As NLRP3 plays a role in the pathogenesis of HPV induced chronic inflammation, miR-223 and miR-22 , both of which regulate inflammasome activation, could be potential therapeutic tools for the treatment of CC. Glioblastoma (GBM) High NLRP3 inflammasome activation and high levels of inflammasome products are found in malignant glioblastoma (GBM) ( Basu et al., 2004 ; Tarassishin et al., 2014 ). Increased IL-1β, a major NLRP3 inflammasome product, was linked to the release of VEGF and MMPs, angiogenic factors, in human astrocytes and GBM cells ( Suh et al., 2013 ). Therefore, it could be suggested that inflammasome activation favors GBM growth and spread. Several miRNAs were shown to regulate inflammasome expression, where decreased miRNA levels could promote GBM growth and invasion. Ding Q. et al. (2018) demonstrated that miR-223 , which is effective at reducing NLRP3 inflammasome levels in several tumors ( Wu et al., 2012 ), was decreased in GBM tissues ( Ding Q. et al., 2018 ). However, a conflicting report from Cheng et al. (2017) indicated that miR-223 is overexpressed in GBM cell lines. Similar findings were also reported in GBM stem like cells and GBM tissues ( Huang B.S. et al., 2017 ). Similarly there are conflicting data regarding miR-223 targets and phenotypic impacts. A miR-223-3p mimic inhibited tumor cell proliferation and migration, effects that were due to a reduction in proinflammatory cytokines IL-1β and IL-18 in GBM cell lines ( Ding Q. et al., 2018 ). Also, nuclear factor I-A (NFIA) was a target of miR-223 in GBM cell lines and was found to decrease tumorigenesis in the CNS ( Glasgow et al., 2013 ). The pro-tumorigenic effect of miR-223 was linked to suppression of the tumor suppressor paired box 6 ( PAX6 ) ( Cheng et al., 2017 ). By targeting PAX6 , miR-223 could promote GBM stem cell chemotherapy resistance ( Huang B.S. et al., 2017 ). The mechanism underlying the diverse effects of miR-223 on GBM growth and metastasis remains largely unknown. However, it could be suggested that the stage of tumorigenesis plays a role in the effect of miR-223 in GBM. Levels of miR-22 and miR-30e , two post-transcriptional regulators of NLRP3 , are low in GBM tissues ( Li W.B. et al., 2013 ; Chakrabarti et al., 2016 ; Chen H. et al., 2016 ). In addition to targeting NLRP3 , miR-22 can also directly target the 3′-UTRs of SIRT1 ( Li W.B. et al., 2013 ), and miR-22 mimics decrease the expression of SIRT1 protein in GBM cell lines ( Li W.B. et al., 2013 ). Interestingly, several studies have demonstrated that SIRT1 can suppress NLRP3 ( Ma et al., 2015 ; Jiang et al., 2016 ; Zhou C.C. et al., 2016 ). It could be proposed that the decreased levels of miR-22 could fail to control NLRP3 expression, which could enable GMB tumorigenesis. DNA Methylation DNA methylation is dynamic and changes during the embryonic development and differentiation ( Berger, 2007 ). It was shown that DNA methylation silences genes to ensure monoallelic expression, prevent endogenous retrovirus expression and transposon actions ( Walsh et al., 1998 ; Bourc'his et al., 2001 ; Bourc'his and Bestor, 2004 ). DNA methylation is essential for normal cell function; however, its role in the pathogenesis of several diseases has also been confirmed ( Wei et al., 2016 ; Vento-Tormo et al., 2017 ). DNA demethylation is often detected near promoters, suggesting that gene overexpression could play role in pathogenesis of many pathologies ( Ryan et al., 2010 ; Bierne et al., 2012 ). NLRP3 inflammasome expression can also be regulated by changes in gene methylation status. For example, NLRP3 gene expression is silenced in health which appears to be essential for inhibiting inflammation ( Ryan et al., 2010 ; Bierne et al., 2012 ; Wei et al., 2016 ). However, demethylation and, subsequent, overexpression of NLRP3 was linked to pathogenesis of cryopyrin-associated periodic syndromes (CAPS) ( Vento-Tormo et al., 2017 ) and Mycobacterium tuberculosis infection ( Wei et al., 2016 ). Histone Modifications The effect of epigenetic modification of histones was studied using several inflammatory models ( Bayarsaihan, 2011 ). Histone acetylation is essential for initiation of an activation phase of inflammation, which is characterized by the release of pro-inflammatory cytokines via CREB, mitogen-activated protein kinases (MAPKs), nuclear factor-κB (NF-κB) and signal transducer and activator of transcription (STAT) factors ( Escobar et al., 2012 ). In contrast, histone deacetylations regulate the late, an attenuation phase of inflammation ( Villagra et al., 2010 ). It appears that inflammasome activation can also be regulated by affecting the acetylation status of histones, as it was recently shown by Liu C.C. et al. (2018) . The authors demonstrated upregulation of NLRP3 in patients diagnosed with painful neuropathy, which could be prevented by inhibition of histone acetylation. Non-coding RNAs In addition to epigenetic modulation non-coding RNAs are also involved in NLRP3 regulation ( Bayarsaihan, 2011 ), as was demonstrated in the setting of inflammation caused by microbial and viral infection ( Li et al., 2010 ; Ryan et al., 2010 ; Bierne et al., 2012 ; Jin et al., 2013a ; Chen and Ichinohe, 2015 ). This inflammation is post-transcriptionally regulated via non-coding RNAs targeting inflammasome components, where mRNA stability and inhibition of translation were most commonly affected ( Bayarsaihan, 2011 ). Post-transcriptional Regulation of NLRP3 Inflammasomes: MicroRNA (miRNA) MicroRNAs are endogenous conservative, single-stranded non-coding RNAs which are 19–24 nucleotides long. Usually, miRNAs are derived from transcripts with a hairpin structure and are loaded into the Argonaute protein within a silencing complex ( Hutvagner and Zamore, 2002 ; Mourelatos et al., 2002 ; Bartel, 2004 ). The inhibitory effect of miRNAs is explained by their binding to the untranslated regions (UTRs) of transcripts which modulates the stability and translation of the target mRNA ( Figure 2 ) ( Ruvkun, 2001 ; Filipowicz et al., 2008 ; Bartel, 2009 ; Coll and O'Neill, 2010 ). miRNAs can modulate the expression of histone modifies including histone deacetylases and DNA methyltransferases resulting in modulation of histone modifications and DNA methylation ( Tuddenham et al., 2006 ; Fabbri et al., 2007 ). FIGURE 2 miRNA regulation of NLRP3 inflammasome expression. (A) Priming signal triggers NLRP3, PC1, IL-1β, and IL-18 transcription and protein synthesis. Activation signal initiates inflammasome formation and release of AC1. AC1 proteolytically cleaves pro-IL-1β and pro-IL-18, liberating active cytokines. (B) Suppression of NLRP3 protein translation and inflammasome formation by miRNA. Priming stimulus triggers NLRP3 transcription; however, miR-223 , miR-22 , miR-30e , and miR-7 bind to the UTR region of NLRP3 mRNA and interrupt protein translation. Absence of NLRP3 protein leads failure of the inflammasome protein complex formation, when the second stimulus present. NLRP3 activation is tightly regulated where two signals are required to initiate functional inflammasome formation. The first signal includes cell priming with TLR ligands ( Bauernfeind et al., 2009 ; Franchi et al., 2009 ). Therefore, it could be suggested that targeting TLR expression will impact the inflammasome activity. Indirect regulation of TLR expression includes modulation of the downstream pathways molecules, which has been shown in injuries, inflammation and cancer ( Coll and O'Neill, 2010 ; Sheedy et al., 2010 ; Nahid et al., 2011 ; Anzola et al., 2018 ; Tan et al., 2018 ; Zhi et al., 2018 ). TLR4 ligands are the most studied priming signals of NLRP3 activation ( Groslambert and Py, 2018 ). It was shown that the TLR ligand binding increases the level of several miRNAs, including miR-155 , miR-146a , miR-21 , and miR-132 , which were linked to inhibition of TLR4/MyD88/NF-κB signaling ( Coll and O'Neill, 2010 ; Sheedy et al., 2010 ; Nahid et al., 2011 ; Anzola et al., 2018 ; Tan et al., 2018 ; Zhi et al., 2018 ). It appears that upregulation of miRNAs is a component of a negative feedback mechanism designed to down-modulate inflammatory cytokine production after response to microbial stimuli ( Ceppi et al., 2009 ). A direct inhibitory effect of let-7 family miRNAs on TLR4 mRNA has been demonstrated ( Chen et al., 2007 ). Let-7 miRNA regulation of TLR4 was shown to occur via post-transcriptional suppression ( Androulidaki et al., 2009 ). It was suggested that let-7 miRNA downregulation of TLR4 could have detrimental effect on host defense against microbes, promoting microbial survival and propagation ( Chen et al., 2005 ; Muxel et al., 2018 ). Post-transcriptional regulation of TLR signaling and its impact on diseases are reviewed by Nahid et al. (2011) . Active inflammasome complex formation requires a second signal, initiating substantial NLRP3 transcription ( Dostert et al., 2008 ; Piccini et al., 2008 ). During this transcriptionally active phase, NLRP3 mRNA could be regulated by miRNA, as was shown by miR-223 ( Bauernfeind et al., 2012 ). According to an in silico analysis, miR-223 can bind to a highly conserved region of the 3′UTR of NLRP3 mRNA and subsequently interfere with protein translation ( Lewis et al., 2005 ). Interestingly, miR-223 appears to be an important NLRP3 regulator in leukocytes ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ), where the miRNA levels have been shown to vary in different leukocyte subsets. For example, this miRNA was found absent in T and B lymphocytes ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ). In contrast, the miR-223 was demonstrated in myeloid cells, where it was highest in neutrophils, followed by macrophages and dendritic cells ( Bauernfeind et al., 2012 ). It has been suggested that this miRNA plays role in granulocyte production and regulation of inflammation ( Johnnidis et al., 2008 ; Neudecker et al., 2017 ). Decreased production of pro-inflammatory cytokines such as IL-1β and IL-18 was demonstrated in cells treated with miR-223 or its mimics ( Neudecker et al., 2017 ; Ding Q. et al., 2018 ). These data suggest that miR-223 could be a potential target for regulation of NLRP3 expression, where increased miRNA could reduce inflammasome activation and, subsequently, abrogate the inflammation ( Bauernfeind et al., 2012 ; Haneklaus et al., 2012 ). Since several miRNAs could regulate expression of a single transcript ( Krek et al., 2005 ), it is likely that in addition to miR-223 , other miRNAs can alter NLRP3 transcription ( Figure 3 ). FIGURE 3 UTR binding sites of NLRP3 for miRNAs responsible for the regulation of inflammation. Numerous studies have identified that pathogens, trauma and cancer can cause abnormal expression of miRNAs which impair NLPR3 inflammasome function disrupt the functional complex formation and its signaling ( Table 1 ). Table 1 Aberrant miRNA expressions linked to inflammasome related diseases. Disease miRNA Regulation of miRNA Target cell Target gene References Inflammatory bowel diseases miR-223 ↑ Intestinal biopsies NLRP3 Neudecker et al., 2017 ↑ Circulating monocytes, neutrophils Johnnidis et al., 2008 ; Bauernfeind et al., 2012 ; Neudecker et al., 2017 ↓ Macrophages Rheumatoid arthritis miR-33 ↑ Macrophages PGC1-α Karunakaran et al., 2015 ; Xie Q. et al., 2018 Type 1 diabetes miR-146a ↓ Macrophages TLR2, TLR4 Bhatt et al., 2016 ; Xie Z. et al., 2018 Type 2 diabetes miR-146a ↓ Macrophages TLR2, TLR4 Balasubramanyam et al., 2011 Systemic lupus erythematosus miR-23b ↓ Inflammatory lesions TAB2, TAB3, IKK-α Zhu et al., 2012 Parkinson's disease miR-7 ↓ Microglia NLRP3 Zhou Y. et al., 2016 miR-30e ↓ NLRP3 Li D. et al., 2018 Atherosclerosis miR-22 ↓ Monocytes, macrophages NLRP3 Huang W.Q. et al., 2017 miR-9 ↓ JAK1 Wang F. et al., 2017 miR-30e-3p ↓ FOXO3 Li P. et al., 2018 Acute lung injury/acute respiratory distress syndrome miR-223 ↑ Ly6G+ neutrophils NLRP3 Feng et al., 2017 Hepatocellular carcinoma miR-223 ↑ Tumor cell line NLRP3, EPB41L3, FOXO1 Li X. et al., 2011 ; Kim et al., 2017 miR-223 ↓ Patient's sera NLRP3 Bhattacharya et al., 2016 miR-30e ↓ NLRP3 Bhattacharya et al., 2016 Colorectal cancer miR-223 Tumor type specific Tumor tissue, tumor cell line NLRP3, FoxO3a Ju et al., 2018 miR-22 ↓ SP-1 Xia et al., 2017 Gastric cancer miR-223 ↑ Tumor tissue NLRP3 Haneklaus et al., 2012 miR-22 ↓ Macrophages NLRP3 Li S. et al., 2018 Oral squamous cell carcinoma miR-223 ↑ Tumor tissue RHOB Manikandan et al., 2016 miR-22 ↓ NLRP3 Feng et al., 2018 Cervical cancer miR-223 ↓ Tumor tissue, tumor cell line FOXO1 Wu et al., 2012 miR-22 ↓ HDAC6 Wongjampa et al., 2018 Glioblastoma miR-223 Controversial Tumor tissue, tumor cell line NFIA, PAX6 Fazi et al., 2005 ; Glasgow et al., 2013 ; Cheng et al., 2017 ; Ding Q. et al., 2018 miR-22 ↓ SIRT1 Li W.B. et al., 2013 miRNA in Regulation of Inflammasome in Infections Inflammasome activation is an important component of infectious pathogens surveillance and antimicrobial immune and inflammatory responses. This inflammasome was shown to be activated by several bacterial pathogens including Staphylococcus aureus , Salmonella typhimurium , Listeria monocytogenes , Mycobacterium , Streptococcus pyogenes , Neisseria gonorrhoeae as well as fungi such as Candida albicans and Aspergillus fumigatus ( Franchi et al., 2006 ; Mariathasan et al., 2006 ; Miao et al., 2006 ; Craven et al., 2009 ; Duncan et al., 2009 ; Harder et al., 2009 ; Hise et al., 2009 ; Joly et al., 2009 ; Munoz-Planillo et al., 2009 ; Broz et al., 2010 ; Carlsson et al., 2010 ; McElvania Tekippe et al., 2010 ; Said-Sadier et al., 2010 ). NAIP/NLRC4 inflammasome can protect against Salmonella Typhimurium and C. rodentium invasion by bacteria expulsion from intestinal epithelial cells together with IL-18 and eicosanoid lipid mediators release ( Nordlander et al., 2014 ; Sellin et al., 2014 ; Rauch et al., 2017 ). It appears that NLRP3 activation is essential for establishing the inflammatory milieu in the target tissue and augmenting the phagocytic capacity of the local macrophages ( Master et al., 2008 ; Melehani and Duncan, 2016 ; Cohen et al., 2018 ). Enhanced macrophage bactericidal activity is the most commonly identified mechanism of inflammasome antimicrobial effect ( Master et al., 2008 ; Cohen et al., 2018 ). Additionally, NLRP3 activation induced death of macrophages was described as an effort to prevent microbial propagation and spread ( Miao et al., 2010 ; Sagulenko et al., 2013 ). However, there is a growing body of evidence suggesting that there is a threshold of NLRP3 activity, which acts as a safeguard mechanism to prevent inflammasome over-activation. It appears that aberrant NLRP3 activation could have a detrimental effect on tissues homeostasis and compromise barrier integrity ( Bortolotti et al., 2018 ; McKenzie et al., 2018 ). It is this detrimental effect of the inflammasome over-activation that is often employed by microbes to ensure spread and propagation ( Duncan et al., 2009 ; Harder et al., 2009 ; Carlsson et al., 2010 ). Microbial virulence factors often act as NLRP3 activators. For example, it was shown that the detrimental (to the host) role of Esx1, a membrane lysis factor of Mycobacterium ( Stanley et al., 2003 ), is linked to inflammasome activation ( Carlsson et al., 2010 ). Two virulence factors of group A Streptococcus (GAS), M protein and streptolysin O, were also identified as contributing into NLRP3 activation and IL-1β production ( Harder et al., 2009 ; Valderrama et al., 2017 ). Both virulence factors are commonly detected in association with invasive GAS infections, including necrotizing fasciitis and toxic shock syndrome. Therefore, NLRP3 activation by virulent factors could promote microbe propagation and aid their escape from immune clearance. Restoring the NLRP3 activation threshold could be a novel therapeutic approach for treatment of invasive infections. In this respect, miRNA may be a tool to regain control over NLRP3. It has been shown that miR-223 expression is consistently high in NLRP3 responsive cells, suggesting the high efficacy of this miRNA in prevention of inflammasome over-activation ( Bauernfeind et al., 2012 ). Dorhoi et al. (2013) demonstrated that miR-223 is upregulated in the blood and lung parenchyma of patients diagnosed with tuberculosis. Also, data collected using animal models confirmed the link between deletion of miR-223 and increased susceptibility to Mycobacterium tuberculosum infection ( Dorhoi et al., 2013 ). Similarly, a protective role of miR-223 in Staphylococcus aureus infection was demonstrated by Fang et al. (2016) . Additionally, the effect of targeting TLR4 for NLRP3 regulation in Listeria monocytogenes infection was demonstrated by Schnitger et al. (2011) . The authors identified that, miR-146a can directly inhibit TLR4 receptor expression, which can downregulate inflammasome activity ( Schnitger et al., 2011 ). Many viruses can activate inflammasomes, including Influenza virus, Hepatitis C virus, Herpes simplex virus-1, etc. ( Delaloye et al., 2009 ; Ichinohe et al., 2010 ; Ito et al., 2012 ; Kaushik et al., 2012 ; Negash et al., 2013 ; Triantafilou et al., 2013a , b ; Wu et al., 2013 ; Ermler et al., 2014 ; Chen and Ichinohe, 2015 ). Inflammasome activation appears to be essential for anti-viral protection, serving as viral genome sensors and triggering innate immune response ( Muruve et al., 2008 ; Lupfer et al., 2015 ). The protective role of inflammasomes was shown in influenza virus infection as an increased viral clearance was NLRP3 dependent ( Allen et al., 2009 ). Also, inflammasome activation improved the survival rate in an animal model of influenza ( Ichinohe et al., 2009 ). Thomas et al. (2009) demonstrated that, the innate immune response activation by NLRP3 inflammasomes is essential for animal protection. However, our understanding of the mechanisms of inflammasome antiviral defense remains limited ( Anand et al., 2011 ). Some viruses were shown to post-transcriptionally regulate inflammasome expression to benefit self-replication and propagation ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). For example, miRNA suppression of inflammasomes was shown in Epstein Barr Virus (EBV) infected cells ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). It appears that, EBV can avert NLRP3 inflammasome activation by expressing miRNAs encoded by three BHRF1 -regions and 40 BART -regions of the viral genome ( Albanese et al., 2016 ; Tagawa et al., 2016 ; Farrell, 2018 ). Additionally, two miRNAs encoded by EBV, miR-BART11-5p and miR-BART15 , were identified by Haneklaus et al. (2012) , which could bind to the 3′-UTR of NLRP3 , the same site targeted by miR-223 , and inhibit the inflammasome. It remains to be determined whether these viral miRNA could be used as therapeutic targets. miRNA Regulation of Inflammasome in Autoimmune Diseases Autoimmune diseases are often the result of a dysregulated immune response, characterized by inflammation and organ damage ( Chang, 2013 ; Yang and Chiang, 2015 ). Chronic inflammation is frequently identified as a predisposing factor for an autoimmune reaction ( Yang and Chiang, 2015 ). Multiple mechanisms were suggested to explain prolonged inflammation leading to autoimmunity; where failure to control inflammasome activation was recently identified in some autoimmune conditions ( Yang and Chiang, 2015 ). It has been established that in addition to inflammation, an increased secretion of IL-1β and IL-18, can stimulate proliferation and organ distribution of the effector T cells, which can cause tissue damage ( Oyanguren-Desez et al., 2011 ; Celhar et al., 2012 ). Therefore, targeting the inflammasome could be suggested to restore control over the inflammatory and immune response. Therapeutic potentials of several NLRP3 targeting miRNAs were investigated in autoimmune diseases such as inflammatory bowel diseases (IBDs) ( Neudecker et al., 2017 ), RA ( Xie Z. et al., 2018 ), type 1 diabetes (T1D) ( Yang and Chiang, 2015 ), type 2 diabetes (T2D) ( Yang and Chiang, 2015 ), and systemic lupus erythematosus (SLE) ( Zhu et al., 2012 ). Inflammatory bowel diseases (IBDs) Inflammatory bowel diseases are characterized by chronic inflammation of the intestine and comprise two disorders Crohn's disease and ulcerative colitis. It is believed that the pathogenesis of IBDs is associated with dysregulation of innate and adaptive immune responses, triggered by microbial antigens. This could result in chronic inflammation of the digestive tract and damage to the intestinal mucosa ( Fiocchi, 1998 ). The role of the inflammasome in intestinal inflammation is controversial. Zaki et al. (2010) reported that, NLRP3 induced production of IL-18 in intestinal epithelial cells can be protective, and contributes to epithelium integrity in experimental colitis. In contrast, Seo et al. (2015) have demonstrated the role of inflammasome in exacerbation of an intestinal pathology. The damaging effect of the inflammasome was also confirmed by Shouval et al. (2016) , who identified that IL-1β inhibition improves the course of IBDs. It appears that increased IL-1β levels and tissue damage in IBDs are linked to NLRP3 activation in myeloid leukocytes infiltrating the gut tissue ( Neudecker et al., 2017 ). The role of the inflammasome in IBDs pathogenesis was also confirmed by using a miR-223 deficient animal model of colitis ( Neudecker et al., 2017 ). miR-223 deficient mice develop experimental colitis manifesting with colonic ulceration, inflammatory leukocyte infiltration and tissue injury which resembles closely IBDs ( Neudecker et al., 2017 ). Tissue injury in these mice was linked to an enhanced NLRP3 expression and elevated IL-1β ( Neudecker et al., 2017 ). Treatment of animals with miR-223 mimetics alleviated symptoms of the colitis which coincided with reduced NLRP3 RNA and IL-1β levels ( Neudecker et al., 2017 ). This data presents miR-223 as a novel biomarker and therapeutic target in subsets of IBDs and colitis ( Polytarchou et al., 2015 ). Rheumatoid arthritis (RA) Rheumatoid arthritis is a chronic, systemic inflammatory disease affecting joints as well as skin, eyes, lungs, heart, and blood vessels ( Scott et al., 2010 ). It was suggested that RA pathogenesis is related to activation of the NLRP3/IL-1β axis, where inflammasome activation was linked to worsening symptoms of the disease ( Xie Q. et al., 2018 ). It was shown that activation of NLRP3 leads to an abundant expression of IL-1β ( Guo et al., 2018 ), which can trigger T helper type 17 (Th17) cell differentiations and osteoclasts activation in RA ( Dayer, 2003 ; McInnes and Schett, 2011 ; Zhang et al., 2015b ). Th17 cells play a central role in RA pathogenesis, by maintaining chronic inflammation, recruiting neutrophils and promoting joint degradation ( Cai et al., 2001 ; Shahrara et al., 2009 ; Leipe et al., 2010 ). Recently, an indirect effect of miR-33 on NLRP3 activation was demonstrated in RA ( Xie Q. et al., 2018 ), which could be explained by miRNA controlled dysregulation of mitochondrial function ( Schroder et al., 2010 ; Zhou et al., 2011 ; Miao et al., 2014 ; Ouimet et al., 2015 ). Xie Q. et al. (2018) suggested that miR-33 increases mitochondrial oxygen consumption and accumulation of reactive oxygen species which upregulates expression of NLRP3 and PCA1 in RA. Also, both miR - 33 expression and NLRP3 inflammasome activity were found to be higher in RA monocytes as compared to controls ( Xie Q. et al., 2018 ). These findings indicate that miR - 33 could play an indirect role in pathogenesis of RA through NLRP3 inflammasome activation. Additional studies will provide more insight into the miRNA regulation of NLRP3 in RA and its therapeutic and prognostic implications. Type 1 diabetes (T1D) Type 1 diabetes is caused by autoimmune targeted elimination of pancreatic β cells islet ( Kloppel et al., 1985 ). It was shown that TLRs play an essential role in the pathogenesis of T1D ( Xie Z. et al., 2018 ). Upregulated expression of TLR4 as well as increased activity of the downstream targets was demonstrated in monocytes from T1D ( Devaraj et al., 2008 ). Increased expression of activated TLRs was explained as a reaction to a high levels of circulating ligands in TID ( Devaraj et al., 2009 ). Also, epigenetic regulation was associated with an aberrant TLR signaling and an increased IL-1β expression in T1D ( Grishman et al., 2012 ). Several miRNAs were found altered in pre-TID patients, where levels of nine miRNAs ( miR-146a , miR-561 , and miR-548a-3p , miR-184 , and miR-200a ) were decreased, and two miRNAs ( miR-30c and miR-487a ) were increased ( Grieco et al., 2018 ). Supporting these results was data published by Wang G. et al. (2018) demonstrating lower levels of miR-150 , miR-146a , and miR-424 compared to controls. One of the most consistent findings was the decreased miR-146a levels in T1D. It appears that miR-146a deficiency could play role in T1D exacerbation and increased IL-1β and IL-18 expression ( Bhatt et al., 2016 ). Increased IL-1β levels could indicate inflammasome activation in T1D, although the role of inflammasome in the disease pathogenesis remains largely unknown. Type 2 diabetes (T2D) Circulating autoantibodies to β cells, self-reactive T cells and the glucose-lowering efficacy of some immunomodulatory therapies are suggestive of the autoimmune nature of the T2D ( Itariu and Stulnig, 2014 ). Interestingly, a role for miRNA regulation of gene expression was demonstrated in T2D, where Balasubramanyam et al. (2011) have shown reduced miR-146a which was associated with increased NF-κB , TNF-α and IL-6 mRNA levels. It is the same miRNA, which was found implicated to pathogenesis of T1D ( Xie Z. et al., 2018 ), indicating potential similarities in the pathogenesis of both diseases. Recently in vivo studies demonstrated that miR-146a deficiency could increase expression of M1 and suppress expression of M2 markers in macrophages collected from patients with diabetes ( Bhatt et al., 2016 ). Macrophage polarization occurs in the presence of IFNγ (M1) or IL-4 (M2) ( Nathan et al., 1983 ; Stein et al., 1992 ) and is linked to pro-inflammatory and anti-inflammatory activities, respectively. M1 macrophages were shown to support inflammation by producing pro-inflammatory cytokines, including the inflammasome product IL-1β ( Bhatt et al., 2016 ). Therefore, a link could be suggested between low miR-146a levels and inflammasome activation in M1 cells. More investigation is required to identify the connection between miR-146a and inflammasome activation and the role of this in T2D pathogenesis. Systemic lupus erythematosus (SLE) Systemic lupus erythematosus is an autoimmune disease caused by the loss of immune tolerance to ubiquitous autoantigens ( Tsokos, 2011 ). Inflammation plays essential role in SLE pathogenesis ( Yang et al., 2014 ; Rose and Dorner, 2017 ), where high levels of circulating proinflammatory cytokines are commonly detected ( Yao et al., 2016 ; Mende et al., 2018 ). Inflammasome activation is proposed as one of the mechanisms underlying increased proinflammatory cytokine level in SLE ( Kahlenberg and Kaplan, 2014 ). This assumption is supported by a report where IL-1β deficient mice were found to be resistant to experimental SLE ( Voronov et al., 2006 ). Also, an increased expression of NLRP3 and AC1 have been reported in SLE nephritis biopsies ( Kahlenberg et al., 2011 ). Kahlenberg and Kaplan (2014) have shown that SLE macrophages are highly reactive to innate immune stimuli, often leading to inflammasome activation. Therefore, targeting inflammasome activity could be a novel approach for SLE treatment. The expression of several miRNAs targeting the inflammasome and its products were found differentially expressed in SLE. For example, Wang et al. (2012) have demonstrated high levels of circulating miR-223 , which was shown to inhibit NLRP3 , in SLE. Also, reduced levels of circulating miR-146a , which regulates priming of TLRs, was found in SLE plasma ( Wang et al., 2012 ). Interestingly, expression of miR-23b , which indirectly inhibits IL-1β responses, was shown to be downregulated in inflammatory lesions of SLE patients and animal model ( Zhu et al., 2012 ). More studies are required to determine the role of miRNAs in pathogenesis of SLE and their therapeutic potential. miRNA Regulation of Inflammasome in Neurodegenerative Disorders Inflammasome products, IL-1β and IL-18, were shown to be essential for the health and functional competence of the nervous system ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ). NLRP3 expression was demonstrated in microglia and astrocytes, which could explain the constitutive level of these cytokines in the brain ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ; Savage et al., 2012 ; Minkiewicz et al., 2013 ; Cho et al., 2014 ; Lu M. et al., 2014 ). Interestingly, higher than normal levels of IL-1β and IL-18 were found in several neurodegenerative disorders, suggesting that over-activation of inflammasomes may play a role in pathogenesis of these diseases ( Cho et al., 2014 ; Lu M. et al., 2014 ; Denes et al., 2015 ; Mamik and Power, 2017 ; Song et al., 2017 ). The significance of miRNA in the regulation of inflammasome activation in the pathogenesis of neurodegenerative diseases remains largely unknown. However, the role of an aberrant miRNA in regulation of NLRP3 expression was previously demonstrated in Parkinson's disease (PD). Parkinson's disease is a neurodegenerative disease which is characterized by progressive loss of dopaminergic neurons in substantia nigra compacta ( Gasser, 2009 ). It is believed that accumulation of α-Syn fibrillary aggregates in the brain, most notably in the nigral dopaminergic neurons, induces the neuroinflammation ( Eriksen et al., 2003 ). According to Zhou Y. et al. (2016) , α-Syn can activate NLRP3 inflammasomes in microglia leading to an increased production of IL-1β. The authors also demonstrated that, miR-7 and miR-30e analogs can inhibit NLRP3 inflammasome mediated neuroinflammation in the brain and protect dopaminergic neurons ( Zhou Y. et al., 2016 ). It appears that the anti-inflammatory effects of miR-7 and miR-30e are associated with their targeting of NLRP3 mRNA in microglial cells. Interestingly, decreased miR-7 and miR-30e expression was demonstrated in PD, which could lead to the loss of the regulatory control of α-Syn induced NLRP3 activation ( Li D. et al., 2018 ). miRNA Regulation of the Inflammasome in Cardiovascular Diseases (CVDs) The physiological significance of inflammation is confirmed as it facilitates elimination of destructive stimuli and pathogens. However, aberrant inflammatory responses could cause tissue damage, tissue fibrosis and chronic diseases ( Liu D. et al., 2018 ). Inflammation is recognized as a major risk factor for CVDs ( Zhou et al., 2018 ), where chronic inflammasome activation was shown to contribute to the pathogenesis of atherosclerosis, ischemic and non-ischemic heart diseases ( Zhou et al., 2018 ). Therefore, regulation of inflammasome activity using miRNA could be used for treatment and prevention of CVDs. Currently, strong evidence for the role of NLRP3 activation has been demonstrated in pathogenesis atherosclerosis. Atherosclerosis is a form of CVD characterized by narrowing of the blood vessel lumen due to plaque formation, continuous dyslipidemia and inflammation ( Ross, 1993 ). Chronic inflammation is commonly found in and around the atherosclerotic plaques which has an adverse effect on the arterial wall structure and function ( Bernhagen et al., 2007 ). It is believed that atherogenic lipid mediators, involved in the formation of chronic inflammation in atherosclerotic plaque ( Chen et al., 2006 ), can trigger peripheral blood monocytes migration and differentiation into macrophages within the intima of the arterial wall ( Chen et al., 2006 ). T cells were also detected in atherosclerotic lesions ( Kleemann et al., 2008 ), where, together with activated macrophages, they were shown to secrete proinflammatory mediators such as interferons, interleukins, and proteases ( Østerud and Bjørklid, 2003 ; Shashkin et al., 2005 ; Tabas, 2005 ; Chen et al., 2006 ). IL-1β expression was identified in the early phase of atherosclerotic plaque formation and this stimulates secretion of additional cytokines and chemokines ( Kleemann et al., 2008 ). Therefore, inflammasome activation in macrophages and T cell within the atherosclerotic lesion contributes to the pathogenesis of chronic inflammation. miR-22 , a miRNA inhibiting NLRP3 , is decreased in peripheral blood mononuclear cells from coronary atherosclerosis ( Chen B. et al., 2016 ), suggesting that upregulation of this miRNA could have therapeutic potential in CVD. Supporting this assumption, Huang W.Q. et al. (2017) investigated the effect of miR-22 on the NLRP3 inflammasome and endothelial cell damage in an in vivo model of coronary heart disease. The authors demonstrated that miR-22 mimics could decrease the release of inflammatory cytokines such as IL-1β and IL-18 by suppressing NLRP3 expression in monocytes and macrophages ( Huang W.Q. et al., 2017 ). Two additional miRNAs, miR-9 and mir-30e-5p were found to indirectly affect inflammasome activation in atherosclerosis ( Wang Y. et al., 2017 ; Li P. et al., 2018 ). It appears that miR-9 could indirectly suppress inflammasome activation by targeting an atherogenic lipid mediator, oxidized low-density lipoprotein (oxLDL), in atherosclerosis ( Liu W. et al., 2014 ). In another report, Wang Y. et al. (2017) reported that miR-9 inhibits NLRP3 inflammasome activation induced by oxLDL in human THP-1 derived macrophages and peripheral blood monocytes in an in vitro atherosclerosis model. miR-9 targets Janus kinase 1 ( JAK1 ) pathway ( Wang Y. et al., 2017 ) inhibiting expression of NF-κB p65 which is required for the first step of NLRP3 inflammasome activation ( Wang Y. et al., 2017 ). In addition, miR-30c-5p was linked to an indirect regulation of NLRP3 expression in atherosclerosis ( Li P. et al., 2018 ). Li P. et al. (2018) reported that miR-30c-5p protects human aortic endothelial cells (HAECs) from the oxLDL insult by targeting FOXO3 . The authors showed that miR-30c-5p can suppress FOXO3 expression and, consequently, decrease levels of NLRP3, AC1, IL-18 and IL-1β in HAECs ( Li P. et al., 2018 ). As evidence emerges supporting the role of NLRP3 in the pathogenesis of atherosclerosis, targeting the inflammasome becomes an attractive therapeutic approach, where miRNAs could be suitable novel tools. miRNA in Regulation of Inflammasome in Cancer The role of the inflammasome in tumorigenesis remains controversial. Some reports indicate that NLRP3 inflammasome activation and IL-18 signaling protect against colorectal cancer ( Karki et al., 2017 ), whereas progression of breast cancer, fibrosarcoma, gastric carcinoma, and lung metastasis were shown to be supported by the inflammasome ( Okamoto et al., 2010 ; Kolb et al., 2014 ). Inflammasome regulation is complex, where multiple factors are implicated, making identification of the key regulatory elements challenging. As the inflammasome involvement in pathogenesis of some malignancies becomes more evident, understanding the regulatory mechanisms could lead to the discovery of novel therapeutic targets for cancer treatment. Hepatocellular carcinoma (HCC) Hepatocellular carcinoma (HCC) is a frequent sequelae of hepatitis B and hepatitis C viral infection ( Perz et al., 2006 ). It is understood that these viruses activate NLRP3 inflammasomes causing hepatocyte pyroptosis, apoptosis and fibrosis ( Kofahi et al., 2016 ). However, HCC tissue analysis failed to detect inflammasome activation; in fact, it was found to be significantly down-regulated when compared to the adjacent normal tissue ( Zhu et al., 2011 ; Wei et al., 2014 ). To explain this inconsistency, Wei et al. (2014) suggested that NLRP3 expression is dynamic changing during the progression of HCC. It appears that NLRP3 expression was increased in liver cells at the early stages of transformation, while inflammasome levels were decreased in malignant cells when compared to adjacent normal tissue ( Wei et al., 2014 ). Interestingly, levels of miR-223 , a negative regulator of NLRP3 , were found to be increased in Hep3B cells derived from HCC ( Wan et al., 2018 ). Increased miR-223 was shown to coincide with tumor growth, suggesting a role in post-transcriptional mechanisms in malignant progression. In addition to NLRP3 , miR-223 was shown to target erythrocyte membrane protein band 4.1 like 3 ( EPB41L3 ) and FOXO1 ( Li and Rana, 2014 ; Kim et al., 2017 ). FOXO1 transcription factor binds to the thioredoxin-interacting protein (TXNIP) and regulates genes involved in cell death as well as the oxidative stress responses ( Kim et al., 2017 ). TXNIP interacts with the NLRP3 inflammasome and activates AC1 in murine β-cells ( Zhou et al., 2010 ). In addition, miR-223 appears to be released systemically, where the level of this miRNA in the plasma was significantly lower in HCC cases ( Giray et al., 2014 ). In addition to miR-223 , decreased circulating miR-30e , which also targets NLRP3 , was found in HCC cases ( Bhattacharya et al., 2016 ). Therefore, it could be suggested that analysis of serum levels of miR-223 and miR-30e could be used for diagnosis of HCC as well as an indicator of the efficacy of anticancer therapeutics. Colorectal cancer (CRC) Data on the role of NLRP3 in colorectal cancer (CRC) pathogenesis is inconsistent, where some evidence suggests a pro-tumorigenic role for the inflammasome, while others identified that the inflammasomes protects against tumor ( Allen et al., 2010 ; Huber et al., 2012 ; Guo et al., 2014 ; Wang et al., 2016 ). Inflammasome expression analysis also demonstrated contradicting results where Wang et al. (2016) reported high NLRP3 in mesenchymal-like colon cancer cells, while Allen et al. (2010) demonstrated decreased inflammasome expression in colitis-associated cancer. Inflammasome contribution to tumorigenesis varies depending on the target cell type in the intestinal tissue ( Lissner and Siegmund, 2011 ). According to Lissner and Siegmund (2011) , inflammasome activation is required to maintain integrity of the epithelium. However, aggravated activation of the inflammasome stimulates intestinal inflammation, which could have a detrimental effect on epithelium permeability and increase its leakage ( Lissner and Siegmund, 2011 ). It was identified that damage to the intestinal epithelium could trigger NLRP3 activation and secretion of IL-18, a proinflammatory cytokine ( Huber et al., 2012 ). Subsequently, it was shown that IL-18 could reduce the expression of IL-22 binding protein (IL-22BP) and increase levels of IL-22 ( Huber et al., 2012 ). Although IL-22 is protective against malignancies, aberrant over expression of IL-22 could trigger gut epithelial cell transformation and CRC development ( Huber et al., 2012 ). Therefore, it is believed that IL-18, a NLRP3 product, has a promoting role in CRC development ( Huber et al., 2012 ). Targeting the inflammasome was suggested as a potential approach for treatment of CRC ( Guo et al., 2014 ). NLRP3 expression was shown to be regulated by multiple miRNAs in various diseases ( Haneklaus et al., 2012 ; Feng et al., 2018 ; Wan et al., 2018 ; Xie Q. et al., 2018 ). However, the role of miRNAs in cancer pathogenesis is not straight forward. There are inconsistent results regarding the expression status of miR-223 , a known regulator of NLRP3 expression, in CRC cell lines and primary tumors. In a clinical study, the expression of miR-223 was found to be significantly higher in stage III/IV patients ( Ding J. et al., 2018 ). However, levels of miR-223 vary significantly in colon tumor derived cell lines ( Ding J. et al., 2018 ). Wu et al. (2012) reported reduced expression of miR-223 in a HCT116, a CRC cell line. In contrast, several research groups demonstrated up-regulation of miR-223 in CRC cell lines and primary tissues ( Wang F. et al., 2017 ; Ju et al., 2018 ; Wei et al., 2018 ). Similar to these results, Ju et al. (2018) demonstrated up-regulation of miR-223 in SW620, a CRC cell line. It was identified that high expression of miR-223 suppresses FoxO3a and enhances cancer cell proliferation ( Ju et al., 2018 ). It appears that the protumorigenic effect of Foxo3a is via NF-κB activation, which is essential for upregulation of the inflammasome linked proinflammatory signaling pathways ( Thompson et al., 2015 ). Unlike miR-223 , data on miR-22 expression status in CRC consistently demonstrates that miR-22 expression is significantly lower in CRC tissues and cell lines ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Also, absence of miR-22 was shown to positively correlate with increased cancer cell proliferation, migration, invasion, and metastasis ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Multiple genes were identified as targets for miR-22 including TIAM1 ( Li B. et al., 2013 ), BTG1 ( Zhang et al., 2015a ), HuR ( Liu Y. et al., 2018 ), and SP-1 ( Xia et al., 2017 ). Among these genes, only SP-1 gene expression was linked to inflammasome regulation ( Hofmann et al., 2015 ). According to Hofmann et al. (2015) , Sp-1 protein could contribute to NLRP3 inflammasome activation in monocytes in chronic recurrent multifocal osteomyelitis. However, the role of Sp-1 in activation of the NLRP3 inflammasome in CRC tumor tissues and monocytes remains largely unknown. Recent finding revealed that, in addition to miR-22 , another negative regulator of NLRP3, miR-30e , is absent in CRC tumors as compared to normal colon tissues ( Laudato et al., 2017 ). However, the role of miR-30e in CRC pathogenesis remains unknown. Gastric cancer (GC) It was shown that NLRP3 inflammasome activation promotes gastric cancer (GC) cells proliferation ( Li S. et al., 2018 ). Over expression of miR-223 supports GC invasion and metastasis in primary GC tumors ( Haneklaus et al., 2012 ). Additionally, Li S. et al. (2018) reported that increased NLRP3 expression in GC tumors and macrophages negatively correlates with miR-22 expression. The authors also demonstrated that constitutive expression of miR-22 dramatically decreases NLRP3 mRNA expression and IL-1β secretion in macrophages ( Li S. et al., 2018 ). Therefore, the effect of targeting NLRP3 expression with miRNAs in tumors and immune cells may vary depending on tumor and/or cell type. Oral squamous cell carcinoma (OSCC) High NLRP3 expression was found in oral squamous cell carcinoma (OSCC) cells and tissues ( Wang H. et al., 2018 ). A role for NLRP3 supporting OSCC proliferation and growth was demonstrated in several reports. Wang G. et al. (2018) demonstrated a positive correlation between NLRP3 expression and tumor size, lymph node status and IL-1β expression in OSCC tissue specimens and in vivo models of OSCC. Also, the authors showed that, silencing of NLRP3 in OSCC cell lines reduced cell proliferation, migration, and invasion in vitro ( Wang H. et al., 2018 ). Additionally, high expression of the NLRP3 inflammasome mediates chemoresistance in OSCC ( Feng et al., 2018 ). Therefore, downregulation of NLRP3 could have a therapeutic potential in OSCC. Surprisingly, high expression of miR-223 , which targets NLRP3 , was found in primary OSCC tissue ( Manikandan et al., 2016 ). In silico analysis identified a Ras Homolog Family Member B ( RHOB ) as a potential target for miR-223 in OSCC ( Manikandan et al., 2016 ). It appears that miR-223 could indirectly suppress NLRP3 and TLR4/NF-κB signaling via RHOB ( Yan et al., 2019 ). These data provide a novel potential target for OSCC treatment, where miR-223 inhibition of NLRP3 could be attained through RHOB. Overexpression of miR-22 in OSCC was shown to reduce NLRP3 activation and decrease OSCC malignancy ( Feng et al., 2018 ). miR-22 levels were shown to be inversely correlated with NLRP3 expression and miR-22 levels were significantly lower in OSCC compared to adjacent non-cancerous tissue ( Feng et al., 2018 ). The inhibitory effect of miR-22 on OSCC migration was confirmed using a lentiviral expression system. As expected an inhibitor of miR-22 promoted OSCC spread ( Feng et al., 2018 ). The 3′-UTR of the NLRP3 gene was identified as a miR-22 target site ( Feng et al., 2018 ). It appears that NLRP3 promotes OSCC growth and tumor spread, which makes miR-22 a potential therapeutic target for cancer treatment. Two miRNAs, miR-223 and miR-22 , were identified as inhibiting the inflammasome and, subsequently, suppressing tumor growth. Therefore, the anti-tumor effect of these molecules in OSCC warrants further investigation. Cervical cancer (CC) Human papillomavirus (HPV) infection and persistent chronic inflammation were identified as fundamental for the pathogenesis of cervical cancer (CC) ( de Castro-Sobrinho et al., 2016 ; Kriek et al., 2016 ). HPV can cause chronic inflammation by inducing TLR4 expression and impairing the TLR4-NF-κB pathway ( Wang et al., 2014 ; He A. et al., 2016 ). Wu et al. (2012) reported reduced expression of miR-223 , which targets NLRP3 , in the CC cell line HeLa. The authors also demonstrated that over-expression of miR-223 inhibits tumor cell proliferation by targeting FOXO1 ( Wu et al., 2012 ). In addition, another direct post-transcriptional regulator of NLRP3 , miR-22, was found to be down-regulated in CC cell lines and tissues ( Xin et al., 2016 ; Wongjampa et al., 2018 ). Furthermore, Wongjampa et al. (2018) reported an inverse correlation between histone deacetylase 6 (HDAC6) and miR-22 . It was previously shown that HDAC6 directly binds to NLRP3 via its ubiquitin-binding domain to regulate NLRP3 inflammasome expression ( Hwang et al., 2015 ). As NLRP3 plays a role in the pathogenesis of HPV induced chronic inflammation, miR-223 and miR-22 , both of which regulate inflammasome activation, could be potential therapeutic tools for the treatment of CC. Glioblastoma (GBM) High NLRP3 inflammasome activation and high levels of inflammasome products are found in malignant glioblastoma (GBM) ( Basu et al., 2004 ; Tarassishin et al., 2014 ). Increased IL-1β, a major NLRP3 inflammasome product, was linked to the release of VEGF and MMPs, angiogenic factors, in human astrocytes and GBM cells ( Suh et al., 2013 ). Therefore, it could be suggested that inflammasome activation favors GBM growth and spread. Several miRNAs were shown to regulate inflammasome expression, where decreased miRNA levels could promote GBM growth and invasion. Ding Q. et al. (2018) demonstrated that miR-223 , which is effective at reducing NLRP3 inflammasome levels in several tumors ( Wu et al., 2012 ), was decreased in GBM tissues ( Ding Q. et al., 2018 ). However, a conflicting report from Cheng et al. (2017) indicated that miR-223 is overexpressed in GBM cell lines. Similar findings were also reported in GBM stem like cells and GBM tissues ( Huang B.S. et al., 2017 ). Similarly there are conflicting data regarding miR-223 targets and phenotypic impacts. A miR-223-3p mimic inhibited tumor cell proliferation and migration, effects that were due to a reduction in proinflammatory cytokines IL-1β and IL-18 in GBM cell lines ( Ding Q. et al., 2018 ). Also, nuclear factor I-A (NFIA) was a target of miR-223 in GBM cell lines and was found to decrease tumorigenesis in the CNS ( Glasgow et al., 2013 ). The pro-tumorigenic effect of miR-223 was linked to suppression of the tumor suppressor paired box 6 ( PAX6 ) ( Cheng et al., 2017 ). By targeting PAX6 , miR-223 could promote GBM stem cell chemotherapy resistance ( Huang B.S. et al., 2017 ). The mechanism underlying the diverse effects of miR-223 on GBM growth and metastasis remains largely unknown. However, it could be suggested that the stage of tumorigenesis plays a role in the effect of miR-223 in GBM. Levels of miR-22 and miR-30e , two post-transcriptional regulators of NLRP3 , are low in GBM tissues ( Li W.B. et al., 2013 ; Chakrabarti et al., 2016 ; Chen H. et al., 2016 ). In addition to targeting NLRP3 , miR-22 can also directly target the 3′-UTRs of SIRT1 ( Li W.B. et al., 2013 ), and miR-22 mimics decrease the expression of SIRT1 protein in GBM cell lines ( Li W.B. et al., 2013 ). Interestingly, several studies have demonstrated that SIRT1 can suppress NLRP3 ( Ma et al., 2015 ; Jiang et al., 2016 ; Zhou C.C. et al., 2016 ). It could be proposed that the decreased levels of miR-22 could fail to control NLRP3 expression, which could enable GMB tumorigenesis. miRNA in Regulation of Inflammasome in Infections Inflammasome activation is an important component of infectious pathogens surveillance and antimicrobial immune and inflammatory responses. This inflammasome was shown to be activated by several bacterial pathogens including Staphylococcus aureus , Salmonella typhimurium , Listeria monocytogenes , Mycobacterium , Streptococcus pyogenes , Neisseria gonorrhoeae as well as fungi such as Candida albicans and Aspergillus fumigatus ( Franchi et al., 2006 ; Mariathasan et al., 2006 ; Miao et al., 2006 ; Craven et al., 2009 ; Duncan et al., 2009 ; Harder et al., 2009 ; Hise et al., 2009 ; Joly et al., 2009 ; Munoz-Planillo et al., 2009 ; Broz et al., 2010 ; Carlsson et al., 2010 ; McElvania Tekippe et al., 2010 ; Said-Sadier et al., 2010 ). NAIP/NLRC4 inflammasome can protect against Salmonella Typhimurium and C. rodentium invasion by bacteria expulsion from intestinal epithelial cells together with IL-18 and eicosanoid lipid mediators release ( Nordlander et al., 2014 ; Sellin et al., 2014 ; Rauch et al., 2017 ). It appears that NLRP3 activation is essential for establishing the inflammatory milieu in the target tissue and augmenting the phagocytic capacity of the local macrophages ( Master et al., 2008 ; Melehani and Duncan, 2016 ; Cohen et al., 2018 ). Enhanced macrophage bactericidal activity is the most commonly identified mechanism of inflammasome antimicrobial effect ( Master et al., 2008 ; Cohen et al., 2018 ). Additionally, NLRP3 activation induced death of macrophages was described as an effort to prevent microbial propagation and spread ( Miao et al., 2010 ; Sagulenko et al., 2013 ). However, there is a growing body of evidence suggesting that there is a threshold of NLRP3 activity, which acts as a safeguard mechanism to prevent inflammasome over-activation. It appears that aberrant NLRP3 activation could have a detrimental effect on tissues homeostasis and compromise barrier integrity ( Bortolotti et al., 2018 ; McKenzie et al., 2018 ). It is this detrimental effect of the inflammasome over-activation that is often employed by microbes to ensure spread and propagation ( Duncan et al., 2009 ; Harder et al., 2009 ; Carlsson et al., 2010 ). Microbial virulence factors often act as NLRP3 activators. For example, it was shown that the detrimental (to the host) role of Esx1, a membrane lysis factor of Mycobacterium ( Stanley et al., 2003 ), is linked to inflammasome activation ( Carlsson et al., 2010 ). Two virulence factors of group A Streptococcus (GAS), M protein and streptolysin O, were also identified as contributing into NLRP3 activation and IL-1β production ( Harder et al., 2009 ; Valderrama et al., 2017 ). Both virulence factors are commonly detected in association with invasive GAS infections, including necrotizing fasciitis and toxic shock syndrome. Therefore, NLRP3 activation by virulent factors could promote microbe propagation and aid their escape from immune clearance. Restoring the NLRP3 activation threshold could be a novel therapeutic approach for treatment of invasive infections. In this respect, miRNA may be a tool to regain control over NLRP3. It has been shown that miR-223 expression is consistently high in NLRP3 responsive cells, suggesting the high efficacy of this miRNA in prevention of inflammasome over-activation ( Bauernfeind et al., 2012 ). Dorhoi et al. (2013) demonstrated that miR-223 is upregulated in the blood and lung parenchyma of patients diagnosed with tuberculosis. Also, data collected using animal models confirmed the link between deletion of miR-223 and increased susceptibility to Mycobacterium tuberculosum infection ( Dorhoi et al., 2013 ). Similarly, a protective role of miR-223 in Staphylococcus aureus infection was demonstrated by Fang et al. (2016) . Additionally, the effect of targeting TLR4 for NLRP3 regulation in Listeria monocytogenes infection was demonstrated by Schnitger et al. (2011) . The authors identified that, miR-146a can directly inhibit TLR4 receptor expression, which can downregulate inflammasome activity ( Schnitger et al., 2011 ). Many viruses can activate inflammasomes, including Influenza virus, Hepatitis C virus, Herpes simplex virus-1, etc. ( Delaloye et al., 2009 ; Ichinohe et al., 2010 ; Ito et al., 2012 ; Kaushik et al., 2012 ; Negash et al., 2013 ; Triantafilou et al., 2013a , b ; Wu et al., 2013 ; Ermler et al., 2014 ; Chen and Ichinohe, 2015 ). Inflammasome activation appears to be essential for anti-viral protection, serving as viral genome sensors and triggering innate immune response ( Muruve et al., 2008 ; Lupfer et al., 2015 ). The protective role of inflammasomes was shown in influenza virus infection as an increased viral clearance was NLRP3 dependent ( Allen et al., 2009 ). Also, inflammasome activation improved the survival rate in an animal model of influenza ( Ichinohe et al., 2009 ). Thomas et al. (2009) demonstrated that, the innate immune response activation by NLRP3 inflammasomes is essential for animal protection. However, our understanding of the mechanisms of inflammasome antiviral defense remains limited ( Anand et al., 2011 ). Some viruses were shown to post-transcriptionally regulate inflammasome expression to benefit self-replication and propagation ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). For example, miRNA suppression of inflammasomes was shown in Epstein Barr Virus (EBV) infected cells ( Kieff and Rickinson, 2007 ; Rickinson and Kieff, 2007 ). It appears that, EBV can avert NLRP3 inflammasome activation by expressing miRNAs encoded by three BHRF1 -regions and 40 BART -regions of the viral genome ( Albanese et al., 2016 ; Tagawa et al., 2016 ; Farrell, 2018 ). Additionally, two miRNAs encoded by EBV, miR-BART11-5p and miR-BART15 , were identified by Haneklaus et al. (2012) , which could bind to the 3′-UTR of NLRP3 , the same site targeted by miR-223 , and inhibit the inflammasome. It remains to be determined whether these viral miRNA could be used as therapeutic targets. miRNA Regulation of Inflammasome in Autoimmune Diseases Autoimmune diseases are often the result of a dysregulated immune response, characterized by inflammation and organ damage ( Chang, 2013 ; Yang and Chiang, 2015 ). Chronic inflammation is frequently identified as a predisposing factor for an autoimmune reaction ( Yang and Chiang, 2015 ). Multiple mechanisms were suggested to explain prolonged inflammation leading to autoimmunity; where failure to control inflammasome activation was recently identified in some autoimmune conditions ( Yang and Chiang, 2015 ). It has been established that in addition to inflammation, an increased secretion of IL-1β and IL-18, can stimulate proliferation and organ distribution of the effector T cells, which can cause tissue damage ( Oyanguren-Desez et al., 2011 ; Celhar et al., 2012 ). Therefore, targeting the inflammasome could be suggested to restore control over the inflammatory and immune response. Therapeutic potentials of several NLRP3 targeting miRNAs were investigated in autoimmune diseases such as inflammatory bowel diseases (IBDs) ( Neudecker et al., 2017 ), RA ( Xie Z. et al., 2018 ), type 1 diabetes (T1D) ( Yang and Chiang, 2015 ), type 2 diabetes (T2D) ( Yang and Chiang, 2015 ), and systemic lupus erythematosus (SLE) ( Zhu et al., 2012 ). Inflammatory bowel diseases (IBDs) Inflammatory bowel diseases are characterized by chronic inflammation of the intestine and comprise two disorders Crohn's disease and ulcerative colitis. It is believed that the pathogenesis of IBDs is associated with dysregulation of innate and adaptive immune responses, triggered by microbial antigens. This could result in chronic inflammation of the digestive tract and damage to the intestinal mucosa ( Fiocchi, 1998 ). The role of the inflammasome in intestinal inflammation is controversial. Zaki et al. (2010) reported that, NLRP3 induced production of IL-18 in intestinal epithelial cells can be protective, and contributes to epithelium integrity in experimental colitis. In contrast, Seo et al. (2015) have demonstrated the role of inflammasome in exacerbation of an intestinal pathology. The damaging effect of the inflammasome was also confirmed by Shouval et al. (2016) , who identified that IL-1β inhibition improves the course of IBDs. It appears that increased IL-1β levels and tissue damage in IBDs are linked to NLRP3 activation in myeloid leukocytes infiltrating the gut tissue ( Neudecker et al., 2017 ). The role of the inflammasome in IBDs pathogenesis was also confirmed by using a miR-223 deficient animal model of colitis ( Neudecker et al., 2017 ). miR-223 deficient mice develop experimental colitis manifesting with colonic ulceration, inflammatory leukocyte infiltration and tissue injury which resembles closely IBDs ( Neudecker et al., 2017 ). Tissue injury in these mice was linked to an enhanced NLRP3 expression and elevated IL-1β ( Neudecker et al., 2017 ). Treatment of animals with miR-223 mimetics alleviated symptoms of the colitis which coincided with reduced NLRP3 RNA and IL-1β levels ( Neudecker et al., 2017 ). This data presents miR-223 as a novel biomarker and therapeutic target in subsets of IBDs and colitis ( Polytarchou et al., 2015 ). Rheumatoid arthritis (RA) Rheumatoid arthritis is a chronic, systemic inflammatory disease affecting joints as well as skin, eyes, lungs, heart, and blood vessels ( Scott et al., 2010 ). It was suggested that RA pathogenesis is related to activation of the NLRP3/IL-1β axis, where inflammasome activation was linked to worsening symptoms of the disease ( Xie Q. et al., 2018 ). It was shown that activation of NLRP3 leads to an abundant expression of IL-1β ( Guo et al., 2018 ), which can trigger T helper type 17 (Th17) cell differentiations and osteoclasts activation in RA ( Dayer, 2003 ; McInnes and Schett, 2011 ; Zhang et al., 2015b ). Th17 cells play a central role in RA pathogenesis, by maintaining chronic inflammation, recruiting neutrophils and promoting joint degradation ( Cai et al., 2001 ; Shahrara et al., 2009 ; Leipe et al., 2010 ). Recently, an indirect effect of miR-33 on NLRP3 activation was demonstrated in RA ( Xie Q. et al., 2018 ), which could be explained by miRNA controlled dysregulation of mitochondrial function ( Schroder et al., 2010 ; Zhou et al., 2011 ; Miao et al., 2014 ; Ouimet et al., 2015 ). Xie Q. et al. (2018) suggested that miR-33 increases mitochondrial oxygen consumption and accumulation of reactive oxygen species which upregulates expression of NLRP3 and PCA1 in RA. Also, both miR - 33 expression and NLRP3 inflammasome activity were found to be higher in RA monocytes as compared to controls ( Xie Q. et al., 2018 ). These findings indicate that miR - 33 could play an indirect role in pathogenesis of RA through NLRP3 inflammasome activation. Additional studies will provide more insight into the miRNA regulation of NLRP3 in RA and its therapeutic and prognostic implications. Type 1 diabetes (T1D) Type 1 diabetes is caused by autoimmune targeted elimination of pancreatic β cells islet ( Kloppel et al., 1985 ). It was shown that TLRs play an essential role in the pathogenesis of T1D ( Xie Z. et al., 2018 ). Upregulated expression of TLR4 as well as increased activity of the downstream targets was demonstrated in monocytes from T1D ( Devaraj et al., 2008 ). Increased expression of activated TLRs was explained as a reaction to a high levels of circulating ligands in TID ( Devaraj et al., 2009 ). Also, epigenetic regulation was associated with an aberrant TLR signaling and an increased IL-1β expression in T1D ( Grishman et al., 2012 ). Several miRNAs were found altered in pre-TID patients, where levels of nine miRNAs ( miR-146a , miR-561 , and miR-548a-3p , miR-184 , and miR-200a ) were decreased, and two miRNAs ( miR-30c and miR-487a ) were increased ( Grieco et al., 2018 ). Supporting these results was data published by Wang G. et al. (2018) demonstrating lower levels of miR-150 , miR-146a , and miR-424 compared to controls. One of the most consistent findings was the decreased miR-146a levels in T1D. It appears that miR-146a deficiency could play role in T1D exacerbation and increased IL-1β and IL-18 expression ( Bhatt et al., 2016 ). Increased IL-1β levels could indicate inflammasome activation in T1D, although the role of inflammasome in the disease pathogenesis remains largely unknown. Type 2 diabetes (T2D) Circulating autoantibodies to β cells, self-reactive T cells and the glucose-lowering efficacy of some immunomodulatory therapies are suggestive of the autoimmune nature of the T2D ( Itariu and Stulnig, 2014 ). Interestingly, a role for miRNA regulation of gene expression was demonstrated in T2D, where Balasubramanyam et al. (2011) have shown reduced miR-146a which was associated with increased NF-κB , TNF-α and IL-6 mRNA levels. It is the same miRNA, which was found implicated to pathogenesis of T1D ( Xie Z. et al., 2018 ), indicating potential similarities in the pathogenesis of both diseases. Recently in vivo studies demonstrated that miR-146a deficiency could increase expression of M1 and suppress expression of M2 markers in macrophages collected from patients with diabetes ( Bhatt et al., 2016 ). Macrophage polarization occurs in the presence of IFNγ (M1) or IL-4 (M2) ( Nathan et al., 1983 ; Stein et al., 1992 ) and is linked to pro-inflammatory and anti-inflammatory activities, respectively. M1 macrophages were shown to support inflammation by producing pro-inflammatory cytokines, including the inflammasome product IL-1β ( Bhatt et al., 2016 ). Therefore, a link could be suggested between low miR-146a levels and inflammasome activation in M1 cells. More investigation is required to identify the connection between miR-146a and inflammasome activation and the role of this in T2D pathogenesis. Systemic lupus erythematosus (SLE) Systemic lupus erythematosus is an autoimmune disease caused by the loss of immune tolerance to ubiquitous autoantigens ( Tsokos, 2011 ). Inflammation plays essential role in SLE pathogenesis ( Yang et al., 2014 ; Rose and Dorner, 2017 ), where high levels of circulating proinflammatory cytokines are commonly detected ( Yao et al., 2016 ; Mende et al., 2018 ). Inflammasome activation is proposed as one of the mechanisms underlying increased proinflammatory cytokine level in SLE ( Kahlenberg and Kaplan, 2014 ). This assumption is supported by a report where IL-1β deficient mice were found to be resistant to experimental SLE ( Voronov et al., 2006 ). Also, an increased expression of NLRP3 and AC1 have been reported in SLE nephritis biopsies ( Kahlenberg et al., 2011 ). Kahlenberg and Kaplan (2014) have shown that SLE macrophages are highly reactive to innate immune stimuli, often leading to inflammasome activation. Therefore, targeting inflammasome activity could be a novel approach for SLE treatment. The expression of several miRNAs targeting the inflammasome and its products were found differentially expressed in SLE. For example, Wang et al. (2012) have demonstrated high levels of circulating miR-223 , which was shown to inhibit NLRP3 , in SLE. Also, reduced levels of circulating miR-146a , which regulates priming of TLRs, was found in SLE plasma ( Wang et al., 2012 ). Interestingly, expression of miR-23b , which indirectly inhibits IL-1β responses, was shown to be downregulated in inflammatory lesions of SLE patients and animal model ( Zhu et al., 2012 ). More studies are required to determine the role of miRNAs in pathogenesis of SLE and their therapeutic potential. Inflammatory bowel diseases (IBDs) Inflammatory bowel diseases are characterized by chronic inflammation of the intestine and comprise two disorders Crohn's disease and ulcerative colitis. It is believed that the pathogenesis of IBDs is associated with dysregulation of innate and adaptive immune responses, triggered by microbial antigens. This could result in chronic inflammation of the digestive tract and damage to the intestinal mucosa ( Fiocchi, 1998 ). The role of the inflammasome in intestinal inflammation is controversial. Zaki et al. (2010) reported that, NLRP3 induced production of IL-18 in intestinal epithelial cells can be protective, and contributes to epithelium integrity in experimental colitis. In contrast, Seo et al. (2015) have demonstrated the role of inflammasome in exacerbation of an intestinal pathology. The damaging effect of the inflammasome was also confirmed by Shouval et al. (2016) , who identified that IL-1β inhibition improves the course of IBDs. It appears that increased IL-1β levels and tissue damage in IBDs are linked to NLRP3 activation in myeloid leukocytes infiltrating the gut tissue ( Neudecker et al., 2017 ). The role of the inflammasome in IBDs pathogenesis was also confirmed by using a miR-223 deficient animal model of colitis ( Neudecker et al., 2017 ). miR-223 deficient mice develop experimental colitis manifesting with colonic ulceration, inflammatory leukocyte infiltration and tissue injury which resembles closely IBDs ( Neudecker et al., 2017 ). Tissue injury in these mice was linked to an enhanced NLRP3 expression and elevated IL-1β ( Neudecker et al., 2017 ). Treatment of animals with miR-223 mimetics alleviated symptoms of the colitis which coincided with reduced NLRP3 RNA and IL-1β levels ( Neudecker et al., 2017 ). This data presents miR-223 as a novel biomarker and therapeutic target in subsets of IBDs and colitis ( Polytarchou et al., 2015 ). Rheumatoid arthritis (RA) Rheumatoid arthritis is a chronic, systemic inflammatory disease affecting joints as well as skin, eyes, lungs, heart, and blood vessels ( Scott et al., 2010 ). It was suggested that RA pathogenesis is related to activation of the NLRP3/IL-1β axis, where inflammasome activation was linked to worsening symptoms of the disease ( Xie Q. et al., 2018 ). It was shown that activation of NLRP3 leads to an abundant expression of IL-1β ( Guo et al., 2018 ), which can trigger T helper type 17 (Th17) cell differentiations and osteoclasts activation in RA ( Dayer, 2003 ; McInnes and Schett, 2011 ; Zhang et al., 2015b ). Th17 cells play a central role in RA pathogenesis, by maintaining chronic inflammation, recruiting neutrophils and promoting joint degradation ( Cai et al., 2001 ; Shahrara et al., 2009 ; Leipe et al., 2010 ). Recently, an indirect effect of miR-33 on NLRP3 activation was demonstrated in RA ( Xie Q. et al., 2018 ), which could be explained by miRNA controlled dysregulation of mitochondrial function ( Schroder et al., 2010 ; Zhou et al., 2011 ; Miao et al., 2014 ; Ouimet et al., 2015 ). Xie Q. et al. (2018) suggested that miR-33 increases mitochondrial oxygen consumption and accumulation of reactive oxygen species which upregulates expression of NLRP3 and PCA1 in RA. Also, both miR - 33 expression and NLRP3 inflammasome activity were found to be higher in RA monocytes as compared to controls ( Xie Q. et al., 2018 ). These findings indicate that miR - 33 could play an indirect role in pathogenesis of RA through NLRP3 inflammasome activation. Additional studies will provide more insight into the miRNA regulation of NLRP3 in RA and its therapeutic and prognostic implications. Type 1 diabetes (T1D) Type 1 diabetes is caused by autoimmune targeted elimination of pancreatic β cells islet ( Kloppel et al., 1985 ). It was shown that TLRs play an essential role in the pathogenesis of T1D ( Xie Z. et al., 2018 ). Upregulated expression of TLR4 as well as increased activity of the downstream targets was demonstrated in monocytes from T1D ( Devaraj et al., 2008 ). Increased expression of activated TLRs was explained as a reaction to a high levels of circulating ligands in TID ( Devaraj et al., 2009 ). Also, epigenetic regulation was associated with an aberrant TLR signaling and an increased IL-1β expression in T1D ( Grishman et al., 2012 ). Several miRNAs were found altered in pre-TID patients, where levels of nine miRNAs ( miR-146a , miR-561 , and miR-548a-3p , miR-184 , and miR-200a ) were decreased, and two miRNAs ( miR-30c and miR-487a ) were increased ( Grieco et al., 2018 ). Supporting these results was data published by Wang G. et al. (2018) demonstrating lower levels of miR-150 , miR-146a , and miR-424 compared to controls. One of the most consistent findings was the decreased miR-146a levels in T1D. It appears that miR-146a deficiency could play role in T1D exacerbation and increased IL-1β and IL-18 expression ( Bhatt et al., 2016 ). Increased IL-1β levels could indicate inflammasome activation in T1D, although the role of inflammasome in the disease pathogenesis remains largely unknown. Type 2 diabetes (T2D) Circulating autoantibodies to β cells, self-reactive T cells and the glucose-lowering efficacy of some immunomodulatory therapies are suggestive of the autoimmune nature of the T2D ( Itariu and Stulnig, 2014 ). Interestingly, a role for miRNA regulation of gene expression was demonstrated in T2D, where Balasubramanyam et al. (2011) have shown reduced miR-146a which was associated with increased NF-κB , TNF-α and IL-6 mRNA levels. It is the same miRNA, which was found implicated to pathogenesis of T1D ( Xie Z. et al., 2018 ), indicating potential similarities in the pathogenesis of both diseases. Recently in vivo studies demonstrated that miR-146a deficiency could increase expression of M1 and suppress expression of M2 markers in macrophages collected from patients with diabetes ( Bhatt et al., 2016 ). Macrophage polarization occurs in the presence of IFNγ (M1) or IL-4 (M2) ( Nathan et al., 1983 ; Stein et al., 1992 ) and is linked to pro-inflammatory and anti-inflammatory activities, respectively. M1 macrophages were shown to support inflammation by producing pro-inflammatory cytokines, including the inflammasome product IL-1β ( Bhatt et al., 2016 ). Therefore, a link could be suggested between low miR-146a levels and inflammasome activation in M1 cells. More investigation is required to identify the connection between miR-146a and inflammasome activation and the role of this in T2D pathogenesis. Systemic lupus erythematosus (SLE) Systemic lupus erythematosus is an autoimmune disease caused by the loss of immune tolerance to ubiquitous autoantigens ( Tsokos, 2011 ). Inflammation plays essential role in SLE pathogenesis ( Yang et al., 2014 ; Rose and Dorner, 2017 ), where high levels of circulating proinflammatory cytokines are commonly detected ( Yao et al., 2016 ; Mende et al., 2018 ). Inflammasome activation is proposed as one of the mechanisms underlying increased proinflammatory cytokine level in SLE ( Kahlenberg and Kaplan, 2014 ). This assumption is supported by a report where IL-1β deficient mice were found to be resistant to experimental SLE ( Voronov et al., 2006 ). Also, an increased expression of NLRP3 and AC1 have been reported in SLE nephritis biopsies ( Kahlenberg et al., 2011 ). Kahlenberg and Kaplan (2014) have shown that SLE macrophages are highly reactive to innate immune stimuli, often leading to inflammasome activation. Therefore, targeting inflammasome activity could be a novel approach for SLE treatment. The expression of several miRNAs targeting the inflammasome and its products were found differentially expressed in SLE. For example, Wang et al. (2012) have demonstrated high levels of circulating miR-223 , which was shown to inhibit NLRP3 , in SLE. Also, reduced levels of circulating miR-146a , which regulates priming of TLRs, was found in SLE plasma ( Wang et al., 2012 ). Interestingly, expression of miR-23b , which indirectly inhibits IL-1β responses, was shown to be downregulated in inflammatory lesions of SLE patients and animal model ( Zhu et al., 2012 ). More studies are required to determine the role of miRNAs in pathogenesis of SLE and their therapeutic potential. miRNA Regulation of Inflammasome in Neurodegenerative Disorders Inflammasome products, IL-1β and IL-18, were shown to be essential for the health and functional competence of the nervous system ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ). NLRP3 expression was demonstrated in microglia and astrocytes, which could explain the constitutive level of these cytokines in the brain ( McAfoose and Baune, 2009 ; Dinarello et al., 2012 ; Savage et al., 2012 ; Minkiewicz et al., 2013 ; Cho et al., 2014 ; Lu M. et al., 2014 ). Interestingly, higher than normal levels of IL-1β and IL-18 were found in several neurodegenerative disorders, suggesting that over-activation of inflammasomes may play a role in pathogenesis of these diseases ( Cho et al., 2014 ; Lu M. et al., 2014 ; Denes et al., 2015 ; Mamik and Power, 2017 ; Song et al., 2017 ). The significance of miRNA in the regulation of inflammasome activation in the pathogenesis of neurodegenerative diseases remains largely unknown. However, the role of an aberrant miRNA in regulation of NLRP3 expression was previously demonstrated in Parkinson's disease (PD). Parkinson's disease is a neurodegenerative disease which is characterized by progressive loss of dopaminergic neurons in substantia nigra compacta ( Gasser, 2009 ). It is believed that accumulation of α-Syn fibrillary aggregates in the brain, most notably in the nigral dopaminergic neurons, induces the neuroinflammation ( Eriksen et al., 2003 ). According to Zhou Y. et al. (2016) , α-Syn can activate NLRP3 inflammasomes in microglia leading to an increased production of IL-1β. The authors also demonstrated that, miR-7 and miR-30e analogs can inhibit NLRP3 inflammasome mediated neuroinflammation in the brain and protect dopaminergic neurons ( Zhou Y. et al., 2016 ). It appears that the anti-inflammatory effects of miR-7 and miR-30e are associated with their targeting of NLRP3 mRNA in microglial cells. Interestingly, decreased miR-7 and miR-30e expression was demonstrated in PD, which could lead to the loss of the regulatory control of α-Syn induced NLRP3 activation ( Li D. et al., 2018 ). miRNA Regulation of the Inflammasome in Cardiovascular Diseases (CVDs) The physiological significance of inflammation is confirmed as it facilitates elimination of destructive stimuli and pathogens. However, aberrant inflammatory responses could cause tissue damage, tissue fibrosis and chronic diseases ( Liu D. et al., 2018 ). Inflammation is recognized as a major risk factor for CVDs ( Zhou et al., 2018 ), where chronic inflammasome activation was shown to contribute to the pathogenesis of atherosclerosis, ischemic and non-ischemic heart diseases ( Zhou et al., 2018 ). Therefore, regulation of inflammasome activity using miRNA could be used for treatment and prevention of CVDs. Currently, strong evidence for the role of NLRP3 activation has been demonstrated in pathogenesis atherosclerosis. Atherosclerosis is a form of CVD characterized by narrowing of the blood vessel lumen due to plaque formation, continuous dyslipidemia and inflammation ( Ross, 1993 ). Chronic inflammation is commonly found in and around the atherosclerotic plaques which has an adverse effect on the arterial wall structure and function ( Bernhagen et al., 2007 ). It is believed that atherogenic lipid mediators, involved in the formation of chronic inflammation in atherosclerotic plaque ( Chen et al., 2006 ), can trigger peripheral blood monocytes migration and differentiation into macrophages within the intima of the arterial wall ( Chen et al., 2006 ). T cells were also detected in atherosclerotic lesions ( Kleemann et al., 2008 ), where, together with activated macrophages, they were shown to secrete proinflammatory mediators such as interferons, interleukins, and proteases ( Østerud and Bjørklid, 2003 ; Shashkin et al., 2005 ; Tabas, 2005 ; Chen et al., 2006 ). IL-1β expression was identified in the early phase of atherosclerotic plaque formation and this stimulates secretion of additional cytokines and chemokines ( Kleemann et al., 2008 ). Therefore, inflammasome activation in macrophages and T cell within the atherosclerotic lesion contributes to the pathogenesis of chronic inflammation. miR-22 , a miRNA inhibiting NLRP3 , is decreased in peripheral blood mononuclear cells from coronary atherosclerosis ( Chen B. et al., 2016 ), suggesting that upregulation of this miRNA could have therapeutic potential in CVD. Supporting this assumption, Huang W.Q. et al. (2017) investigated the effect of miR-22 on the NLRP3 inflammasome and endothelial cell damage in an in vivo model of coronary heart disease. The authors demonstrated that miR-22 mimics could decrease the release of inflammatory cytokines such as IL-1β and IL-18 by suppressing NLRP3 expression in monocytes and macrophages ( Huang W.Q. et al., 2017 ). Two additional miRNAs, miR-9 and mir-30e-5p were found to indirectly affect inflammasome activation in atherosclerosis ( Wang Y. et al., 2017 ; Li P. et al., 2018 ). It appears that miR-9 could indirectly suppress inflammasome activation by targeting an atherogenic lipid mediator, oxidized low-density lipoprotein (oxLDL), in atherosclerosis ( Liu W. et al., 2014 ). In another report, Wang Y. et al. (2017) reported that miR-9 inhibits NLRP3 inflammasome activation induced by oxLDL in human THP-1 derived macrophages and peripheral blood monocytes in an in vitro atherosclerosis model. miR-9 targets Janus kinase 1 ( JAK1 ) pathway ( Wang Y. et al., 2017 ) inhibiting expression of NF-κB p65 which is required for the first step of NLRP3 inflammasome activation ( Wang Y. et al., 2017 ). In addition, miR-30c-5p was linked to an indirect regulation of NLRP3 expression in atherosclerosis ( Li P. et al., 2018 ). Li P. et al. (2018) reported that miR-30c-5p protects human aortic endothelial cells (HAECs) from the oxLDL insult by targeting FOXO3 . The authors showed that miR-30c-5p can suppress FOXO3 expression and, consequently, decrease levels of NLRP3, AC1, IL-18 and IL-1β in HAECs ( Li P. et al., 2018 ). As evidence emerges supporting the role of NLRP3 in the pathogenesis of atherosclerosis, targeting the inflammasome becomes an attractive therapeutic approach, where miRNAs could be suitable novel tools. miRNA in Regulation of Inflammasome in Cancer The role of the inflammasome in tumorigenesis remains controversial. Some reports indicate that NLRP3 inflammasome activation and IL-18 signaling protect against colorectal cancer ( Karki et al., 2017 ), whereas progression of breast cancer, fibrosarcoma, gastric carcinoma, and lung metastasis were shown to be supported by the inflammasome ( Okamoto et al., 2010 ; Kolb et al., 2014 ). Inflammasome regulation is complex, where multiple factors are implicated, making identification of the key regulatory elements challenging. As the inflammasome involvement in pathogenesis of some malignancies becomes more evident, understanding the regulatory mechanisms could lead to the discovery of novel therapeutic targets for cancer treatment. Hepatocellular carcinoma (HCC) Hepatocellular carcinoma (HCC) is a frequent sequelae of hepatitis B and hepatitis C viral infection ( Perz et al., 2006 ). It is understood that these viruses activate NLRP3 inflammasomes causing hepatocyte pyroptosis, apoptosis and fibrosis ( Kofahi et al., 2016 ). However, HCC tissue analysis failed to detect inflammasome activation; in fact, it was found to be significantly down-regulated when compared to the adjacent normal tissue ( Zhu et al., 2011 ; Wei et al., 2014 ). To explain this inconsistency, Wei et al. (2014) suggested that NLRP3 expression is dynamic changing during the progression of HCC. It appears that NLRP3 expression was increased in liver cells at the early stages of transformation, while inflammasome levels were decreased in malignant cells when compared to adjacent normal tissue ( Wei et al., 2014 ). Interestingly, levels of miR-223 , a negative regulator of NLRP3 , were found to be increased in Hep3B cells derived from HCC ( Wan et al., 2018 ). Increased miR-223 was shown to coincide with tumor growth, suggesting a role in post-transcriptional mechanisms in malignant progression. In addition to NLRP3 , miR-223 was shown to target erythrocyte membrane protein band 4.1 like 3 ( EPB41L3 ) and FOXO1 ( Li and Rana, 2014 ; Kim et al., 2017 ). FOXO1 transcription factor binds to the thioredoxin-interacting protein (TXNIP) and regulates genes involved in cell death as well as the oxidative stress responses ( Kim et al., 2017 ). TXNIP interacts with the NLRP3 inflammasome and activates AC1 in murine β-cells ( Zhou et al., 2010 ). In addition, miR-223 appears to be released systemically, where the level of this miRNA in the plasma was significantly lower in HCC cases ( Giray et al., 2014 ). In addition to miR-223 , decreased circulating miR-30e , which also targets NLRP3 , was found in HCC cases ( Bhattacharya et al., 2016 ). Therefore, it could be suggested that analysis of serum levels of miR-223 and miR-30e could be used for diagnosis of HCC as well as an indicator of the efficacy of anticancer therapeutics. Colorectal cancer (CRC) Data on the role of NLRP3 in colorectal cancer (CRC) pathogenesis is inconsistent, where some evidence suggests a pro-tumorigenic role for the inflammasome, while others identified that the inflammasomes protects against tumor ( Allen et al., 2010 ; Huber et al., 2012 ; Guo et al., 2014 ; Wang et al., 2016 ). Inflammasome expression analysis also demonstrated contradicting results where Wang et al. (2016) reported high NLRP3 in mesenchymal-like colon cancer cells, while Allen et al. (2010) demonstrated decreased inflammasome expression in colitis-associated cancer. Inflammasome contribution to tumorigenesis varies depending on the target cell type in the intestinal tissue ( Lissner and Siegmund, 2011 ). According to Lissner and Siegmund (2011) , inflammasome activation is required to maintain integrity of the epithelium. However, aggravated activation of the inflammasome stimulates intestinal inflammation, which could have a detrimental effect on epithelium permeability and increase its leakage ( Lissner and Siegmund, 2011 ). It was identified that damage to the intestinal epithelium could trigger NLRP3 activation and secretion of IL-18, a proinflammatory cytokine ( Huber et al., 2012 ). Subsequently, it was shown that IL-18 could reduce the expression of IL-22 binding protein (IL-22BP) and increase levels of IL-22 ( Huber et al., 2012 ). Although IL-22 is protective against malignancies, aberrant over expression of IL-22 could trigger gut epithelial cell transformation and CRC development ( Huber et al., 2012 ). Therefore, it is believed that IL-18, a NLRP3 product, has a promoting role in CRC development ( Huber et al., 2012 ). Targeting the inflammasome was suggested as a potential approach for treatment of CRC ( Guo et al., 2014 ). NLRP3 expression was shown to be regulated by multiple miRNAs in various diseases ( Haneklaus et al., 2012 ; Feng et al., 2018 ; Wan et al., 2018 ; Xie Q. et al., 2018 ). However, the role of miRNAs in cancer pathogenesis is not straight forward. There are inconsistent results regarding the expression status of miR-223 , a known regulator of NLRP3 expression, in CRC cell lines and primary tumors. In a clinical study, the expression of miR-223 was found to be significantly higher in stage III/IV patients ( Ding J. et al., 2018 ). However, levels of miR-223 vary significantly in colon tumor derived cell lines ( Ding J. et al., 2018 ). Wu et al. (2012) reported reduced expression of miR-223 in a HCT116, a CRC cell line. In contrast, several research groups demonstrated up-regulation of miR-223 in CRC cell lines and primary tissues ( Wang F. et al., 2017 ; Ju et al., 2018 ; Wei et al., 2018 ). Similar to these results, Ju et al. (2018) demonstrated up-regulation of miR-223 in SW620, a CRC cell line. It was identified that high expression of miR-223 suppresses FoxO3a and enhances cancer cell proliferation ( Ju et al., 2018 ). It appears that the protumorigenic effect of Foxo3a is via NF-κB activation, which is essential for upregulation of the inflammasome linked proinflammatory signaling pathways ( Thompson et al., 2015 ). Unlike miR-223 , data on miR-22 expression status in CRC consistently demonstrates that miR-22 expression is significantly lower in CRC tissues and cell lines ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Also, absence of miR-22 was shown to positively correlate with increased cancer cell proliferation, migration, invasion, and metastasis ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Multiple genes were identified as targets for miR-22 including TIAM1 ( Li B. et al., 2013 ), BTG1 ( Zhang et al., 2015a ), HuR ( Liu Y. et al., 2018 ), and SP-1 ( Xia et al., 2017 ). Among these genes, only SP-1 gene expression was linked to inflammasome regulation ( Hofmann et al., 2015 ). According to Hofmann et al. (2015) , Sp-1 protein could contribute to NLRP3 inflammasome activation in monocytes in chronic recurrent multifocal osteomyelitis. However, the role of Sp-1 in activation of the NLRP3 inflammasome in CRC tumor tissues and monocytes remains largely unknown. Recent finding revealed that, in addition to miR-22 , another negative regulator of NLRP3, miR-30e , is absent in CRC tumors as compared to normal colon tissues ( Laudato et al., 2017 ). However, the role of miR-30e in CRC pathogenesis remains unknown. Gastric cancer (GC) It was shown that NLRP3 inflammasome activation promotes gastric cancer (GC) cells proliferation ( Li S. et al., 2018 ). Over expression of miR-223 supports GC invasion and metastasis in primary GC tumors ( Haneklaus et al., 2012 ). Additionally, Li S. et al. (2018) reported that increased NLRP3 expression in GC tumors and macrophages negatively correlates with miR-22 expression. The authors also demonstrated that constitutive expression of miR-22 dramatically decreases NLRP3 mRNA expression and IL-1β secretion in macrophages ( Li S. et al., 2018 ). Therefore, the effect of targeting NLRP3 expression with miRNAs in tumors and immune cells may vary depending on tumor and/or cell type. Oral squamous cell carcinoma (OSCC) High NLRP3 expression was found in oral squamous cell carcinoma (OSCC) cells and tissues ( Wang H. et al., 2018 ). A role for NLRP3 supporting OSCC proliferation and growth was demonstrated in several reports. Wang G. et al. (2018) demonstrated a positive correlation between NLRP3 expression and tumor size, lymph node status and IL-1β expression in OSCC tissue specimens and in vivo models of OSCC. Also, the authors showed that, silencing of NLRP3 in OSCC cell lines reduced cell proliferation, migration, and invasion in vitro ( Wang H. et al., 2018 ). Additionally, high expression of the NLRP3 inflammasome mediates chemoresistance in OSCC ( Feng et al., 2018 ). Therefore, downregulation of NLRP3 could have a therapeutic potential in OSCC. Surprisingly, high expression of miR-223 , which targets NLRP3 , was found in primary OSCC tissue ( Manikandan et al., 2016 ). In silico analysis identified a Ras Homolog Family Member B ( RHOB ) as a potential target for miR-223 in OSCC ( Manikandan et al., 2016 ). It appears that miR-223 could indirectly suppress NLRP3 and TLR4/NF-κB signaling via RHOB ( Yan et al., 2019 ). These data provide a novel potential target for OSCC treatment, where miR-223 inhibition of NLRP3 could be attained through RHOB. Overexpression of miR-22 in OSCC was shown to reduce NLRP3 activation and decrease OSCC malignancy ( Feng et al., 2018 ). miR-22 levels were shown to be inversely correlated with NLRP3 expression and miR-22 levels were significantly lower in OSCC compared to adjacent non-cancerous tissue ( Feng et al., 2018 ). The inhibitory effect of miR-22 on OSCC migration was confirmed using a lentiviral expression system. As expected an inhibitor of miR-22 promoted OSCC spread ( Feng et al., 2018 ). The 3′-UTR of the NLRP3 gene was identified as a miR-22 target site ( Feng et al., 2018 ). It appears that NLRP3 promotes OSCC growth and tumor spread, which makes miR-22 a potential therapeutic target for cancer treatment. Two miRNAs, miR-223 and miR-22 , were identified as inhibiting the inflammasome and, subsequently, suppressing tumor growth. Therefore, the anti-tumor effect of these molecules in OSCC warrants further investigation. Cervical cancer (CC) Human papillomavirus (HPV) infection and persistent chronic inflammation were identified as fundamental for the pathogenesis of cervical cancer (CC) ( de Castro-Sobrinho et al., 2016 ; Kriek et al., 2016 ). HPV can cause chronic inflammation by inducing TLR4 expression and impairing the TLR4-NF-κB pathway ( Wang et al., 2014 ; He A. et al., 2016 ). Wu et al. (2012) reported reduced expression of miR-223 , which targets NLRP3 , in the CC cell line HeLa. The authors also demonstrated that over-expression of miR-223 inhibits tumor cell proliferation by targeting FOXO1 ( Wu et al., 2012 ). In addition, another direct post-transcriptional regulator of NLRP3 , miR-22, was found to be down-regulated in CC cell lines and tissues ( Xin et al., 2016 ; Wongjampa et al., 2018 ). Furthermore, Wongjampa et al. (2018) reported an inverse correlation between histone deacetylase 6 (HDAC6) and miR-22 . It was previously shown that HDAC6 directly binds to NLRP3 via its ubiquitin-binding domain to regulate NLRP3 inflammasome expression ( Hwang et al., 2015 ). As NLRP3 plays a role in the pathogenesis of HPV induced chronic inflammation, miR-223 and miR-22 , both of which regulate inflammasome activation, could be potential therapeutic tools for the treatment of CC. Glioblastoma (GBM) High NLRP3 inflammasome activation and high levels of inflammasome products are found in malignant glioblastoma (GBM) ( Basu et al., 2004 ; Tarassishin et al., 2014 ). Increased IL-1β, a major NLRP3 inflammasome product, was linked to the release of VEGF and MMPs, angiogenic factors, in human astrocytes and GBM cells ( Suh et al., 2013 ). Therefore, it could be suggested that inflammasome activation favors GBM growth and spread. Several miRNAs were shown to regulate inflammasome expression, where decreased miRNA levels could promote GBM growth and invasion. Ding Q. et al. (2018) demonstrated that miR-223 , which is effective at reducing NLRP3 inflammasome levels in several tumors ( Wu et al., 2012 ), was decreased in GBM tissues ( Ding Q. et al., 2018 ). However, a conflicting report from Cheng et al. (2017) indicated that miR-223 is overexpressed in GBM cell lines. Similar findings were also reported in GBM stem like cells and GBM tissues ( Huang B.S. et al., 2017 ). Similarly there are conflicting data regarding miR-223 targets and phenotypic impacts. A miR-223-3p mimic inhibited tumor cell proliferation and migration, effects that were due to a reduction in proinflammatory cytokines IL-1β and IL-18 in GBM cell lines ( Ding Q. et al., 2018 ). Also, nuclear factor I-A (NFIA) was a target of miR-223 in GBM cell lines and was found to decrease tumorigenesis in the CNS ( Glasgow et al., 2013 ). The pro-tumorigenic effect of miR-223 was linked to suppression of the tumor suppressor paired box 6 ( PAX6 ) ( Cheng et al., 2017 ). By targeting PAX6 , miR-223 could promote GBM stem cell chemotherapy resistance ( Huang B.S. et al., 2017 ). The mechanism underlying the diverse effects of miR-223 on GBM growth and metastasis remains largely unknown. However, it could be suggested that the stage of tumorigenesis plays a role in the effect of miR-223 in GBM. Levels of miR-22 and miR-30e , two post-transcriptional regulators of NLRP3 , are low in GBM tissues ( Li W.B. et al., 2013 ; Chakrabarti et al., 2016 ; Chen H. et al., 2016 ). In addition to targeting NLRP3 , miR-22 can also directly target the 3′-UTRs of SIRT1 ( Li W.B. et al., 2013 ), and miR-22 mimics decrease the expression of SIRT1 protein in GBM cell lines ( Li W.B. et al., 2013 ). Interestingly, several studies have demonstrated that SIRT1 can suppress NLRP3 ( Ma et al., 2015 ; Jiang et al., 2016 ; Zhou C.C. et al., 2016 ). It could be proposed that the decreased levels of miR-22 could fail to control NLRP3 expression, which could enable GMB tumorigenesis. Hepatocellular carcinoma (HCC) Hepatocellular carcinoma (HCC) is a frequent sequelae of hepatitis B and hepatitis C viral infection ( Perz et al., 2006 ). It is understood that these viruses activate NLRP3 inflammasomes causing hepatocyte pyroptosis, apoptosis and fibrosis ( Kofahi et al., 2016 ). However, HCC tissue analysis failed to detect inflammasome activation; in fact, it was found to be significantly down-regulated when compared to the adjacent normal tissue ( Zhu et al., 2011 ; Wei et al., 2014 ). To explain this inconsistency, Wei et al. (2014) suggested that NLRP3 expression is dynamic changing during the progression of HCC. It appears that NLRP3 expression was increased in liver cells at the early stages of transformation, while inflammasome levels were decreased in malignant cells when compared to adjacent normal tissue ( Wei et al., 2014 ). Interestingly, levels of miR-223 , a negative regulator of NLRP3 , were found to be increased in Hep3B cells derived from HCC ( Wan et al., 2018 ). Increased miR-223 was shown to coincide with tumor growth, suggesting a role in post-transcriptional mechanisms in malignant progression. In addition to NLRP3 , miR-223 was shown to target erythrocyte membrane protein band 4.1 like 3 ( EPB41L3 ) and FOXO1 ( Li and Rana, 2014 ; Kim et al., 2017 ). FOXO1 transcription factor binds to the thioredoxin-interacting protein (TXNIP) and regulates genes involved in cell death as well as the oxidative stress responses ( Kim et al., 2017 ). TXNIP interacts with the NLRP3 inflammasome and activates AC1 in murine β-cells ( Zhou et al., 2010 ). In addition, miR-223 appears to be released systemically, where the level of this miRNA in the plasma was significantly lower in HCC cases ( Giray et al., 2014 ). In addition to miR-223 , decreased circulating miR-30e , which also targets NLRP3 , was found in HCC cases ( Bhattacharya et al., 2016 ). Therefore, it could be suggested that analysis of serum levels of miR-223 and miR-30e could be used for diagnosis of HCC as well as an indicator of the efficacy of anticancer therapeutics. Colorectal cancer (CRC) Data on the role of NLRP3 in colorectal cancer (CRC) pathogenesis is inconsistent, where some evidence suggests a pro-tumorigenic role for the inflammasome, while others identified that the inflammasomes protects against tumor ( Allen et al., 2010 ; Huber et al., 2012 ; Guo et al., 2014 ; Wang et al., 2016 ). Inflammasome expression analysis also demonstrated contradicting results where Wang et al. (2016) reported high NLRP3 in mesenchymal-like colon cancer cells, while Allen et al. (2010) demonstrated decreased inflammasome expression in colitis-associated cancer. Inflammasome contribution to tumorigenesis varies depending on the target cell type in the intestinal tissue ( Lissner and Siegmund, 2011 ). According to Lissner and Siegmund (2011) , inflammasome activation is required to maintain integrity of the epithelium. However, aggravated activation of the inflammasome stimulates intestinal inflammation, which could have a detrimental effect on epithelium permeability and increase its leakage ( Lissner and Siegmund, 2011 ). It was identified that damage to the intestinal epithelium could trigger NLRP3 activation and secretion of IL-18, a proinflammatory cytokine ( Huber et al., 2012 ). Subsequently, it was shown that IL-18 could reduce the expression of IL-22 binding protein (IL-22BP) and increase levels of IL-22 ( Huber et al., 2012 ). Although IL-22 is protective against malignancies, aberrant over expression of IL-22 could trigger gut epithelial cell transformation and CRC development ( Huber et al., 2012 ). Therefore, it is believed that IL-18, a NLRP3 product, has a promoting role in CRC development ( Huber et al., 2012 ). Targeting the inflammasome was suggested as a potential approach for treatment of CRC ( Guo et al., 2014 ). NLRP3 expression was shown to be regulated by multiple miRNAs in various diseases ( Haneklaus et al., 2012 ; Feng et al., 2018 ; Wan et al., 2018 ; Xie Q. et al., 2018 ). However, the role of miRNAs in cancer pathogenesis is not straight forward. There are inconsistent results regarding the expression status of miR-223 , a known regulator of NLRP3 expression, in CRC cell lines and primary tumors. In a clinical study, the expression of miR-223 was found to be significantly higher in stage III/IV patients ( Ding J. et al., 2018 ). However, levels of miR-223 vary significantly in colon tumor derived cell lines ( Ding J. et al., 2018 ). Wu et al. (2012) reported reduced expression of miR-223 in a HCT116, a CRC cell line. In contrast, several research groups demonstrated up-regulation of miR-223 in CRC cell lines and primary tissues ( Wang F. et al., 2017 ; Ju et al., 2018 ; Wei et al., 2018 ). Similar to these results, Ju et al. (2018) demonstrated up-regulation of miR-223 in SW620, a CRC cell line. It was identified that high expression of miR-223 suppresses FoxO3a and enhances cancer cell proliferation ( Ju et al., 2018 ). It appears that the protumorigenic effect of Foxo3a is via NF-κB activation, which is essential for upregulation of the inflammasome linked proinflammatory signaling pathways ( Thompson et al., 2015 ). Unlike miR-223 , data on miR-22 expression status in CRC consistently demonstrates that miR-22 expression is significantly lower in CRC tissues and cell lines ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Also, absence of miR-22 was shown to positively correlate with increased cancer cell proliferation, migration, invasion, and metastasis ( Zhang et al., 2012 , 2015a ; Li B. et al., 2013 ; Xia et al., 2017 ; Liu Y. et al., 2018 ). Multiple genes were identified as targets for miR-22 including TIAM1 ( Li B. et al., 2013 ), BTG1 ( Zhang et al., 2015a ), HuR ( Liu Y. et al., 2018 ), and SP-1 ( Xia et al., 2017 ). Among these genes, only SP-1 gene expression was linked to inflammasome regulation ( Hofmann et al., 2015 ). According to Hofmann et al. (2015) , Sp-1 protein could contribute to NLRP3 inflammasome activation in monocytes in chronic recurrent multifocal osteomyelitis. However, the role of Sp-1 in activation of the NLRP3 inflammasome in CRC tumor tissues and monocytes remains largely unknown. Recent finding revealed that, in addition to miR-22 , another negative regulator of NLRP3, miR-30e , is absent in CRC tumors as compared to normal colon tissues ( Laudato et al., 2017 ). However, the role of miR-30e in CRC pathogenesis remains unknown. Gastric cancer (GC) It was shown that NLRP3 inflammasome activation promotes gastric cancer (GC) cells proliferation ( Li S. et al., 2018 ). Over expression of miR-223 supports GC invasion and metastasis in primary GC tumors ( Haneklaus et al., 2012 ). Additionally, Li S. et al. (2018) reported that increased NLRP3 expression in GC tumors and macrophages negatively correlates with miR-22 expression. The authors also demonstrated that constitutive expression of miR-22 dramatically decreases NLRP3 mRNA expression and IL-1β secretion in macrophages ( Li S. et al., 2018 ). Therefore, the effect of targeting NLRP3 expression with miRNAs in tumors and immune cells may vary depending on tumor and/or cell type. Oral squamous cell carcinoma (OSCC) High NLRP3 expression was found in oral squamous cell carcinoma (OSCC) cells and tissues ( Wang H. et al., 2018 ). A role for NLRP3 supporting OSCC proliferation and growth was demonstrated in several reports. Wang G. et al. (2018) demonstrated a positive correlation between NLRP3 expression and tumor size, lymph node status and IL-1β expression in OSCC tissue specimens and in vivo models of OSCC. Also, the authors showed that, silencing of NLRP3 in OSCC cell lines reduced cell proliferation, migration, and invasion in vitro ( Wang H. et al., 2018 ). Additionally, high expression of the NLRP3 inflammasome mediates chemoresistance in OSCC ( Feng et al., 2018 ). Therefore, downregulation of NLRP3 could have a therapeutic potential in OSCC. Surprisingly, high expression of miR-223 , which targets NLRP3 , was found in primary OSCC tissue ( Manikandan et al., 2016 ). In silico analysis identified a Ras Homolog Family Member B ( RHOB ) as a potential target for miR-223 in OSCC ( Manikandan et al., 2016 ). It appears that miR-223 could indirectly suppress NLRP3 and TLR4/NF-κB signaling via RHOB ( Yan et al., 2019 ). These data provide a novel potential target for OSCC treatment, where miR-223 inhibition of NLRP3 could be attained through RHOB. Overexpression of miR-22 in OSCC was shown to reduce NLRP3 activation and decrease OSCC malignancy ( Feng et al., 2018 ). miR-22 levels were shown to be inversely correlated with NLRP3 expression and miR-22 levels were significantly lower in OSCC compared to adjacent non-cancerous tissue ( Feng et al., 2018 ). The inhibitory effect of miR-22 on OSCC migration was confirmed using a lentiviral expression system. As expected an inhibitor of miR-22 promoted OSCC spread ( Feng et al., 2018 ). The 3′-UTR of the NLRP3 gene was identified as a miR-22 target site ( Feng et al., 2018 ). It appears that NLRP3 promotes OSCC growth and tumor spread, which makes miR-22 a potential therapeutic target for cancer treatment. Two miRNAs, miR-223 and miR-22 , were identified as inhibiting the inflammasome and, subsequently, suppressing tumor growth. Therefore, the anti-tumor effect of these molecules in OSCC warrants further investigation. Cervical cancer (CC) Human papillomavirus (HPV) infection and persistent chronic inflammation were identified as fundamental for the pathogenesis of cervical cancer (CC) ( de Castro-Sobrinho et al., 2016 ; Kriek et al., 2016 ). HPV can cause chronic inflammation by inducing TLR4 expression and impairing the TLR4-NF-κB pathway ( Wang et al., 2014 ; He A. et al., 2016 ). Wu et al. (2012) reported reduced expression of miR-223 , which targets NLRP3 , in the CC cell line HeLa. The authors also demonstrated that over-expression of miR-223 inhibits tumor cell proliferation by targeting FOXO1 ( Wu et al., 2012 ). In addition, another direct post-transcriptional regulator of NLRP3 , miR-22, was found to be down-regulated in CC cell lines and tissues ( Xin et al., 2016 ; Wongjampa et al., 2018 ). Furthermore, Wongjampa et al. (2018) reported an inverse correlation between histone deacetylase 6 (HDAC6) and miR-22 . It was previously shown that HDAC6 directly binds to NLRP3 via its ubiquitin-binding domain to regulate NLRP3 inflammasome expression ( Hwang et al., 2015 ). As NLRP3 plays a role in the pathogenesis of HPV induced chronic inflammation, miR-223 and miR-22 , both of which regulate inflammasome activation, could be potential therapeutic tools for the treatment of CC. Glioblastoma (GBM) High NLRP3 inflammasome activation and high levels of inflammasome products are found in malignant glioblastoma (GBM) ( Basu et al., 2004 ; Tarassishin et al., 2014 ). Increased IL-1β, a major NLRP3 inflammasome product, was linked to the release of VEGF and MMPs, angiogenic factors, in human astrocytes and GBM cells ( Suh et al., 2013 ). Therefore, it could be suggested that inflammasome activation favors GBM growth and spread. Several miRNAs were shown to regulate inflammasome expression, where decreased miRNA levels could promote GBM growth and invasion. Ding Q. et al. (2018) demonstrated that miR-223 , which is effective at reducing NLRP3 inflammasome levels in several tumors ( Wu et al., 2012 ), was decreased in GBM tissues ( Ding Q. et al., 2018 ). However, a conflicting report from Cheng et al. (2017) indicated that miR-223 is overexpressed in GBM cell lines. Similar findings were also reported in GBM stem like cells and GBM tissues ( Huang B.S. et al., 2017 ). Similarly there are conflicting data regarding miR-223 targets and phenotypic impacts. A miR-223-3p mimic inhibited tumor cell proliferation and migration, effects that were due to a reduction in proinflammatory cytokines IL-1β and IL-18 in GBM cell lines ( Ding Q. et al., 2018 ). Also, nuclear factor I-A (NFIA) was a target of miR-223 in GBM cell lines and was found to decrease tumorigenesis in the CNS ( Glasgow et al., 2013 ). The pro-tumorigenic effect of miR-223 was linked to suppression of the tumor suppressor paired box 6 ( PAX6 ) ( Cheng et al., 2017 ). By targeting PAX6 , miR-223 could promote GBM stem cell chemotherapy resistance ( Huang B.S. et al., 2017 ). The mechanism underlying the diverse effects of miR-223 on GBM growth and metastasis remains largely unknown. However, it could be suggested that the stage of tumorigenesis plays a role in the effect of miR-223 in GBM. Levels of miR-22 and miR-30e , two post-transcriptional regulators of NLRP3 , are low in GBM tissues ( Li W.B. et al., 2013 ; Chakrabarti et al., 2016 ; Chen H. et al., 2016 ). In addition to targeting NLRP3 , miR-22 can also directly target the 3′-UTRs of SIRT1 ( Li W.B. et al., 2013 ), and miR-22 mimics decrease the expression of SIRT1 protein in GBM cell lines ( Li W.B. et al., 2013 ). Interestingly, several studies have demonstrated that SIRT1 can suppress NLRP3 ( Ma et al., 2015 ; Jiang et al., 2016 ; Zhou C.C. et al., 2016 ). It could be proposed that the decreased levels of miR-22 could fail to control NLRP3 expression, which could enable GMB tumorigenesis. Future Aspects for Clinical Approaches The role of the NLRP3 inflammasome in the pathogenesis of several diseases was demonstrated, including CAPS, autoimmune disorders and cancers ( Aganna et al., 2002 ; Martinon et al., 2006 ; Masters et al., 2009 ; Bauer et al., 2010 ; Wen et al., 2011 ). An increased IL-1β level, commonly found in these diseases, is a strong indicator of NLRP3 inflammasome activation. Also, the body of evidence suggests that IL-1β plays a central role in disease pathogenesis. Therefore, targeting IL-1β, a NLRP3 inflammasome product, appears to be a rational therapeutic approach. The efficacy of anti-IL-1β therapy was demonstrated in CAPS, where both the symptoms and severity of the disease were alleviated using either an IL-1β receptor antagonist or anti-IL-1β antibodies ( Hoffman et al., 2008 ; Dinarello, 2009 ; Lachmann et al., 2009 ). A similar approach targeting IL-1β was successfully applied to treat NLRP3 inflammasome associated autoimmune diseases and cancer ( Larsen et al., 2007 ; Lust et al., 2009 ). These data provide compelling evidence for the NLRP3 inflammasome as a potential therapeutic target for treatment of the diseases associated with an elevated level of IL-1β. In this respect, miRNAs have therapeutic potentials as they could target NLRP3 preventing its expression and, consequently, averting IL-1β production. miRNA based replacement and silencing therapeutic approaches were tested in several preclinical and clinical studies ( Li and Rana, 2014 ). miRNAs and miRNA-targeting oligonucleotides approaches (mimic and/or anti-miR technologies) appear to be more effective when compared to small-molecule drugs due to their ability to effect concurrently multiple gene targets ( Li and Rana, 2014 ). Anti- miR-122 oligonucleotide, Miravirsen, was the first miRNA-based therapeutic used to treat hepatitis c infection ( Lindow and Kauppinen, 2012 ; van der Ree et al., 2016 ). Currently Miravirsen is in a phase II clinical trial ( van der Ree et al., 2016 ). Several phase I clinical trials and pre-clinical studies using miRNA-targeting oligonucleotide technologies targeted to Let-7 , miR-10b , miR-21 , miR-34 , miR-155 , miR-221 , and others , have demonstrated positive results ( Moles, 2017 ). miRNA-targeting oligonucleotides are designed to bind to their targeted miRNA ( Li and Rana, 2014 ). miRNAs generally target more than one gene in the same signaling pathway ( Li Z. et al., 2011 ; Li and Rana, 2014 ). This feature of miRNAs makes them valuable as therapeutic candidates ( Li and Rana, 2014 ). However, there are still multiple obstacles to overcome, including target specificity and the potential toxicity of miRNA-targeting oligonucleotides ( Merhautova et al., 2016 ). First, the limited specificity, anti-miRs generally target nucleotide sequences on miRNAs which can be present on multiple miRNAs within the same family ( Hogan et al., 2014 ). Chemical modifications of anti-miRs have been suggested to improve their specificity ( Hogan et al., 2014 ). Second, when administered without a carrier molecule, their effect may be limited and they can be cleared by the liver and kidney ( Bennett and Swayze, 2010 ). Third, anti-miRs can be sensed and eliminated by receptors of the innate and adaptive immune responses ( Diebold et al., 2004 ; Heil et al., 2004 ). To overcome this limitation, tissue specific antibody coated chemically engineered polymer-based nanoparticles and carrier proteins have been developed to improve the specificity and efficacy of delivery. For example, the therapeutic efficiency of miR-223 was improved by using nanoparticle lipid emulsions as a delivery method, in animal model of colitis ( Neudecker et al., 2017 ). These exciting results demonstrate great potential for miRNA-based treatments of diseases linked to NLRP3 dysfunction. Our understanding of the role of the inflammasome in disease pathogenesis is still limited and is hampering development of the miRNA targeting therapeutics against the inflammasome. However, exciting discoveries in fundamental and preclinical research in recent years have demonstrated great potential for miRNA targeting in the treatment of diseases linked to NLRP3 dysfunction. Author Contributions GT and SK contributed to the conception and design of the study. ZG organized the database. SK wrote the first draft of the manuscript. GT, EM, ZG, AM, AR, and SK wrote sections of the manuscript. All authors contributed to manuscript revision, read and approved the submitted version. Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8463397/

Climatic Factors Influencing the Anthrax Outbreak of 2016 in Siberia, Russia

In 2016, an outbreak of anthrax killing thousands of reindeer and affecting dozens of humans occurred on the Yamal peninsula, Northwest Siberia, after 70 years of epidemiological situation without outbreaks. The trigger of the outbreak has been ascribed to the activation of spores due to permafrost thaw that was accelerated during the summer heat wave. The focus of our study is on the dynamics of local environmental factors in connection with the observed anthrax revival. We show that permafrost was thawing rapidly for already 6 years before the outbreak. During 2011–2016, relatively warm years were followed by cold years with a thick snow cover, preventing freezing of the soil. Furthermore, the spread of anthrax was likely intensified by an extremely dry summer of 2016. Concurrent with the long-term decreasing trend in the regional annual precipitation, the rainfall in July 2016 was less than 10% of its 30-year mean value. We conclude that epidemiological situation of anthrax in the previously contaminated Arctic regions requires monitoring of climatic factors such as warming and precipitation extremes. Supplementary Information The online version contains supplementary material available at 10.1007/s10393-021-01549-5. Supplementary Information The online version contains supplementary material available at 10.1007/s10393-021-01549-5. Introduction Anthrax has been known since ancient times with the first descriptions dating back to Hippocrates, fifth century BC (Schwartz 2009 ), and it is endemic to all the continents except Antarctica (Dragon and Rennie 1995 ; WHO 2008 ; Malkhazova et al. 2019 ; Carlson et al. 2019 ). The disease is caused by the soil bacteria Bacillus anthracis . The bacteria are sensitive to the moisture, acidity and organic content of soils, and their life cycles are influenced by climatic factors, such as ambient temperature and precipitation (Dragon and Rennie 1995 ; WHO 2008 ; Waits et al. 2018 ; Walsh et al. 2018 ; Malkhazova et al. 2019 ; Carlson et al. 2019 ). Therefore, some regions are more affected by anthrax than others (for a compilation map based on outbreaks from 2005 to 2016, see Carlson et al. 2019 ). In the regions endemic for anthrax, high incidences occur during dry and warm periods following intensive precipitation, explaining localities of the major outbreaks in countries with warm climates such as Turkey, Ethiopia, South Africa etc. (Malkhazova et al. 2019 ; Carlson et al. 2019 ). In spite of that, anthrax can survive cold climates as well. The vast geographical range of anthrax and risk of recurrence after years or even decades (Dragon and Rennie 1995 ) is due to high resistivity of spores to unfavorable conditions and their ability to effectively reproduce themselves (Driks 2009 ). In the beginning of the twentieth century, a northern part of Western Siberia was experiencing severe and recurring epizootics of anthrax: more than one million reindeer died (Popova et al. 2016 ). The affected territories include a large area on Yamal peninsula (Yamal district), somewhat smaller areas in the north of Nadym district and in the central Tazovsky district, and many small sites in Priuralsky, Nadym and Pur districts (Popova and Kulichenko 2017 ). Since the 1940s and up to 2007, vaccination of the reindeer population effectively eliminated the disease (Popova et al. 2016 ; Arkhangelskaya 2016 ). In 2007, a decade before the outbreak, the vaccination of the reindeer was halted (Popova et al. 2016 ; Arkhangelskaya 2016 ). During this decade, more than 200,000 samples of soil from 32 of known anthrax-contaminated areas of Yamal-Nenets Autonomous District were analyzed and none showed signs of B. anthracis (Shestakova 2016 ; Selyaninov et al. 2016 ). Vulnerability of bacteria to repetitive freeze–thaw cycles (Malkhazova et al. 2019 ) could lead to an eventual sanitation of the soil, given decades of epidemiological stability. Supporting the sanitation hypothesis, Cherkassky ( 2003 ) investigated 360 soil probes from contaminated areas on Yamal peninsula in 1968 and found that the soil pH was in the range of 3–5, i.e., below the threshold value of 6 (e.g., Van Ness 1971 ), and that the soil was poor in organics (humus content below 3%). Although absent near the soil surface, the anthrax spores could remain intact in the carcasses of dead infected animals buried in permafrost. Its thawing due to a warming climate might revive previously frozen bacteria back to life (Popova et al. 2016 ; Goudarzi 2016 ; Coghlan 2016 ). A recent study found that B. anthracis strains isolated at Yamal were close to those isolated from permafrost in Yakutia (Timofeev et al. 2019 ), supporting the hypothesis about permafrost thawing as a trigger for the outbreak. Here, we considered regional (acting on the scales of ~ 100 km) weather and climate parameters, connected their recent dynamics to the dynamics of permafrost and addressed the existing hypotheses about the anthrax outbreak on Yamal. We started with active layer thickness (ALT) dynamics characterizing permafrost thawing in the sites near the outbreak location. Furthermore, we studied dynamics of the mean annual air temperature (MAAT), as well as snow depth and temperature-based indices, aiming to explain ALT dynamics. To account for the joint effect of warm and preceding cold season on ALT, we used frost numbers and found correlations between those and ALT. Finally, we outlined large-scale phenomena and processes potentially relevant for anthrax outbreaks in cold climates. Methods We used meteorological data sets and data on active layer thickness from several sites in Yamal-Nenets Autonomous District and Komi Republic, Russia, close to the location of the outbreak (Fig. 1 ; Popova et al. 2016 ). The medical geography data set giving some quantitative information on the outbreak can be found in Supplementary Information, Table S1. We started with the hypothesis that permafrost thawing could trigger an outbreak and first considered recent dynamics of permafrost and its link to climatic parameters. Then, we looked into summer precipitation, a parameter that could influence spreading of anthrax. Figure 1 a Location of meteorological stations and Circumpolar Active Layer Monitoring (CALM) in Yamal-Nenets Autonomous District. Black curves separate continuous from discontinuous permafrost areas (continuous permafrost to the north). Red circles show areas of major anthrax epizootics, with 2650 reindeer and 36 human cases registered in Yamal district (near Novy Port) and 1 reindeer case—in Tazovsky district (near Antipayuta). b Median snow depth and mean annual air temperature at sites to the east of Ural mountains. Thin dashed and solid curves represent 25th and 75th quartiles. No increase in snow depth in 2014–2015. c Median snow depth and mean annual air temperature at the sites near the Gulf of Ob. Note an increase in snow depth in 2014–2015. The active layer thickness, which is the maximum annual thaw depth, is the parameter characterizing thawing of permafrost. We estimated the effect of air temperature and snow on the soil surface temperature which determines the state and dynamics of permafrost. Temperature-based indices that prove useful in estimates of ALT are freezing and thawing indices (or degree-day sums) corresponding to the sums of daily mean temperatures during cold and warm seasons, respectively (absolute value in case of cold season, see subsection on freezing and thawing indices below). Cold and warm seasons are defined as the seasons when daily mean temperature is stably below and above 0°C, respectively. Air freezing and thawing indices can be derived from the air temperature, whereas surface freezing and thawing indices are calculated based on the surface temperature. The soil surface temperature is not routinely measured at meteorological stations, so we modelled the surface indices. From MAAT and snow depth, we estimated a freezing n -factor (e.g., Klene et al. 2001 )—the ratio between the freezing indices for surface and air, which characterizes heat transfer between soil surface and air. In winter, the soil surface temperature is higher than the air temperature due to snow insulation, and the corresponding freezing indices are lower. Deep snow corresponds to n -factors close to zero, compared to no-snow conditions with n -factor equal to unity. We calculated the surface freezing index from the air freezing index and a freezing n -factor. The formula can be found in subsection ' n -factors.' After that, we used freezing and thawing indices for air and surface to construct air and surface frost numbers and studied their correlation with ALT. Active Layer Thickness The ALT data used in this study were taken from the open database of Circumpolar Active Layer Monitoring network (CALM). Five measurement sites are located in Yamal-Nenets Autonomous District (Fig. 1 , Table S2): one near Nadym (monitoring site R1, Bobrik et al. 2015 ), and four sites near Vaskiny Dachi (sites R5, R5A, R5B, R5C, Leibman et al. 2015 ). One site, Ayach-Yakha near Vorkuta (site R2, Mazhitova and Kaverin 2007 ), is located in the European part of Russia, on the western side of Ural Mountains (Table S2). Nadym and Vorkuta are both located in the area of discontinuous permafrost, while Vaskiny Dachi is in the area of continuous permafrost. MAAT and Snow Depth In order to estimate the influence of meteorological factors on permafrost dynamics, we used measurements from nine meteorological stations operated by Roshydromet (all connected to World Meteorological Organization network) in Yamal-Nenets Autonomous District (Fig. 1 , Table S3). Time series cover period of 2006–2018. The data were downloaded from the site www.rp5.ru (last access February 7, 2019) and were quality controlled (SI). The variables used for the analysis included the air temperature at two meters height, precipitation and snow depth. Using meteorological data, we calculated the mean annual air temperature (MAAT) and mean snow depth during freezing seasons for different years. The MAAT was calculated as the sum of daily mean temperatures divided by the number of days in a year. The mean snow depth for a cold season was calculated as the sum of snow depths, measured on the days when they exceeded zero, divided by the total number of these days. Freezing and Thawing Indices Freezing and thawing indices (or degree-day sums), I f and I t respectively, were calculated using the following formulas: 1 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ I_{{ ext{f}}} = - \mathop \sum \limits_{{\overline{T} 0}} \overline{T} $$\end{document} I f = - ∑ T ¯ 0 T ¯ where the sum is taken over all days during the freezing/thawing season, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$\overline{T}$$\end{document} T ¯ is the daily mean temperature. The freezing and thawing indices calculated from the air temperature, characterize the annual heat balance indicative of air being cooled or heated during the year. In what follows, we reserve notations I f and I t for the freezing and thawing indices calculated based on air temperature. The mean annual air temperature can be calculated as MAAT = ( I t − I f )/ P , where P is the number of days in the year. Air freezing and thawing indices were calculated directly from air temperature using formula ( 1 ). Surface freezing ( I f surf ) and thawing ( I t surf ) indices can be calculated from Eq. ( 1 ) if surface temperature is used instead of air temperature. However, surface temperature is not measured routinely at meteorological stations. n -Factors The surface freezing and thawing indices can be calculated multiplying the air indices with the corresponding n -factors. Freezing ( n f ) and thawing ( n t ) n -factors, defined as n t(f) = I t(f) surf / I t(f), are the bulk coefficients characterizing heat transfer from air to soil surface on the seasonal time scale and accounting for snow and soil properties in winter and vegetation and soil effects in summer (Klene et al. 2001 ). We used n t = 0.8 as a thawing n -factor based on the measurements from boreholes in tundra near Nadym (Kukkonen et al. 2020 ), see also Jorgensen and Kreig ( 1988 ). The freezing n -factor n f exhibits a greater variability (Kukkonen et al. 2020 ) and needs to be quantified separately for different years. We derived n f as a function of the mean snow depth and MAAT using model calculations by Smith and Riseborough ( 2002 ). Similar method was used in a recent study mapping permafrost boundaries (Obu et al. 2019 ). We calculated time series of freezing n -factors for three meteorological stations: Novy Port, Antipayuta and Vorkuta (Fig. 3 a, S2a, S2b). Note that snow depth was not always measured at the station Novy Port which is the closest to the area of anthrax outbreak. For the years when measurements were not available (Table S3, Fig. S1b), we used the mean snow depth calculated from the data of three closest meteorological stations: Yangi-Yugan, Nyda and Nadym (Fig. 1 ). Frost Number Frost number F (Nelson and Outcalt 1987 ) is a combination of freezing and thawing indices, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$F = rac{{\sqrt {I_{{ ext{f}}} } }}{{\sqrt {I_{{ ext{f}}} } + \sqrt {I_{{ ext{t}}} } }} $$\end{document} F = I f I f + I t , and therefore, accounts for temperature regime during both freezing and thawing seasons. As a reference, the surface frost number of 0.67 was used to mark the boundary between continuous and discontinuous permafrost zones (Nelson and Outcalt 1987 ). Air and surface frost numbers were calculated from air and surface freezing and thawing indices, respectively. The locations of CALM sites do not coincide with locations of meteorological stations; therefore, to calculate frost numbers, we used the data from closest stations (Nadym for Nadym, Antipayuta for Vaskiny Dachi). Summer Precipitation As a factor influencing spread of the disease, we calculated monthly precipitation as sums of daily precipitation in summer and compared their mean values over periods 2005–2013 and 2014–2018 to the climatological normal of monthly precipitation for 1981–2010. The normals were taken from the Climate Assessment Database (CADBv2) provided by NOAA (last access 15 July 2021). Active Layer Thickness The ALT data used in this study were taken from the open database of Circumpolar Active Layer Monitoring network (CALM). Five measurement sites are located in Yamal-Nenets Autonomous District (Fig. 1 , Table S2): one near Nadym (monitoring site R1, Bobrik et al. 2015 ), and four sites near Vaskiny Dachi (sites R5, R5A, R5B, R5C, Leibman et al. 2015 ). One site, Ayach-Yakha near Vorkuta (site R2, Mazhitova and Kaverin 2007 ), is located in the European part of Russia, on the western side of Ural Mountains (Table S2). Nadym and Vorkuta are both located in the area of discontinuous permafrost, while Vaskiny Dachi is in the area of continuous permafrost. MAAT and Snow Depth In order to estimate the influence of meteorological factors on permafrost dynamics, we used measurements from nine meteorological stations operated by Roshydromet (all connected to World Meteorological Organization network) in Yamal-Nenets Autonomous District (Fig. 1 , Table S3). Time series cover period of 2006–2018. The data were downloaded from the site www.rp5.ru (last access February 7, 2019) and were quality controlled (SI). The variables used for the analysis included the air temperature at two meters height, precipitation and snow depth. Using meteorological data, we calculated the mean annual air temperature (MAAT) and mean snow depth during freezing seasons for different years. The MAAT was calculated as the sum of daily mean temperatures divided by the number of days in a year. The mean snow depth for a cold season was calculated as the sum of snow depths, measured on the days when they exceeded zero, divided by the total number of these days. Freezing and Thawing Indices Freezing and thawing indices (or degree-day sums), I f and I t respectively, were calculated using the following formulas: 1 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ I_{{ ext{f}}} = - \mathop \sum \limits_{{\overline{T} 0}} \overline{T} $$\end{document} I f = - ∑ T ¯ 0 T ¯ where the sum is taken over all days during the freezing/thawing season, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$\overline{T}$$\end{document} T ¯ is the daily mean temperature. The freezing and thawing indices calculated from the air temperature, characterize the annual heat balance indicative of air being cooled or heated during the year. In what follows, we reserve notations I f and I t for the freezing and thawing indices calculated based on air temperature. The mean annual air temperature can be calculated as MAAT = ( I t − I f )/ P , where P is the number of days in the year. Air freezing and thawing indices were calculated directly from air temperature using formula ( 1 ). Surface freezing ( I f surf ) and thawing ( I t surf ) indices can be calculated from Eq. ( 1 ) if surface temperature is used instead of air temperature. However, surface temperature is not measured routinely at meteorological stations. n -Factors The surface freezing and thawing indices can be calculated multiplying the air indices with the corresponding n -factors. Freezing ( n f ) and thawing ( n t ) n -factors, defined as n t(f) = I t(f) surf / I t(f), are the bulk coefficients characterizing heat transfer from air to soil surface on the seasonal time scale and accounting for snow and soil properties in winter and vegetation and soil effects in summer (Klene et al. 2001 ). We used n t = 0.8 as a thawing n -factor based on the measurements from boreholes in tundra near Nadym (Kukkonen et al. 2020 ), see also Jorgensen and Kreig ( 1988 ). The freezing n -factor n f exhibits a greater variability (Kukkonen et al. 2020 ) and needs to be quantified separately for different years. We derived n f as a function of the mean snow depth and MAAT using model calculations by Smith and Riseborough ( 2002 ). Similar method was used in a recent study mapping permafrost boundaries (Obu et al. 2019 ). We calculated time series of freezing n -factors for three meteorological stations: Novy Port, Antipayuta and Vorkuta (Fig. 3 a, S2a, S2b). Note that snow depth was not always measured at the station Novy Port which is the closest to the area of anthrax outbreak. For the years when measurements were not available (Table S3, Fig. S1b), we used the mean snow depth calculated from the data of three closest meteorological stations: Yangi-Yugan, Nyda and Nadym (Fig. 1 ). Frost Number Frost number F (Nelson and Outcalt 1987 ) is a combination of freezing and thawing indices, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$F = rac{{\sqrt {I_{{ ext{f}}} } }}{{\sqrt {I_{{ ext{f}}} } + \sqrt {I_{{ ext{t}}} } }} $$\end{document} F = I f I f + I t , and therefore, accounts for temperature regime during both freezing and thawing seasons. As a reference, the surface frost number of 0.67 was used to mark the boundary between continuous and discontinuous permafrost zones (Nelson and Outcalt 1987 ). Air and surface frost numbers were calculated from air and surface freezing and thawing indices, respectively. The locations of CALM sites do not coincide with locations of meteorological stations; therefore, to calculate frost numbers, we used the data from closest stations (Nadym for Nadym, Antipayuta for Vaskiny Dachi). Summer Precipitation As a factor influencing spread of the disease, we calculated monthly precipitation as sums of daily precipitation in summer and compared their mean values over periods 2005–2013 and 2014–2018 to the climatological normal of monthly precipitation for 1981–2010. The normals were taken from the Climate Assessment Database (CADBv2) provided by NOAA (last access 15 July 2021). Results Dynamics of ALT Near Outbreak Location The geographical location of the site of a major reindeer epizootic and human infection cases (Popova et al. 2016 ) is shown in Figure 1 a (see also Supplementary Information, Table S1). All these cases occurred close to the boundary separating zones of continuous and discontinuous permafrost (Kotlyakov and Khromova 2002 ; Obu et al. 2019 ). We analyzed the behavior of ALT from three Circumpolar Active Layer Monitoring (CALM) sites closest (200–400 km) to the outbreak location (Novy Port, Fig. 1 a): one to the north (Vaskiny Dachi, continuous permafrost), one to the south (Nadym, discontinuous permafrost) and one to the west (Vorkuta, discontinuous permafrost). In contrast to the heat wave hypothesis, the rate of thawing was not enhanced in any of the sites in 2016 (Fig. 2 a, Fig. S1c). The maximum values of ALT in Vaskiny Dachi and Nadym followed the trends already started earlier: according to Pettitt's test, the change point is 2011, p value is 0.02. ALT was deepening continuously in Nadym from the minimum in 2010 up to 2016, reaching more than 40% higher level compared to the average value of ALT during 1997–2010. In Vaskiny Dachi, the continuous thawing was interrupted by freezing in 2014, so that the ALT increase was somewhat smaller. The dynamics of the active layer in Vaskiny Dachi show general agreement with the behavior of the mean annual air temperature (MAAT, see Methods) (Fig. 2 b), as expected for the cold sites underlain by continuous permafrost (Smith and Riseborough 2002 ). However, the dynamics of ALT in Nadym cannot be explained by warming alone. Figure 2 a Time series of active layer thickness from Circumpolar Active Layer Monitoring sites. Relative increase in active layer thickness in 2016 as compared to its mean value before 2010 is 43% in the area of discontinuous permafrost (Nadym) and 26% in the area of continuous permafrost (Vaskiny Dachi). Measurements from 2009 are marked as unreliable by the data PIs, therefore dashed line is used between 2008 and 2010. b Mean annual air temperature for all sites. Note similar dynamics of active layer thickness in Vaskiny Dachi and mean annual air temperature. Another important factor for permafrost thaw in the discontinuous permafrost area is the snow depth (Williams and Smith 1989 ; Stieglitz et al. 2003 ). Acting as an insulator, snow prevents heat transfer between the cold air and soil surface, thus suppressing freezing of soil in winter. We identified a regional pattern in the snow depth near the Gulf of Ob (Fig. 1 c), having roughly 50% higher values in 2014–2015 compared with the period 2006–2013. The average snow depth and the maximum snow depth reached 85 cm (Fig. S1b) and 160 cm, respectively, at one of the stations in 2014. The outbreak of anthrax occurred in the area of deep snow, close to the Gulf of Ob (Fig. 1 ). Contrary to this, there was no increase in the snow depth during 2014–2015 in the sites closer to Ural Mountains, its mean value fluctuating around 40 cm (Fig. 1 b). A snow depth should not exceed 50 cm for permafrost occurrence in tundra near Vorkuta (Mazhitova and Kaverin, 2007 ; Shamanova 1970 ). The anthrax outbreak occurred at higher latitudes, near Novy Port, so the corresponding critical snow depth for permafrost occurrence is higher (Smith and Riseborough 2002 ). However, the drastic snow cover increase of 2014–2015 exceeded the critical values for permafrost occurrence even for the lower MAAT of − 6..− 7°C observed at Novy Port. While this short-term increase in the snow depth did not lead immediately to permafrost degradation, such deep snow certainly kept the soil warm during several winters in a row. More heat energy of the following thawing seasons could be expended directly to thaw permafrost, rather than to warm deeply frozen soil after the cold winter. Joint Effect of Snow and Temperature on Permafrost Thawing Freezing indices for surface and air, together with estimated n -factors for Novy Port, are shown in Figure 3 a. The difference between the air and surface freezing indices is remarkable. The air freezing index was characterized by small-amplitude oscillations near the constant mean value, 3400°C day, with an exception of a maximum in 2010. The soil surface freezing index was relatively constant before 2010, after which it was lower by approximately one third in 5 years out of eight, due to either warm winters or thick snow. The air freezing index was equal or higher than its mean value after 2010 during 6 years out of 8, whereas the surface index was lower than its mean value after 2010 in 5 years out of 8. In 2012 and 2016 when the snow depth was close to 40–50 cm (Fig. 1 c, S1b), the n -factor was relatively high but the air index was low (winters were warm). Oppositely, in 2014–2015, the air index was high, but the n -factor was low due to deep snow (Fig. 1 c, S1b). These two factors, warm winters and deep snow, caused persistently lower surface freezing index. These dynamics of the surface freezing index are in qualitative agreement with the ALT dynamics in Nadym. A rapid deepening of the active layer was observed in 2014 (Fig. 2 ), characterized by the largest snow depth in this region. Thawing indices (Fig. 3 b) also increased by ca. 15% as compared to the mean value before 2010, indicating warmer summers. Figure 3 a Air freezing index (light blue bars), surface freezing index (cyan bars) and freezing n -factor (blue curve) at Novy Port. Straight dashed lines indicate mean surface freezing index in 2006–2009 and decreased index in 2012, 2014–2016. b Air thawing index (light red bars) and surface thawing index (red bars). In Antipayuta, the most northern meteorological station considered here and closest to Vaskiny Dachi, the soil surface freezing index after 2010 was characterized by strong oscillations with large amplitudes (Fig. S2a). However, the lowest surface freezing indices were presumably associated with warm winters rather than with deep snow. Strong MAAT oscillations lead to destabilization of permafrost state, especially if warmer winters are followed by warm summers. Oppositely, for Vorkuta, we found that the soil surface freezing index oscillated near the constant mean value, 730 °C day, during the whole period of observations (Fig. S2b). Note the high freezing n -factor in 2016 due to the thin snow cover (Fig. S1b), causing a deep freezing of soil in winter and moderating the effect of the heat wave on permafrost. However, ALT in Vorkuta is generally steady (Fig. S1c), suggesting specific soil properties, most probably characterized by an ice-enriched upper permafrost layer (Mazhitova and Kaverin 2007 ). In this case, the 'zero curtain' (phase transition) periods can be long, leading to a lower sensitivity of permafrost to ambient conditions. Another explanation for a steady ALT in Vorkuta is soil subsidence due to thermokarst (Mazhitova and Kaverin 2007 ). The combination of freezing and thawing indices, a frost number (Methods), was used for linking meteorological parameters influencing permafrost dynamics and ALT. From Figure 4 , ALT in Nadym was significantly and strongly correlated with the surface frost number (Pearson R =  − 0.73, p = 0.005) rather than the air frost number, indicating that snow was an important factor. Oppositely, ALT in Vaskiny Dachi was significantly and moderately correlated with the air frost number (Pearson R =  − 0.65, p = 0.017), suggesting air temperature as the major driving factor of permafrost dynamics (see also Fig. 2 ). The latter could be a consequence of the snow distribution characteristic of Central Yamal (Vaskiny Dachi), which is highly uneven due to low tundra vegetation and strong winds (Leibman et al. 2015 ). The region of outbreak was located between Vaskiny Dachi and Nadym. The dynamics of permafrost there could carry features of either site, but the most effective thaw could be naturally expected at the sites accumulating snow. The increase in ALT between 2010 and 2016 due to the temperature effect alone was 26% (estimate for Vaskiny Dachi), but deep snow enhanced this effect resulting in 43% increase in ALT (estimate for Nadym). Additional characterization of permafrost thaw using temperature on the top of permafrost for three sites is given in SI (Sec. S2, Fig. S5). Figure 4 Correlations of active layer thickness with air and surface frost numbers at Nadym (R1) ( a , b ) and Vaskiny Dachi (R5) ( c , d ). Summer Precipitation Oppositely to snow, the summer precipitation in Novy Port decreased during the latest years (Fig. S3). A decreasing trend in precipitation was previously observed at the station Mys Kamenny (Frey and Smith 2003 ), 70 km to the north from Novy Port. Note that summer precipitation at the three stations within the distance of only 200 km from each other (Nadym, Nyda and Novy Port, Fig. 1 ) showed different dynamics during the recent years (Fig. S3). Before 2005, precipitation dynamics in Nyda and Novy Port were similar, suggesting the same mechanisms governing precipitation, and at both sites precipitation decreased. However, after 2005, precipitation patterns were completely different with a rise in Nyda compared to a substantial decline in Novy Port. One explanation for this difference could be a change in vegetation in response to warming near Nyda where a greening trend was reported during 2000–2014 (Miles and Esau 2016 ). In Novy Port, monthly precipitation in July 2016 was only 5% of the climatological normal for 1981–2010 (Fig. 5 ), and in Antipayuta, it was below 20% (Fig. S4). When drought occurred, plants extracted water from deeper soils which could bring anthrax spores to the surface (Hugh-Jones and Blackburn 2009 ). In addition, the lack of precipitation increased the probability of reindeer infection through chewing dry grass and contributed to the high transmissivity of the disease by blood-sucking insects such as tabanids, known to be more active in warm and dry weather (Gainer 2016 ). Figure 5 a Month-to-month variability of precipitation and NOAA climatological normal of precipitation for 1981–2010 in Novy Port. Note that for 2005–2012, data quality is not sufficient (i.e., up to 50% of measurements per month could be missing), nevertheless the monthly precipitation values in summer exceed those from 2014 to 2018. Data quality for 2014–2018 is acceptable (i.e., less than 15% of data for each month was missing, Table S4). Note decrease in summer precipitation during years 2014–2018 in comparison with climatological normal (see also Fig. S3). Dynamics of precipitation is in accordance with the Second Roshydromet Assessment Report on Climate Change in the Russian Federation, which identifies this region as one with the strongest decreasing trends in annual precipitation (100 mm during 1936–2010). b Monthly precipitation in 2016–2018. Dynamics of ALT Near Outbreak Location The geographical location of the site of a major reindeer epizootic and human infection cases (Popova et al. 2016 ) is shown in Figure 1 a (see also Supplementary Information, Table S1). All these cases occurred close to the boundary separating zones of continuous and discontinuous permafrost (Kotlyakov and Khromova 2002 ; Obu et al. 2019 ). We analyzed the behavior of ALT from three Circumpolar Active Layer Monitoring (CALM) sites closest (200–400 km) to the outbreak location (Novy Port, Fig. 1 a): one to the north (Vaskiny Dachi, continuous permafrost), one to the south (Nadym, discontinuous permafrost) and one to the west (Vorkuta, discontinuous permafrost). In contrast to the heat wave hypothesis, the rate of thawing was not enhanced in any of the sites in 2016 (Fig. 2 a, Fig. S1c). The maximum values of ALT in Vaskiny Dachi and Nadym followed the trends already started earlier: according to Pettitt's test, the change point is 2011, p value is 0.02. ALT was deepening continuously in Nadym from the minimum in 2010 up to 2016, reaching more than 40% higher level compared to the average value of ALT during 1997–2010. In Vaskiny Dachi, the continuous thawing was interrupted by freezing in 2014, so that the ALT increase was somewhat smaller. The dynamics of the active layer in Vaskiny Dachi show general agreement with the behavior of the mean annual air temperature (MAAT, see Methods) (Fig. 2 b), as expected for the cold sites underlain by continuous permafrost (Smith and Riseborough 2002 ). However, the dynamics of ALT in Nadym cannot be explained by warming alone. Figure 2 a Time series of active layer thickness from Circumpolar Active Layer Monitoring sites. Relative increase in active layer thickness in 2016 as compared to its mean value before 2010 is 43% in the area of discontinuous permafrost (Nadym) and 26% in the area of continuous permafrost (Vaskiny Dachi). Measurements from 2009 are marked as unreliable by the data PIs, therefore dashed line is used between 2008 and 2010. b Mean annual air temperature for all sites. Note similar dynamics of active layer thickness in Vaskiny Dachi and mean annual air temperature. Another important factor for permafrost thaw in the discontinuous permafrost area is the snow depth (Williams and Smith 1989 ; Stieglitz et al. 2003 ). Acting as an insulator, snow prevents heat transfer between the cold air and soil surface, thus suppressing freezing of soil in winter. We identified a regional pattern in the snow depth near the Gulf of Ob (Fig. 1 c), having roughly 50% higher values in 2014–2015 compared with the period 2006–2013. The average snow depth and the maximum snow depth reached 85 cm (Fig. S1b) and 160 cm, respectively, at one of the stations in 2014. The outbreak of anthrax occurred in the area of deep snow, close to the Gulf of Ob (Fig. 1 ). Contrary to this, there was no increase in the snow depth during 2014–2015 in the sites closer to Ural Mountains, its mean value fluctuating around 40 cm (Fig. 1 b). A snow depth should not exceed 50 cm for permafrost occurrence in tundra near Vorkuta (Mazhitova and Kaverin, 2007 ; Shamanova 1970 ). The anthrax outbreak occurred at higher latitudes, near Novy Port, so the corresponding critical snow depth for permafrost occurrence is higher (Smith and Riseborough 2002 ). However, the drastic snow cover increase of 2014–2015 exceeded the critical values for permafrost occurrence even for the lower MAAT of − 6..− 7°C observed at Novy Port. While this short-term increase in the snow depth did not lead immediately to permafrost degradation, such deep snow certainly kept the soil warm during several winters in a row. More heat energy of the following thawing seasons could be expended directly to thaw permafrost, rather than to warm deeply frozen soil after the cold winter. Joint Effect of Snow and Temperature on Permafrost Thawing Freezing indices for surface and air, together with estimated n -factors for Novy Port, are shown in Figure 3 a. The difference between the air and surface freezing indices is remarkable. The air freezing index was characterized by small-amplitude oscillations near the constant mean value, 3400°C day, with an exception of a maximum in 2010. The soil surface freezing index was relatively constant before 2010, after which it was lower by approximately one third in 5 years out of eight, due to either warm winters or thick snow. The air freezing index was equal or higher than its mean value after 2010 during 6 years out of 8, whereas the surface index was lower than its mean value after 2010 in 5 years out of 8. In 2012 and 2016 when the snow depth was close to 40–50 cm (Fig. 1 c, S1b), the n -factor was relatively high but the air index was low (winters were warm). Oppositely, in 2014–2015, the air index was high, but the n -factor was low due to deep snow (Fig. 1 c, S1b). These two factors, warm winters and deep snow, caused persistently lower surface freezing index. These dynamics of the surface freezing index are in qualitative agreement with the ALT dynamics in Nadym. A rapid deepening of the active layer was observed in 2014 (Fig. 2 ), characterized by the largest snow depth in this region. Thawing indices (Fig. 3 b) also increased by ca. 15% as compared to the mean value before 2010, indicating warmer summers. Figure 3 a Air freezing index (light blue bars), surface freezing index (cyan bars) and freezing n -factor (blue curve) at Novy Port. Straight dashed lines indicate mean surface freezing index in 2006–2009 and decreased index in 2012, 2014–2016. b Air thawing index (light red bars) and surface thawing index (red bars). In Antipayuta, the most northern meteorological station considered here and closest to Vaskiny Dachi, the soil surface freezing index after 2010 was characterized by strong oscillations with large amplitudes (Fig. S2a). However, the lowest surface freezing indices were presumably associated with warm winters rather than with deep snow. Strong MAAT oscillations lead to destabilization of permafrost state, especially if warmer winters are followed by warm summers. Oppositely, for Vorkuta, we found that the soil surface freezing index oscillated near the constant mean value, 730 °C day, during the whole period of observations (Fig. S2b). Note the high freezing n -factor in 2016 due to the thin snow cover (Fig. S1b), causing a deep freezing of soil in winter and moderating the effect of the heat wave on permafrost. However, ALT in Vorkuta is generally steady (Fig. S1c), suggesting specific soil properties, most probably characterized by an ice-enriched upper permafrost layer (Mazhitova and Kaverin 2007 ). In this case, the 'zero curtain' (phase transition) periods can be long, leading to a lower sensitivity of permafrost to ambient conditions. Another explanation for a steady ALT in Vorkuta is soil subsidence due to thermokarst (Mazhitova and Kaverin 2007 ). The combination of freezing and thawing indices, a frost number (Methods), was used for linking meteorological parameters influencing permafrost dynamics and ALT. From Figure 4 , ALT in Nadym was significantly and strongly correlated with the surface frost number (Pearson R =  − 0.73, p = 0.005) rather than the air frost number, indicating that snow was an important factor. Oppositely, ALT in Vaskiny Dachi was significantly and moderately correlated with the air frost number (Pearson R =  − 0.65, p = 0.017), suggesting air temperature as the major driving factor of permafrost dynamics (see also Fig. 2 ). The latter could be a consequence of the snow distribution characteristic of Central Yamal (Vaskiny Dachi), which is highly uneven due to low tundra vegetation and strong winds (Leibman et al. 2015 ). The region of outbreak was located between Vaskiny Dachi and Nadym. The dynamics of permafrost there could carry features of either site, but the most effective thaw could be naturally expected at the sites accumulating snow. The increase in ALT between 2010 and 2016 due to the temperature effect alone was 26% (estimate for Vaskiny Dachi), but deep snow enhanced this effect resulting in 43% increase in ALT (estimate for Nadym). Additional characterization of permafrost thaw using temperature on the top of permafrost for three sites is given in SI (Sec. S2, Fig. S5). Figure 4 Correlations of active layer thickness with air and surface frost numbers at Nadym (R1) ( a , b ) and Vaskiny Dachi (R5) ( c , d ). Summer Precipitation Oppositely to snow, the summer precipitation in Novy Port decreased during the latest years (Fig. S3). A decreasing trend in precipitation was previously observed at the station Mys Kamenny (Frey and Smith 2003 ), 70 km to the north from Novy Port. Note that summer precipitation at the three stations within the distance of only 200 km from each other (Nadym, Nyda and Novy Port, Fig. 1 ) showed different dynamics during the recent years (Fig. S3). Before 2005, precipitation dynamics in Nyda and Novy Port were similar, suggesting the same mechanisms governing precipitation, and at both sites precipitation decreased. However, after 2005, precipitation patterns were completely different with a rise in Nyda compared to a substantial decline in Novy Port. One explanation for this difference could be a change in vegetation in response to warming near Nyda where a greening trend was reported during 2000–2014 (Miles and Esau 2016 ). In Novy Port, monthly precipitation in July 2016 was only 5% of the climatological normal for 1981–2010 (Fig. 5 ), and in Antipayuta, it was below 20% (Fig. S4). When drought occurred, plants extracted water from deeper soils which could bring anthrax spores to the surface (Hugh-Jones and Blackburn 2009 ). In addition, the lack of precipitation increased the probability of reindeer infection through chewing dry grass and contributed to the high transmissivity of the disease by blood-sucking insects such as tabanids, known to be more active in warm and dry weather (Gainer 2016 ). Figure 5 a Month-to-month variability of precipitation and NOAA climatological normal of precipitation for 1981–2010 in Novy Port. Note that for 2005–2012, data quality is not sufficient (i.e., up to 50% of measurements per month could be missing), nevertheless the monthly precipitation values in summer exceed those from 2014 to 2018. Data quality for 2014–2018 is acceptable (i.e., less than 15% of data for each month was missing, Table S4). Note decrease in summer precipitation during years 2014–2018 in comparison with climatological normal (see also Fig. S3). Dynamics of precipitation is in accordance with the Second Roshydromet Assessment Report on Climate Change in the Russian Federation, which identifies this region as one with the strongest decreasing trends in annual precipitation (100 mm during 1936–2010). b Monthly precipitation in 2016–2018. Discussion and Conclusion How Climatic Factors Contribute to the Anthrax Outbreak: An Outline Overall, the schematic of how different factors could contribute to anthrax outbreak in cold climates is shown in Figure 6 . For a fixed population of livestock (reindeer), the availability of spores in soil (previous contamination) is of utmost importance, whereas timing, dispersion and incidence rates are connected to climatic factors. Some factors drive permafrost thaw, thereby acting as a trigger of outbreak (warming, deep snow), while others contribute to the spread of disease (lack of summer precipitation). Figure 6 Outline of the connections between climatic factors and anthrax outbreak in the Arctic. Arctic amplification, Arctic oscillation and sea ice retreat determine temperature dynamics in the Arctic on annual scale. Sea ice retreat introduces an increasing trend into the winter precipitation dynamics, whereas local weather patterns and extreme events contribute to its variability. Summer precipitation is determined by the local underlying surface properties, evapotranspiration and convective patterns. Warming climate and winter precipitation dynamics influence active layer deepening which can trigger anthrax outbreak via revival of old bacteria. Dry summer boosts spread of the disease and intensifies the outbreak. Vaccination is a preventive measure to control the spread of disease. The increasing trend in the mean annual temperature (Frey and Smith 2003 ) reflects enhanced warming of Arctic as compared to lower latitudes (Arctic amplification, Box et al. 2019 ). Fluctuations in the mean annual temperature, showing a similar behavior at all the stations (Fig. S1), are largely driven by the dynamics of the Arctic oscillation (Frey and Smith 2003 ) which determines winter-time synoptic activity in the region. Arctic oscillation is subject to a random variability and has been mainly in its positive phase since 1990s (NOAA, Climate Indicators), which has led to warmer winters in northern Eurasia, including West Siberia. Another influencing factor is precipitation (Waits et al. 2018 ). There has been a general smooth increase in winter precipitation related to sea ice retreat (Cohen et al. 2014 ). However, the rapid permafrost thaw described here happened due to an anomalously deep snow, pointing at extreme weather contribution, also mentioned by Hedlund et al. ( 2014 ) as the factor increasing risk of infectious diseases in Arctic and subarctic regions. In agreement with the previous study showing a high spatial variability in annual precipitation (Frey and Smith 2003 ), we found a high variability in the snow depth at different meteorological stations over the spatial scale of ca 300 km (Fig. 1 ). Summer precipitation is dependent on local circulation patterns and properties of the underlying surface. As compared to winter precipitation, it varies on the scale of only 100 km (Fig. 5 , S3). Finally, a major anthropogenic factor influencing the probability of outbreak is vaccination. Vaccination of reindeer had long been a successful preventive measure in this region (Popova et al. 2016 ; Kolonin 1969 ). Its role is to prevent occasional infection and to stop a spreading of the disease (WHO 2008 ). Results in the Context of Literature and Implications for Future Research Climate change introduces a risk of the global anthrax outbreak, both in lower and higher latitudes (Kangbai and Momoh 2017 ). A recent study mapping global distribution of B. Anthracis (Carlson et al. 2019 ) admits lack of the data on the outbreaks in northern latitudes. Here, we provide the summary of the data characterizing the anthrax outbreak of 2016 in Siberia, and we present the analysis of climatic factors leading to the outbreak. Notably, it was suggested (Carlson et al. 2019 ) that a set of climatic factors causing outbreaks in the cold climates could differ from that in the warm and dry ones. Based on the present case study, we identified winter precipitation as an additional factor, which has not been considered before. However, we admit lack of replicate examples for anthrax outbreaks in the cold regions, which is among the limitations of this study. Another previous study considering risk factors for outbreak in high latitudes (Hueffer et al. 2020 ) admitted the importance of vaccination and reindeer number but provided only a limited analysis of climatic factors, focusing on summer temperatures in Salekhard. We performed an extensive analysis of recent temperature and precipitation and linked it to the dynamics of active layer thickness characterizing permafrost thawing. The major hypothesis about the trigger of outbreak is related to thawing permafrost. There are also questions on how the outbreak has become so widespread. Given that ALT was increasing since 2010, bacteria could be released from permafrost even earlier than in 2016. The ability of bacteria to undergo the whole life cycle in the soil is a subject of debate (Hugh-Jones and Blackburn 2009 ). Recently, earthworms, plants and amoebae have been shown to interact with B. anthracis, demonstrating a possibility of bacteria's life cycle outside the host, although in laboratory conditions (Carlson et al. 2018 ). The contaminated areas in Siberia were considered to be prone to sanitation (Popova and Kulichenko 2017 ) due to unfavorable soil environment and weather conditions that could abrupt the life cycle of bacteria during an eventual vegetative stage. However, the situation has changed. Since 1968 when Cherkassky examined soil probes (Cherkassky 2003 ), the growing season (the period with mean daily temperature above 5°C) in Novy Port has lengthened by almost a month and the mean temperature of the growing season has increased by 1°C (Sizov et al. 2021 ). The climate has become less harsh. In addition, the current vegetation trends show active greening in Yamal district (Miles and Ezau 2016 ; Sizov et al. 2021 ), which could enrich soil with organics. Milder cooling of soils during two years prior to the outbreak, together with warmer and longer summers, could create conditions for bacteria to complete their life cycle by increasing the amount of spores if those were released from permafrost earlier than in 2016. Our analysis points at the importance of the local climatic factors for the outbreak. A combination of climate factors acting for several years in a row caused a strong regional effect over a spatial scale of only 100 km. These scales represent a challenge for the studies based on global models (Walsh et al. 2018 ; Carlson et al. 2019 ) hindering the use of large-scale models for prognostic purposes. Regional models with higher resolution can therefore be recommended for the monitoring and forecasts of weather conditions causing unfavorable epidemiological situations in cold climates. In addition, mathematical models can be applied on a local scale (Friedman and Yakubu 2013 ; Saad-Roy et al. 2017 ; Gomez et al. 2018 ; Stella et al. 2020 ). Remarkably, long-term climate dynamics indicated risks long before the outbreak. In Northwest Siberia, the mean annual air temperature and snow depth increased by 0.4–0.6°C and 4–10 cm per decade, respectively, during 50–60 years before 2012, while annual precipitation decreased by 50–100 mm in 75 years (1936–2010) (Katsov et al. 2014 ). Thus, in the absence of vaccination, the dramatic consequences could likely be a question of time. The risk of anthrax outbreaks associated with climate change was pointed out for East Siberia by Revich and Podolnaya ( 2011 ). A proper information campaign on the importance of vaccination and prognosis on unfavorable meteorological and climatological conditions is essential to prevent outbreaks in future. How Climatic Factors Contribute to the Anthrax Outbreak: An Outline Overall, the schematic of how different factors could contribute to anthrax outbreak in cold climates is shown in Figure 6 . For a fixed population of livestock (reindeer), the availability of spores in soil (previous contamination) is of utmost importance, whereas timing, dispersion and incidence rates are connected to climatic factors. Some factors drive permafrost thaw, thereby acting as a trigger of outbreak (warming, deep snow), while others contribute to the spread of disease (lack of summer precipitation). Figure 6 Outline of the connections between climatic factors and anthrax outbreak in the Arctic. Arctic amplification, Arctic oscillation and sea ice retreat determine temperature dynamics in the Arctic on annual scale. Sea ice retreat introduces an increasing trend into the winter precipitation dynamics, whereas local weather patterns and extreme events contribute to its variability. Summer precipitation is determined by the local underlying surface properties, evapotranspiration and convective patterns. Warming climate and winter precipitation dynamics influence active layer deepening which can trigger anthrax outbreak via revival of old bacteria. Dry summer boosts spread of the disease and intensifies the outbreak. Vaccination is a preventive measure to control the spread of disease. The increasing trend in the mean annual temperature (Frey and Smith 2003 ) reflects enhanced warming of Arctic as compared to lower latitudes (Arctic amplification, Box et al. 2019 ). Fluctuations in the mean annual temperature, showing a similar behavior at all the stations (Fig. S1), are largely driven by the dynamics of the Arctic oscillation (Frey and Smith 2003 ) which determines winter-time synoptic activity in the region. Arctic oscillation is subject to a random variability and has been mainly in its positive phase since 1990s (NOAA, Climate Indicators), which has led to warmer winters in northern Eurasia, including West Siberia. Another influencing factor is precipitation (Waits et al. 2018 ). There has been a general smooth increase in winter precipitation related to sea ice retreat (Cohen et al. 2014 ). However, the rapid permafrost thaw described here happened due to an anomalously deep snow, pointing at extreme weather contribution, also mentioned by Hedlund et al. ( 2014 ) as the factor increasing risk of infectious diseases in Arctic and subarctic regions. In agreement with the previous study showing a high spatial variability in annual precipitation (Frey and Smith 2003 ), we found a high variability in the snow depth at different meteorological stations over the spatial scale of ca 300 km (Fig. 1 ). Summer precipitation is dependent on local circulation patterns and properties of the underlying surface. As compared to winter precipitation, it varies on the scale of only 100 km (Fig. 5 , S3). Finally, a major anthropogenic factor influencing the probability of outbreak is vaccination. Vaccination of reindeer had long been a successful preventive measure in this region (Popova et al. 2016 ; Kolonin 1969 ). Its role is to prevent occasional infection and to stop a spreading of the disease (WHO 2008 ). Results in the Context of Literature and Implications for Future Research Climate change introduces a risk of the global anthrax outbreak, both in lower and higher latitudes (Kangbai and Momoh 2017 ). A recent study mapping global distribution of B. Anthracis (Carlson et al. 2019 ) admits lack of the data on the outbreaks in northern latitudes. Here, we provide the summary of the data characterizing the anthrax outbreak of 2016 in Siberia, and we present the analysis of climatic factors leading to the outbreak. Notably, it was suggested (Carlson et al. 2019 ) that a set of climatic factors causing outbreaks in the cold climates could differ from that in the warm and dry ones. Based on the present case study, we identified winter precipitation as an additional factor, which has not been considered before. However, we admit lack of replicate examples for anthrax outbreaks in the cold regions, which is among the limitations of this study. Another previous study considering risk factors for outbreak in high latitudes (Hueffer et al. 2020 ) admitted the importance of vaccination and reindeer number but provided only a limited analysis of climatic factors, focusing on summer temperatures in Salekhard. We performed an extensive analysis of recent temperature and precipitation and linked it to the dynamics of active layer thickness characterizing permafrost thawing. The major hypothesis about the trigger of outbreak is related to thawing permafrost. There are also questions on how the outbreak has become so widespread. Given that ALT was increasing since 2010, bacteria could be released from permafrost even earlier than in 2016. The ability of bacteria to undergo the whole life cycle in the soil is a subject of debate (Hugh-Jones and Blackburn 2009 ). Recently, earthworms, plants and amoebae have been shown to interact with B. anthracis, demonstrating a possibility of bacteria's life cycle outside the host, although in laboratory conditions (Carlson et al. 2018 ). The contaminated areas in Siberia were considered to be prone to sanitation (Popova and Kulichenko 2017 ) due to unfavorable soil environment and weather conditions that could abrupt the life cycle of bacteria during an eventual vegetative stage. However, the situation has changed. Since 1968 when Cherkassky examined soil probes (Cherkassky 2003 ), the growing season (the period with mean daily temperature above 5°C) in Novy Port has lengthened by almost a month and the mean temperature of the growing season has increased by 1°C (Sizov et al. 2021 ). The climate has become less harsh. In addition, the current vegetation trends show active greening in Yamal district (Miles and Ezau 2016 ; Sizov et al. 2021 ), which could enrich soil with organics. Milder cooling of soils during two years prior to the outbreak, together with warmer and longer summers, could create conditions for bacteria to complete their life cycle by increasing the amount of spores if those were released from permafrost earlier than in 2016. Our analysis points at the importance of the local climatic factors for the outbreak. A combination of climate factors acting for several years in a row caused a strong regional effect over a spatial scale of only 100 km. These scales represent a challenge for the studies based on global models (Walsh et al. 2018 ; Carlson et al. 2019 ) hindering the use of large-scale models for prognostic purposes. Regional models with higher resolution can therefore be recommended for the monitoring and forecasts of weather conditions causing unfavorable epidemiological situations in cold climates. In addition, mathematical models can be applied on a local scale (Friedman and Yakubu 2013 ; Saad-Roy et al. 2017 ; Gomez et al. 2018 ; Stella et al. 2020 ). Remarkably, long-term climate dynamics indicated risks long before the outbreak. In Northwest Siberia, the mean annual air temperature and snow depth increased by 0.4–0.6°C and 4–10 cm per decade, respectively, during 50–60 years before 2012, while annual precipitation decreased by 50–100 mm in 75 years (1936–2010) (Katsov et al. 2014 ). Thus, in the absence of vaccination, the dramatic consequences could likely be a question of time. The risk of anthrax outbreaks associated with climate change was pointed out for East Siberia by Revich and Podolnaya ( 2011 ). A proper information campaign on the importance of vaccination and prognosis on unfavorable meteorological and climatological conditions is essential to prevent outbreaks in future. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 961 KB) Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 961 KB)

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9958924/

Aggregation-Based Bacterial Separation with Gram-Positive Selectivity by Using a Benzoxaborole-Modified Dendrimer

Antimicrobial-resistant (AMR) bacteria have become a critical global issue in recent years. The inefficacy of antimicrobial agents against AMR bacteria has led to increased difficulty in treating many infectious diseases. Analyses of the environmental distribution of bacteria are important for monitoring the AMR problem, and a rapid as well as viable pH- and temperature-independent bacterial separation method is required for collecting and concentrating bacteria from environmental samples. Thus, we aimed to develop a useful and selective bacterial separation method using a chemically synthesized nanoprobe. The metal-free benzoxaborole-based dendrimer probe BenzoB-PAMAM(+), which was synthesized from carboxy-benzoxaborole and a poly(amidoamine) (PAMAM) dendrimer, could help achieve Gram-positive bacterial separation by recognizing Gram-positive bacterial surfaces over a wide pH range, leading to the formation of large aggregations. The recognition site of benzoxaborole has a desirable high acidity and may therefore be responsible for the improved Gram-positive selectivity. The Gram-positive bacterial aggregation was then successfully collected by using a 10 μm membrane filter, with Gram-negative bacteria remaining in the filtrate solution. BenzoB-PAMAM(+) will thus be useful for application in biological analyses and could contribute to further investigations of bacterial distributions in environmental soil or water. 1. Introduction The proliferation of antimicrobial-resistant (AMR) bacteria has become a critical global concern [ 1 , 2 , 3 ] from the standpoint of achieving the Sustainable Development Goals [ 4 , 5 ]. The World Health Organization (WHO) has indicated that AMR is one of the top 10 global public health threats facing humanity. Several drugs, such as antibiotics, which are essential in treating infectious diseases, promote the development of AMR bacteria. When exposed to antibiotics AMR bacteria survive longer than other bacteria, and genetic mutations are thus passed on to the following generations. The misuse or overuse of antimicrobial agents and medicines is a driving force for AMR because improper antibiotic use accelerates genetic mutations. New AMR bacterial strains no longer respond to antibiotics and other antimicrobial drugs, rendering infections increasingly difficult or impossible to treat, increasing the risk of disease, severe illness, and death. High numbers of antibiotic-resistant bacteria have already been observed globally, suggesting a decrease in the number of effective antibiotics, which will result in reduced efficacy in treating infections. According to the Global Antimicrobial Resistance and Use Surveillance System (GLASS), the rate of resistance to ciprofloxacin, an antibiotic commonly used to treat urinary tract infections, varied from 8.4 to 92.9% for Escherichia coli [ 6 ]. This high resistance rate suggests that ciprofloxacin is now ineffective in over 50% of the population in many countries; however, AMR bacteria have been infecting humans, animals, plants, and the environment, and have spread worldwide. Therefore, investigating the bacteria that live in environments such as water [ 7 , 8 ] or soil [ 9 , 10 ] is crucial in monitoring the development and spread of AMR bacteria. Monitoring is also useful for solving other problems, such as agricultural pesticide resistance [ 11 ]. One study reported that the lack of countermeasures would result in global losses of approximately USD 100 trillion annually [ 12 ]. Neither the economic impacts nor the health problems associated with AMR can therefore be ignored [ 13 ]. Metagenomic analyses, which are one of the most powerful tools for the analysis of environmental samples, are often used to study microbiomes [ 14 ]. The methods used for the recognition and collection of environmental bacteria differ in terms of efficacy and are thus important in terms of monitoring. Since impurities, such as mud or other organisms, must be separated from samples, size exclusion filters are often used before bacterial analyses; however, because filters are not selective, impurities of the same or smaller diameter are not removed. Hence, bacterial collection methods that can selectively recognize bacteria are required. The problem has attracted the interest of many researchers, with biologists and chemists attempting to develop novel recognition or separation methods. Biological analysis methods, such as an immunosensing [ 15 , 16 ], enzyme-linked immunosorbent assay [ 17 ] or the polymerase chain reaction [ 18 , 19 , 20 ], which have high sensitivity and are species-specific because of their biological properties, have all been utilized with the aim of selective bacterial recognition, along with other methods, such as surface-enhanced Raman spectroscopy [ 21 , 22 ] or electrochemical detection [ 23 ]. However, to the best of our knowledge, these methods require advanced technical skills and expensive instruments, and are thus unsuitable in many situations. Extraction methods, such as polymer-based separations, have also been investigated worldwide [ 24 , 25 ]. Magnetic nanoprobes, magnetic beads that are labeled with specific bacterial recognition sites, are particularly useful for separating and collecting bacteria [ 26 ]. In this method, magnetic-nanoprobe-bound bacteria are attracted to a magnet placed near the sample tube. The solution is removed, and the bacteria are obtained with the magnetic probe. Research that developed both magnetic probes and specific devices has also been reported [ 27 ]. Biological materials, including immunogenic materials, are generally used to recognize and select bacteria under this method [ 28 ]; however, problems such as nanoprobe self-aggregation, the limited experimental conditions (such as pH and temperature) under which specific biological materials can be used, or the need for expensive reagents or devices, such as specific antibodies, mean that easier, more reasonable, and useful methods with which to recognize and collect bacteria are required. Our research group is focused on separation by aggregation in addition to recognition by boronic acid compounds. In 2019 we reported a chemically modified poly(amidoamine) (PAMAM) dendrimer, which has a few phenylboronic acids on its terminal structure that could recognize and form aggregates of Gram-positive bacteria, such as Staphylococcus aureus ( Figure 1 and Figure 2 ) [ 29 ], as confirmed by the naked eye. The core PAMAM dendrimers forming the nanoprobe, which comprise a group of synthetic spherical polymers [ 30 ], have desirable characteristics that allow bacterial extraction, water solubility, biocompatibility, and tolerance to self-aggregation, contrary to other nanoprobe cores, such as silica nanoparticles [ 31 ]. Boronic acid compounds are known as saccharide chemosensors [ 32 , 33 ] because they form boronate esters with the cis -diols in saccharides [ 34 ]. Boronic acid-based chemosensors are preferable for bacterial recognition because they are metal-free compounds that can rapidly recognize a target, such as bacterial saccharides, at room temperature (rt). We observed that our phenylboronic acid-modified PAMAM dendrimer (B-PAMAM) recognized lipoteichoic acid (LTA) [ 35 ], which is a surface structure that is specific to Gram-positive bacteria [ 36 ] and includes cis -diols [ 37 ]. B-PAMAM recognized LTA saccharides with the help of an electrostatic interaction between the positively charged amino terminus of B-PAMAM and the bacterial anionic surface [ 35 , 38 ]; however, the recognition ability was insufficient for obtaining selective aggregation and succeeding in bacterial separation. A turbidity measurement proved that only approximately half of the bacteria in the suspension were included in the aggregation owing to B-PAMAM exposure [ 29 ]. Therefore, in this study another boronic acid modification with stronger recognition ability is required to improve the selectivity and yield of Gram-positive bacteria that can be obtained using this method. Based on these requirements, carboxy-benzoxaborole, which should have a stronger affinity for saccharides rather than carboxy-phenylboronic acid, was prepared [ 39 ]. The improved benzoxaborole-modified PAMAM dendrimer, BenzoB-PAMAM(+) ( Figure 2 ), was then used to investigate bacterial selectivity and pH tolerance, as well as its use in further applications. 2. Results and Discussion 2.1. Characteristics of the Boronic Acid-Based BenzoB-PAMAMs Nanoprobes 2.1.1. Structure and Synthesis of BenzoB-PAMAMs First, we focused on benzoxaborole-modified PAMAM dendrimers (BenzoB-PAMAMs) instead of B-PAMAM. Since saccharide recognition, which results from the formation of boronate esters between boronic acids and cis -diols, mainly proceeds from conjugate tetrahedral boronate anions, boronic acids with large acid dissociation constants ( K a s) are preferred. Benzoxaborole is a boronic acid analog known for its low p K a value. For instance, the p K a value of phenylboronic acid is approximately 8.9 [ 40 ], whereas that of benzoxaborole is 7.5 [ 41 ]. BenzoB-PAMAM might thus have a stronger affinity and better pH tolerance than B-PAMAM. The 4-carboxy-benzoxaborole segment was then synthesized and subjected to condensation with the amine-terminated PAMAM(G4) dendrimer ( Scheme 1 ) to form BenzoB-PAMAM(+). Anionic BenzoB-PAMAM(−) was synthesized from the carboxylic acid-terminated PAMAM(G3.5) dendrimer for comparison with BenzoB-PAMAM(+) ( Scheme 2 ) in the same manner as BenzoB-PAMAM(+). The amine-modified segment 5 was synthesized from a carboxy-benzoxaborole segment and N -Boc-diaminoethane (4-amino-benzoxaborole was not used) to produce a molecule with the same p K a value as BenzoB-PAMAM(+). 2.1.2. Surface Properties of BenzoB-PAMAMs The zeta potential was measured to estimate the electrostatic interaction between BenzoB-PAMAMs and the bacterial surface ( Figure 3 ). We confirmed that the BenzoB-PAMAMs were successfully synthesized and that the desired charges were obtained by using the PAMAM dendrimer cores. BenzoB-PAMAM(+), with an amine terminus, is positively charged ( Figure 3 A), whereas BenzoB-PAMAM(−), with a carboxylic acid terminus, shows a relatively anionic surface ( Figure 3 B). The results obtained by using BenzoB-PAMAM(+) indicate that the zeta potential changed from positive to negative, while the pH increased from 8 to 9. This change may result from either the terminal primary amines or benzoxaborole modification. The assignment is further discussed below. As bacteria have negatively charged surfaces, the attraction effect of BenzoB-PAMAM(+) was likely electrostatic. In contrast to BenzoB-PAMAM(+), the negatively charged BenzoB-PAMAM(−) might be affected by electrostatic repulsion. 2.2. Bacterial Recognition by BenzoB-PAMAMs 2.2.1. Recognition Confirmed by a Turbidity Measurement and Direct Observation Turbidity was measured ( Figure 4 ) to elucidate the effects of electrostatic interaction on bacterial recognition. When a probe recognizes bacterial saccharides, the complexes formed by bacteria and the probe should result in changes to the turbidity, as has been reported previously [ 29 ]. Measurements were recorded between pHs of 4.0 and 11.0 to cover all biologically relevant pH conditions. Researchers have reported the growth of Gram-positive bacteria, S. aureus , in environments with pHs of 4.0−9.8 [ 42 ] and Gram-negative bacteria, E. coli , at pHs of 4.5−9.0 [ 43 ]. The results in Figure 4 A indicate that turbidity does not change under all pH conditions for BenzoB-PAMAM(−). As S. aureus [ 44 , 45 ] and E. coli [ 46 , 47 ] have negatively charged surfaces, in the same manner as general bacteria, the results indicate that electrostatic repulsion between the anionic terminus of BenzoB-PAMAM(−) and the bacterial surface disturbed any saccharide recognition. In contrast, BenzoB-PAMAM(+) selectively recognized Gram-positive S. aureus under wide pH conditions, as seen in Figure 4 B (from a pH of 5 or 6 to 10). We have already reported that electrostatic interaction between cationic dendrimers and negatively charged bacterial surfaces enhanced recognition ability, whereas electrostatic repulsion disturbed bacterial recognition [ 35 , 38 ]. The resulting aggregation was easily confirmed by the naked eye ( Figure 5 ). Images revealed no aggregation at a pH of 4.0 ( Figure 5 A); however, it is visible at pHs of 5.0 ( Figure 5 B) and 6.0 ( Figure 5 C), as suggested by the results of the turbidity measurements. Considering the previous results, that phenylboronic acid-modified B-PAMAM led to selectivity from a pH of 6−8 or 7−9 [ 38 ], BenzoB-PAMAM(+) has a significantly improved recognition ability compared with that of B-PAMAM. The results suggest that the lower p K a value of boronic acid results in a stronger recognition force towards the LTA of Gram-positive bacteria. Aggregation was also confirmed through microscopic observation to obtain further information ( Figure 6 ). Compared with bacterial images ( Figure 6 B,D), images depicting BenzoB-PAMAM(+) suspensions revealed complexes resulting from the interaction between bacteria and BenzoB-PAMAM(+) ( Figure 6 A,C); however, the aggregation size is far more apparent for Gram-positive S. aureus and Gram-negative E. coli . We considered that the aggregation size was instrumental in changes in turbidity and the presence of visible precipitation. Although the E. coli suspension formed minute aggregations with BenzoB-PAMAM(+), no structure was observed at the bottom of the sample tube. We concluded that the aggregation could be extracted via scooping. Notably, some aggregated bacteria appeared dead, as we have previously reported [ 48 ]; however, this was not a problem because metagenomic analyses, which are conducted after bacterial extraction, are not affected by the state of bacteria. 2.2.2. Improved Recognition in Association with the Desirable p K a Value of Benzoxaborole The p K a value of the benzoxaborole recognition sites on BenzoB-PAMAM(+) was investigated using benzoxaborole with an amine terminal. UV–Vis absorption spectra were measured between pHs of 4.0 and 11.0 ( Figure 7 A). Absorbance at 266 nm, which produces large differences under different pH conditions, was used to determine the p K a value ( Figure 7 B). By the inflexion point, a p K a value of 6.0 was observed for the benzoxaborole comprising BenzoB-PAMAM(+). This value is consistent with the results of the turbidity measurements, with the probe recognizing bacteria at pHs of 5−6. Benzoxaborole can easily exist as a tetrahedral anion because of its low p K a value, and its character is desirable for saccharide recognition, which starts preferentially in the initial boronate anions. It should be noted that the change in the zeta potential at pHs around 8−9 was not caused by benzoxaborole, according to its p K a value of 6.0, and may have been caused by the terminal primary amines on BenzoB-PAMAM(+). Researchers have reported p K a values of 8−9 for the terminal primary amines on PAMAM(G4) dendrimers [ 49 ], matching the pH range, which shows changes in line with the zeta potential at pHs of 8−9 ( Figure 3 B). In summary, BenzoB-PAMAM(+) had the desired low p K a value, which enhanced selective bacterial recognition, and the zeta potential of the probe was mainly determined by terminal amine groups rather than the benzoxaborole recognition site. 2.2.3. Bacterial Selectivity Using BenzoB-PAMAM(+) Changes in the turbidity were also measured using eight different bacteria to investigate bacterial selectivity ( Figure 8 ). The results indicate that BenzoB-PAMAM selectively recognized Gram-positive bacteria, with a decrease in turbidity. Compared with the results obtained using B-PAMAM, which showed a decrease of approximately 50% in turbidity for a Gram-positive bacterial suspension [ 29 ], a decrease of almost 100% for BenzoB-PAMAM(+), with its low p K a value, indicates its suitability for bacterial separation. A turbidity of approximately zero means that almost all of the bacteria in a suspension have formed aggregations with probes and can be extracted with a sufficient yield. We confirmed that the improved recognition and Gram-positive selectivity were derived from LTA recognition ( Figures S1 and S2 ). From the perspective of bacterial extraction by aggregation, the size of an aggregation is significant. Typically, Gram-positive bacterial aggregates that are significantly larger than bacteria (or some minute aggregation of Gram-negative bacteria) could be obtained by scooping up the precipitation followed by washing through a membrane filter of an appropriate pore size. Eight bacteria used in Figure 8 were observed via microscopy to confirm the size of the aggregations ( Figure 9 and Figure 10 ). The microscopy images demonstrate that Gram-positive bacteria, which result in dramatic turbidity decreases, formed large aggregates of approximately 50−200 μm ( Figure 9 ). We also confirmed that a low concentration of bacteria (10 6 CFU·mL −1 ) formed approximately 10 μm aggregates ( Figure S3 ). Gram-negative bacteria did not produce such large aggregates, even E. coli ATCC25922, which formed minute aggregates ( Figure 10 ). As the aggregation size of E. coli ATCC25922 was smaller than 10 μm, we concluded that a membrane with a filter of a 10 μm pore size was suitable for extracting and washing Gram-positive bacterial aggregations by trapping Gram-positive bacteria while allowing Gram-negative bacteria to pass through. 2.2.4. Filtration for Separating Aggregations We attempted to separate the Gram-positive bacterial aggregates from the probes by filtration with a 10 µm pore size membrane filter. The obtained materials on the filter and in the filtrate were inspected separately by using a microscope ( Figure 11 ). Suspensions that include S. aureus IAM1011 typically exhibit large aggregations on the filters ( Figure 11 A,C,D). The sample without S. aureus IAM1011 revealed bacteria in the filtrate solution with no aggregation on the filter ( Figure 11 B). These results show that the membrane filter could separate the aggregation by S. aureus IAM1011 from the sample solution as desired. Although the aggregation contained a certain amount of E. coli K12W3110 stained with EB ( Figure 11 C,D), E. coli K12W3110 itself did not aggregate, as seen in Figure 11 B. We also confirmed that excess E. coli K12W3110 did not disturb the aggregation process ( Figure 11 D). Concordantly, E. coli K12W3110 was observed in the filtrate solution, meaning that only a small number of E. coli were involved in the aggregation. Another bacterial mixture, comprising S. aureus ATCC25923 and E. coli ATCC25922, also exhibited aggregation on the filter and dispersed bacteria in the filtrate solution ( Figure S4 ). In summary, Gram-positive bacteria were successfully separated by filtering the aggregates through a membrane filter with a 10 µm pore size. 2.1. Characteristics of the Boronic Acid-Based BenzoB-PAMAMs Nanoprobes 2.1.1. Structure and Synthesis of BenzoB-PAMAMs First, we focused on benzoxaborole-modified PAMAM dendrimers (BenzoB-PAMAMs) instead of B-PAMAM. Since saccharide recognition, which results from the formation of boronate esters between boronic acids and cis -diols, mainly proceeds from conjugate tetrahedral boronate anions, boronic acids with large acid dissociation constants ( K a s) are preferred. Benzoxaborole is a boronic acid analog known for its low p K a value. For instance, the p K a value of phenylboronic acid is approximately 8.9 [ 40 ], whereas that of benzoxaborole is 7.5 [ 41 ]. BenzoB-PAMAM might thus have a stronger affinity and better pH tolerance than B-PAMAM. The 4-carboxy-benzoxaborole segment was then synthesized and subjected to condensation with the amine-terminated PAMAM(G4) dendrimer ( Scheme 1 ) to form BenzoB-PAMAM(+). Anionic BenzoB-PAMAM(−) was synthesized from the carboxylic acid-terminated PAMAM(G3.5) dendrimer for comparison with BenzoB-PAMAM(+) ( Scheme 2 ) in the same manner as BenzoB-PAMAM(+). The amine-modified segment 5 was synthesized from a carboxy-benzoxaborole segment and N -Boc-diaminoethane (4-amino-benzoxaborole was not used) to produce a molecule with the same p K a value as BenzoB-PAMAM(+). 2.1.2. Surface Properties of BenzoB-PAMAMs The zeta potential was measured to estimate the electrostatic interaction between BenzoB-PAMAMs and the bacterial surface ( Figure 3 ). We confirmed that the BenzoB-PAMAMs were successfully synthesized and that the desired charges were obtained by using the PAMAM dendrimer cores. BenzoB-PAMAM(+), with an amine terminus, is positively charged ( Figure 3 A), whereas BenzoB-PAMAM(−), with a carboxylic acid terminus, shows a relatively anionic surface ( Figure 3 B). The results obtained by using BenzoB-PAMAM(+) indicate that the zeta potential changed from positive to negative, while the pH increased from 8 to 9. This change may result from either the terminal primary amines or benzoxaborole modification. The assignment is further discussed below. As bacteria have negatively charged surfaces, the attraction effect of BenzoB-PAMAM(+) was likely electrostatic. In contrast to BenzoB-PAMAM(+), the negatively charged BenzoB-PAMAM(−) might be affected by electrostatic repulsion. 2.1.1. Structure and Synthesis of BenzoB-PAMAMs First, we focused on benzoxaborole-modified PAMAM dendrimers (BenzoB-PAMAMs) instead of B-PAMAM. Since saccharide recognition, which results from the formation of boronate esters between boronic acids and cis -diols, mainly proceeds from conjugate tetrahedral boronate anions, boronic acids with large acid dissociation constants ( K a s) are preferred. Benzoxaborole is a boronic acid analog known for its low p K a value. For instance, the p K a value of phenylboronic acid is approximately 8.9 [ 40 ], whereas that of benzoxaborole is 7.5 [ 41 ]. BenzoB-PAMAM might thus have a stronger affinity and better pH tolerance than B-PAMAM. The 4-carboxy-benzoxaborole segment was then synthesized and subjected to condensation with the amine-terminated PAMAM(G4) dendrimer ( Scheme 1 ) to form BenzoB-PAMAM(+). Anionic BenzoB-PAMAM(−) was synthesized from the carboxylic acid-terminated PAMAM(G3.5) dendrimer for comparison with BenzoB-PAMAM(+) ( Scheme 2 ) in the same manner as BenzoB-PAMAM(+). The amine-modified segment 5 was synthesized from a carboxy-benzoxaborole segment and N -Boc-diaminoethane (4-amino-benzoxaborole was not used) to produce a molecule with the same p K a value as BenzoB-PAMAM(+). 2.1.2. Surface Properties of BenzoB-PAMAMs The zeta potential was measured to estimate the electrostatic interaction between BenzoB-PAMAMs and the bacterial surface ( Figure 3 ). We confirmed that the BenzoB-PAMAMs were successfully synthesized and that the desired charges were obtained by using the PAMAM dendrimer cores. BenzoB-PAMAM(+), with an amine terminus, is positively charged ( Figure 3 A), whereas BenzoB-PAMAM(−), with a carboxylic acid terminus, shows a relatively anionic surface ( Figure 3 B). The results obtained by using BenzoB-PAMAM(+) indicate that the zeta potential changed from positive to negative, while the pH increased from 8 to 9. This change may result from either the terminal primary amines or benzoxaborole modification. The assignment is further discussed below. As bacteria have negatively charged surfaces, the attraction effect of BenzoB-PAMAM(+) was likely electrostatic. In contrast to BenzoB-PAMAM(+), the negatively charged BenzoB-PAMAM(−) might be affected by electrostatic repulsion. 2.2. Bacterial Recognition by BenzoB-PAMAMs 2.2.1. Recognition Confirmed by a Turbidity Measurement and Direct Observation Turbidity was measured ( Figure 4 ) to elucidate the effects of electrostatic interaction on bacterial recognition. When a probe recognizes bacterial saccharides, the complexes formed by bacteria and the probe should result in changes to the turbidity, as has been reported previously [ 29 ]. Measurements were recorded between pHs of 4.0 and 11.0 to cover all biologically relevant pH conditions. Researchers have reported the growth of Gram-positive bacteria, S. aureus , in environments with pHs of 4.0−9.8 [ 42 ] and Gram-negative bacteria, E. coli , at pHs of 4.5−9.0 [ 43 ]. The results in Figure 4 A indicate that turbidity does not change under all pH conditions for BenzoB-PAMAM(−). As S. aureus [ 44 , 45 ] and E. coli [ 46 , 47 ] have negatively charged surfaces, in the same manner as general bacteria, the results indicate that electrostatic repulsion between the anionic terminus of BenzoB-PAMAM(−) and the bacterial surface disturbed any saccharide recognition. In contrast, BenzoB-PAMAM(+) selectively recognized Gram-positive S. aureus under wide pH conditions, as seen in Figure 4 B (from a pH of 5 or 6 to 10). We have already reported that electrostatic interaction between cationic dendrimers and negatively charged bacterial surfaces enhanced recognition ability, whereas electrostatic repulsion disturbed bacterial recognition [ 35 , 38 ]. The resulting aggregation was easily confirmed by the naked eye ( Figure 5 ). Images revealed no aggregation at a pH of 4.0 ( Figure 5 A); however, it is visible at pHs of 5.0 ( Figure 5 B) and 6.0 ( Figure 5 C), as suggested by the results of the turbidity measurements. Considering the previous results, that phenylboronic acid-modified B-PAMAM led to selectivity from a pH of 6−8 or 7−9 [ 38 ], BenzoB-PAMAM(+) has a significantly improved recognition ability compared with that of B-PAMAM. The results suggest that the lower p K a value of boronic acid results in a stronger recognition force towards the LTA of Gram-positive bacteria. Aggregation was also confirmed through microscopic observation to obtain further information ( Figure 6 ). Compared with bacterial images ( Figure 6 B,D), images depicting BenzoB-PAMAM(+) suspensions revealed complexes resulting from the interaction between bacteria and BenzoB-PAMAM(+) ( Figure 6 A,C); however, the aggregation size is far more apparent for Gram-positive S. aureus and Gram-negative E. coli . We considered that the aggregation size was instrumental in changes in turbidity and the presence of visible precipitation. Although the E. coli suspension formed minute aggregations with BenzoB-PAMAM(+), no structure was observed at the bottom of the sample tube. We concluded that the aggregation could be extracted via scooping. Notably, some aggregated bacteria appeared dead, as we have previously reported [ 48 ]; however, this was not a problem because metagenomic analyses, which are conducted after bacterial extraction, are not affected by the state of bacteria. 2.2.2. Improved Recognition in Association with the Desirable p K a Value of Benzoxaborole The p K a value of the benzoxaborole recognition sites on BenzoB-PAMAM(+) was investigated using benzoxaborole with an amine terminal. UV–Vis absorption spectra were measured between pHs of 4.0 and 11.0 ( Figure 7 A). Absorbance at 266 nm, which produces large differences under different pH conditions, was used to determine the p K a value ( Figure 7 B). By the inflexion point, a p K a value of 6.0 was observed for the benzoxaborole comprising BenzoB-PAMAM(+). This value is consistent with the results of the turbidity measurements, with the probe recognizing bacteria at pHs of 5−6. Benzoxaborole can easily exist as a tetrahedral anion because of its low p K a value, and its character is desirable for saccharide recognition, which starts preferentially in the initial boronate anions. It should be noted that the change in the zeta potential at pHs around 8−9 was not caused by benzoxaborole, according to its p K a value of 6.0, and may have been caused by the terminal primary amines on BenzoB-PAMAM(+). Researchers have reported p K a values of 8−9 for the terminal primary amines on PAMAM(G4) dendrimers [ 49 ], matching the pH range, which shows changes in line with the zeta potential at pHs of 8−9 ( Figure 3 B). In summary, BenzoB-PAMAM(+) had the desired low p K a value, which enhanced selective bacterial recognition, and the zeta potential of the probe was mainly determined by terminal amine groups rather than the benzoxaborole recognition site. 2.2.3. Bacterial Selectivity Using BenzoB-PAMAM(+) Changes in the turbidity were also measured using eight different bacteria to investigate bacterial selectivity ( Figure 8 ). The results indicate that BenzoB-PAMAM selectively recognized Gram-positive bacteria, with a decrease in turbidity. Compared with the results obtained using B-PAMAM, which showed a decrease of approximately 50% in turbidity for a Gram-positive bacterial suspension [ 29 ], a decrease of almost 100% for BenzoB-PAMAM(+), with its low p K a value, indicates its suitability for bacterial separation. A turbidity of approximately zero means that almost all of the bacteria in a suspension have formed aggregations with probes and can be extracted with a sufficient yield. We confirmed that the improved recognition and Gram-positive selectivity were derived from LTA recognition ( Figures S1 and S2 ). From the perspective of bacterial extraction by aggregation, the size of an aggregation is significant. Typically, Gram-positive bacterial aggregates that are significantly larger than bacteria (or some minute aggregation of Gram-negative bacteria) could be obtained by scooping up the precipitation followed by washing through a membrane filter of an appropriate pore size. Eight bacteria used in Figure 8 were observed via microscopy to confirm the size of the aggregations ( Figure 9 and Figure 10 ). The microscopy images demonstrate that Gram-positive bacteria, which result in dramatic turbidity decreases, formed large aggregates of approximately 50−200 μm ( Figure 9 ). We also confirmed that a low concentration of bacteria (10 6 CFU·mL −1 ) formed approximately 10 μm aggregates ( Figure S3 ). Gram-negative bacteria did not produce such large aggregates, even E. coli ATCC25922, which formed minute aggregates ( Figure 10 ). As the aggregation size of E. coli ATCC25922 was smaller than 10 μm, we concluded that a membrane with a filter of a 10 μm pore size was suitable for extracting and washing Gram-positive bacterial aggregations by trapping Gram-positive bacteria while allowing Gram-negative bacteria to pass through. 2.2.4. Filtration for Separating Aggregations We attempted to separate the Gram-positive bacterial aggregates from the probes by filtration with a 10 µm pore size membrane filter. The obtained materials on the filter and in the filtrate were inspected separately by using a microscope ( Figure 11 ). Suspensions that include S. aureus IAM1011 typically exhibit large aggregations on the filters ( Figure 11 A,C,D). The sample without S. aureus IAM1011 revealed bacteria in the filtrate solution with no aggregation on the filter ( Figure 11 B). These results show that the membrane filter could separate the aggregation by S. aureus IAM1011 from the sample solution as desired. Although the aggregation contained a certain amount of E. coli K12W3110 stained with EB ( Figure 11 C,D), E. coli K12W3110 itself did not aggregate, as seen in Figure 11 B. We also confirmed that excess E. coli K12W3110 did not disturb the aggregation process ( Figure 11 D). Concordantly, E. coli K12W3110 was observed in the filtrate solution, meaning that only a small number of E. coli were involved in the aggregation. Another bacterial mixture, comprising S. aureus ATCC25923 and E. coli ATCC25922, also exhibited aggregation on the filter and dispersed bacteria in the filtrate solution ( Figure S4 ). In summary, Gram-positive bacteria were successfully separated by filtering the aggregates through a membrane filter with a 10 µm pore size. 2.2.1. Recognition Confirmed by a Turbidity Measurement and Direct Observation Turbidity was measured ( Figure 4 ) to elucidate the effects of electrostatic interaction on bacterial recognition. When a probe recognizes bacterial saccharides, the complexes formed by bacteria and the probe should result in changes to the turbidity, as has been reported previously [ 29 ]. Measurements were recorded between pHs of 4.0 and 11.0 to cover all biologically relevant pH conditions. Researchers have reported the growth of Gram-positive bacteria, S. aureus , in environments with pHs of 4.0−9.8 [ 42 ] and Gram-negative bacteria, E. coli , at pHs of 4.5−9.0 [ 43 ]. The results in Figure 4 A indicate that turbidity does not change under all pH conditions for BenzoB-PAMAM(−). As S. aureus [ 44 , 45 ] and E. coli [ 46 , 47 ] have negatively charged surfaces, in the same manner as general bacteria, the results indicate that electrostatic repulsion between the anionic terminus of BenzoB-PAMAM(−) and the bacterial surface disturbed any saccharide recognition. In contrast, BenzoB-PAMAM(+) selectively recognized Gram-positive S. aureus under wide pH conditions, as seen in Figure 4 B (from a pH of 5 or 6 to 10). We have already reported that electrostatic interaction between cationic dendrimers and negatively charged bacterial surfaces enhanced recognition ability, whereas electrostatic repulsion disturbed bacterial recognition [ 35 , 38 ]. The resulting aggregation was easily confirmed by the naked eye ( Figure 5 ). Images revealed no aggregation at a pH of 4.0 ( Figure 5 A); however, it is visible at pHs of 5.0 ( Figure 5 B) and 6.0 ( Figure 5 C), as suggested by the results of the turbidity measurements. Considering the previous results, that phenylboronic acid-modified B-PAMAM led to selectivity from a pH of 6−8 or 7−9 [ 38 ], BenzoB-PAMAM(+) has a significantly improved recognition ability compared with that of B-PAMAM. The results suggest that the lower p K a value of boronic acid results in a stronger recognition force towards the LTA of Gram-positive bacteria. Aggregation was also confirmed through microscopic observation to obtain further information ( Figure 6 ). Compared with bacterial images ( Figure 6 B,D), images depicting BenzoB-PAMAM(+) suspensions revealed complexes resulting from the interaction between bacteria and BenzoB-PAMAM(+) ( Figure 6 A,C); however, the aggregation size is far more apparent for Gram-positive S. aureus and Gram-negative E. coli . We considered that the aggregation size was instrumental in changes in turbidity and the presence of visible precipitation. Although the E. coli suspension formed minute aggregations with BenzoB-PAMAM(+), no structure was observed at the bottom of the sample tube. We concluded that the aggregation could be extracted via scooping. Notably, some aggregated bacteria appeared dead, as we have previously reported [ 48 ]; however, this was not a problem because metagenomic analyses, which are conducted after bacterial extraction, are not affected by the state of bacteria. 2.2.2. Improved Recognition in Association with the Desirable p K a Value of Benzoxaborole The p K a value of the benzoxaborole recognition sites on BenzoB-PAMAM(+) was investigated using benzoxaborole with an amine terminal. UV–Vis absorption spectra were measured between pHs of 4.0 and 11.0 ( Figure 7 A). Absorbance at 266 nm, which produces large differences under different pH conditions, was used to determine the p K a value ( Figure 7 B). By the inflexion point, a p K a value of 6.0 was observed for the benzoxaborole comprising BenzoB-PAMAM(+). This value is consistent with the results of the turbidity measurements, with the probe recognizing bacteria at pHs of 5−6. Benzoxaborole can easily exist as a tetrahedral anion because of its low p K a value, and its character is desirable for saccharide recognition, which starts preferentially in the initial boronate anions. It should be noted that the change in the zeta potential at pHs around 8−9 was not caused by benzoxaborole, according to its p K a value of 6.0, and may have been caused by the terminal primary amines on BenzoB-PAMAM(+). Researchers have reported p K a values of 8−9 for the terminal primary amines on PAMAM(G4) dendrimers [ 49 ], matching the pH range, which shows changes in line with the zeta potential at pHs of 8−9 ( Figure 3 B). In summary, BenzoB-PAMAM(+) had the desired low p K a value, which enhanced selective bacterial recognition, and the zeta potential of the probe was mainly determined by terminal amine groups rather than the benzoxaborole recognition site. 2.2.3. Bacterial Selectivity Using BenzoB-PAMAM(+) Changes in the turbidity were also measured using eight different bacteria to investigate bacterial selectivity ( Figure 8 ). The results indicate that BenzoB-PAMAM selectively recognized Gram-positive bacteria, with a decrease in turbidity. Compared with the results obtained using B-PAMAM, which showed a decrease of approximately 50% in turbidity for a Gram-positive bacterial suspension [ 29 ], a decrease of almost 100% for BenzoB-PAMAM(+), with its low p K a value, indicates its suitability for bacterial separation. A turbidity of approximately zero means that almost all of the bacteria in a suspension have formed aggregations with probes and can be extracted with a sufficient yield. We confirmed that the improved recognition and Gram-positive selectivity were derived from LTA recognition ( Figures S1 and S2 ). From the perspective of bacterial extraction by aggregation, the size of an aggregation is significant. Typically, Gram-positive bacterial aggregates that are significantly larger than bacteria (or some minute aggregation of Gram-negative bacteria) could be obtained by scooping up the precipitation followed by washing through a membrane filter of an appropriate pore size. Eight bacteria used in Figure 8 were observed via microscopy to confirm the size of the aggregations ( Figure 9 and Figure 10 ). The microscopy images demonstrate that Gram-positive bacteria, which result in dramatic turbidity decreases, formed large aggregates of approximately 50−200 μm ( Figure 9 ). We also confirmed that a low concentration of bacteria (10 6 CFU·mL −1 ) formed approximately 10 μm aggregates ( Figure S3 ). Gram-negative bacteria did not produce such large aggregates, even E. coli ATCC25922, which formed minute aggregates ( Figure 10 ). As the aggregation size of E. coli ATCC25922 was smaller than 10 μm, we concluded that a membrane with a filter of a 10 μm pore size was suitable for extracting and washing Gram-positive bacterial aggregations by trapping Gram-positive bacteria while allowing Gram-negative bacteria to pass through. 2.2.4. Filtration for Separating Aggregations We attempted to separate the Gram-positive bacterial aggregates from the probes by filtration with a 10 µm pore size membrane filter. The obtained materials on the filter and in the filtrate were inspected separately by using a microscope ( Figure 11 ). Suspensions that include S. aureus IAM1011 typically exhibit large aggregations on the filters ( Figure 11 A,C,D). The sample without S. aureus IAM1011 revealed bacteria in the filtrate solution with no aggregation on the filter ( Figure 11 B). These results show that the membrane filter could separate the aggregation by S. aureus IAM1011 from the sample solution as desired. Although the aggregation contained a certain amount of E. coli K12W3110 stained with EB ( Figure 11 C,D), E. coli K12W3110 itself did not aggregate, as seen in Figure 11 B. We also confirmed that excess E. coli K12W3110 did not disturb the aggregation process ( Figure 11 D). Concordantly, E. coli K12W3110 was observed in the filtrate solution, meaning that only a small number of E. coli were involved in the aggregation. Another bacterial mixture, comprising S. aureus ATCC25923 and E. coli ATCC25922, also exhibited aggregation on the filter and dispersed bacteria in the filtrate solution ( Figure S4 ). In summary, Gram-positive bacteria were successfully separated by filtering the aggregates through a membrane filter with a 10 µm pore size. 3. Materials and Methods 3.1. Reagents and Apparatus 3.1.1. Chemical Reagents All reagents and solvents were purchased from commercial suppliers and used without further purification unless otherwise specified. Acetone, super dehydrated (016-23465), agar (018-15811), ammonium acetate (NH 4 OAc, 016-02845), n -bromosuccinimide (NBS, 021-07232), dichloromethane (DCM, 139-02444),1,4-dioxane, super dehydrated (040-31651), ethyl acetate (EtOAc, 051-00351), hydrochloric acid (080-01066), magnesium sulfate, anhydrous (137-12335), potassium acetate (KOAc, 166-01372), 30 % potassium hydroxide solution (32904-00), sodium chloride (191-01665), sodium periodate (197-02402), and sodium sulfate (197-03345) were purchased from the FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). 4′,6-Diamidino-2-phenylindole dihydrochloride n -hydrate (DAPI, D523), ethidium bromide solution (EB, BS06), and propidium iodide solution (PI, P378) were purchased from DOJINDO LABORATORIES (Kumamoto, Japan). 2,2′-azobis(isobutyronitrile) (AIBN, A0566), [1,1′-bis(diphenylphosphino) ferrocene]palladium(II) dichloride dichloromethane adduct (Pd(dppf)Cl 2 /CH 2 Cl 2 , B2064), bis(pinacolato)diboron (B 2 pin 2 , B1964), 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMT-MM, D2919), methyl 4-bromo-3-methylbenzoate ( 1 , B2256), and N -(tert-butoxycarbonyl)-1,2-diaminoethane ( N -Boc-diaminoethane, A1371) were purchased from the Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). Acetonitrile dehydrated (MeCN, 01837-95), dimethyl sulfoxide-d 6 (DMSO-d 6 , 11560-33), methanol (MeOH, 25183-70), methanol-d 4 (99.8 atom%D, 25183-70), and deuterium water (D 2 O, 99.8 atom%D, 10086-23) were purchased from the Kanto Chemical Co. Inc. (Tokyo, Japan). Omnipore membrane 10 μm × Φ13 mm (JCWP01300), poly(amidoamine) (PAMAM) dendrimer, ethylenediamine core, generation 4.0 (412449) solution, PAMAM dendrimer, ethylenediamine core, generation 3.5 solution, 10 wt.% in methanol (412430), and Swinnex Filter Holder Φ13 mm (SX0001300) were purchased from the Sigma-Aldrich Japan Co. LLC. (Tokyo, Japan). Bacto yeast extract (212750) and Bacto Tryptone (211705) were purchased from Nippon Becton, Dickinson Co. Ltd. (Tokyo, Japan). Phosphate-buffered saline (PBS, 2101) was purchased from the Cell Science & Technology Institute Inc. (Miyagi, Japan). Water was distilled and deionized twice by using a Milli-Q water system (WG222, Yamato Scientific Co. Ltd., Tokyo, Japan and Milli-Q Advantage, Merck Millipore, Burlington, MA, USA) before use. A Spectra/Por 6 dialysis bag (MW cutoff = 1000) was purchased from the Repligen Co. (Waltham, MA, USA). 3.1.2. Apparatus 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a JEOL JNM-ECA 500 spectrometer (500 MHz) by JEOL (Tokyo, Japan) at 300 K or a Bruker Avance III HD 400 MHz by Bruker (Billerica, MA, USA) at 300 K using a deuterated solvent. All pH values were recorded by using Horiba F-52 and F-72 pH meters (Horiba, Ltd., Kyoto, Japan). Ultraviolet−visible (UV−Vis) absorption spectra were measured by using a JASCO V-570 or V-760 UV−Vis spectrophotometer (JASCO Co., Tokyo, Japan) equipped with a Peltier Thermo controller and a 10 mm quartz cell. Zeta potential measurements were carried out at 25 °C by using a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, Worcestershire, U.K.). Samples were shaken by using a MULTI SHAKER MS-300 (AS ONE Co., Osaka, Japan). Centrifugation was conducted by using a CF15RN (Hitachi High-Technologies, Co., Tokyo, Japan). The p K a value was determined by using Igor Pro v5.0.3.0 (Wavemetrics inc., Portland, OR, USA) based on the acid dissociation model of monobasic acids. 3.2. Preparation of Dendrimer Probes 3.2.1. Synthesis of Methyl 3-Methyl-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 2 ) Compound 2 was synthesized by using commercially available compound 1 [ 39 ] under an argon atmosphere. Bis(pinacolato)diboron (1.1 eq, 1.68 g, and 6.6 mmol), KOAc (3.1 eq, 1.85 g, and 19 mmol), and Pd(dppf)Cl 2 /CH 2 Cl 2 (2 mol%, 0.11 g, and 0.14 mmol) were added to the solution produced by dissolving compound 1 (1.0 eq, 1.38 g, and 6.0 mmol) in 1,4-dioxane (10.0 mL). The resulting orange reaction solution was stirred vigorously at 95 °C for 22 h before undesired precipitates were removed through filtration. The filtrate was then evaporated, and the product was extracted three times with DCM/water. The combined organic layers were dried over MgSO 4 and concentrated in vacuo. Purification by silica gel column chromatography (DCM 100%) was then used to generate product 2 as a yellow oil (684 mg, 2.5 mmol, and 40%). The structure was confirmed from the resulting 1 H NMR spectrum ( Figure S5 ). 3.2.2. Synthesis of Methyl 3-(Bromomethyl)-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 3 ) Compound 3 was synthesized by using compound 2 under an argon atmosphere [ 39 ]. NBS (1.3 eq, 695 mg, and 3.9 mmol) and AIBN (2 mol%, 10 mg, and 0.06 mmol) were added to the solution generated by dissolving compound 2 (1.0 eq, 861 mg, and 3.1 mmol) in MeCN (15.0 mL). The solution was stirred at 90 °C for 2.5 h before the solvent was evaporated. DCM was added to the concentrated solution to remove the undesired white precipitation. Purification by silica gel column chromatography (DCM 100%) generated product 3 in the form of a yellowish oil (580 mg, 1.6 mmol, and 52%). The structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S6 ). 3.2.3. Synthesis of 1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxylic Acid ( 4 ) Compound 4 was synthesized by using compound 3 [ 39 ], which (1.0 eq, 580 mg, and 1.6 mmol) was dissolved in an acetone/water mixture at 1/1 ( v / v ). NaIO 4 (4.9 eq, 1.74 g, and 8.1 mmol) and NH 4 OAc (5.0 eq, 630 mg, and 8.1 mmol) were added to the produced solution, and the mixture was stirred at rt for 1 day. The solvent was then evaporated, and the product was extracted three times with EtOAc/water. The combined organic layers were then dried over MgSO 4 and concentrated in vacuo. Afterwards, 15% KOH aq. (3.0 mL) was added to the obtained product and stirred for 1.5 h at rt, followed by acidification via HCl aq. The obtained white precipitation, compound 4 (152 mg, 0.83 mmol, and 52%), was extracted by filtration, and the structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S7 ). 3.2.4. Synthesis of BenzoB-PAMAM(+) Compound 4 (8.0 eq, 11.3 mg, and 64 μmol) and DMT-MM (40.0 eq, 88.2 mg, and 0.32 mmol) were dissolved in methanol (10.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G4) dendrimer (1.4 mL, 1.0 eq, and 8.0 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days before the purified product was lyophilized to produce white flocks (124.5 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S8 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.2.5. Synthesis of N -(2-Aminoethyl)-1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxamide ( 5 ) N -Boc-diaminoethane (96 mg, 1.2 eq, and 0.6 mmol) and synthetic product 4 (89 mg, 1.0 eq, and 0.5 mmol) were dissolved in methanol (3 mL). The mixture was stirred at rt for 30 min [ 35 ]. DMT-MM (550 mg, 4.0 eq, and 2.0 mmol) was added to the solution, and a condensation reaction was performed at rt for 4 days. The product of the reaction was extracted three times with EtOAc, washed with brine, and dried over Na 2 SO 4 . The solvent was removed by evaporation, and HCl in methanol (4 M) was added to the residue. The solution was stirred at rt for 3 h to induce deprotection. The solvent was evaporated, and the final product 5 was obtained as a white solid (212 mg). The undesired byproducts were excluded by dialysis in the following synthesis. The identity was confirmed by ESI-HRMS spectral measurement and the resulting 1 H and 13 C NMR ( Figures S9 and S10 ) spectra. The ESI-HRMS ( m / z ) calculated for C 11 H 16 B 1 N 2 O 3 [M + CH 3 ] + 235.1254 found a figure of 235.1277. 3.2.6. Synthesis of BenzoB-PAMAM(−) Compound 5 (8.0 eq, 11.7 mg, and 49.5 μmol) and DMT-MM (32.0 eq, 54.6 mg, and 0.20 mmol) were dissolved in methanol (5.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G3.5) dendrimer (1.0 mL, 1.0 eq, and 6.2 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was then transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days. The purified product was lyophilized to give white flocks (62.2 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S11 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.3. Biological Experiments 3.3.1. Bacterial Culture The lysogeny broth (LB) was composed of 2 g of Bacto Tryptone, 1 g of Bacto yeast extract, and 2 g of NaCl in 200 mL of distilled water. S. aureus IAM1011, S. aureus ATCC25923, S. aureus ATCC29213, S. pseudintermedius 2012-S-27, S. epidermidis ATCC12228, Enterococcus faecalis ATCC29212, E. coli K12W3110, E. coli ATCC25922, Pseudomonas aeruginosa ATCC27853, and Salmonella enteritidis ATCC13311 were provided by RIKEN BRC (Ibaraki, Japan). All bacteria were cultured at 37 °C overnight on LB agar plates. A mixture of 200 mL of LB and 3 g of agar was used to prepare each LB plate. Cultured colonies were selected and isolated overnight in LB at 37 °C. The suspension was centrifuged (10,000 rpm, 1 min) and washed twice with distilled water. The corresponding buffer was added to the washed cells, and the bacterial suspension was centrifuged. The concentration of the bacterial suspension was adjusted by measuring the optical density at 600 nm (OD 600 ) after vortex mixing. The 4.5 × 10 8 CFU·mL −1 S. aureus IAM1011 suspension gave OD 600 = 1.0, and the 1.0 × 10 9 CFU·mL −1 E. coli K12W3110 suspension gave OD 600 = 1.0. The generated bacterial cultures were used in the following experiments. 3.3.2. Bacterial Recognition PAMAM dendrimer probes (0.75 mL, 6.6 × 10 −6 CFU·mL −1 ) and bacterial cells in a buffer solution (0.75 mL, 4.5 × 10 8 CFU·mL −1 unless otherwise noted) were mixed, and the OD 600 was measured as a reference. Mixing was performed for 10 min at 2000 rpm. A turbidity measurement or fluorescence microscopy observation was conducted after standing for 10 min. For fluorescence microscopy, the DAPI solution was mixed with cultured bacteria in PBS, and excess dye was carefully washed off before mixing for 10 min. The PI solution was then mixed again and observed via microscopy. The change in turbidity was calculated from the difference in the optical density (ΔOD 600 ) of a sample before and after mixing. ΔOD 600 = OD 600 (after) − OD 600 (before). Turbidity change = ΔOD 600 /OD 600 (before). 3.3.3. Bacterial Separation Each bacterial suspension was mixed with a probe solution, and a bacterial recognition protocol was conducted before separation. The sample was then removed by a syringe and filtered by using a 10 μm Omnipore membrane. The resultant filtrate solution and the filter were observed separately via a microscope. 3.1. Reagents and Apparatus 3.1.1. Chemical Reagents All reagents and solvents were purchased from commercial suppliers and used without further purification unless otherwise specified. Acetone, super dehydrated (016-23465), agar (018-15811), ammonium acetate (NH 4 OAc, 016-02845), n -bromosuccinimide (NBS, 021-07232), dichloromethane (DCM, 139-02444),1,4-dioxane, super dehydrated (040-31651), ethyl acetate (EtOAc, 051-00351), hydrochloric acid (080-01066), magnesium sulfate, anhydrous (137-12335), potassium acetate (KOAc, 166-01372), 30 % potassium hydroxide solution (32904-00), sodium chloride (191-01665), sodium periodate (197-02402), and sodium sulfate (197-03345) were purchased from the FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). 4′,6-Diamidino-2-phenylindole dihydrochloride n -hydrate (DAPI, D523), ethidium bromide solution (EB, BS06), and propidium iodide solution (PI, P378) were purchased from DOJINDO LABORATORIES (Kumamoto, Japan). 2,2′-azobis(isobutyronitrile) (AIBN, A0566), [1,1′-bis(diphenylphosphino) ferrocene]palladium(II) dichloride dichloromethane adduct (Pd(dppf)Cl 2 /CH 2 Cl 2 , B2064), bis(pinacolato)diboron (B 2 pin 2 , B1964), 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMT-MM, D2919), methyl 4-bromo-3-methylbenzoate ( 1 , B2256), and N -(tert-butoxycarbonyl)-1,2-diaminoethane ( N -Boc-diaminoethane, A1371) were purchased from the Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). Acetonitrile dehydrated (MeCN, 01837-95), dimethyl sulfoxide-d 6 (DMSO-d 6 , 11560-33), methanol (MeOH, 25183-70), methanol-d 4 (99.8 atom%D, 25183-70), and deuterium water (D 2 O, 99.8 atom%D, 10086-23) were purchased from the Kanto Chemical Co. Inc. (Tokyo, Japan). Omnipore membrane 10 μm × Φ13 mm (JCWP01300), poly(amidoamine) (PAMAM) dendrimer, ethylenediamine core, generation 4.0 (412449) solution, PAMAM dendrimer, ethylenediamine core, generation 3.5 solution, 10 wt.% in methanol (412430), and Swinnex Filter Holder Φ13 mm (SX0001300) were purchased from the Sigma-Aldrich Japan Co. LLC. (Tokyo, Japan). Bacto yeast extract (212750) and Bacto Tryptone (211705) were purchased from Nippon Becton, Dickinson Co. Ltd. (Tokyo, Japan). Phosphate-buffered saline (PBS, 2101) was purchased from the Cell Science & Technology Institute Inc. (Miyagi, Japan). Water was distilled and deionized twice by using a Milli-Q water system (WG222, Yamato Scientific Co. Ltd., Tokyo, Japan and Milli-Q Advantage, Merck Millipore, Burlington, MA, USA) before use. A Spectra/Por 6 dialysis bag (MW cutoff = 1000) was purchased from the Repligen Co. (Waltham, MA, USA). 3.1.2. Apparatus 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a JEOL JNM-ECA 500 spectrometer (500 MHz) by JEOL (Tokyo, Japan) at 300 K or a Bruker Avance III HD 400 MHz by Bruker (Billerica, MA, USA) at 300 K using a deuterated solvent. All pH values were recorded by using Horiba F-52 and F-72 pH meters (Horiba, Ltd., Kyoto, Japan). Ultraviolet−visible (UV−Vis) absorption spectra were measured by using a JASCO V-570 or V-760 UV−Vis spectrophotometer (JASCO Co., Tokyo, Japan) equipped with a Peltier Thermo controller and a 10 mm quartz cell. Zeta potential measurements were carried out at 25 °C by using a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, Worcestershire, U.K.). Samples were shaken by using a MULTI SHAKER MS-300 (AS ONE Co., Osaka, Japan). Centrifugation was conducted by using a CF15RN (Hitachi High-Technologies, Co., Tokyo, Japan). The p K a value was determined by using Igor Pro v5.0.3.0 (Wavemetrics inc., Portland, OR, USA) based on the acid dissociation model of monobasic acids. 3.1.1. Chemical Reagents All reagents and solvents were purchased from commercial suppliers and used without further purification unless otherwise specified. Acetone, super dehydrated (016-23465), agar (018-15811), ammonium acetate (NH 4 OAc, 016-02845), n -bromosuccinimide (NBS, 021-07232), dichloromethane (DCM, 139-02444),1,4-dioxane, super dehydrated (040-31651), ethyl acetate (EtOAc, 051-00351), hydrochloric acid (080-01066), magnesium sulfate, anhydrous (137-12335), potassium acetate (KOAc, 166-01372), 30 % potassium hydroxide solution (32904-00), sodium chloride (191-01665), sodium periodate (197-02402), and sodium sulfate (197-03345) were purchased from the FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). 4′,6-Diamidino-2-phenylindole dihydrochloride n -hydrate (DAPI, D523), ethidium bromide solution (EB, BS06), and propidium iodide solution (PI, P378) were purchased from DOJINDO LABORATORIES (Kumamoto, Japan). 2,2′-azobis(isobutyronitrile) (AIBN, A0566), [1,1′-bis(diphenylphosphino) ferrocene]palladium(II) dichloride dichloromethane adduct (Pd(dppf)Cl 2 /CH 2 Cl 2 , B2064), bis(pinacolato)diboron (B 2 pin 2 , B1964), 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMT-MM, D2919), methyl 4-bromo-3-methylbenzoate ( 1 , B2256), and N -(tert-butoxycarbonyl)-1,2-diaminoethane ( N -Boc-diaminoethane, A1371) were purchased from the Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). Acetonitrile dehydrated (MeCN, 01837-95), dimethyl sulfoxide-d 6 (DMSO-d 6 , 11560-33), methanol (MeOH, 25183-70), methanol-d 4 (99.8 atom%D, 25183-70), and deuterium water (D 2 O, 99.8 atom%D, 10086-23) were purchased from the Kanto Chemical Co. Inc. (Tokyo, Japan). Omnipore membrane 10 μm × Φ13 mm (JCWP01300), poly(amidoamine) (PAMAM) dendrimer, ethylenediamine core, generation 4.0 (412449) solution, PAMAM dendrimer, ethylenediamine core, generation 3.5 solution, 10 wt.% in methanol (412430), and Swinnex Filter Holder Φ13 mm (SX0001300) were purchased from the Sigma-Aldrich Japan Co. LLC. (Tokyo, Japan). Bacto yeast extract (212750) and Bacto Tryptone (211705) were purchased from Nippon Becton, Dickinson Co. Ltd. (Tokyo, Japan). Phosphate-buffered saline (PBS, 2101) was purchased from the Cell Science & Technology Institute Inc. (Miyagi, Japan). Water was distilled and deionized twice by using a Milli-Q water system (WG222, Yamato Scientific Co. Ltd., Tokyo, Japan and Milli-Q Advantage, Merck Millipore, Burlington, MA, USA) before use. A Spectra/Por 6 dialysis bag (MW cutoff = 1000) was purchased from the Repligen Co. (Waltham, MA, USA). 3.1.2. Apparatus 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a JEOL JNM-ECA 500 spectrometer (500 MHz) by JEOL (Tokyo, Japan) at 300 K or a Bruker Avance III HD 400 MHz by Bruker (Billerica, MA, USA) at 300 K using a deuterated solvent. All pH values were recorded by using Horiba F-52 and F-72 pH meters (Horiba, Ltd., Kyoto, Japan). Ultraviolet−visible (UV−Vis) absorption spectra were measured by using a JASCO V-570 or V-760 UV−Vis spectrophotometer (JASCO Co., Tokyo, Japan) equipped with a Peltier Thermo controller and a 10 mm quartz cell. Zeta potential measurements were carried out at 25 °C by using a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, Worcestershire, U.K.). Samples were shaken by using a MULTI SHAKER MS-300 (AS ONE Co., Osaka, Japan). Centrifugation was conducted by using a CF15RN (Hitachi High-Technologies, Co., Tokyo, Japan). The p K a value was determined by using Igor Pro v5.0.3.0 (Wavemetrics inc., Portland, OR, USA) based on the acid dissociation model of monobasic acids. 3.2. Preparation of Dendrimer Probes 3.2.1. Synthesis of Methyl 3-Methyl-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 2 ) Compound 2 was synthesized by using commercially available compound 1 [ 39 ] under an argon atmosphere. Bis(pinacolato)diboron (1.1 eq, 1.68 g, and 6.6 mmol), KOAc (3.1 eq, 1.85 g, and 19 mmol), and Pd(dppf)Cl 2 /CH 2 Cl 2 (2 mol%, 0.11 g, and 0.14 mmol) were added to the solution produced by dissolving compound 1 (1.0 eq, 1.38 g, and 6.0 mmol) in 1,4-dioxane (10.0 mL). The resulting orange reaction solution was stirred vigorously at 95 °C for 22 h before undesired precipitates were removed through filtration. The filtrate was then evaporated, and the product was extracted three times with DCM/water. The combined organic layers were dried over MgSO 4 and concentrated in vacuo. Purification by silica gel column chromatography (DCM 100%) was then used to generate product 2 as a yellow oil (684 mg, 2.5 mmol, and 40%). The structure was confirmed from the resulting 1 H NMR spectrum ( Figure S5 ). 3.2.2. Synthesis of Methyl 3-(Bromomethyl)-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 3 ) Compound 3 was synthesized by using compound 2 under an argon atmosphere [ 39 ]. NBS (1.3 eq, 695 mg, and 3.9 mmol) and AIBN (2 mol%, 10 mg, and 0.06 mmol) were added to the solution generated by dissolving compound 2 (1.0 eq, 861 mg, and 3.1 mmol) in MeCN (15.0 mL). The solution was stirred at 90 °C for 2.5 h before the solvent was evaporated. DCM was added to the concentrated solution to remove the undesired white precipitation. Purification by silica gel column chromatography (DCM 100%) generated product 3 in the form of a yellowish oil (580 mg, 1.6 mmol, and 52%). The structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S6 ). 3.2.3. Synthesis of 1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxylic Acid ( 4 ) Compound 4 was synthesized by using compound 3 [ 39 ], which (1.0 eq, 580 mg, and 1.6 mmol) was dissolved in an acetone/water mixture at 1/1 ( v / v ). NaIO 4 (4.9 eq, 1.74 g, and 8.1 mmol) and NH 4 OAc (5.0 eq, 630 mg, and 8.1 mmol) were added to the produced solution, and the mixture was stirred at rt for 1 day. The solvent was then evaporated, and the product was extracted three times with EtOAc/water. The combined organic layers were then dried over MgSO 4 and concentrated in vacuo. Afterwards, 15% KOH aq. (3.0 mL) was added to the obtained product and stirred for 1.5 h at rt, followed by acidification via HCl aq. The obtained white precipitation, compound 4 (152 mg, 0.83 mmol, and 52%), was extracted by filtration, and the structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S7 ). 3.2.4. Synthesis of BenzoB-PAMAM(+) Compound 4 (8.0 eq, 11.3 mg, and 64 μmol) and DMT-MM (40.0 eq, 88.2 mg, and 0.32 mmol) were dissolved in methanol (10.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G4) dendrimer (1.4 mL, 1.0 eq, and 8.0 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days before the purified product was lyophilized to produce white flocks (124.5 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S8 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.2.5. Synthesis of N -(2-Aminoethyl)-1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxamide ( 5 ) N -Boc-diaminoethane (96 mg, 1.2 eq, and 0.6 mmol) and synthetic product 4 (89 mg, 1.0 eq, and 0.5 mmol) were dissolved in methanol (3 mL). The mixture was stirred at rt for 30 min [ 35 ]. DMT-MM (550 mg, 4.0 eq, and 2.0 mmol) was added to the solution, and a condensation reaction was performed at rt for 4 days. The product of the reaction was extracted three times with EtOAc, washed with brine, and dried over Na 2 SO 4 . The solvent was removed by evaporation, and HCl in methanol (4 M) was added to the residue. The solution was stirred at rt for 3 h to induce deprotection. The solvent was evaporated, and the final product 5 was obtained as a white solid (212 mg). The undesired byproducts were excluded by dialysis in the following synthesis. The identity was confirmed by ESI-HRMS spectral measurement and the resulting 1 H and 13 C NMR ( Figures S9 and S10 ) spectra. The ESI-HRMS ( m / z ) calculated for C 11 H 16 B 1 N 2 O 3 [M + CH 3 ] + 235.1254 found a figure of 235.1277. 3.2.6. Synthesis of BenzoB-PAMAM(−) Compound 5 (8.0 eq, 11.7 mg, and 49.5 μmol) and DMT-MM (32.0 eq, 54.6 mg, and 0.20 mmol) were dissolved in methanol (5.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G3.5) dendrimer (1.0 mL, 1.0 eq, and 6.2 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was then transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days. The purified product was lyophilized to give white flocks (62.2 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S11 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.2.1. Synthesis of Methyl 3-Methyl-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 2 ) Compound 2 was synthesized by using commercially available compound 1 [ 39 ] under an argon atmosphere. Bis(pinacolato)diboron (1.1 eq, 1.68 g, and 6.6 mmol), KOAc (3.1 eq, 1.85 g, and 19 mmol), and Pd(dppf)Cl 2 /CH 2 Cl 2 (2 mol%, 0.11 g, and 0.14 mmol) were added to the solution produced by dissolving compound 1 (1.0 eq, 1.38 g, and 6.0 mmol) in 1,4-dioxane (10.0 mL). The resulting orange reaction solution was stirred vigorously at 95 °C for 22 h before undesired precipitates were removed through filtration. The filtrate was then evaporated, and the product was extracted three times with DCM/water. The combined organic layers were dried over MgSO 4 and concentrated in vacuo. Purification by silica gel column chromatography (DCM 100%) was then used to generate product 2 as a yellow oil (684 mg, 2.5 mmol, and 40%). The structure was confirmed from the resulting 1 H NMR spectrum ( Figure S5 ). 3.2.2. Synthesis of Methyl 3-(Bromomethyl)-4-(4,4,5,5-Tetramethyl-1,3,2-Dioxaborolan-2-Yl)Benzoate ( 3 ) Compound 3 was synthesized by using compound 2 under an argon atmosphere [ 39 ]. NBS (1.3 eq, 695 mg, and 3.9 mmol) and AIBN (2 mol%, 10 mg, and 0.06 mmol) were added to the solution generated by dissolving compound 2 (1.0 eq, 861 mg, and 3.1 mmol) in MeCN (15.0 mL). The solution was stirred at 90 °C for 2.5 h before the solvent was evaporated. DCM was added to the concentrated solution to remove the undesired white precipitation. Purification by silica gel column chromatography (DCM 100%) generated product 3 in the form of a yellowish oil (580 mg, 1.6 mmol, and 52%). The structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S6 ). 3.2.3. Synthesis of 1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxylic Acid ( 4 ) Compound 4 was synthesized by using compound 3 [ 39 ], which (1.0 eq, 580 mg, and 1.6 mmol) was dissolved in an acetone/water mixture at 1/1 ( v / v ). NaIO 4 (4.9 eq, 1.74 g, and 8.1 mmol) and NH 4 OAc (5.0 eq, 630 mg, and 8.1 mmol) were added to the produced solution, and the mixture was stirred at rt for 1 day. The solvent was then evaporated, and the product was extracted three times with EtOAc/water. The combined organic layers were then dried over MgSO 4 and concentrated in vacuo. Afterwards, 15% KOH aq. (3.0 mL) was added to the obtained product and stirred for 1.5 h at rt, followed by acidification via HCl aq. The obtained white precipitation, compound 4 (152 mg, 0.83 mmol, and 52%), was extracted by filtration, and the structure was confirmed by using the resulting 1 H NMR spectrum ( Figure S7 ). 3.2.4. Synthesis of BenzoB-PAMAM(+) Compound 4 (8.0 eq, 11.3 mg, and 64 μmol) and DMT-MM (40.0 eq, 88.2 mg, and 0.32 mmol) were dissolved in methanol (10.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G4) dendrimer (1.4 mL, 1.0 eq, and 8.0 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days before the purified product was lyophilized to produce white flocks (124.5 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S8 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.2.5. Synthesis of N -(2-Aminoethyl)-1-Hydroxy-1,3-Dihydro-2,1-Benzoxaborole-5-Carboxamide ( 5 ) N -Boc-diaminoethane (96 mg, 1.2 eq, and 0.6 mmol) and synthetic product 4 (89 mg, 1.0 eq, and 0.5 mmol) were dissolved in methanol (3 mL). The mixture was stirred at rt for 30 min [ 35 ]. DMT-MM (550 mg, 4.0 eq, and 2.0 mmol) was added to the solution, and a condensation reaction was performed at rt for 4 days. The product of the reaction was extracted three times with EtOAc, washed with brine, and dried over Na 2 SO 4 . The solvent was removed by evaporation, and HCl in methanol (4 M) was added to the residue. The solution was stirred at rt for 3 h to induce deprotection. The solvent was evaporated, and the final product 5 was obtained as a white solid (212 mg). The undesired byproducts were excluded by dialysis in the following synthesis. The identity was confirmed by ESI-HRMS spectral measurement and the resulting 1 H and 13 C NMR ( Figures S9 and S10 ) spectra. The ESI-HRMS ( m / z ) calculated for C 11 H 16 B 1 N 2 O 3 [M + CH 3 ] + 235.1254 found a figure of 235.1277. 3.2.6. Synthesis of BenzoB-PAMAM(−) Compound 5 (8.0 eq, 11.7 mg, and 49.5 μmol) and DMT-MM (32.0 eq, 54.6 mg, and 0.20 mmol) were dissolved in methanol (5.0 mL), and the reaction mixture was stirred at rt for 30 min [ 35 ]. The PAMAM(G3.5) dendrimer (1.0 mL, 1.0 eq, and 6.2 μmol) was added, and the reaction mixture was stirred at rt for 2 days. The reaction mixture was then transferred into a Spectra/Por 6 dialysis bag and dialyzed against methanol as well as distilled water for several days. The purified product was lyophilized to give white flocks (62.2 mg), the chemical structure of which was confirmed by a 1 H NMR measurement ( Figure S11 ). The number of benzoxaborole substituents was estimated from the corresponding peak area in the 1 H NMR spectrum. 3.3. Biological Experiments 3.3.1. Bacterial Culture The lysogeny broth (LB) was composed of 2 g of Bacto Tryptone, 1 g of Bacto yeast extract, and 2 g of NaCl in 200 mL of distilled water. S. aureus IAM1011, S. aureus ATCC25923, S. aureus ATCC29213, S. pseudintermedius 2012-S-27, S. epidermidis ATCC12228, Enterococcus faecalis ATCC29212, E. coli K12W3110, E. coli ATCC25922, Pseudomonas aeruginosa ATCC27853, and Salmonella enteritidis ATCC13311 were provided by RIKEN BRC (Ibaraki, Japan). All bacteria were cultured at 37 °C overnight on LB agar plates. A mixture of 200 mL of LB and 3 g of agar was used to prepare each LB plate. Cultured colonies were selected and isolated overnight in LB at 37 °C. The suspension was centrifuged (10,000 rpm, 1 min) and washed twice with distilled water. The corresponding buffer was added to the washed cells, and the bacterial suspension was centrifuged. The concentration of the bacterial suspension was adjusted by measuring the optical density at 600 nm (OD 600 ) after vortex mixing. The 4.5 × 10 8 CFU·mL −1 S. aureus IAM1011 suspension gave OD 600 = 1.0, and the 1.0 × 10 9 CFU·mL −1 E. coli K12W3110 suspension gave OD 600 = 1.0. The generated bacterial cultures were used in the following experiments. 3.3.2. Bacterial Recognition PAMAM dendrimer probes (0.75 mL, 6.6 × 10 −6 CFU·mL −1 ) and bacterial cells in a buffer solution (0.75 mL, 4.5 × 10 8 CFU·mL −1 unless otherwise noted) were mixed, and the OD 600 was measured as a reference. Mixing was performed for 10 min at 2000 rpm. A turbidity measurement or fluorescence microscopy observation was conducted after standing for 10 min. For fluorescence microscopy, the DAPI solution was mixed with cultured bacteria in PBS, and excess dye was carefully washed off before mixing for 10 min. The PI solution was then mixed again and observed via microscopy. The change in turbidity was calculated from the difference in the optical density (ΔOD 600 ) of a sample before and after mixing. ΔOD 600 = OD 600 (after) − OD 600 (before). Turbidity change = ΔOD 600 /OD 600 (before). 3.3.3. Bacterial Separation Each bacterial suspension was mixed with a probe solution, and a bacterial recognition protocol was conducted before separation. The sample was then removed by a syringe and filtered by using a 10 μm Omnipore membrane. The resultant filtrate solution and the filter were observed separately via a microscope. 3.3.1. Bacterial Culture The lysogeny broth (LB) was composed of 2 g of Bacto Tryptone, 1 g of Bacto yeast extract, and 2 g of NaCl in 200 mL of distilled water. S. aureus IAM1011, S. aureus ATCC25923, S. aureus ATCC29213, S. pseudintermedius 2012-S-27, S. epidermidis ATCC12228, Enterococcus faecalis ATCC29212, E. coli K12W3110, E. coli ATCC25922, Pseudomonas aeruginosa ATCC27853, and Salmonella enteritidis ATCC13311 were provided by RIKEN BRC (Ibaraki, Japan). All bacteria were cultured at 37 °C overnight on LB agar plates. A mixture of 200 mL of LB and 3 g of agar was used to prepare each LB plate. Cultured colonies were selected and isolated overnight in LB at 37 °C. The suspension was centrifuged (10,000 rpm, 1 min) and washed twice with distilled water. The corresponding buffer was added to the washed cells, and the bacterial suspension was centrifuged. The concentration of the bacterial suspension was adjusted by measuring the optical density at 600 nm (OD 600 ) after vortex mixing. The 4.5 × 10 8 CFU·mL −1 S. aureus IAM1011 suspension gave OD 600 = 1.0, and the 1.0 × 10 9 CFU·mL −1 E. coli K12W3110 suspension gave OD 600 = 1.0. The generated bacterial cultures were used in the following experiments. 3.3.2. Bacterial Recognition PAMAM dendrimer probes (0.75 mL, 6.6 × 10 −6 CFU·mL −1 ) and bacterial cells in a buffer solution (0.75 mL, 4.5 × 10 8 CFU·mL −1 unless otherwise noted) were mixed, and the OD 600 was measured as a reference. Mixing was performed for 10 min at 2000 rpm. A turbidity measurement or fluorescence microscopy observation was conducted after standing for 10 min. For fluorescence microscopy, the DAPI solution was mixed with cultured bacteria in PBS, and excess dye was carefully washed off before mixing for 10 min. The PI solution was then mixed again and observed via microscopy. The change in turbidity was calculated from the difference in the optical density (ΔOD 600 ) of a sample before and after mixing. ΔOD 600 = OD 600 (after) − OD 600 (before). Turbidity change = ΔOD 600 /OD 600 (before). 3.3.3. Bacterial Separation Each bacterial suspension was mixed with a probe solution, and a bacterial recognition protocol was conducted before separation. The sample was then removed by a syringe and filtered by using a 10 μm Omnipore membrane. The resultant filtrate solution and the filter were observed separately via a microscope. 4. Conclusions Bacterial separation was reported by using the benzoxaborole-based dendrimer probe BenzoB-PAMAM(+), which selectively recognized Gram-positive bacteria. BenzoB-PAMAM(+) was newly synthesized by the condensation of carboxy-benzoxaborole and the PAMAM(G4) dendrimer, and was found to recognize the bacterial saccharide that is part of LTA on a Gram-positive bacterial surface over a wide pH range with the help of an electrostatic interaction. The benzoxaborole recognition site showed a desirable low p K a value, and might thus result in good selectivity. BenzoB-PAMAM(+) led to the development of large aggregations of Gram-positive bacteria, whereas aggregation was either not observed or minute in size (>10 μm) for Gram-negative bacteria. The selectivity and size of the Gram-positive bacterial aggregations enabled separation by using a 10 μm membrane filter. The presence of Gram-negative bacteria in the filtrate solution was also confirmed. It is true that a small number of Gram-negative bacteria were involved in the aggregation, but most Gram-negative bacteria were successfully separated by filtration. The collection efficacy for each bacterial species needs to be investigated in future studies. In summary, BenzoB-PAMAM(+), which was designed as a novel aggregation-based and metal-free separation method, successfully recognized and collected Gram-positive bacteria, demonstrating its potential for application in bacterial separation and concentration from environmental soil or water. We believe that these findings contribute significantly to the study of AMR bacterial distributions in the environment. Figures and Schemes Figure 1 Schematic image of selective bacterial recognition by boronic acid-modified PAMAM dendrimer probes. Figure 2 Structures of boronic acid-based PAMAM dendrimer probes. molecules-28-01704-sch001_Scheme 1 Scheme 1 Synthesis of the benzoxaborole segment and BenzoB-PAMAM(+). molecules-28-01704-sch002_Scheme 2 Scheme 2 Synthesis of BenzoB-PAMAM(−). Figure 3 Zeta potential measurement of BenzoB-PAMAMs in PBS ( n = 5). [probe] = 6.6 × 10 −6 M. Bars express standard deviations (SDs) from five independent experiments. ( A ) BenzoB-PAMAM(−). ( B ) BenzoB-PAMAM(+). Figure 4 Turbidity changes in bacterial suspensions containing BenzoB-PAMAMs in PBS ( n = 3). [BenzoB-PAMAM(−) or BenzoB-PAMAM(+)] = 3.3 × 10 −6 M. [ S. aureus IAM1011 or E. coli K12W3110] = 2.3 × 10 8 CFU·mL −1 . Bars represent the SD in three separate experiments. ( A ) BenzoB-PAMAM(−). ( B ) BenzoB-PAMAM(+). Figure 5 Aggregate formation for BenzoB-PAMAM(+) in PBS. [BenzoB-PAMAM(+)] = 3.3 × 10 −6 M, [ S. aureus IAM1011] = 2.3 × 10 8 CFU·mL −1 ; aggregates are highlighted using red squares. ( A ) pH of 4.0; ( B ) pH of 5.0; and ( C ) pH of 6.0. Figure 6 Microscope observation of S. aureus IAM1011 and E. coli K12W3110 at a pH of 7.4, adjusted with PBS. [BenzoB-PAMAM(+)] = 3.3 × 10 −6 M or none; bacterial concentration was set at 2.3 × 10 8 CFU·mL −1 for S. aureus IAM1011 and 5.0 × 10 8 CFU·mL −1 for E. coli K12W3110. From left to right: DIC, DAPI, PI, and merged microscopy images. Scale bar = 200 µm in ( A ) and 100 µm in ( B – D ). Figure 7 UV–Vis absorbance results for benzoxaborole 5 in PBS from pHs of 4.0 to 11.0. [compound 5 ] = 10.0 × 10 −6 M. ( A ) UV–Vis absorption spectra under various pH conditions. ( B ) Summary at 266 nm from which the p K a value was calculated. Figure 8 Turbidity experiments using various bacteria at a pH of 7.4, adjusted with PBS ( n = 3). [BenzoB-PAMAM(+)] = none or 3.3 × 10 −6 M; bacterial concentration was set to OD 600 = 0.3. Bars represent SDs from three independent experiments. Figure 9 Microscope images of Gram-positive bacteria BenzoB-PAMAM(+) at a pH of 7.4, adjusted with PBS. [BenzoB-PAMAM(+)] = 3.3 × 10 −6 M; bacterial concentration was set at OD 600 = 0.3. From left to right: DIC, DAPI, and merged microscopy images. Scale bar = 100 µm. ( A ) S. aureus ATCC25923. ( B ) S. aureus ATCC29213. ( C ) S. pseudintermedius 2012-S-27. ( D ) S. epidermidis ATCC12228. ( E ) E. faecalis ATCC29212. Figure 10 Microscope images of Gram-negative bacteria obtained using BenzoB-PAMAM(+) at a pH of 7.4, adjusted with PBS. [BenzoB-PAMAM(+)] = 3.3 × 10 −6 M; bacterial concentration was set at OD 600 = 0.3. From left to right: DIC, DAPI, and merged microscopy images. Scale bar = 100 µm. ( A ) E. coli ATCC25922. ( B ) P. aeruginosa ATCC27853. ( C ) S. enteritidis ATCC13311. Figure 11 Microscope images of filter and filtrate using BenzoB-PAMAM(+) at a pH of 7.4, adjusted with PBS. [BenzoB-PAMAM(+)] = 3.3 × 10 −6 M. From left to right: merged microscopy images of filter and filtrate samples. ( A ) 1.4 × 10 8 CFU·mL −1 of S. aureus IAM1011 stained with DAPI. ( B ) 3.0 × 10 8 CFU·mL −1 E. coli K12W3110 stained with DAPI. ( C ) 1.0 × 10 8 CFU·mL −1 S. aureus IAM1011 (DAPI) and E. coli K12W3110 (EB). ( D ) 1.4 × 10 8 CFU·mL −1 S. aureus IAM1011 (DAPI) and 3.0 × 10 8 CFU·mL −1 E. coli K12W3110 (EB).

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5785212/

Surface display of OmpC of Salmonella serovar Pullorum on Bacillus subtilis spores

Salmonellosis is a major public health problem throughout the world. Thus, there is a huge need for diversified control strategies for Salmonella infections. In this work, we have assessed the potential use of Bacillus subtilis ( B . subtilis ) spores for the expression of a major protective antigen of Salmonella serovar Pullorum, OmpC. The expression of OmpC on the surface of spores was determined by immunofluorescence microscopy. Mice immunized with recombinant spores expressing the OmpC antigen presented significant levels of OmpC-specific serum IgG and mucosal SIgA antibodies than in mice immunized with non-recombinant spores ( p <0.01). In addition, oral immunization with recombinant spores was able to induce a significant level of protection in mice against lethal challenge with Salmonella serovar Typhimurium. These results suggest that B . subtilis spores have promising potential in the development of mucosal vaccines against Salmonella infections. Introduction Pullorum disease (PD) is a worldwide poultry disease that caused enormous economic losses throughout large areas of the world. Salmonella serovar Pullorum ( S . Pullorum) is the etiological agent of PD, which is an acute septicemic disease that results in anorexia, diarrhea, dehydration, weakness and high mortality in chicks and poults, and loss of weight, decreased egg production and hatchability, diarrhea, and lesions and abnormalities of the reproductive tract in mature fowl [ 1 , 2 ]. Although different strategies to control Salmonella infections are currently available, continuous emergence of multi-drug resistance [ 3 , 4 ] and novel Salmonella variants [ 5 , 6 ] are still a mortal threat to the poultry industry and public health. Thus, it is urgently needed to develop new control strategies for PD in poultry. Since most pathogenic organisms, including Salmonella , penetrate to the host through the mucosal membranes, effective mucosal vaccines that induce immunity at the site of infection are able to confer more efficient protective immune responses against Salmonella [ 7 , 8 ]. There is now much evidence that OmpC, an outer membrane protein (porin) from Salmonella , is a promising candidate antigen that efficiently stimulates innate and adaptive immune responses [ 9 , 10 ]. However, without suitable antigen delivery systems and adjuvants, most protein antigens are poorly immunogenic when mucosal immunization [ 11 ]. B . subtilis is a non-pathogenic aerobic Gram-positive and endospore-forming bacterium that generally regarded as safe (GRAS) [ 12 ]. B . subtilis spores, an extremely stable form under harsh life conditions, are covered with a multilayered coat, which composed of at least 70 different protein species [ 13 ]. Previous studies have confirmed that spore coat protein can be employed as fusion partner for expression and display of vaccine antigens on the spore surface, and protective systemic and mucosal immune responses were elicited following oral or intranasal administration of recombinant B . subtilis spores without adjuvants [ 14 – 18 ]. Moreover, the spores presenting antigens also have the ability to induce both antigen specific CD4+ and CD8+ T cell cellular immune responses [ 18 – 20 ]. Excellent resistance properties and safety, together with several other attractive advantages, such as a readily genetic manipulation, low production cost, and easy administration, transport and storage, make B . subtilis spore an ideal candidate for the expression and delivery of vaccine antigens to complex and rigorous mucosal surfaces where antigens are sampled. In the present study, we successfully constructed a recombinant B . subtilis spores expressing immunogenic antigen OmpC of S . Pullorum on the spore surface. Mice were orally immunized with recombinant spores to evaluate the mucosal and systemic antigen-specific immune responses. In addition, the protective efficacy of recombinant spores was also investigated in mice using a challenge experiment. This work indicates that spore-based expression and delivery system of vaccine antigens, as a novel strategy, has promising future in protection of mucosal surfaces against invasion by Salmonella . Materials and methods Ethics statement All animal experiment procedures were carried out in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the Chinese Association for Laboratory Animal Sciences ( http://www.calas.org.cn/ ). The Institutional Animal Care and Use Committee (IACUC) at the Sichuan Agricultural University approved all the procedures used in this study (Protocol NO. DKY-S20123517). Mice Female BALB/c mice (specific pathogen-free, SPF) aged approximately 6–8 weeks were obtained from the Vital River, Beijing, China. Mice were raised under SPF conditions at 24±2°C with a light-controlled regimen (12-hour light/dark cycle). Mice were anesthetized intraperitoneally with ketamine (40 mg/kg), and all efforts were made to minimize suffering. In the challenge experiment, mice were closely monitored for signs of Salmonella infection, and mice that developed clear clinical symptoms or any signs of infections include diarrhea, fever, and abdominal cramps were humanely euthanized. Bacterial strains S . Pullorum strain CVCC533 and S . Typhimurium strain SL1344 were issued by China Institute of Veterinary Drugs Control (Beijing, China). The integrative vector pDG364 and B . subtilis strain 168 ( trp ‾ ) were obtained from Bacillus Genetic Stock Center (BGSC) (Columbus, OH). The subcloning vector pMD19-T and Escherichia coli strain DH5α were purchased from TaKaRa (Dalian, China). Construction of integrative recombinant plasmid The integrative recombinant plasmid was constructed by introducing cotC :: ompC gene fusion into the pDG364 ( Fig 1A ). Firstly, a 1078 bp DNA fragment coding for OmpC (GenBank accession NO. CP_003047) was PCR amplified using S . Pullorum chromosome as a template and oligonucleotides ompC-F and ompC-R as primers ( Table 1 ). The PCR product was sequentially digested with Hind ⠢ and Eco RI and cloned into the pDG364 previously digested with the same enzymes, yielding plasmid pDG364- ompC . A purified cotC gene (GenBank accession NO. NC_000964) containing the promoter sequence and the whole coding sequence was amplified by PCR using the B . subtilis chromosomal as template and oligonucleotides cotC-F and cotC-R as primers ( Table 1 ). The PCR product was sequentially digested with Bam HI and Hind ⠢ and cloned in frame to the 5‵ end of the ompC gene carried by plasmid pDG364- ompC , yielding plasmid pDG364- cotC - ompC . 10.1371/journal.pone.0191627.g001 Fig 1 Schematic representation of the construction of the recombinant spores. The cotC :: ompC gene fusion and cat (chloramphenicol-resistance gene) gene carried by plasmid pDG364- cotC - ompC were integrated into the amyE gene locus of B . subtilis 168 chromosome by double cross-over recombination events. Arrows indicate direction of transcription. 10.1371/journal.pone.0191627.t001 Table 1 Oligonucleotides list. Name Sequence a Restriction site cotC-F CG GGATCC TGTAGGATAAATCGTTTGGGC Bam HI cotC-R GGGGGGG AAGCTT GTAGTGTTTTTTATGCTTTTTATAC Hind ⠢ ompC-F CCC AAGCTT AATAAAGACGGCAACAAATTAGACC Hind ⠢ ompC-R C GAATTC TTAGAACTGGTAAACCAGACCC Eco RI amyE-F CCAATGAGGTTAAGAGTATTCC Null amyE-R CGAGAAGCTATCACCGCCCAGC Null a The underlined letters indicate the introduction of restriction sites. Chromosomal integration Plasmid pDG364- cotC - ompC was linearize by digestion with Xba I and used to transform competent cells of the B . subtilis strain 168 under previously described procedures [ 21 ]. Chloramphenicol-resistant recombinant B . subtilis were obtained by double-crossover recombination event ( Fig 1B ), and several clones for each transformation were tested by amylase activity analysis [ 21 ] and then confirmed by PCR using chromosomal DNA as template and four oligonucleotides (ompC-F and ompC-R, amyE-F and amyE-R, amyE-F and ompC-R, ompC-F and amyE-R) ( Table 1 ) to prime the reaction. Immunofluorescence microscopy The expression and display of fusion protein CotC-OmpC on the spore surface was confirmed using immunofluorescence microscopy, about 10 μl of spore suspension was prepared and fixed on slides as methods adapted from previous report [ 22 , 23 ]. Samples were blocked for 30min with 3% (w/v) bovine serum albumin (BSA) in PBS (pH7.4) at room temperature (RT) and then washed ten times with PBS. The samples were incubated overnight at 4°C with OmpC antiserum (raised in chicken), washed ten times and then incubated further with Cy3-labeled goat anti-chicken IgG (Sangon Biotech, 1:2000 in PBST) for 45 min at RT. After washing procedures, samples were viewed and photographed under fluorescent microscope (Eclipse, TE2000U, Nikon). Immunization in mice A total of 120 mice were randomly divided into 5 groups (24 for each): Group A and group B were dosed orally (0.2 ml) with spores of either recombinant B . subtilis or isogenic wild-type B . subtilis 168 by gavage (1.0×10 10 CFU spores/time/ mouse) at 0, 15 days respectively; group C and group D were administrated orally by feeding the diet mixed with spores of either recombinant B . subtilis or isogenic wild-type B . subtilis 168 (1.0×10 6 CFU spores/g); a naïve, untreated control group was also included (group E). Indirect ELISA for detection of OmpC-specific serum and mucosal antibodies On 1 day before and 22 days after the first dose, mice from each group were anesthetized with an intraperitoneal ketamine injection at 40 mg/kg body weight, bled retro-orbitally, and sacrificed by cervical dislocation. Intestinal content samples were immediately collected from each mouse in the sterile operation. Samples were tested by ELISA as previously described [ 24 – 26 ] with the following modifications: Flat-bottom microtiter plates (high-binding capacity; Sangon Biotech, China) were coated with 100μl of purified S . Pullorum strain CVCC533 (1×10 9 CFU/ml) in carbonate buffer solution (pH9.6) overnight at RT. After the plates were blocked for 1h at 37°C with 3% BSA, serum samples were applied as a 1/50 dilution in diluted buffer. Replicate samples were utilized together with a negative control (pre-immune serum). After incubation at 37°C for 1h, the plates were washed and then incubated with HRP-conjugate rabbit anti-mouse IgG (Santa Cruz Biotechnology). Plates were incubated for 1h at RT and then developed using TMB substrate. Reactions were stopped using 2M H 2 SO 4 , and optical densities (ODs) were read at 450 nm. ELISA of small intestinal content samples was tested with a similar method, and IgA was detected by using HRP-conjugate goat anti-mouse IgA (Santa Cruz Biotechnology). Protection studies 22 days after the first dose, 16 mice of each group mentioned above were randomly chosen and divided into two groups and were challenged with approximately 4×10 8 CFU (2×LD 50 ) and 2×10 9 CFU (10×LD 50 ) [ 27 ]of S . Typhimurium strain SL1344 by the intra-peritoneal rout, respectively. The animals were strictly monitored 72h after challenge, and individuals showing clear clinical symptoms or any signs of S . Typhimurium infections include diarrhea, fever, and abdominal cramps were considered unprotected and humanely euthanized by overdose of ketamine. Statistical analysis Samples were tested individually, and results were expressed as the arithmetic mean ± S.D. (standard deviation). Statistical significance of the difference between group means was assessed by unpaired Student's t -test, and the significance level was set at P ≤0.05. Statistical analysis was performed with SPSS for Windows version 16.0. Ethics statement All animal experiment procedures were carried out in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the Chinese Association for Laboratory Animal Sciences ( http://www.calas.org.cn/ ). The Institutional Animal Care and Use Committee (IACUC) at the Sichuan Agricultural University approved all the procedures used in this study (Protocol NO. DKY-S20123517). Mice Female BALB/c mice (specific pathogen-free, SPF) aged approximately 6–8 weeks were obtained from the Vital River, Beijing, China. Mice were raised under SPF conditions at 24±2°C with a light-controlled regimen (12-hour light/dark cycle). Mice were anesthetized intraperitoneally with ketamine (40 mg/kg), and all efforts were made to minimize suffering. In the challenge experiment, mice were closely monitored for signs of Salmonella infection, and mice that developed clear clinical symptoms or any signs of infections include diarrhea, fever, and abdominal cramps were humanely euthanized. Bacterial strains S . Pullorum strain CVCC533 and S . Typhimurium strain SL1344 were issued by China Institute of Veterinary Drugs Control (Beijing, China). The integrative vector pDG364 and B . subtilis strain 168 ( trp ‾ ) were obtained from Bacillus Genetic Stock Center (BGSC) (Columbus, OH). The subcloning vector pMD19-T and Escherichia coli strain DH5α were purchased from TaKaRa (Dalian, China). Construction of integrative recombinant plasmid The integrative recombinant plasmid was constructed by introducing cotC :: ompC gene fusion into the pDG364 ( Fig 1A ). Firstly, a 1078 bp DNA fragment coding for OmpC (GenBank accession NO. CP_003047) was PCR amplified using S . Pullorum chromosome as a template and oligonucleotides ompC-F and ompC-R as primers ( Table 1 ). The PCR product was sequentially digested with Hind ⠢ and Eco RI and cloned into the pDG364 previously digested with the same enzymes, yielding plasmid pDG364- ompC . A purified cotC gene (GenBank accession NO. NC_000964) containing the promoter sequence and the whole coding sequence was amplified by PCR using the B . subtilis chromosomal as template and oligonucleotides cotC-F and cotC-R as primers ( Table 1 ). The PCR product was sequentially digested with Bam HI and Hind ⠢ and cloned in frame to the 5‵ end of the ompC gene carried by plasmid pDG364- ompC , yielding plasmid pDG364- cotC - ompC . 10.1371/journal.pone.0191627.g001 Fig 1 Schematic representation of the construction of the recombinant spores. The cotC :: ompC gene fusion and cat (chloramphenicol-resistance gene) gene carried by plasmid pDG364- cotC - ompC were integrated into the amyE gene locus of B . subtilis 168 chromosome by double cross-over recombination events. Arrows indicate direction of transcription. 10.1371/journal.pone.0191627.t001 Table 1 Oligonucleotides list. Name Sequence a Restriction site cotC-F CG GGATCC TGTAGGATAAATCGTTTGGGC Bam HI cotC-R GGGGGGG AAGCTT GTAGTGTTTTTTATGCTTTTTATAC Hind ⠢ ompC-F CCC AAGCTT AATAAAGACGGCAACAAATTAGACC Hind ⠢ ompC-R C GAATTC TTAGAACTGGTAAACCAGACCC Eco RI amyE-F CCAATGAGGTTAAGAGTATTCC Null amyE-R CGAGAAGCTATCACCGCCCAGC Null a The underlined letters indicate the introduction of restriction sites. Chromosomal integration Plasmid pDG364- cotC - ompC was linearize by digestion with Xba I and used to transform competent cells of the B . subtilis strain 168 under previously described procedures [ 21 ]. Chloramphenicol-resistant recombinant B . subtilis were obtained by double-crossover recombination event ( Fig 1B ), and several clones for each transformation were tested by amylase activity analysis [ 21 ] and then confirmed by PCR using chromosomal DNA as template and four oligonucleotides (ompC-F and ompC-R, amyE-F and amyE-R, amyE-F and ompC-R, ompC-F and amyE-R) ( Table 1 ) to prime the reaction. Immunofluorescence microscopy The expression and display of fusion protein CotC-OmpC on the spore surface was confirmed using immunofluorescence microscopy, about 10 μl of spore suspension was prepared and fixed on slides as methods adapted from previous report [ 22 , 23 ]. Samples were blocked for 30min with 3% (w/v) bovine serum albumin (BSA) in PBS (pH7.4) at room temperature (RT) and then washed ten times with PBS. The samples were incubated overnight at 4°C with OmpC antiserum (raised in chicken), washed ten times and then incubated further with Cy3-labeled goat anti-chicken IgG (Sangon Biotech, 1:2000 in PBST) for 45 min at RT. After washing procedures, samples were viewed and photographed under fluorescent microscope (Eclipse, TE2000U, Nikon). Immunization in mice A total of 120 mice were randomly divided into 5 groups (24 for each): Group A and group B were dosed orally (0.2 ml) with spores of either recombinant B . subtilis or isogenic wild-type B . subtilis 168 by gavage (1.0×10 10 CFU spores/time/ mouse) at 0, 15 days respectively; group C and group D were administrated orally by feeding the diet mixed with spores of either recombinant B . subtilis or isogenic wild-type B . subtilis 168 (1.0×10 6 CFU spores/g); a naïve, untreated control group was also included (group E). Indirect ELISA for detection of OmpC-specific serum and mucosal antibodies On 1 day before and 22 days after the first dose, mice from each group were anesthetized with an intraperitoneal ketamine injection at 40 mg/kg body weight, bled retro-orbitally, and sacrificed by cervical dislocation. Intestinal content samples were immediately collected from each mouse in the sterile operation. Samples were tested by ELISA as previously described [ 24 – 26 ] with the following modifications: Flat-bottom microtiter plates (high-binding capacity; Sangon Biotech, China) were coated with 100μl of purified S . Pullorum strain CVCC533 (1×10 9 CFU/ml) in carbonate buffer solution (pH9.6) overnight at RT. After the plates were blocked for 1h at 37°C with 3% BSA, serum samples were applied as a 1/50 dilution in diluted buffer. Replicate samples were utilized together with a negative control (pre-immune serum). After incubation at 37°C for 1h, the plates were washed and then incubated with HRP-conjugate rabbit anti-mouse IgG (Santa Cruz Biotechnology). Plates were incubated for 1h at RT and then developed using TMB substrate. Reactions were stopped using 2M H 2 SO 4 , and optical densities (ODs) were read at 450 nm. ELISA of small intestinal content samples was tested with a similar method, and IgA was detected by using HRP-conjugate goat anti-mouse IgA (Santa Cruz Biotechnology). Protection studies 22 days after the first dose, 16 mice of each group mentioned above were randomly chosen and divided into two groups and were challenged with approximately 4×10 8 CFU (2×LD 50 ) and 2×10 9 CFU (10×LD 50 ) [ 27 ]of S . Typhimurium strain SL1344 by the intra-peritoneal rout, respectively. The animals were strictly monitored 72h after challenge, and individuals showing clear clinical symptoms or any signs of S . Typhimurium infections include diarrhea, fever, and abdominal cramps were considered unprotected and humanely euthanized by overdose of ketamine. Statistical analysis Samples were tested individually, and results were expressed as the arithmetic mean ± S.D. (standard deviation). Statistical significance of the difference between group means was assessed by unpaired Student's t -test, and the significance level was set at P ≤0.05. Statistical analysis was performed with SPSS for Windows version 16.0. Results Construction and chromosomal integration of cotC :: ompC gene fusion To obtain recombinant B . subtilis spores expressing OmpC on their surface, a recombinant plasmid pDG364- cotC - ompC containing the cotC :: ompC gene fusion for double cross-over with B . subtilis chromosome was constructed by fusing the ompC gene into frame of the coding part of cotC gene ( Fig 1A ). And then cotC :: ompC gene fusion was integrated into the B . subtilis chromosome at the nonessential amyE gene locus by double cross-over event ( Fig 1B ) Individual clones for each transformation were first tested by amylase activity analysis and named B . subtilis SE2. Integration of cotC :: ompC gene fusion at the amyE locus can interrupt the expression and secretion of amylase. As a result, no white halo was noted around the recombinant clones on a starch-containing plate stained by iodine ( Fig 2 ). 10.1371/journal.pone.0191627.g002 Fig 2 Amylase activity analysis. Recombinant and nonrecombinant strains grew on the starch-containing LB plate before (A) and after (B) being stained by iodine. The integration of cotC :: ompC gene fusion disrupts amyE gene and made the strain amylase deficient, transparent halo was produced around the B . subtilis 168, but in the recombinant B . subtilis SE2 clones, no transparent halo was produced. To further verify that cotC :: ompC gene fusion were localized at the amyE locus, individual clones for each transformation were tested by PCR using different primer pairs. Primers amyE-F and amyE-R amplified a 556 bp wild-type fragment and a 2,500 bp integrated fragment ( Fig 3 ). Moreover, primers ompC-F and ompC-R, amyE-F and ompC-R, ompC-F and amyE-R produced fragments around 1,078 bp, 1,600 bp, and 1,700 bp using the recombinant B . Subtilis SE2 chromosome as a template, respectively. However, no PCR products were observed in wild-type B . Subtilis chromosome ( Fig 3 ). All these results above indicated that the occurrence of correct chromosomal integration of cotC :: ompC gene fusion. 10.1371/journal.pone.0191627.g003 Fig 3 PCR analysis with different primer pairs. Site-directed PCR analysis of cotC :: ompC gene fusion integrated into the chromosome of B . subtilis 168. lane WT, B . subtilis 168; lane RT, recombinant B . subtilis SE2; primer pairs used in PCR are labeled above. Surface display of OmpC on the recombinant spores To verify the expression of OmpC antigen on the surface of the spores, intact spores purified from wild-type and isogenic recombinant strains were probed by fluorescence immunoassay with OmpC-specific primary antibodies and Cy3-labeled goat anti-chicken IgG secondary antibodies. Distinctive red fluorescent staining was detected on the recombinant B . subtilis SE2 spores surface ( Fig 4 ). However, no fluorescence signal was detected on the surface of wild type B . subtilis 168 spores ( Fig 4 ). The results indicated that OmpC was successfully expressed on the surface of the recombinant spores. 10.1371/journal.pone.0191627.g004 Fig 4 Immunofluorescence microscopy analysis. Sporulation of B . subtilis strains was induced by the exhaustion method, and spores were collected in DSM after 48h. Spores were labeled with OmpC antiserum followed by Cy3-labeled goat anti-chicken IgG. Spores were visualized by phase-contrast (PC) and immunofluorescence (IF) microscopy. Oral delivery of the OmpC displayed on spores induces systemic and mucosal antibodies responses To evaluate the immune responses derived by recombinant spores administrated by different immunization strategies, serum and small intestinal content samples were collected from mice before immunization and 22 days after the first dose and were tested by ELISA. The levels of both OmpC-specific serum IgG and intestinal mucosal SIgA antibodies in mice dosed with recombinant B . Subtilis SE2 spores (group A and C) were significantly higher ( p <0.01) than those of mice dosed with nonrecombinant spores (group B and D), or the naïve control group (group E) ( Fig 5 ). Moreover, mice dosed by feeding the diet mixed with recombinant spores (group C) presented significant levels of anti-OmpC serum IgG and intestinal mucosal SIgA responses ( p <0.05) than mice dosed by gavage with recombinant spores of the same strains (group A) ( Fig 5 ). 10.1371/journal.pone.0191627.g005 Fig 5 Serum IgG and intestinal mucosal SIgA antibodies responses. PI, pre-immune samples. Samples obtained from the mice immunized with recombinant spores (A and C) and wild-type spores (B and D); E, samples obtained from a naïve, untreated control group. ** indicates p <0.01. Protection in mice In order to assess the protection potential of recombinant B . subtilis spores, mice immunized with spores were challenged with S . Typhimurium strain SL1344 by the intra-peritoneal route as described in Material and Methods. The results presented here show that mice orally dosed with recombinant spores (group A and C) can induce 100% protection against a challenge dose of 2×LD 50 , and when mice were challenged with 10×LD 50 , only 50% protection was observed in the group A and 75% protection was observed in the groups C ( Table 2 ). In comparison, 100% mortality occurred in the mice immunized with wild type B . subtilis spores (group B and D) and all naïve mice (group E) ( Table 2 ). 10.1371/journal.pone.0191627.t002 Table 2 Protection of mice against challenge with S . Typhimurium strain SL1344. Groups 2×LD 50 10×LD 50 Dead/Total No. Protection Dead/Total No. Protection A 0/8 100% 4/8 50% B 8/8 0 8/8 0 C 0/8 100% 2/8 75% D 8/8 0 8/8 0 E 8/8 0 8/8 0 Construction and chromosomal integration of cotC :: ompC gene fusion To obtain recombinant B . subtilis spores expressing OmpC on their surface, a recombinant plasmid pDG364- cotC - ompC containing the cotC :: ompC gene fusion for double cross-over with B . subtilis chromosome was constructed by fusing the ompC gene into frame of the coding part of cotC gene ( Fig 1A ). And then cotC :: ompC gene fusion was integrated into the B . subtilis chromosome at the nonessential amyE gene locus by double cross-over event ( Fig 1B ) Individual clones for each transformation were first tested by amylase activity analysis and named B . subtilis SE2. Integration of cotC :: ompC gene fusion at the amyE locus can interrupt the expression and secretion of amylase. As a result, no white halo was noted around the recombinant clones on a starch-containing plate stained by iodine ( Fig 2 ). 10.1371/journal.pone.0191627.g002 Fig 2 Amylase activity analysis. Recombinant and nonrecombinant strains grew on the starch-containing LB plate before (A) and after (B) being stained by iodine. The integration of cotC :: ompC gene fusion disrupts amyE gene and made the strain amylase deficient, transparent halo was produced around the B . subtilis 168, but in the recombinant B . subtilis SE2 clones, no transparent halo was produced. To further verify that cotC :: ompC gene fusion were localized at the amyE locus, individual clones for each transformation were tested by PCR using different primer pairs. Primers amyE-F and amyE-R amplified a 556 bp wild-type fragment and a 2,500 bp integrated fragment ( Fig 3 ). Moreover, primers ompC-F and ompC-R, amyE-F and ompC-R, ompC-F and amyE-R produced fragments around 1,078 bp, 1,600 bp, and 1,700 bp using the recombinant B . Subtilis SE2 chromosome as a template, respectively. However, no PCR products were observed in wild-type B . Subtilis chromosome ( Fig 3 ). All these results above indicated that the occurrence of correct chromosomal integration of cotC :: ompC gene fusion. 10.1371/journal.pone.0191627.g003 Fig 3 PCR analysis with different primer pairs. Site-directed PCR analysis of cotC :: ompC gene fusion integrated into the chromosome of B . subtilis 168. lane WT, B . subtilis 168; lane RT, recombinant B . subtilis SE2; primer pairs used in PCR are labeled above. Surface display of OmpC on the recombinant spores To verify the expression of OmpC antigen on the surface of the spores, intact spores purified from wild-type and isogenic recombinant strains were probed by fluorescence immunoassay with OmpC-specific primary antibodies and Cy3-labeled goat anti-chicken IgG secondary antibodies. Distinctive red fluorescent staining was detected on the recombinant B . subtilis SE2 spores surface ( Fig 4 ). However, no fluorescence signal was detected on the surface of wild type B . subtilis 168 spores ( Fig 4 ). The results indicated that OmpC was successfully expressed on the surface of the recombinant spores. 10.1371/journal.pone.0191627.g004 Fig 4 Immunofluorescence microscopy analysis. Sporulation of B . subtilis strains was induced by the exhaustion method, and spores were collected in DSM after 48h. Spores were labeled with OmpC antiserum followed by Cy3-labeled goat anti-chicken IgG. Spores were visualized by phase-contrast (PC) and immunofluorescence (IF) microscopy. Oral delivery of the OmpC displayed on spores induces systemic and mucosal antibodies responses To evaluate the immune responses derived by recombinant spores administrated by different immunization strategies, serum and small intestinal content samples were collected from mice before immunization and 22 days after the first dose and were tested by ELISA. The levels of both OmpC-specific serum IgG and intestinal mucosal SIgA antibodies in mice dosed with recombinant B . Subtilis SE2 spores (group A and C) were significantly higher ( p <0.01) than those of mice dosed with nonrecombinant spores (group B and D), or the naïve control group (group E) ( Fig 5 ). Moreover, mice dosed by feeding the diet mixed with recombinant spores (group C) presented significant levels of anti-OmpC serum IgG and intestinal mucosal SIgA responses ( p <0.05) than mice dosed by gavage with recombinant spores of the same strains (group A) ( Fig 5 ). 10.1371/journal.pone.0191627.g005 Fig 5 Serum IgG and intestinal mucosal SIgA antibodies responses. PI, pre-immune samples. Samples obtained from the mice immunized with recombinant spores (A and C) and wild-type spores (B and D); E, samples obtained from a naïve, untreated control group. ** indicates p <0.01. Protection in mice In order to assess the protection potential of recombinant B . subtilis spores, mice immunized with spores were challenged with S . Typhimurium strain SL1344 by the intra-peritoneal route as described in Material and Methods. The results presented here show that mice orally dosed with recombinant spores (group A and C) can induce 100% protection against a challenge dose of 2×LD 50 , and when mice were challenged with 10×LD 50 , only 50% protection was observed in the group A and 75% protection was observed in the groups C ( Table 2 ). In comparison, 100% mortality occurred in the mice immunized with wild type B . subtilis spores (group B and D) and all naïve mice (group E) ( Table 2 ). 10.1371/journal.pone.0191627.t002 Table 2 Protection of mice against challenge with S . Typhimurium strain SL1344. Groups 2×LD 50 10×LD 50 Dead/Total No. Protection Dead/Total No. Protection A 0/8 100% 4/8 50% B 8/8 0 8/8 0 C 0/8 100% 2/8 75% D 8/8 0 8/8 0 E 8/8 0 8/8 0 Discussion Since Salmonellae are facultative intracellular pathogens that are capable of subverting host immune defenses by adopting various strategies [ 28 ], prevention and elimination of Salmonella have become the enormous challenge. Thus, to develop new mucosal candidate vaccines by employing more immunogenic and protective components of Salmonella , such as outer membrane proteins, is a significant issue. The immune response of the host is usually directed initially to surface-associated components of the bacterial cell [ 29 ]. In addition, recent evidence indicates that surface-associated antigens are promising antigens candidates for the induction of both humoral and cellular immunity to Salmonella [ 30 ]. The porin OmpC of Salmonella was identified as the major surface antigen with unique exposed immune epitopes, and can induce both innate and adaptive immunity in the host [ 9 , 10 , 31 ]. Moreover, the nucleotide sequence analysis of different Salmonella serotypes indicates that ompC gene is highly conserved [ 32 , 33 ]. Thus, OmpC has considerable potential in development of vaccines against mucosal infection by different Salmonella serovars. However, protein antigens are poorly immunogenic when delivered by oral immunization due to enzymatic or chemical degradation in the gastro-intestinal tract. B . subtilis spores are extremely resistant to environmental stresses, such as degradation in the protease-rich condition of gastro-intestinal tract, and oral immunization with spores carrying vaccine antigens can induce mucosal and systematic immune response in the absence of adjuvants [ 14 , 15 ]. In this study, we have shown that the outer coat component CotC of B . subtilis spores can be used as a fusion partner for surface display of OmpC on the spore surface. Mice orally immunized with recombinant B . subtilis SE2 spores were able to stimulate appreciable anti-OmpC serum IgG and intestinal mucosal SIgA responses. The local SIgA immune response observed in this study is critical and necessary for mucosal immunity against invasion by Salmonella , and demonstrates that spores have a potential to be developed into an effective delivery system to enhance protective immune responses at mucosal surfaces. Moreover, a striking observation from this work is that oral immunization with recombinant spore expressing OmpC of S . Pullorum was able to confer a significant level of protection in mice against lethal challenge with S . Typhimurium. The most likely reason for this cross-protection activity is that the pore-functioning porins of Enterobacteriaceae are highly conserved in the course of evolution and have similar biological structure and immunogenicity [ 34 , 35 ]. There is no doubt that spore-based vaccine antigen delivery systems are still in their infancy. Despite several unique advantages, as compared to other delivery systems, the low expression efficiency of vaccine antigens by recombinant spores is a major obstacle for spore-based vaccine development. Many efforts are made to improve the expression efficiency, such as exploring different promoter to regulate the expression of heterologous proteins on the surface of spores [ 36 , 37 ], selection of an appropriate anchor protein and peptide linker according to structural and functional properties of heterologous proteins [ 38 ]. In conclusion, our study indicates that OmpC can be successfully expressed and displayed on the surface of B . subtilis spores by using CotC as a fusion partner. Oral immunization in mice with recombinant spores can induce both antigen-specific systemic IgG and mucosal SIgA responses. Moreover, a significant level of protection against lethal challenge with S . Typhimurium is also detected in mice orally immunized with recombinant spores. Overall, these results suggest that B . subtilis spores have great potential as an attractive delivery system for heterologous antigens to the mucosal site of infection.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2289981/

Zoonoses Likely to Be Used in Bioterrorism

SYNOPSIS Bioterrorism is the deliberate release of viruses, bacteria, or other agents used to cause illness or death in people, animals, or plants. Only modest microbiologic skills are needed to produce and effectively use biologic weapons. And biological warfare has afflicted campaigns throughout military history, at times playing an important role in determining their outcomes. There is a long list of potential pathogens for use by terrorists, but only a few are easy to prepare and disperse. Of the infectious diseases, the vast majority are zoonoses. The Centers for Disease Control and Prevention's highest-priority bioterrorism agents are in Category A. The only disease that does not affect animals in Category A is smallpox, which was eliminated by a worldwide vaccination program in the late 1970s. Because these diseases can infect animals and humans, the medical and veterinary communities should work closely together in clinical, public health, and research settings.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7123169/

Design and Simulation of Isolation Room for a Hospital

Heating, ventilation and air conditioning (HVAC) of hospitals is a highly specialized field and critical care units like isolation rooms and operation theatres deserve special attention, as infected patients must be isolated from ambient environment in order to prevent the infection from spreading and to save the life of the patient. This manuscript aims to optimize the ventilation strategy towards contaminant suppression in the isolation room. 3D Navier-Stokes and energy equation using finite volume method (FVM) with a domain of isolation room is solved for appropriate boundary conditions. The patient's body is approximated as a semi-cylindrical shape resting on a bed and is treated as a constant heat source. Velocity and temperature profile inside the isolation room for various configurations are simulated. Our results suggest that immune-suppressed patients should be kept near the air supply and infectious patients near the exhaust. Introduction The later stage of twentieth century witnessed an alarming spread of Tuberculosis (TB) across the European nations, and this sparked an extensive research in the field of infection control techniques [ 1 ]. This led to development in the field of infection suppression in order to reduce the spread of disease. A logical solution to the problem was isolation of the infected patient. Deadlier threats like AIDS, bioterrorism (anthrax attacks, Tokyo Metro Sarin Gas attack etc.) and SARS have triggered many concerns about the control of the indoor environment and unveiled the role of ventilation system design in isolating an infected or vulnerable patient [ 2 ]. More recent global outbreaks like swine flu, bird flu and Ebola virus have shown that even though we have advanced technologically, there is still a lot left to be done in the field of contaminant suppression [ 3 ]. Medical facilities are places where relatively high levels of pathogenic (disease-causing) micro-organisms are generated and concentrated by an infected patient population or by procedures that handle infected human tissues and bodily fluids. These pathogens are spread by a number of contact and noncontact (airborne) causes. Hospital-acquired infections (HAIs, also referred to as nosocomial infections) have a significant impact on patient care. Mortality rates from HAIs are significant and affect the overall cost of healthcare delivery. It is agreed that 80–90% of HAIs are transmitted by direct contact, with 10–20% resulting from an airborne transmission (representing 0.4–1% of admitted patients) [ 4 ]. It is, therefore, imperative that we take a closer look at the design of healthcare facilities with respect to suppression of contaminants while at the same time providing patient comfort to ensure that the patient does not get worse after admission and leaves the facility healthy. For achieving this goal, the design of the heating, ventilation and cooling (herein referred to as HVAC) system of a healthcare facility is important and has to be economically viable. The term is often used interchangeably with the more common 'air conditioning'. The term air conditioning refers to a system that controls temperature, moisture in the air (humidity), supply of outside air for ventilation, filtration of airborne particles, and air movement in the occupied space. HVAC systems in healthcare facilities provide a broad range of services in support of populations who are uniquely vulnerable to an elevated risk of health, fire and safety hazard. The role of the HVAC system in life safety and infection control becomes more important with increasing complexity of the medical services provided and the acuity of illness of the patient population [ 5 ]. Isolation refers to the separation of a seriously ill patient to stop the spread of infection or protect the patient from irritating factors. A patient known or suspected to harbour transmissible micro-organisms should be placed in a single room. A single room with appropriate air handling and ventilation is particularly important for reducing the risk of airborne transmission of micro-organisms from a source patient to susceptible patients and other persons in hospitals. Airborne transmission is just one of the means by which nosocomial infection spreads. Other modes of infection include contact, droplet, common vehicle and vector-borne. An isolation room is designed and provided with all the facilities that are required to ensure protection from all the above-mentioned modes of transmission of infectious microbes. Standard isolation rooms have a pan sanitizer near the room, a staff hand-wash basin within the room, an en suite bathroom, a self-closing door, a label as a standard isolation room. The most predominant mode of infection communication, i.e. airborne transmission is taken care of by an appropriate HVAC system. Broadly speaking, there are two classes of isolation rooms: Class N—negative pressure room and Class P—positive pressure room [ 6 ]. Negative pressure rooms are for patients who are suffering from infectious or contagious diseases like TB and SARS. The aim of placing infected persons in negative pressure rooms is to reduce transmission of disease via the airborne route. Conroy et al. [ 7 ] came up with a list of features, which should exist in a negative pressure room to effectively isolate the patient. Positive pressure rooms are installed in certain facilities to isolate profoundly immune-compromised patients, such as those suffering from AIDS. A lot of research has hence been focussed towards the design of isolation wards with respect to ventilation strategy for contaminant suppression and patient comfort. A patient in an uncomfortable environment is exposed to thermal stress that may hinder the body's ability to properly regulate body heat, interfere with rest and be psychologically harmful. As per healthcare guidelines [ 8 , 9 ], ventilation strategy in an isolation room should be such that there are no stagnant and under-ventilated areas where infectious particles might be accumulated. In addition, healthcare providers subjected to an uncomfortable environment may not function at peak performance levels [ 5 ]. A study [ 8 ] of isolated patients concluded that healthcare workers visited patients in contact isolation less often than they visited regular patients, which might compromise patient safety and healthcare quality. Keeping this in mind, a proper ventilation strategy must be devised which provides for infection control and mitigation, comfort of both patients and healthcare providers, and efficient power consumption. As such, there are many challenges to overcome in the design of an efficient HVAC system for contaminant suppression. The initial design problem of an isolation room is to fix a location of the supply and exhaust vents inside the isolation room. Cheong and Phua [ 10 ] proposed a ventilation strategy for effective removal of pollutant from hospital isolation wards and inferred that a low-lying exhaust together with a ceiling supply duct delivering laminar air was the best combination. The role of air changes per hour (ACH) in order to reduce the effective residence time of contaminant was also of prime research interest. Memarzadeh and Xu [ 11 ] investigated the role of air changes per hour in possible transmission of airborne infections by simulating various ventilation system configurations and concluded that the dominant factor that affects the transmission and control of contaminants is the path between the contaminant source and exhaust, and not the airflow ventilation rate as was expected. This observation was also corroborated experimentally by Mousavi and Grosskopf [ 12 ] and Walker et al. [ 13 ]. More in-depth studies on the path of a contaminant inside an isolation ward and its response to various strategies have been carried out. Tung et al. [ 14 ] carried out a study on contaminant (sulphur hexafluoride, SF6) dispersion inside isolation room by varying the room negative pressure (from −2.5 to −15.0 Pa), the ventilation rates (12 and 24 h −1 ), by factoring in the local air quality index, and the exposure index of hospital workers. Derakshan and Shaker [ 15 ], in an attempt to reduce energy consumption by increasing the efficiency of natural ventilation, investigated volumetric flow rate in a building for various aspect ratios of the windows and their locations for various wind directions. Wan et al. [ 16 ] proposed a new method for selecting optimum indoor temperature and relative humidity to achieve minimum energy consumption for a required indoor thermal comfort level, which is evaluated with effective temperature. The research into contaminant suppression can be further extended to general hospital wards in order to reduce HAIs. Beggs et al. [ 17 ] conducted research into the need of ventilation system inside general multi-bed hospital wards. They found that bio-aerosol concentration was least when the supply and exhaust were ceiling mounted as opposed to other strategies. Chung et al. [ 18 ] studied the different ventilation patterns arranged by two outlet and two inlet diffusers at various locations. A three-dimensional k-ε turbulence model using a finite difference code (UNIC), which was developed by Engineering Science Inc., USA, was adopted in this work. The numerical simulation was first adopted to predict the airflow patterns and ventilation performance and later, a laboratory experiment at room air and contaminant distribution in a full-scale test chamber was conducted to validate the simulation results. The results indicated that the location of inlet and outlet diffusers severely affected the ventilation efficiency. The arrangement of inlet/outlet diffusers can cause different flow recirculation patterns, and this may change the mean age of indoor contaminant, thus deciding how quickly the indoor contaminant can be removed. Yu et al. [ 19 ] conducted a study on ventilation strategies for hospitals to mitigate the risk of infections. They studied the effects of location of infected patients and air change per hour to minimize deposition and floating time of airborne virus particles while maximizing energy efficiency. Although there is a plethora of research [ 7 – 15 , 17 , 18 , 20 – 23 ] on making contaminant removal systems efficient and optimizing power consumption in buildings, a comprehensive research that amalgamates effective ventilation systems with infection control in isolation wards of hospitals has a long way to go. In this background, the manuscript is to be used as a basis to set up a more advanced study on the subject of contaminant suppression and HVAC design. Here, we intend to understand the effect of placement of air supply and exhaust vents on the fluid flow and temperature profile inside the isolation ward. Our research intends to develop the existing theories and put forward an inclusive solution to the problem of HVAC design of an isolation ward keeping all the above factors in mind. The model for various orientation of air supply and exhaust vents of the isolated room was developed and simulation was carried out using in-house CFD solver. Details of CFD solver were reported elsewhere [ 24 ]. Computational Model and Solution Methodology The physical domains in this work are for an isolation room of a hospital. It can be any ordinary room in a hospital, which is spacious enough for a patient, care provider/s, various equipment, etc. Figure 1 shows the complete flow diagram of our study. In unidirectional flow, the air supply and exhaust vent are on opposite walls. However, the position of patient bed is varied. In multi-directional flow, air supply inlets are placed on the ceiling, and exhaust on the sidewall close to the patient bed. Figure 2 shows the schematic diagram of various ventilation schemes, which are to be analysed. The isolation room studied in this work has the dimensions 4.88 m × 3.60 m × 3.05 m. The ventilation rate is taken as twelve air change per hour (ACH) which is equivalent to 0.1812 m 3 s −1 as recommended in literature [ 25 – 29 ]. For the case of simplicity, the infiltration of air through the door and windows and the radiation effects have not been taken into consideration, and as such it is considered logical to hypothesize the room as one without any door or window. Fig. 1 Flow diagram of study carried in present work Fig. 2 Schematic diagram of isolation room with bed and body of the patient. a Case-1: unidirectional flow and bed close to air supply inlet. b Case-2: unidirectional flow and bed close to air outlet and c Case-3: multi-directional flow bed position close to the outlet For the unidirectional flow cases, high sidewall supply opening is provided on one wall and the exhaust is positioned 0.3048 m above the floor on the opposite sidewall. The dimensions of inlet and exhaust vents for the unidirectional flow case are 1.0668 m × 0.4572 m. This ensures an inlet velocity of 0.3716 m s −1 [ 29 ]. The patient's body is approximated as a cylinder 0.3048 m in diameter. The bed is placed at a height of 0.4572 m from the ground. In case-1, the bed is 0.9144 m away from the air supply sidewall, whereas in case-2, the bed is near the exhaust sidewall as shown in Fig. 2 a, b, respectively. Furthermore, the location of the air supply duct in cases 1 and 2 is lowered along the wall and the changes in flow pattern and temperature profile across the room are analysed. In this manner, two sub-cases for two locations of the air supply duct are studied in both the bed configurations of the isolation room. In all the cases, the temperature of the supply air is held constant at 300 K; the inlet velocities are constant for the cases 1 and 2. The third case is that of a mixed-ventilation system—multi-directional flow. It consists of multiple inlet vents strategically located on the ceiling and an exhaust vent near the patient's bed as shown in Fig. 2 c. The inlet vents are of size 0.6096 m × 0.2286 m. The ACH rating is maintained at 12 air changes. However, to ensure proper mixing of air inside the domain, high momentum air is to be supplied. That is, the inlet velocity is 0.46 m s −1 . The patient's body is directly under one of the supply openings, but it is ensured that the face/head of the patient does not receive the impinging jet of the supply air. The other two vents are located symmetrically on two sides of the central air supply vent. Approximations for Patient's Body: The patient's body is a constant source of sensible heat at roughly 70 watts of strength. This amounts to around 590 W m −3 . It has been simulated as a cylinder 1.76 m in length and 0.3048 m in diameter. The density is taken as 1800 kg m −3 , specific heat as 3470 J Kg −1 K −1 and thermal conductivity as 0.34 W m −1 K −1 to simulate a material close enough to human flesh. The source of this data is the website of School of physics, University of Sydney [ 30 ]. For the sake of simplicity, no other heat source has been considered in the present work. The present study aims to simulate the human body in the presence of externally supplied conditioned air at 300 K and 50% relative humidity [ 31 ]. The problem is rendered transient and as such is solved until steady state is attained in terms of net heat exchange across the entire domain (Table 1 ). Table 1 Thermo-physical properties of the computational domain Domain Material Density Kg m −3 C P J Kg −1 K −1 Thermal conductivity W m −1 K −1 Viscosity Kg m −1 s −1 Fluid Air 1.225 1006.43 0.0242 1.789E−5 Walls Calcium carbonate 2800 856 2.25 – Bed Wood 700 2310 0.173 – Body Approximate to human bones 1800 3470 0.45 – Mathematical Model The k-ε model is amongst the most widely used turbulence models. The standard k-ε model is the simplest 'complete model' of turbulence in which the solution of two separate transport equations allows the turbulent velocity and length scales to be independently determined and it has become the workhorse of practical engineering flow calculations in the time since Launder and Spalding [ 32 ] proposed it. Robustness, economy and reasonable accuracy for a wide range of turbulent flows explain its popularity in industrial flow and heat transfer simulations. The various conservation equations used in standard k-ε model are given hereunder: 1 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{{\partial u_{i} }}{{\partial x_{i} }} = 0 $$\end{document} 2 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{{\partial u_{i} }}{\partial t} + rac{{\partial u_{i} u_{j} }}{{\partial x_{j} }} = - rac{\partial }{{\partial x_{i} }}\left( { rac{P}{ ho } + rac{2}{3}k} ight) + rac{\partial }{{\partial x_{j} }}\left[ { u_{t} + \left( { rac{{\partial u_{i} }}{{\partial x_{j} }} + rac{{\partial u_{j} }}{{\partial x_{i} }}} ight)} ight] $$\end{document} 3 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{\partial k}{\partial t} + rac{{\partial ku_{j} }}{{\partial x{}_{j}}} = rac{\partial }{{\partial x_{j} }}\left( { rac{{ u_{t} }}{{\sigma_{1} }} rac{\partial k}{{\partial x_{j} }}} ight) + u_{t} S - arepsilon $$\end{document} 4 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{\partial arepsilon }{\partial t} + rac{{\partial arepsilon u_{j} }}{{\partial x{}_{j}}} = rac{\partial }{{\partial x_{j} }}\left( { rac{{ u_{t} }}{{\sigma_{2} }} rac{\partial arepsilon }{{\partial x_{j} }}} ight) + C_{1} rac{ arepsilon }{k} u_{t} S - C_{2} rac{{ arepsilon^{2} }}{k} $$\end{document} 5 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ u_{t} = C_{\mu } rac{{k^{2} }}{ arepsilon } $$\end{document} 6 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ S = \left( { rac{{\partial u_{i} }}{{\partial x_{j} }} + rac{{\partial u_{j} }}{{\partial x_{i} }}} ight) rac{{\partial u_{i} }}{{\partial x_{j} }}; $$\end{document} where \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \sigma_{1} = 1.0,\sigma_{2} = 1.3,\sigma_{3} = 1.0,C_{\mu } = 0.09,C_{1} = 1.44,C_{2} = 1.92 $$\end{document} . This family of equations is generally known as the k-ε model. In the present work, k-ε turbulence (two-equation) model is used for 3D simulation. The standard k-ε is a semi-empirical model based on model transport equations for the turbulence kinetic energy ( k ) and its dissipation rate ( ε ). In this technique, the Reynolds stress is modelled in terms of two turbulence parameters, the turbulent kinetic energy, k and the turbulent energy dissipation rate, ε are defined below: 7 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ k = rac{1}{2}\left( {\overline{{u^{\prime}}^{2}} + \overline{{v^{\prime}}^{2}} + \overline{{w^{\prime}}^{2}} } ight) $$\end{document} 8 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ egin{aligned} arepsilon = & u \left[ {\left( { rac{{\partial u^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial u^{\prime}}}{\partial y}} ight)^{2} + \left( { rac{{\partial u^{\prime}}}{\partial z}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial y}} ight)^{2} } ight. \ & \quad \quad \left. { + \left( { rac{{\partial v^{\prime}}}{\partial z}} ight)^{2} + \left( { rac{{\partial w^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial y}} ight)^{2} + \left( { rac{{\partial w^{\prime}}}{\partial z}} ight)^{2} } ight] \ \end{aligned} $$\end{document} where ( u ′, v ′, w ′) is the fluctuating velocity vector. k , the turbulent kinetic energy, is defined as the variance of the fluctuations in velocity. It has dimensions of (L2 T-2). ε , the turbulence eddy dissipation rate, is the rate at which the velocity fluctuations dissipate in a turbulent flow. It has dimensions of turbulent kinetic energy per unit time (L2 T-3). The kinetic energy is zero for laminar flow and can be as large as 5% of the kinetic energy of the mean flow in a highly turbulent case. The finite volume method (FVM) is adopted to solve the aforementioned Eqs. ( 1 – 8 ). The details of this method were reported by authors elsewhere [ 24 ]. In this method, the physical domain is subdivided into a number of contiguous control volumes on an unstructured tetrahedral grid, as shown in Fig. 3 . Unstructured tetrahedral grids are constructed by dividing the computational domain into small cells that have a tetrahedral shape. The strength of this type of grid lies in its flexibility for handling very complex geometries. A commercial software package (ICEMCFD) was used to generate meshes. The generated mesh was made very fine near the patient body (0.01 m) and coarse towards the isolation room walls (0.2 m). A refined mesh was also generated at the air supply inlet and the exhaust. To effectively capture the air flow around the patient body, a mesh density box was generated with elements of size 0.025 m as can be seen from Fig. 3 b. Equations ( 1 – 8 ) were discretized using FVM Techniques. Convective term is discretized using hybrid upwind scheme. The diffusion terms are discretized using second-order central difference. Weighted average of surrounding cell-centred properties is used for determining nodal quantities. SIMPLE like algorithm is used for handling pressure velocity decoupling. Collocated grid arrangement is applied, and movement interpolation is used to determine the face flux of each cells. System of algebraic equation is solved using BICGStab. Fig. 3 Mesh on computational domain a isometric view b sectional view at z = 1.525 m Mathematical Model The k-ε model is amongst the most widely used turbulence models. The standard k-ε model is the simplest 'complete model' of turbulence in which the solution of two separate transport equations allows the turbulent velocity and length scales to be independently determined and it has become the workhorse of practical engineering flow calculations in the time since Launder and Spalding [ 32 ] proposed it. Robustness, economy and reasonable accuracy for a wide range of turbulent flows explain its popularity in industrial flow and heat transfer simulations. The various conservation equations used in standard k-ε model are given hereunder: 1 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{{\partial u_{i} }}{{\partial x_{i} }} = 0 $$\end{document} 2 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{{\partial u_{i} }}{\partial t} + rac{{\partial u_{i} u_{j} }}{{\partial x_{j} }} = - rac{\partial }{{\partial x_{i} }}\left( { rac{P}{ ho } + rac{2}{3}k} ight) + rac{\partial }{{\partial x_{j} }}\left[ { u_{t} + \left( { rac{{\partial u_{i} }}{{\partial x_{j} }} + rac{{\partial u_{j} }}{{\partial x_{i} }}} ight)} ight] $$\end{document} 3 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{\partial k}{\partial t} + rac{{\partial ku_{j} }}{{\partial x{}_{j}}} = rac{\partial }{{\partial x_{j} }}\left( { rac{{ u_{t} }}{{\sigma_{1} }} rac{\partial k}{{\partial x_{j} }}} ight) + u_{t} S - arepsilon $$\end{document} 4 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ rac{\partial arepsilon }{\partial t} + rac{{\partial arepsilon u_{j} }}{{\partial x{}_{j}}} = rac{\partial }{{\partial x_{j} }}\left( { rac{{ u_{t} }}{{\sigma_{2} }} rac{\partial arepsilon }{{\partial x_{j} }}} ight) + C_{1} rac{ arepsilon }{k} u_{t} S - C_{2} rac{{ arepsilon^{2} }}{k} $$\end{document} 5 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ u_{t} = C_{\mu } rac{{k^{2} }}{ arepsilon } $$\end{document} 6 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ S = \left( { rac{{\partial u_{i} }}{{\partial x_{j} }} + rac{{\partial u_{j} }}{{\partial x_{i} }}} ight) rac{{\partial u_{i} }}{{\partial x_{j} }}; $$\end{document} where \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \sigma_{1} = 1.0,\sigma_{2} = 1.3,\sigma_{3} = 1.0,C_{\mu } = 0.09,C_{1} = 1.44,C_{2} = 1.92 $$\end{document} . This family of equations is generally known as the k-ε model. In the present work, k-ε turbulence (two-equation) model is used for 3D simulation. The standard k-ε is a semi-empirical model based on model transport equations for the turbulence kinetic energy ( k ) and its dissipation rate ( ε ). In this technique, the Reynolds stress is modelled in terms of two turbulence parameters, the turbulent kinetic energy, k and the turbulent energy dissipation rate, ε are defined below: 7 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ k = rac{1}{2}\left( {\overline{{u^{\prime}}^{2}} + \overline{{v^{\prime}}^{2}} + \overline{{w^{\prime}}^{2}} } ight) $$\end{document} 8 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ egin{aligned} arepsilon = & u \left[ {\left( { rac{{\partial u^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial u^{\prime}}}{\partial y}} ight)^{2} + \left( { rac{{\partial u^{\prime}}}{\partial z}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial y}} ight)^{2} } ight. \ & \quad \quad \left. { + \left( { rac{{\partial v^{\prime}}}{\partial z}} ight)^{2} + \left( { rac{{\partial w^{\prime}}}{\partial x}} ight)^{2} + \left( { rac{{\partial v^{\prime}}}{\partial y}} ight)^{2} + \left( { rac{{\partial w^{\prime}}}{\partial z}} ight)^{2} } ight] \ \end{aligned} $$\end{document} where ( u ′, v ′, w ′) is the fluctuating velocity vector. k , the turbulent kinetic energy, is defined as the variance of the fluctuations in velocity. It has dimensions of (L2 T-2). ε , the turbulence eddy dissipation rate, is the rate at which the velocity fluctuations dissipate in a turbulent flow. It has dimensions of turbulent kinetic energy per unit time (L2 T-3). The kinetic energy is zero for laminar flow and can be as large as 5% of the kinetic energy of the mean flow in a highly turbulent case. The finite volume method (FVM) is adopted to solve the aforementioned Eqs. ( 1 – 8 ). The details of this method were reported by authors elsewhere [ 24 ]. In this method, the physical domain is subdivided into a number of contiguous control volumes on an unstructured tetrahedral grid, as shown in Fig. 3 . Unstructured tetrahedral grids are constructed by dividing the computational domain into small cells that have a tetrahedral shape. The strength of this type of grid lies in its flexibility for handling very complex geometries. A commercial software package (ICEMCFD) was used to generate meshes. The generated mesh was made very fine near the patient body (0.01 m) and coarse towards the isolation room walls (0.2 m). A refined mesh was also generated at the air supply inlet and the exhaust. To effectively capture the air flow around the patient body, a mesh density box was generated with elements of size 0.025 m as can be seen from Fig. 3 b. Equations ( 1 – 8 ) were discretized using FVM Techniques. Convective term is discretized using hybrid upwind scheme. The diffusion terms are discretized using second-order central difference. Weighted average of surrounding cell-centred properties is used for determining nodal quantities. SIMPLE like algorithm is used for handling pressure velocity decoupling. Collocated grid arrangement is applied, and movement interpolation is used to determine the face flux of each cells. System of algebraic equation is solved using BICGStab. Fig. 3 Mesh on computational domain a isometric view b sectional view at z = 1.525 m Result and Discussion Before analysing the flow pattern of aforementioned configuration of inlet and outlet of isolated room, the present numerical methodology is validated against data available in the literature [ 18 ]. Figure 4 shows the velocity at various location of room, and it has good agreement against the experiment. Fig. 4 Velocity distribution at various locations inside chamber (4 m × 3 m × 2.5 m) for inlet supply velocity 1.36 m s −1 Post-validation, simulations were carried out for different cases of the isolation room. Each case has a fixed location of patient's bed in the room. In the first two cases, air supply inlet and exhaust were placed on opposite walls of the isolation room as shown earlier in Fig. 2 a, b and the position of the patient bed was moved from the air supply sidewall to the exhaust sidewall. The third case had multiple air supply inlets on the ceiling of the isolation room with the exhaust on the sidewall, as shown in Fig. 2 c. Further, two more cases were studied, which were sub-cases of the first two cases. In these sub-cases, the height of the air supply inlet was varied and its effect on the ventilation inside the isolation room was studied. The two inlet heights studied were 0.4572 m from the ceiling and 0.6096 m from the ceiling. In all cases, the height of the exhaust vent was 0.3048 m from the floor of the room. The results of the isolation room were plotted at various planes such as x - y plane ( z = 1.225 m) and z - y plane ( x = 1.4 and 2.44 m) for knowing the temperature and velocity distribution close and away to the patient. The patient body was modelled as a semi-cylindrical heat source. Figure 5 shows the velocity and temperature contours in x - y plane and z - y plane of the first case study. The air supply inlet is lowered to 0.4572 m below the ceiling. The position of the bed of patient was near the air supply sidewall. Figure 6 shows the flow pattern in the form of velocity vectors in the in the x - y plane and in the z - y plane. Fig. 5 X-velocity and temperature contour of isolation room in the x - y plane ( z = 1.525 m) and the z - y plane ( x = 1.4 m). The height of air supply inlet is 0.4572 m from the ceiling and the patient is close to the air supply sidewall. a , c show u velocity and temperature contour in x - y plane and b , d u velocity and temperature contour in the z - y plane Fig. 6 Flow pattern ( u and v velocity vectors) of air in a x - y plane ( z = 1.525) and b z - y plane ( x = 1.4 m): inlet at 0.4572 m from the ceiling It can be observed that the zone between the patient's bed and the air jet is stagnant. However, below the bed there is comparatively better airflow due to the recirculation of the air jet that hits the exhaust sidewall, as shown in Fig. 6 . Such a location of bed would be hazardous for an infectious patient because the stagnant air above the patient's bed would permit the bacteria to proliferate. However, this situation is favourable for the immune-suppressed patient since he needs to be protected from the supply air directly hitting his body. Such an airflow field serves as a defence shield for an immune-suppressed patient. As far as the rest of the room is concerned, it is under the sole effect of supply air having a temperature of 300 K. The air in the vicinity of the patient's body is slightly warmer than that in the rest of the space in room, as shown in Fig. 5 c, d. Figure 7 shows the velocity and temperature contour in the x - y plane and the z - y plane. The air supply inlet is lowered to 0.6096 m below the ceiling. The position of the bed of patient was near the air supply sidewall. Figure 8 shows the flow pattern in the x - y plane and the z - y plane for this case. It can be observed that the recirculation zones above and below the bed have decreased and stagnation zones have formed. This ventilation strategy as such is not recommended for an immune-suppressed patient because he might start complaining about draught sensation. Placing an infectious patient in the room is not recommended as well due to significant stagnation zones that exist inside the room. Another recirculation zone is formed above the jet, as can be seen from Fig. 8 a, and this could possibly lead to a stagnation zone above the jet further deteriorating the ventilation of the room. These effects can be linked to the short-circuiting of the airflow due to reduced vent height. The temperature inside the room is more evenly distributed than in the previous case. Fig. 7 X-velocity and temperature contour of isolation room in the x - y plane ( z = 1.525 m) and the z - y plane ( x = 1.4 m). The height of air supply inlet is 0.6096 m from the ceiling and the patient is close to the air supply sidewall. a , c show u velocity and temperature contour in the x - y plane and b , d u velocity and temperature contour in the z - y plane Fig. 8 Flow pattern ( u and v velocity vectors) of air in a x - y plane ( z = 1.525) and b z - y plane ( x = 1.4 m): inlet at 0.6096 m from ceiling Figure 9 shows the velocity and temperature contours in the x - y plane and z - y plane of the third case. The air supply inlet is located 0.4572 m below the ceiling. The position of the bed of patient was near the exhaust sidewall. Figure 10 shows the flow pattern in the form of velocity vectors in the x - y plane and the z - y plane. Fig. 9 X-velocity and temperature contour of the isolation room in the x - y plane ( z = 1.525 m) and the z - y plane ( x = 1.4 m). The height of air supply inlet is 0.4572 m from the ceiling, and the patient is close to exhaust sidewall. a , c show u velocity and temperature contour in the x - y plane and b , d u velocity and temperature contour in the z - y plane Fig. 10 Flow pattern ( u and v velocity vectors) of air in a x - y plane ( z = 1.525) and b z - y plane ( x = 1.4 m): inlet at 0.4572 m from the ceiling The height of the air supply inlet is 0.6096 m from the ceiling and the patient is close to the exhaust sidewall. Figure 11 a, c shows u velocity and temperature contour in the x - y plane and Fig. 11 b, d u velocity and temperature contour in the z - y plane. Fig. 11 x -velocity and temperature contour of the isolation room in the x - y plane ( z = 1.525 m) and the z - y plane ( x = 1.4 m) The patient's body starts receiving direct air jet, which deflects down from the exhaust sidewall. However, the feet continue to be in slightly stagnant zone. There is appreciable recirculation in both the upper and lower zones of the isolation room leading to better overall ventilation. This configuration is ideal for immune-suppressed patients as the patients would typically be in a region of stagnation as shown in Fig. 9 b, which would protect them from any possible infections. The draught rating would also not be very high leading to optimal configuration for the immune-suppressed patient. Figure 12 shows the velocity and temperature contour in the x - y and the z - y plane of the fourth case. The air supply inlet was lowered to 0.6096 m below the ceiling. The position of the bed of patient was near the exhaust sidewall. Figure 13 shows the velocity vectors in the in the x - y plane and the z - y plane. Fig. 12 Flow pattern ( u and v velocity vectors) of air in a x - y plane ( z = 1.525) and b z - y plane ( x = 1.4 m): inlet at 0.6096 m from ceiling Fig. 13 y-velocity and temperature contour of the isolation room in the x - y plane ( z = 1.525 m) and the z - y plane ( x = 2.44 m). Three air supply inlets are located on the ceiling and the patient is close to exhaust sidewall. a , c show v velocity and temperature contour in the x - y plane and b , d v velocity and temperature contour in the z - y plane As the inlet height is lowered with the patient near exhaust sidewall, the room becomes very poorly ventilated. Although the space near the patient is agitated, there is a considerable fraction of the whole space that is experiencing stagnation in this strategy. Short-circuiting of flow would also be noticed in this case. Figure 14 shows the velocity and temperature contour in x - y and z - y plane for the fifth case. Multiple air supply inlets are located on the ceiling of the isolation room. The position of the bed of patient was near the exhaust sidewall. Figure 15 shows the flow pattern in the x - y and the y - z plane with the air supply mounted on the ceiling. Fig. 14 Flow pattern ( u and v velocity vectors) of air in a x - y plane ( z = 1.525) and b z - y plane ( x = 2.44 m) Fig. 15 y-velocity profile across the across the isolation room ( y = 0.3 m to y = 3.3 m) Figure 15 shows the y-velocity profile as it develops across different levels in the z - x plane. An infectious patient is advised to be kept in an isolation room provided with a mixed-ventilation system. It has multiple laminar diffusers on the ceiling, and the patient's bed is strategically located with respect to the diffusers so that the bacteria concentration is diluted effectively. The upper zone between the two symmetrically located inlet vents shows some stagnation but in the occupied zone, the air is well mixed due to the presence of two supply openings in the farther corners of the room. With multiple inlet vents, a mixed-ventilation system offers improved control over the whole airflow field inside the domain. The upper zone between the two symmetrically located inlet vents shows some stagnation but in the occupied zone, the air is well mixed due to the presence of two supply openings towards the corners of the room. This arrangement makes the whole room well ventilated. However, higher draught ratings must be expected in such a ventilation system. Conclusion The position of the patient bed and air supply inlet of isolation room for two kinds of patients: infectious and immune-suppressed were simulated using in-house CFD solver. An immune-suppressed patient is one who is vulnerable to infection from any contaminant that accompanies the supply air. The location of the bed and the ventilation system for such a system should be designed to provide the patient maximum protection against the possible contamination in the supply air by ensuring that contaminant resident time inside the room is minimum. The infectious patient on the other hand is one who produces infectious micro-organisms by means of breathing, coughing and sneezing. A TB or a SARS patient would be the suitable example. The ventilation system for such a patient should perform the function of flushing out the infectious bacteria as soon as they are generated. This is possible when the air flows from the less contaminated zone to the most contaminated zone. Against this background, a mathematical model of hospital isolation room was developed and simulated. Based on the simulation results, the following conclusions can be made: The locations for the immune-suppressed patient inside the isolation room should be near the supply sidewall. An infectious patient's bed should be located near the ventilation exhaust. To get rid of the stagnation zones, it would be reasonable to have some constant heat sources like lamps and medical equipment inside the room so that the temperature difference causes the incoming cool air to settle down and displace the warmer air upwards via convection. In this fashion, stagnation observed in the velocity profiles for immune-suppressed patient would not be as severe. Mixed ventilation is a good solution for the infectious patients. It offers better control over the HVAC and ventilation parameters due to the multiple inlet vents. In this study, many simplifications/assumptions were taken that render our results only approximate. For our research to aide in the practical design of a hospital isolation ward or operation theatre, these approximations must be reduced and a model as exact as computationally possible, must be developed.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC547061/

Murine Macrophage Transcriptional Responses to Bacillus anthracis Infection and Intoxication

Interactions between Bacillus anthracis and host macrophages represent critical early events in anthrax pathogenesis, but their details are not clearly understood. Here we report the first genomewide characterization of the transcriptional changes within macrophages infected with B. anthracis and the identification of several hundred host genes that were differentially expressed during this intracellular stage of infection. These loci included both genes that are known to be regulated differentially in response to many other bacterial pathogens and those that appear to be differentially regulated in response to B. anthracis but not other bacterial species that have been tested. These data provide a transcriptional basis for a variety of physiological changes observed during infection, including the induction of apoptosis caused by the infecting bacteria. The expression patterns underlying B. anthracis -induced apoptosis led us to test further the importance of one very highly induced macrophage gene, that for ornithine decarboxylase. Our data show that this enzyme plays an important and previously unrecognized role in suppressing apoptosis in B. anthracis -infected cells. We have also characterized the transcriptional response to anthrax lethal toxin in activated macrophages and found that, following toxin treatment, many of the host inflammatory response pathways are dampened. These data provide insights into B. anthracis pathogenesis as well as potential leads for the development of new diagnostic and therapeutic options.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7512050/

Convalescent Plasma: A Challenging Tool to Treat COVID-19 Patients—A Lesson from the Past and New Perspectives

On March 11 th , 2020, the World Health Organization declared COVID-19 infection as a pandemic. Since it is a novel virus, there are basically no proven drugs or therapies; although many laboratories in different countries are working to develop a vaccine, it will take time to make it available. Passive immunization is the therapy born from the intuition of Behring and Kisato in the late 19 th century. It was widely used for the treatment of bacterial infections until the discovery of antibiotics, as well as during the viral pandemics of the 20 th century and of the beginning of the 21 st ; it still has clinical applications (e.g., tetanus prevention). This paper summarizes the basic principles of passive immunization, with particular reference to convalescent plasma. The literature concerning its use during past epidemics and the results of the first clinical studies concerning its use during the current pandemic are discussed too. A large section is dedicated to the analysis of the possible, although rare, side effects. Recently, in 2017, the WHO Blood Regulators Network (BRN) published a position paper, recommending convalescent plasma as the first-choice treatment to be tested in the absence of authorized drugs; however, this strategy has not been followed. In the current epidemic, the principle of passive immunization through convalescent plasma has been applied in several circumstances and particularly in patients with serious complications. The first reported results are encouraging and confirm the effectiveness of plasma therapy and its safety. Also, the FDA has proposed plasma treatment in order to face the increasingly complex situation and manage patients with serious or immediately life-threatening COVID-19 disease. Several studies and clinical programs are still ongoing. 1. Introduction On March 11 th , 2020, the World Health Organization (WHO) declared COVID-19 infection as a pandemic [ 1 ]. The virus causing COVID-19 infection is a coronavirus called SARS-CoV-2; it began to scare the world since the first days of 2020 during its initial outbreak in China, because of the characteristics of contagion (high rate of contagiousness associated with high lethality) [ 2 ]. Since it is a novel virus, there are basically no proven drugs or therapies. In hospitals all over the world, there are many ongoing clinical studies. Many attempts have been made to treat seriously sick patients, using off-label drugs already known; nevertheless, to date, there is no effective targeted antiviral therapy. In most cases, drug administration has been authorized for a compassionate purpose [ 3 ]. In fact, WHO management of COVID-19 has been mainly focused on infection prevention, case detection, and monitoring; supportive care and nonspecific anti-SARS-CoV-2 treatment have been recommended [ 4 ]. Extensive vaccination is the only strategy to prevent pandemic transmission of SARS-CoV-2. Major efforts are currently being made by many laboratories in several countries to develop a vaccine; however, it will still take time before the vaccine is widely available to the population [ 5 ]. Encouraging news about passive immunization arrived from China at the end of February, and some authors reported them in their scientific publications. Cai et al. cited two official sources (National Health and Health Commission, Health Bureau of the Logistics Support Department of the Central Military Commission; Chinese Society of Blood Transfusion) reporting significant improvements in patients affected by COVID-19 and treated with plasma donated by recovered patients [ 5 ]. Anecdotal experiences are reported by Keith et al. [ 6 ] and from Cunningham et al. [ 7 ]; in particular, the latter group reported that Biotec Group Co. announced that 10 seriously ill patients, treated with immunoglobulin therapy, demonstrated improved oxygenation and reduced inflammation and viral load [ 7 ]. In addition, Casadevall and Pirofski, referring to the news from the Xinhua news source, reported that convalescent serum was used for the therapy of 245 patients with COVID-19 in China during the first outbreak. Although few details are available and published studies involve a small number of patients, the authors concluded that convalescent serum is safe and reduces viral load [ 8 ]. Additionally, convalescent plasma could potentially be used to prevent disease in high-risk cases (vulnerable individuals with underlying medical condition, health care providers, and individuals exposed to confirmed cases of COVID-19 [ 8 ]. The literature about convalescent plasma is rapidly growing. The uniqueness of this work is that it presents all the main aspects about convalescent plasma in a single body. The already published review articles often focused on single aspects of convalescent plasma, and to the best of the author's knowledge, there are no available works covering all the main topics concerning convalescent plasma use. The present manuscript is intended to be a complete and update guide for doctors and institutions. 2. Methods of Research A systematic search was conducted in major electronic databases (PubMed and MEDLINE) and Google Scholar; the applied query was "plasma OR convalescent plasma" AND "COVID-19 OR Sars-Cov-2". 3. Passive Immunization: Basic Principles Virus neutralization by antibodies is the principle behind the functioning of plasma of patients recovered from SARS-CoV-2; high-titer-specific antibodies bind to SARS-CoV-2 neutralizing the viral particles, blocking access to cells, and activating potent effector mechanisms, such as complement activation and phagocytosis [ 7 ]. There are several ways to achieve passive immunization. Antibodies can be delivered to the recipient by (i) human whole blood, (ii) human or animal plasma or serum, (iii) pooled human immunoglobulin for intravenous (IVIG) or intramuscular (IG) use, (iv) high-titer human immunoglobulin for intravenous or intramuscular use from immunized or convalescing donors, and (v) monoclonal antibodies (MAb) [ 9 , 10 ]. Transfusion of whole blood to provide convalescent plasma should be avoided unless its use is clinically indicated; collection of convalescent plasma should be performed only by apheresis to avoid unnecessary red cell loss in the donor [ 11 ]. Plasma administration is the preferred method to provide passive immunity in pandemic scenarios at least in the immediate term; usually, immunoglobulins are prepared by fractionating large pools of human plasma collected from approximately 10,000–40,000 donors [ 12 ]. Moreover, plasma administration represents a rapid and effective therapy, has lower costs than other methods [ 13 ], and presents a broader spectrum response. Studies suggest that it not only neutralizes the pathogen but also provides passive immunomodulatory mediators allowing the recipient to control the excessive inflammatory cascade induced by the infectious agent [ 14 ]. Animal plasma collection should be avoided if possible as it can cause side effects collectively called "serum sickness" [ 15 ]. 4. Convalescent Plasma: A Lesson from Past and Current Applications Convalescent plasma is not a novel therapy; it is a therapy widely used in the past, both for bacterial and viral pathologies. Behering and Kisato were the first in 1890 to provide the basis of passive immunization, then known as serum therapy. Despite limited knowledge on the structural and functional complexity of antibodies, they demonstrated that not previously immunized animals can be protected from sublethal doses of diphtheria and tetanus toxin with serum therapy. The discovery was so important that in 1901, Behring earned his Noble Prize for it. Given the early success in the 1900s, passive immunization was rapidly expanded; it was used to treat several bacterial infections including Corynebacterium diphtheriae, Streptococcus pneumoniae, Streptococcus pyogenes, Clostridium tetani, Haemophilus influenzae, and Neisseria meningitidis; type-specific antipneumococcal serum was used as the first-line treatment for lobar pneumonia [ 15 ]. During the first half of the 20 th century, serum therapies were successfully used to treat patients affected by many infectious diseases (anthrax, plague, scarlet fever, measles, tularemia, diphtheria, dysentery, meningococcal meningitis, rabies, and pneumococcal pneumonia) [ 16 ]. However, serum therapy for bacterial diseases suddenly stopped after the discovery of antibiotics. Moreover, the tools for the correct selection of plasma were still missing; the risk of serum disease was very high in those years, as plasma was frequently prepared from the blood of hyperimmunized animals [ 15 ]. Regarding viral diseases, studies about convalescent plasma use to fight pandemics of the last century are available; the reported results were positive [ 8 ]. In the early 20 th century, convalescent serum was used to fight outbreaks of viral diseases such as poliomyelitis, measles, mumps, and Spanish influenza [ 8 ]. A meta-analysis by Luke and colleagues reported eight studies involving 1,703 patients with 1918 influenza pneumonia from 1918 to 1925. Patients were often selected among the most serious ones and received an infusion of influenza convalescent human blood products; the outcome of the treated patients was compared to that of the untreated influenza pneumonia controls. The study showed a pooled absolute reduction of 21% in the mortality rate compared to controls. Unfortunately, the included studies were few and with methodologic limitations (no study was a blinded, randomized, or placebo-controlled trial; moreover, convalescent sera were developed and used in many cases without measuring antibody titers or without knowledge about viral serotypes); anyway, this treatment received consensus at the time, and it was applied in several countries [ 4 , 8 , 17 ]. In the modern era, the treatment of Argentine hemorrhagic fever (Junin virus) with convalescent immune plasma was applied as part of a nationally organized response; patients treated with immune plasma had a much lower mortality than those given normal plasma [ 16 , 18 ]. Convalescent plasma or immunoglobulins were administered as a last chance to reduce the mortality rate of patients with SARS; several studies showed a shorter hospital stay and lower mortality in patients treated with convalescent plasma compared to those not treated with convalescent plasma. The largest study involved the treatment of 80 patients showing clinical deterioration despite treatment with methylprednisolone. Earlier plasma administration was more likely to be effective: patients treated before the 14 th day had better prognosis compared to those treated later. In addition, patients who were PCR positive and seronegative for coronavirus at the time of therapy had improved prognosis. No immediate adverse reactions were observed [ 4 , 8 , 19 ]. Positive evidence was reported for the treatment of influenza A (H5N1) too [ 20 ]. Regarding the pandemic 2009 influenza A H1N1, the results from the prospective cohort study by Hung and colleagues showed that plasma treatment reduced mortality (the patients involved in the study were seriously ill and required intensive care); no adverse events were observed [ 4 , 8 , 20 ]. A second trial by Hung and colleagues was conducted on 35 patients affected by severe influenza A H1N1 during 2010 and 2011, using immunoglobulins fractionated from plasma of patients recovered from the previous 2009 influenza A H1N1. Treated patients showed a lower viral load and reduced mortality rate within 5 days of symptom onset [ 4 , 8 , 21 ]. A meta-analysis by Mair-Jenkins and colleagues, including 32 studies of SARS coronavirus and severe influenza, reported that convalescent plasma reduced mortality and it was safe (no relevant adverse events or complications after treatment were reported). The reduction of mortality was higher when convalescent plasma was administered earlier after symptom onset [ 22 ]. Regarding Ebola disease, the use of convalescent plasma was recommended by the WHO in 2014 as an empirical treatment during the outbreaks [ 8 ]. The first use of convalescent plasma for Ebola dates back to previous times. The study of Mupapa et al. involved eight patients during an outbreak in 1995; among these, seven survived [ 23 ]. The Ebola-Tx clinical trial tested the efficacy of convalescent plasma as a treatment for Ebola in Guinea: the trial confirmed convalescent plasma safety, but unfortunately, the efficacy was not proven. No association with the dose of neutralizing antibodies was apparently found, even if the levels of neutralizing antibodies were low in many plasma donations. The authors concluded that further studies were needed to assess the effectiveness of antibody doses higher than those used in their study [ 24 ]. Sahr et al. used convalescent serum in Sierra Leone for Ebola treatment: their study revealed a significantly lower fatality rate for patients treated with convalescent whole blood with respect to those receiving standard treatments [ 13 ]. A protocol for convalescent plasma in the treatment of Middle East Respiratory Syndrome (MERS) caused by a coronavirus was established in 2015. Three patients with MERS in South Korea were treated with convalescent serum, but only two showed neutralizing activity. The authors concluded that high antibody titer (≥1 : 80) should be needed to achieve good neutralization activity [ 4 , 8 , 25 ]. Keller and Stiehm listed all the pathologies for which passive immunization has been or is currently being used. For each pathology, they specified when passive immunization is to be used for prevention versus treatment and if the efficacy has been demonstrated (they also pointed out when there is no recommendation to use the passive immunization tool, even in the case of demonstrated efficacy). More than 30 infectious pathologies were analysed. The efficacy of passive immunization in the prevention of infectious diseases has been proven for tetanus, Clostridium botulinum, hepatitis A, hepatitis B, RSV (respiratory syncytial virus), CMV (cytomegalovirus), VZV (varicella zoster virus), rabies, measles, and vaccinia. In addition, passive immunization has been proven but not recommended for the treatment of respiratory infections (Streptococcus, Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae) or for enterovirus infection. The efficacy of passive immunization in the treatment of infectious disease has been proven for diphtheria, tetanus, Clostridium botulinum, and vaccinia and has been proven but not recommended for respiratory infections (Streptococcus, Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae), parvovirus, and enterovirus [ 9 ]. 5. Safeness of Convalescent Plasma: Exploring the Rare Side Effects of Plasma Therapy As already discussed, previous studies on convalescent plasma in pandemic scenarios have shown that plasma is a safe treatment [ 4 , 8 , 13 , 17 – 25 ]. Anyway, side effects are possible for any medication; MacLennan and Barbara analysed the possible side effects of generic plasma administration (not only convalescent plasma) in a recipient. Several factors can lead to adverse events (donor-related factors, which testing is performed on plasma, any treatment or modification to which it has been subjected, interaction between donor factors, and the patient's immune system). Possible adverse reactions can be classified into three groups: immune reactions (anaphylactic/anaphylactoid reactions, mild allergic reactions, haemolysis, and transfusion-related acute lung injury), physicochemical reactions (fluid overload, citrate toxicity, and chemicals), and infectious risks. Anaphylactic reactions are uncommon, but severe and potentially life-threatening (in 2003, the incidence in the UK was approximately 0.002%). IgE mediates anaphylactic reactions; the term "anaphylactoid" describes a similar reaction not mediated by IgE. Less severe allergic reactions are much more common and usually characterized by cutaneous symptoms, ranging from mild pruritus to urticaria and flushing. Haemolysis can occur following transfusion of plasma containing high-titer anti-A or anti-B haemolysins to an A or B recipient. It could be very serious, and deaths have also been reported. To avoid this reaction, plasma should always be ABO compatible; if not possible, the plasma should be tested for haemolysins and found negative for high titer of them. Concomitant transfer of antibodies against other red cell antigens might occasionally cause haemolysis in the recipient. Thus, donor screening procedures to detect clinically significant antibodies are essential to minimize this risk [ 26 ]. TRALI is an acute respiratory reaction, indistinguishable from the adult respiratory distress syndrome (ARDS), occurring in association with transfusion of blood components; the incidence was reported variously, ranging from 1 in 5,000 to 1 in 50,000. It is caused by the presence of antibodies against leucocyte antigens (HLA antigens seem the most frequent [ 11 ]) in donor plasma [ 26 ]. To avoid this risk, preference should be given to the use of plasma from male donors or from females who have never been pregnant, including abortions. This measure lowers the possibility to find antibodies against HLA or granulocyte antigens causing TRALI in the donor plasma [ 11 ]. Moreover, it was reported that the presence of certain antibodies may cause immune enhancement of pathogenicity, termed ADE (antibody-dependent enhancement), for several viral diseases, such as dengue virus and SARS [ 27 ]. Physiochemical reactions can sometimes be severe, but usually not life-threatening: Fluid overload is one of the most common complications of transfusion and can lead to pulmonary oedema Citrate toxicity depends on the action of citrate in binding calcium and therefore in reducing the availability of ionised calcium for normal neuromuscular function; it is not frequent because citrate is rapidly metabolised by the liver Some units of plasma might contain chemicals (e.g., drugs) derived from the donor to which the recipient might react Modern technologies allow to minimize infectious risk. Firstly, bacterial transmission is not a significant risk factor as the plasma is frozen within hours of collection and processing. Secondly, an accurate selection of donors and pathogen reduction processes can be applied to minimize viral infection risk (for single-unit components, methylene blue is used; for plasma pools, solvent detergent is used) [ 26 ]. Specific side effects are identified for single plasma components; immunoglobulins have been associated with thrombotic events, renal toxicity, and aseptic meningitis [ 26 ]. Tamburello and Marando reported that treatment with human immunoglobulin during the SARS-CoV-2 pandemic was associated with a significantly increased risk of same-day thrombotic events (from 0.04 to 14.9%) [ 3 ]. However, the estimated risk of serious adverse events is less than 5% [ 26 ]. A recent work by Joyner and colleagues explores the safeness of the use of convalescent plasma in 20,000 critically ill COVID-19 patients. The cohort studied is very huge; thus, the reported results should be considered particularly reliable. Serious adverse events within 4 hours of completion of COVID-19 plasma transfusion were 146 (less than 1% of all transfusions). 50 events were judged surely unrelated to plasma transfusion. Among the other events, there were 83 nonmortality events reported (37 reports of transfusion-associated circulatory overload, 20 reports of transfusion acute lung injury, and 26 reports of severe allergic transfusion reaction). 13 mortality events happened; they were judged only as possibly, not definitely, related to the transfusion of COVID-19 convalescent plasma. Notably, the vast majority of other serious adverse events, which happened within seven days of completion of the convalescent plasma transfusion, were judged to be unrelated to the plasma transfusion [ 28 ]. 6. Convalescent Plasma: Recommended Therapy in the Context of a Pandemic Recently, in 2017, the WHO Blood Regulators Network (BRN) published a position paper, recommending the need for healthcare systems to prepare adequate infrastructures to deal with the emergence of any pandemic caused by new emerging viruses; in that paper, the BRN suggested plasma from recovered patients as the first-choice treatment to be tested. Based on the evidence from past experience in passive immunization, the BRN explained that there was a considerable possibility that the application of whole blood (as well as plasma, serum, or immunoglobulin concentrates) from convalescent persons could be effective in the treatment/prevention of infectious disease. Thus, in the absence of effective vaccines and antiviral therapies for the emerging pathogen, an organized program to collect convalescent plasma or serum from disease survivors could provide a potentially valuable empirical intervention, while data on the effectiveness and safety of its use are obtained through orderly scientific studies [ 16 ]. In fact, any blood derivative should be considered a drug, and if administered for different indications from the authorized ones, it must be tested for the specific new application [ 29 ]. Epstein and Burnouf updated the BRN recommendations to the current pandemic caused by SARS-CoV-2 [ 11 ]. 7. Results of the First Studies regarding Convalescent Plasma for the Treatment of COVID-19 Several case reports and case series reporting convalescent plasma for COVID-19 patient treatment are available; most of them show positive results in terms of efficacy and safety of the convalescent plasma for treating COVID-19 [ 30 ]. The first case series was published by Shen et al. at the end of March 2020. The authors reported a case series of 5 critically ill patients with acute respiratory distress syndrome (ARDS) under mechanical ventilation. The study compared the clinical outcomes before and after the transfusion. It was observed that ARDS resolved in 4 patients at 12 days after transfusion; in addition, 3 patients were discharged, and 2 patients were in stable conditions [ 31 ]. Unfortunately, data from case reports and case series are observational, and they are not sufficient for a definitive validation of the treatment. Some controlled trials are already available too; they confirm the outcomes of the first case series. A randomized control trial out of Wuhan was the first to be published: 103 patients with severe or life-threatening COVID-19 (52 in the convalescent plasma-treating group and 51 in the control group) were enrolled, but unfortunately, the study had an early termination due to low patient enrollment as the regional outbreak waned. Contrary to expectations, the study failed to detect a statistically significant difference in the evaluated outcomes (time to clinical improvement, 28-day mortality, and time from randomization to discharge). However, convalescent plasma was demonstrated to be associated with antiviral activity in patients with COVID-19 (convalescent plasma treatment was associated with higher rates of negative SARS-CoV-2 viral PCR results from nasopharyngeal swabs at 24, 48, and 72 hours); a statistically significant improvement was noted for the convalescent plasma treatment group compared to controls in the subgroup of patients without life-threatening COVID-19 (91% improvement in the plasma group compared to 68% in the control arm). The median between the onset of symptoms and the beginning of the treatment was 30 days. This time window could have affected the study to detect a clinically important benefit of the convalescent plasma therapy, in addition to the early termination of the study and to the type of patient conditions (only severe or life-threatening disease) [ 32 ]. The researches of Hartman et al., of Liu et al., and of Salazar et al. confirmed that early convalescent plasma administration is of greater clinical benefit than delaying transfusion in patients with severe disease. Hartman et al. described a series of 31 patients (16 patients with severe disease and 15 patients with life-threatening disease). They demonstrated that convalescent plasma is associated with reducing ventilatory requirements in patients with both severe and life-threatening diseases [ 33 ]. These results are consistent also with a recent cohort study by Liu and colleagues. In this study, 39 patients were treated with convalescent plasma and were compared to 156 control patients. The authors reported a lower mortality rate among patients with severe or worse disease who received convalescent plasma and significantly better outcomes among patients transfused prior to mechanical ventilation [ 34 ]. Salazar et al. enrolled 387 patients (136 transfused patients and 251 nontransfused control COVID-19 patients); they found that patients transfused within 72 h of hospital admission had decreased mortality within 28 days, whereas patients transfused after 72 h of hospital admission did not. These data demonstrate that early convalescent plasma transfusion after hospital admission reduces mortality within 28 days posttransfusion [ 35 ]. Two other articles deserve to be mentioned, the one of Duan et al. and the other of Joyner et al. Duan et al. compared 10 severely ill patients treated with convalescent plasma to a historical control group of 10 severely ill patients not treated with convalescent plasma. The COVID-19-transfused patients' group showed better clinical outcomes than the historical control group. All enrolled severe COVID-19 patients had improvement of clinical symptoms and showed different degrees of absorption of the pulmonary lesions after convalescent plasma transfusion. The authors showed amelioration of routine laboratory criteria and pulmonary function (lymphocytopenia, an important index for prognosis in COVID-19, tended to be improved after convalescent plasma transfusion). Increase of neutralizing antibody titers was demonstrated [ 36 ]. Joyner et al. studied the effects of convalescent plasma use in a very huge cohort of 20,000 critically ill hospitalized patients. The aim of the paper differs from the ones previously mentioned: the study was designed to demonstrate the safety of convalescent plasma. Anyway, the seven-day mortality rate in this extremely high-risk cohort of patients was 8.6% only. The authors anticipated the intent to create a control comparator group using patients hospitalized with COVID-19 during the same period; they will discuss potential convalescent plasma efficacy in a future publication. Given the large number of observations, it is expected that the results of this study will have significant importance in evaluating the efficacy of the treatment and its reliability [ 28 ]. Moreover, there are several ongoing randomized controlled trials on the role of convalescent plasma to treat COVID-19 (Zheng et al. estimated that the main underway trials are 48 in the world). Also in this case, a positive confirmation of the results in terms of the efficacy of the treatment for COVID-19 is strongly expected [ 37 ]. 8. Convalescent Plasma for COVID-19: Practical Points Preparation requirements for convalescent plasma follow the standard operating procedures for plasma collection and all applicable regulations. Thus, health system requirements are the same for routine plasma collection procedures via plasmapheresis. During plasma donation procedure, the blood cells and plasma are removed from the body and separated by a plasmapheresis machine; then, the blood cells are returned to the donor while plasma is collected. Plasma products are stored as fresh-frozen plasma, until usage. Recently, approved serological assays are necessary to detect SARS-CoV-2 (RT-PCR test) in serum and virologic assays [ 35 ]. Regarding plasma treatment in the context of the current pandemic, the following points are worth remarking. Plasma should only be collected from selected recovered individuals diagnosed with COVID-19 for at least 3 weeks. At least 14 days must have elapsed since complete recovery, in order to minimize the possible risk of SARS-CoV-2 in the blood. The titer of anti-SARS-CoV-2 IgG should be determined, and virus inactivation procedures should be strictly attended before using plasma [ 5 ]. Actually, the recommended viral neutralization titer cut-off for COVID-19 convalescent plasma is at least ≥1 : 160. This corresponds to a receptor binding domain IgG titer ≥ 1 : 1350 [ 35 ]. A titer of 1 : 80 may be considered acceptable if an alternative matched unit is not available [ 38 ]. Although largely experimental, the optimal dose of convalescent plasma to be administered to a COVID-19 patient ranges between 200 and 500 ml [ 28 , 31 – 33 , 35 , 36 ]. Treatment effectiveness is expected to be better when immune plasma is collected from patients of the same city, or surrounding area, since it is assumed that these donors have defeated the same virus (virus genome can mutate); likewise, lifestyle, diet, and environment play an important role in the development of specific antibodies against the virus [ 39 ]. As a principle, convalescent plasma should be used as soon as possible in the acute stage of the disease of the recipient [ 4 , 5 ], and it is important that the titer of anti-SARS-CoV-2 antibodies be high. It is essential to ensure ABO compatibility between donor and recipient; theoretically, transfusion of plasma from at least two donors may be better to achieve more effective immune protection from delivery of diverse antibodies. Standard selection criteria for plasma donation, according to local requirements, must always be followed, as well as standard postdonation treatment of plasma [ 11 ]. 9. Perspectives Unfortunately, the BRN recommendations have been disregarded. At the beginning of the present pandemic, no healthcare system had already organized programs to collect convalescent plasma or serum from recovered patients to fight a potential new emerging viral pathogen. Currently, some clinical studies and programs have started, but unfortunately, the procedures are very slow [ 7 , 40 , 41 ]. The first reported results are very encouraging and confirm the effectiveness of plasma therapy and its safety [ 42 – 44 ]. The classical process to approve a new drug for clinical use is long, but a terrible pandemic emergency is underway and the time to wrest the fate of many people from death is very short. It is essential that governments follow the strategy recommended by the BRN in 2017. Recently, the Food and Drug Administration (FDA) has also moved on this path, in consideration of the numerous evidences of efficacy and safety of plasma therapy coming from past experiences and the first scientific confirmations in the current pandemic. The administration has remarked the importance to study the safety and efficacy of COVID-19 convalescent plasma enrolling patients in clinical trials. In addition, two other strategies have been authorized to allow patients to access treatment. Firstly, it provides an expanded access for the use of COVID-19 convalescent plasma dedicated to patients with serious or immediately life-threatening COVID-19 disease, who are not eligible or unable to participate in randomized clinical trials (21 CFR 312.305). Secondly, given the public health emergency, FDA facilitates access to COVID-19 convalescent plasma for patients with serious or immediately life-threatening COVID-19 infections: the patient's physician can request a single emergency Investigational New Drug Application to obtain expanded access for an individual patient (21 CFR 312.310) [ 29 ]. This strategy could allow to save as many lives as possible. In addition, the FDA is promoting an awareness campaign to invite patients recovered from COVID-19 to donate plasma [ 45 ]. Expected results from ongoing trials should definitely support the researches in finding the best criteria for including/excluding convalescent plasma in COVID-19 patients' treatment. As detailed in the available studies, performed analysis suggests convalescent plasma for the most serious cases and at their early stage. Thus, the early recognition of the COVID-19 patients who may develop critical illness is the key question for convalescent plasma treatment. They are the patients to be treated with convalescent plasma. It is known that most mild COVID-19 patients can be self-recovered, and convalescent plasma may be inappropriate therapy for them. And for end-stage COVID-19 patients, the convalescent plasma treatment may not be able to regress the poor outcome as demonstrated by the current studies [ 46 ]. 10. Conclusions The strategy of the FDA seems the most correct, since convalescent plasma appears to be a safe and effective therapy, and a vaccine requires a long time to get ready. To date, there are no other authorized therapies against SARS-CoV-2. Furthermore, it should not be forgotten that the other currently applied therapies (e.g., antiviral drugs and hydroxychloroquine) have remarkable side effects and are administered for compassionate use [ 7 , 47 ]. Another key point is that convalescent plasma should be hyperimmune and contain high antibody titers against SARS-CoV-2. It is still unknown how long patients have good antibody levels in their blood [ 48 ]; therefore, at least hypothetically, the time window for convalescent plasma donation is limited to the first period after a patient's full recovery. Fortunately, many patients in the world are recovering from COVID-19 infection. This should be the right time to donate plasma to treat seriously ill patients. Governments should be aware of this opportunity and start organizing appropriate plasma donation campaigns and adequate plasma collection programs. Conflicts of Interest The author denies any conflict of interest.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7476041/

Biomedical Waste Management by Using Nanophotocatalysts: The Need for New Options

Biomedical waste management is getting significant consideration among treatment technologies, since insufficient management can cause danger to medicinal service specialists, patients, and their environmental conditions. The improvement of waste administration protocols, plans, and policies are surveyed, despite setting up training programs on legitimate waste administration for all healthcare service staff. Most biomedical waste substances do not degrade in the environment, and may also not be thoroughly removed through treatment processes. Therefore, the long-lasting persistence of biomedical waste can effectively have adverse impact on wildlife and human beings, as well. Hence, photocatalysis is gaining increasing attention for eradication of pollutants and for improving the safety and clearness of the environment due to its great potential as a green and eco-friendly process. In this regard, nanostructured photocatalysts, in contrast to their regular counterparts, exhibit significant attributes such as non-toxicity, low cost and higher absorption efficiency in a wider range of the solar spectrum, making them the best candidate to employ for photodegradation. Due to these unique properties of nanophotocatalysts for biomedical waste management, we aim to critically evaluate various aspects of these materials in the present review and highlight their importance in healthcare service settings. 1. Introduction Biomedical waste management has recently risen as one of the major challenges that developing countries are confronting. The amount of biomedical waste produced has considerably increased as the worldwide populace has expanded, and accessible assets are not sufficient to deal with it [ 1 ]. Disposal and post treatment of the waste produced in the healthcare system may indirectly cause health hazardous through the release of pathogens and toxic pollutants into the environment. Discarding the untreated healthcare wastes in landfills, if the landfill is not properly constructed, can lead to the contamination of surface, drinking and ground water resources. Additionally, treatment of healthcare wastes with chemical disinfectants can result in the release of chemical substances into the environment if those substances are not handled, stored and disposed of in an environmentally sound manner. Incineration of waste has also been widely practiced; however, insufficient incineration or the incineration of inappropriate materials pollute the air and generate ash residue. Lack of knowledge about the health-related hazards of healthcare waste, inadequate training in proper waste management, absence of waste management and sufficient disposal systems, insufficient financial and human resources and the low priority given to the topic are the most common problems associated with healthcare waste. Many countries either do not have appropriate regulations, or do not enforce them. Successful and efficient management of biomedical waste requires the use of different treatment practices and techniques, such as incineration, autoclave, hydroclave, and microwave treatments [ 2 ]. It is essential that every single applied innovation assures both the environment and public health protection [ 3 ]. Waste management and waste engineering advances have turned out to be extremely vital because of the significant increase in rate and diversity, both in the quality and quantity of the waste that is being produced every day, so using the most financially plausible strategies has become even more crucial than before [ 4 ]. Since waste products cannot be completely eradicated, the choice of waste treatment has become specifically important as its management. The other commonly used methods to treat biomedical waste are mechanical treatments such as granulation, pulverization, shredding, grinding, mixing, agitation, and crushing. This type of treatment has the advantage of reducing the bulk volume of the waste materials by 60 percent or more. Although mechanical treatment does not remove the pathogens or disinfect equipment, it reduces the waste volume to facilitate further treatment or disposal. Equipment involved in mechanical treatment includes but is not limited to crushers, millers, shattering machines and splinterers. These treatment methods can alter the appearance of the waste, which can be useful in reducing the psychological impact of the waste on human observers. In addition to reducing the volume of bulk disposal, mechanical treatment can increase the surface area of the solid pieces before subsequent chemical or heat treatment. Chemical disinfection, such as through the use of chlorine compounds, has been widely used to eliminate the microorganisms in medical waste, as well as oxidizing hazardous chemical constituents. For instance, chlorine bleach has been used to disinfect swimming pools and reduce the risk of disease transmission. Another example of chemical disinfectant compound is ethylene oxide treatment, which is used to disinfect materials and sometimes to treatment of medical waste. Ethylene oxide (EtO) treatment is used to sterilize the equipment that will be frequently used. This disinfectant chemical is not cost effective for use on equipment or treatment of waste that will be disposed of in a landfill. EtO gas can be used to kill microorganisms and disinfect products during packaging processes. Microwave radiation has recently been employed to treat wastewater sludge and to generate heat for treating medical waste. This method of waste treatment can be employed either on-site or mobile by using treatment vehicles. To enhance the efficacy of the microwave treatment and reduce the volume of the final product, the waste will go through a shredding process first. In the case of using dry waste, the waste is wetted with water and then introduced into the microwave chamber, as this method of disinfection is effective only when the waste is damp. Therefore, the microwave treatment units are usually supplied with a humidifier. Although the whole disinfection time is determined by the manufacturer and experience of the operators, approximately 20 min per each batch is required. Microwave should not be confused with irradiation such as gamma rays (from radioactive elements) or electrons, as these two methods are completely different. Gamma irradiation is a means of sterilization by exposing waste to gamma rays, as it breaks down bacterial DNA. To generate gamma rays, radioactive isotopes of cobalt are employed, this is the same radiation source used for the radiation treatment of cancer. However, in cancer treatment, the radiation is intended to kill the malignant cells, whereas to sterilize equipment or treat waste, pathogens are targeted. In contrast, the ultraviolet (UV) radiation used to treat wastewater is not capable of killing microbes so much as it is able to break down chemicals. The efficiency of irradiation as a means of sterilization is highly dependent on the total energy delivered, but even then, this method of treatment suffers from the shadowing effect, which means that waste surfaces facing the radiation source are more sterile than the waste on the shaded side. Therefore, waste with odd shapes, and the sides of contaminated surfaces facing away from the cobalt source, may not be adequately exposed to the radiation. Heat treatment, by contrast, brings every piece of waste to an adequate temperature for sterilization, if done properly. Although vitrification—the means production of glass—has rarely been used, it could be an effective treatment for medical waste. The high temperature kills pathogens and some combustible material via burning or pyrolysis, which results in an off-gas. The remaining by-product is encapsulated in glass, which has a very low diffusivity. However, this method of disinfection might become dangerous if significant quantities of the encapsulated hazardous material leach out of the glass. Ultimately, the vitrified waste can be disposed of in a landfill with confidence. Despite the development of plasma treatment as an alternative to incineration for medical waste treatment, it has not been widely implemented [ 5 ]. To decrease the expense of waste treatment through cost-effective strategies, various methods based on the exploitation of sunlight have been proposed for both solid and liquid waste management [ 6 ]. Among these, photocatalysis is a remarkable technique with a variety of applications, including the debasement of different contaminations in wastewater [ 7 ], antibacterial functions [ 8 ], cleansing of air [ 9 ], and generation of hydrogen [ 10 ]. The photocatalytic procedure is attracting more focus in the field of ecological and environmental safety, as there is a need to achieve the utmost degradation of contaminants attainable under states of mild pressure and temperature. The significant highlight of these procedures is the incorporation of cost-effective near-UV (from 400 nm down to 300 nm) light, with sunlight as an alternative source of irradiation. The term photocatalysis refers to a chemical reaction using light in the presence of a catalyst that assimilates light quanta and is associated with the chemical transformations of the reactants [ 11 ]. The optimal treatment system depends on many factors, such local conditions, availability of resources including technical expertise, waste characteristics and volume, relevant national regulations and safety requirements, technical requirements for installation, operation and maintenance of the treatment system, environmental factors and cost considerations. Nonetheless, waste management systems can be changed and improved only within the financial and technical capacity of a given health-care system, which may then require making small decisions towards an incremental improvement, as well as planning for the attainment of long-term improvements, once certain conditions have been met. Nanophotocatalysts have been widely used to treat waste in the field of environmental and ecological safety, as they have numerous benefits, such as those of low cost, superb stability, high photocatalytic activity, innocuousness to humans, etc. [ 12 ]. Different methods, including ion exchange microorganisms and adsorption, have been used to treat sewage. However, these methods are restricted due to their complex technology, high cost, risk of second contamination, and poor degradation effectiveness [ 13 ]. Compared to other methods for biomedical waste management, nanophotocatalysts are considered one of the most particular strategies with respect to energy consumption, environmental and ecological issues. The advantages of the photocatalysis strategy are as follows [ 14 ]: (a) photocatalysis offers a decent alternative to the traditional energy-concentrated treatment techniques (e.g., ultrafiltration and reverse osmosis) with the capability of using pollution-free and renewable solar energy; (b) it prompts the creation of innocuous products, in contrast to traditional treatment methods in which pollutants only transfer from one phase to another; (c) the procedure can be used for the decimation of an assortment of risky and hazardous compounds in various wastewater streams; (d) it requires less chemical input and can be operated under mild reaction conditions with modest reaction time; (e) minimum generation of secondary waste; (f) it can also be applied to a solid phase (soil), gaseous phase (hydrogen generation), and aqueous treatments. The dependency of the photocatalytic activity on the following criteria has hampered their application [ 15 ]: (a) charge separation; (b) interfacial charge transfer needs to be improved; (c) charge carrier recombination can be inhibited. Although approaches to removing pollutants based on nanostructured catalytic membranes, nanosorbents and nanophotocatalyst are eco-friendly and efficient, in order to purify the waste, they require more energy and sufficient investment. There are many challenges involved in biomedical waste treatment; some precautions are required to keep hazardous waste away from ecological and health issues. New modern equipment for waste treatment is required to be flexible, low cost and efficient for commercialization purposes. Recently, with advancements in nanomaterials such as nanophotocatalysts, nanomotors, nanomembranes, nanosorbents and imprinted polymers, the decontamination of biomedical waste has been effectively revolutionized. However, there has not been a systematic characterization of the risk and hazards related to nanomaterials, and there is a lack of safety regulations for using such catalysts. Overall, nearly all nanocatalysts have toxic effects both in vitro and in vivo at certain concentrations. For instance, ROS generation and cell signaling perturbations are widely accepted causes of nanotoxicity. Furthermore, the toxicity of nanoparticles is also determined by factors such as particle size and surface functionalization [ 16 ]. Although to some degree, the toxicity and adverse effects of commonly used nanocatalysts have been realized, a comprehensive investigation is still required [ 17 ]. This review aims to introduce different types of nanophotocatalysts and provide the main principles, mechanisms, and operating parameters of the photocatalysis process and emphasize its importance in biomedical waste management in detail (see Figure 1 ). 2. Fundamentals and Mechanism of Photocatalytic Reactions Photocatalysis is defined as a series of chemical reactions that is usually initiated by electromagnetic irradiation. The photocatalysis process can be divided into two main stages of reduction and oxidation. When a material is irradiated with photons with energy equal to or higher than its bandgap, the excited electrons in the conduction band (CB) will jump to the valence band (VB) through the bandgap leaving positive holes, which is called reduction. As a result, the generated electrons and holes lead to the formation of reactive oxygen spices (ROS) such as O 2 and OH (oxidation). The kind of ROS depends on the type of material and irradiated photons. A conventional photocatalysis procedure is demonstrated in Figure 2 . The formation of ROS is the significant outcome of photocatalysis that can lead to various effects such as degradation of dye and antibacterial activity. The efficiency of the photocatalysis process can be estimated by its impact on its surroundings, such as degradation, reduction, adoption, or antibacterial activity. The conventional method for assessing the efficiency of the photocatalytic process is to compare between the initial concentrations of the undesirable substances with the concentration of these substances after the photocatalytic reactions. The efficiency of photocatalytic activity is defined by the yield of electrons and holes created. The recombination of the electron–hole pair is one of the main factors that diminishes the efficiency of the photocatalytic activity. In the photocatalytic process, the catalyst and light are simultaneously used to accelerate a chemical reaction. Thus, photocatalysis can be defined as the catalysis-driven speed-up of a light-actuated reaction. Photocatalysts are categorized into two classifications: hetero- and homogeneous procedures [ 19 ]. When comparing the two processes, the heterogeneous procedure is, in fact, a feasible strategy that can be exploited to reduce a variety of wastewater pollutions, and which has some benefits over other competing methods [ 20 ]. This procedure does not suffer from waste disposal problems, but offers complete mineralization and low cost, with the only requirement being mild pressure and temperature conditions. Homogeneous photocatalysis is mostly used with metal compounds as catalysts (e.g., copper, chromium, and iron transition metal complexes). In this procedure, thermal and photon conditions are used concurrently to generate hydroxyl radicals from the higher oxidation state of metal ion complexes. Hence, the reaction between these hydroxyl radicals with organic matters deactivates toxic compounds [ 21 ]. Photocatalytic response fundamentally relies upon the catalyst and energy of the light (photon). Semiconducting materials are generally used as catalysts for sensitization; the light irradiation stimulates the redox process because of their characteristic electron structure by a vacant conduction band and a filled valence band [ 22 ]. The influence of different factors on photocatalytic efficiency has been investigated and further optimized using the response surface methodology (RSM). The RSM is a combination of both statistical and mathematical methods to optimize a complicated process. To optimize the process more accurately, this statistical design of experiments considers the interaction effects between the studied parameters, and can determine the combination of levels. Furthermore, a central composite design (CCD) based on RSM can be successfully applied in the optimization of photodegradation of various organics [ 23 ]. Since the adsorption of pollutants on the surface of solids plays a critical role in heterogeneous photocatalysis, the effect of adsorption on photocatalytic process has been investigated. Simultaneous adsorption and photocatalytic processes has been used for degradation of phenol pollutants with noticeable recyclability and stability of the photocatalyst [ 24 ], bacterial disinfection [ 25 ], photocatalytic water purification [ 26 , 27 , 28 ], and photocatalytic mineralization of phenolic compounds. 3. Types and Characteristics of Nanophotocatalysts Nano-sized photocatalyst particles demonstrate a significantly intensified reactivity compared to larger particles or bulk materials due to their large surface area [ 29 , 30 ]. Novel nanophotocatalysts have been developed with the ability to exploit solar energy to synthesize organic compounds under controlled conditions [ 31 ]. Nanophotocatalysts are mainly classified as surface plasmon resonance-mediated, metal-organic charge-transfer-based, and semiconductor-based nanophotocatalysts, capable of driving various organic reactions photocatalytically. Accordingly, they can be categorized as graphene semiconductors, composites of two semiconductors, core–shell composites, non-metal-doped semiconductor materials, and metal-modified semiconductors. Metal-free catalysis, such as with graphitic carbon nitride nanocomposites, has also been used in photodegradation of aqueous phase organic pollutants [ 32 ]; likewise, polymeric graphitic carbon nitride-based Z-scheme photocatalytic systems, magnetic graphitic carbon nitride photocatalyst, and carbon quantum dot-supported graphitic carbon nitride have been respectively employed for sustainable photocatalytic water purification [ 28 ], degradation of oxytetracycline antibiotic [ 33 ], and photodegradation of 2,4-dinitrophenol [ 34 ]. Among different kinds of photocatalysts, metal oxide semiconductors, including TiO 2 , ZnO, α-Fe 2 O 3 , and WO 3 are the most suitable ones, since they are photocorrosion resistant and have a wide band gap energy. TiO 2 is currently used as the most efficient photocatalyst, and is widely applied in wastewater treatment, since it can promote the oxidation of organic compounds while being thermally stable, non-toxic, cost-effective, and chemically and biologically inert. Structural and surface properties, including surface area, porosity, crystal composition, particle size distribution, and band gap energy, are able to affect the photocatalytic activity of the catalyst [ 35 ]. The following characteristics typically represent nanophotocatalysts with specific favorable benefits over bulk materials [ 36 ]. First, their large surface area to volume ratio results in a high particle fraction, and subsequently a high division of active sites on the catalyst surface. Second, their valence band–conduction band energy gap strongly depends on the size of the nanoparticles [ 37 ]. Moreover, changing the size of the nanocatalyst makes it possible to adjust the absorbance wavelength. Additionally, the optical and electronic characteristics of the nanocatalyst can be modified by tuning their sizes [ 38 ]. Due to all these favorable properties, nanophotocatalysts have been employed in a wide range of reactions, for instance, in organic synthesis, splitting water to hydrogen fuel generation, inactivation of cancer cells, and dye degradation [ 39 , 40 ]. Primary studies on the development of nanophotocatalysts have mainly focused on their degradation capabilities of pollutants and dyes. Given these properties, there is an increasing interest in the use of nanophotocatalysts as catalysts for different organic reactions, providing options in green chemistry as alternatives to the regular techniques used in research laboratories and industry, which apply thermal energy to achieve the same goals [ 41 ]. The efficacy rate in photocatalysis methods is primarily dependent on the size, shape, crystal structure and surface area of the photocatalyst, as well as the morphology, which can also act as a significant factor affecting the final degradation throughput. The amount of catalyst is also directly proportional to the overall rate of photocatalytic reaction. Disinfection efficiency of the photocatalyst can be improved by an increase of its doses [ 42 ]. The pH of the solution is another effective factor, as it determines the photocatalyst surface charge properties. Furthermore, various pH values may impact the efficiency of the disinfection process, values ranging from 6.0 to 8.0 have shown the highest impact. Different studies have evidenced that the microorganisms have pH sensitivity at around 6.5 and 8.0, the range in which photocatalytic activity has been demonstrated to be excellent. This is due to the fact that as pH moves away from neutral, the effectiveness of the overall process declines when pH reaches 7 [ 43 ]. The optimum range of reaction temperature which primarily depends on the activation energies of the materials in the photocatalytic reaction [ 44 ], the light intensity, which mainly influences the degradation rate of photocatalytic reaction [ 45 ], and the nature and concentration of pollutants [ 42 ] can also affect the performance of nanophotocatalysts. Furthermore, inorganic ions such as iron, magnesium, copper, zinc, phosphate, bicarbonate, chloride, sulfate and nitrate can change the rate of photocatalytic degradation of the organic pollutants since they can be adsorbed onto the nanophotocatalyst surface [ 46 ]. 4. Biomedical Applications of Nanophotocatalysts As photocatalysts have a superior capability in the deactivation of various destructive microorganisms, they could reasonably be used as alternatives to conventional techniques (e.g., chlorination), which can generate harmful and undesirable by-products [ 47 ]. Photocatalysis is a flexible and successful procedure that can be adopted in numerous cleansing applications in both air and water frameworks [ 48 ]. Furthermore, photocatalytic surfaces have been used on account of their self-sanitizing attributes. Photocatalytic applications have been recently developed specifically in the contexts of environmental health and indoor air, plant protection, effluents, wastewater and drinking water disinfection, dye removal, the pharmaceutical and food industries, laboratories and hospitals, and biological and medical applications. Due to the low energy consumption and feasible accessibility to solar energy and decreased treatment time, the overall cost for photodegradation of hazardous compounds and pollutants is lower, and hence beneficial [ 49 ]. Applications of nanophotocatalysts are summarized in Figure 3 . With respect to the title of the study, we focus on biomedical applications of nanophotocatalysts in the following sections. 4.1. Laboratory and Hospital 4.1.1. Basics and Fundamentals Hospital waste can be described as a combination of both biological and non-biological waste that is discarded and not intended to be used again. Hospital and laboratory waste can be broadly defined into two categories: hazardous (risk waste) and non-hazardous (non-risk) waste ( Figure 4 ). Risk waste includes sharps, pathological, pharmaceutical, chemical, and radioactive waste, whereas non-risk waste is equivalent to general domestic garbage and introduce no greater risk than normal home waste (e.g., paper, packaging, and food waste). A portion of hospital and laboratory waste management processes includes segregation, handling, transportation, disinfection, mutilation, storage, and final disposal as the foundation for essential progress towards scientifically safe waste management [ 50 ]. There are two main steps involved in the waste disposal process: treatment and final disposal. In the treatment step, depending on the nature of the waste, procedures such as incineration, autoclaving, chemical disinfection, encapsulation, and microwave irradiation are usually applied. The disposed wastes are likely to wind up in landfill, buried inside the premises, discharged into the sewer, etc. Segregation (separation) has a key role in an efficient waste identification and management process. Sorting the waste based on the color of the container is the most proper method for identifying waste categories [ 51 ]. The last treatment applied to the wastes can be performed by autoclave, hydroclave, incineration, or microwave technologies. Waste disposed of in landfill must be properly designated and managed, despite being the cheapest and most readily available way. On the other hand, shredding involves all types of bulk plastic waste, including risk waste, which are first disinfected, and then cut into small pieces and converted into a compact form. This approach is cost effective, and is much cheaper than an incinerator, while causing comparatively no pollution. Autoclaving, however, as the last type of waste disposal is more expensive, but more promising than shredding. People who are exposed to hazardous hospital waste are potentially at higher risk—either those who handle the waste at any stage or are exposed as a consequence of careless management [ 52 ]. The most to least favorable options of waste management are demonstrated in Figure 5 . 4.1.2. Photocatalytic Point of View Photocatalysis is an efficient, affordable strategy for the decontamination and disinfection of hospital waste that applies ultraviolet rays or solar energy to disintegrate antibiotics and disinfect microbes from the waste at the point of origin (see Figure 6 ). In demanding cases such as medical care systems and especially in microbiological labs, frequent and thorough cleaning of surfaces is required to inhibit bacterial transmission and diminish the emergence of microscopic organisms ( Table 1 ). Traditional sterilization strategies are time- and staff-consuming, and do not usually have long-term effectiveness. Additionally, the use of ultraviolet (UV) light instead of aggressive and dangerous chemicals can cause severe occupational health risks [ 53 ]. Titanium dioxide-coated surfaces can perform photocatalytic oxidation as an alternative to the conventional techniques of surface cleansing. The disinfection attributes of anchored titanium thin films on solid surfaces have been investigated in some previous studies [ 53 , 54 ]. The feasibility of this procedure for hygiene has been exhibited using bacteria, for example, Staphylococcus aureus ( S. aureus ), Pseudomonas aeruginosa ( P. aeruginosa ), Enterococcus faecium ( E. faecium ), and Escherichia coli ( E. coli ) [ 53 ]. Another example of self-disinfecting surface application is the inactivation of the deposited E. coli (ATCC8739) cells on membrane filters during fluorescent light irradiation [ 56 ]. In another study, a novel flame-assisted chemical vapor deposition (CVD) technique was used to test the antimicrobial activity of the stainless steel coated with TiO 2 thin films on E. coli [ 57 ]. Due to the extensive applications of the self-sterilizing material stainless steel, because of corrosion resistance, has been used for sterilization of Bacillus pumilus (B. pumilus) and has shown higher photocatalytic activity compared to the glass substrates coated with self-sterilizing materials [ 58 ]. Surfaces coated with CuO-doped Titania photocatalysts were also assessed for their biocidal activity and synergistic impact of photocatalysis and lethality of copper to deactivate bacteriophage T4 and E. coli [ 59 ]. Nitrogen-doped TiO 2 photocatalysts have been used owing to their visible light-induced bactericidal activity against human pathogens [ 60 , 70 ]. Visible light photocatalytic disinfection offers a continuous cleansing of surfaces that are constantly in contact with people, such as push buttons or door handles. This quality has been applied in the inactivation of E. coli using nitrogen/sulfur-co-doped Titania [ 61 , 62 , 63 , 71 ], which presents new disinfectant applications in public places that are consistently exposed to the transmission of pathogens (e.g., hospitals, schools, stations, hotels, airports, public toilets, and public transportation) [ 72 , 73 ]. In another study, prions, an infectious agent of a group of transmissible, fatal neurodegenerative disorders influencing both animals and humans, were inactivated using photocatalysts [ 48 ]. Photocatalytic oxidation, which is the cause of prion inactivation, can decrease the risk of spread, and has shown a significant effect on the pragmatic employment of this approach towards disinfection and cleansing of contaminated objects and surfaces, since these prions can be transmitted by ingestion of contaminated food or during medical treatments using contaminated surgical tools or biological materials. Controlling Legionnaire's disease, which is related to the contamination of hot water distribution systems with Legionella bacteria, is another application of photocatalysts in a laboratory or hospital [ 64 ]. It has also been shown that TiO 2 /UV photocatalytic oxidation has the capacity to mineralize the four strains of L. pneumophilia serogroup 1 cells (strains 977, 1004, 1009, and ATCC 33153) in laboratory-scale surveys, which proves that this could be used as a reasonable procedure for the routine disinfection and sterilization in order to control Legionella bacteria species in the hot water systems of emergency clinics and hospitals (e.g., hyperchlorination and thermal eradication) [ 48 ]. The application of TiO 2 as an effective treatment has also been reported to remove pharmaceutical contaminants from water with high efficiency. Due to their high photolysis sensitivity in aquatic systems, they have been used, for instance, to treat diclofenac [ 74 , 75 ], and triclosan [ 76 ]. 4.2. Biological and Medical 4.2.1. Categories and Conventional Methods Biomedical waste refers to any type of waste generated during treatment, diagnosis, or research activities—testing biological products and the immunization of human beings or animals. Biomedical waste management has become a general medical issue, as it is not only a legal necessity, but also a social responsibility. An improvement of biomedical waste management starts with the reduction of the waste produced [ 77 ]. According to an earlier report, around half of the world's population is in danger of incompatible biomedical waste management, which can impact both work and public places [ 78 ]. About 75% to 90% of total biomedical waste is general non-hazardous waste. The remaining 10% to 25% is dangerous and hazardous (including infectious waste, sharp waste, pathological waste, cytotoxic waste, pharmaceutical waste, liquid infectious waste, chemical waste, radioactive waste, and general health-care waste), which leads to a wide range of issues such as environmental and health risks. The main risks associated with biomedical waste include the microbial, systemic, and local infections caused by exposure to biomedical waste, among which pesticides, disinfectants, and mercury have multiple impacts in different ways, while inappropriate handling of sharps can lead to needle stick injuries and cause infections with blood-borne pathogens such as human immunodeficiency viruses (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) [ 79 ]. The essential needs in biomedical waste management include reducing, waste storage and transport, recycling and reusing, expenses in the annual budget, storage management, separate chemical and pharmaceutical waste segregation, separate storage zones, and documentation related to biomedical waste management [ 80 ]. Waste treatment technologies can be summarized as thermal (autoclaves, steam, microwave, dry heat treatment technologies) [ 80 , 81 ], chemical processes (sodium hypochlorite (NaOCl, 1%–12%), chlorine dioxide, calcium hydroxide, glutaraldehyde and peracetic acid) [ 82 ], incineration [ 83 ], encapsulation and inertization, irradiation technologies, biological processes, membrane bioreactors [ 84 ], disinfection and sterilization, as well as emerging technologies such as alkaline hydrolysis, plasma pyrolysis, superheated steam, ozone and promession [ 85 ]. Other upcoming technologies for the destruction of biomedical waste include base-catalyzed decomposition, gas-phase chemical reduction, sodium reduction, supercritical water oxidation, verification, superheated steam reforming, Fe-tetra-amido macrocyclic ligand (TAML)/peroxide treatment (pharmaceutical waste), biodegradation (using mealworm or bacteria to eat plastics), mechanochemical treatment, sonic technology, electrochemical technologies, phytotechnology, and solvated electron technology [ 86 ]. 4.2.2. Photocatalytic Strategies Photocatalytic processes are used in biomedical applications due to their disinfection abilities. It has been investigated that Staphylococcus aureus ( S. aureus ), a typical pathogenic bacterium in implant-related infection, shows photocatalytic activity using TiO 2 film on titanium substrates and stainless steel [ 87 ]. TiO 2 coatings have been used on bioimplants to apply photocatalysis for antibacterial purposes. The coatings exhibited a bactericidal impact upon UV irradiation, so the implementation of these photocatalytic-coated substrates are a valuable system for controlling infections related to biomedical implants [ 57 , 65 , 66 ]. As recently suggested for air purification systems, photocatalysis can be used to extract dangerous airborne biological risks, such as Anthrax, for example. Therefore, it is a versatile procedure for circumventing the spread of airborne biological threats and preventing bioterror dangers [ 67 ]. Reducing the number of bacteria and preventing their dissemination is essential and can be achieved through disinfection of surfaces in microbiological laboratories, food processing plants, veterinary medicine clinics, and hospitals. Traditional disinfection methods are generally time-consuming and tedious, and not usually sufficient (e.g., cleansing with chemical disinfectants) [ 88 ]. A significant alternative to conventional disinfection methods is the photocatalytic process by coating the surfaces with a thin layer of metal oxide nanostructures [ 89 ]. The high bactericidal property has added value for the practical application of photocatalysis in the treatment of microorganisms (e.g., Pseudomonas aeruginosa , E. coli , Enterococcus faecium , and S. aureus ) [ 68 , 90 ], and plays a crucial role in public health protection [ 69 , 91 , 92 ]. Metal Oxides The development of nanostructured metal oxides, semiconducting oxides, conducting oxides, composites, and polymers have been broadly investigated for quantification and detection of different hazardous biochemicals and chemicals [ 93 ]. It has been shown that metal oxides can photo-oxidize a wide range of organic substances such as alkenes, alkanes, surfactants, aromatics and pesticides [ 94 ]. Several metal oxides, including TiO 2 , ZnO, α-Fe 2 O 3 , MoO 3 , WO 3 , ZrO 2 and SnO 2 can be applied as photocatalysts, which are of great interest in current studies [ 95 ]. Titanium Dioxide (TiO 2 ) TiO 2 is desirable for photocatalysis due to its stability, inertness, and low cost. It is also recyclable and self-regenerating. One of the most important industrial applications of TiO 2 -based photocatalysts is the degradation of expired drugs and pharmaceutical compounds [ 96 ], dyes in textile industries [ 97 ], toxic compounds spills (e.g., pesticides) [ 98 ], natural toxins (e.g., cyanobacterial toxin microcystin-LR) [ 99 ], and a series of parabens as personal care products [ 100 ]. Another application of these nanophotocatalysts is in the treatment of winery wastewater by a photocatalytic reactor [ 101 ]. It has been evidenced in numerous studies that properties such as surface adsorption and photocatalytic reactions of nanocrystalline semiconductor particles are different from those of bulk materials owing to the increased reactive surface area. Application of Nano-Sized TiO 2 has been approved as a component of photocatalytic film covering scalpels [ 102 ], surgical masks [ 103 ], and catheters [ 104 ]. The UV-based disinfection effectiveness of nano-TiO 2 -coated catheters is three times greater than uncoated ones. Another TiO 2 /UV application with similar results has been demonstrated in infected dental implants [ 105 , 106 ]. The TiO 2 /UV process also showed high bactericidal efficiency in orthopedics, and cosmetic surgery in the presence of S. aureus on nano-TiO 2 coated implants, which is a valuable method for decreasing bacterial infection caused by the application of implants in medicine and biomedical fields [ 107 ]. Conventional TiO 2 photocatalysts, however, cannot provide purified and sufficiently safe drinkable water, since water remains toxic following treatment with TiO 2 nanoparticles. Therefore, the novel three-dimensional (3D) structured TiO 2 nanophotocatalyst can be replaced with TiO 2 nanoparticles by which both the safety level and efficiency of purification of the final purified water are preserved. These structures are suitable for environmental and biomedical applications, as they meet the human key safety conditions, according to an in vitro cytotoxicity test of well-purified water by eco-TiO 2 [ 13 , 108 ]. Various types of TiO 2 photocatalytic degradation of organic and inorganic pollutants are summarized in Table 2 . 2. Zinc Oxide (ZnO) ZnO shows efficient activity in photocatalytic degradation of organic contaminants as compared to TiO 2 [ 109 ]. Among different forms of zinc oxides (i.e. ZnO and ZnO 2 ), ZnO can form stable, protective coatings, which act as smart materials. Three different crystalline phases of ZnO include zinc-blende, wurtzite, and rock-salt. Due to these multi-functional qualities, ZnO is extensively used for various applications including photocatalysis [ 110 ], light-emitting diodes [ 111 ], biosensors, solar-cells [ 112 ], field-emission and gas sensing [ 113 ]. Different morphologies of ZnO have been reported, such as nanorods [ 114 ], nanonails, nanopencils [ 115 , 116 ], nanowires [ 117 ], nanotubes, nanobullets [ 118 , 119 ], nanocomb-like structures, nanobelts [ 120 ], nanoribbons [ 121 , 122 , 123 ], nanohelices [ 123 ], nanoneedles [ 124 ], and nanopins [ 125 ]. It has been proved that the properties of these materials are strongly dependent on the size and shape of the ZnO nanoparticles. The effective photocatalytic degradation of acridine orange up to about 90% after 80 min of exposure to UV light by ZnO nanocapsule is reported [ 126 , 127 ]. Photocatalytic degradation of methyl orange using ZnO as the photocatalyst has also been achieved—about 99.7% removal of the azo dye in 180 min [ 128 ]. In another study, almost complete degradation of methylene blue was obtained within 85 min of irradiation time using ZnO nanoparticles synthesized by hydrothermal treatment [ 129 ]. Additionally, ZnO-CeO 2 nanoparticles, which are discussed in the binary metal oxide category, have been applied to remove methylene blue and acridine orange [ 130 ]. 3. Iron Oxide (Fe 2 O 3 ) Iron oxides play an important role in many biological and geological processes. They are increasingly applied as pigments, iron ores, catalysts, and as hemoglobin in the blood. Freely dispersed bulk-, sonic-, and nano-Fe 2 O 3 have been used for photocatalytic oxidation of water under visible and UV irradiation [ 131 ]. Iron oxide (α-Fe 2 O 3 ) exhibits desirable efficiency as an important photocatalyst [ 132 ], with low cost, simple preparation, and n-type semiconducting behavior [ 133 ] with no secondary pollution [ 132 ]. Due to its various applications as sensors, pigments, actuators, and catalysts, it has attracted considerable attention in recent studies [ 134 , 135 , 136 ]. Photocatalytic degradation of organic pollutants via Fe 2 O 3 has been investigated. However, the corresponding photocatalytic mechanism has not been described in detail. The valence electrons of Fe 2 O 3 compared to those of TiO 2 can be excited to the conduction band at wavelengths shorter than 560 nm, which can extensively enhance the efficiency of the sunlight use. The maximum degradation efficiency of 94% for dibutyl phthalate in wastewater (as an excellent plasticizer in different resins, especially nitrocellulose and resins and also an vital additive in special paints and adhesives with about 20 years of hydrolysis half-life) was obtained using Fe 2 O 3 in a photocatalytic process [ 137 ]. Compared to α-Fe 2 O 3 powders, porous α-Fe 2 O 3 films exhibit better photocatalytic activity by water splitting under UV radiation for hydrogen generation [ 138 ]. Photocatalytic oxidation of aniline to azobenzene by Fe 2 O 3 under UV irradiation and natural sunlight in aprotic and protic solvents has also been reported [ 139 ]. 4. Gadolinium Oxide (Gd 2 O 3 ) The global interest in using rare earth metals is increasing, due to their distinctive magnetic and electronic attributes in the fashioning of interfaces and surfaces compared to common bulk materials. Gadolinia (gadolinium (III) oxide) is the most widely available derivative form of gadolinium, and is a potential contrast agent in magnetic resonance imaging (MRI). The Gd 2 O 3 -modified bismuth vanadate (BiVO 4 ) composite, as a photocatalyst, exhibits significantly greater visible-light photocatalytic activity than pure BiVO 4 for methyl orange degradation under visible light irradiation [ 140 ]. Gd 2 O 3 nanorods used to detect ethanol by facile hydrothermal routes demonstrated a lower detection limit with higher sensitivity and shorter response time [ 141 ] compared to the annealed Gd 2 O 3 nanostructures [ 142 ]. Moreover, a moderate photocatalytic activity was evaluated for degradation of methyl orange by uniform Gd 2 O 3 hollow microspheres [ 143 ]. The degradation of about 90% of 4-chlorophenol using modified Gd 2 O 3 photocatalyst prepared by the sol–gel method was measured after 4 h of UV light irradiation [ 144 ]. In another study, a challenging photocatalyst of Gd 2 O 3 nanorods was designed for the degradation of neurotoxic chloramphenicol drugs [ 145 ]. 5. Antimony Oxide (Sb 2 O 4 ) Antimony oxide is classified based on its oxidation states, Sb (III) and Sb (V). Antimony has been applied as a pacifier in enamel, flame retardants, paint and glass art crafts and for making bullets and bullet tracers. It has been used as an alloy for the synthesis of plain bearings, batteries, and solders, as well as as a stabilizer and a catalyst for the preparation of polyethylene terephthalate. The photocatalytic activity of α-Sb 2 O 4 has been demonstrated, with almost 52% degradation of acridine orange in 170 min, with low detection limit, good sensitivity, long linear dynamic range with good linearity in a very short response time [ 146 ], as well as for the removal of heavy metals (e.g., mercury) from waste water [ 147 ], while it has also been reported that synthetic Uranyl Selective Polymeric Membrane sensors based on p-tert-butylbiscalix4arene can be used for the determination of Thorium [ 148 ]. The unique characteristics of nanostructures, such as their large surface area, excellent adsorbing and absorbing activity, bio-friendly nature, and high electron exchange could be reasons for the good sensitivity of these systems [ 149 ]. Binary Metal Oxides In addition to metal oxides, some other metal oxides have also been studied previously for use in the field of photocatalysis, because of their unique benefits and wide range of applications as catalysts, semiconductors, superconductors, ceramics, antifungal agents, adsorbents, and their specific applications in medicines. Many metal oxide semiconductors (e.g., WO 3 , ZrO 2 , ZnO, and Fe 2 O 3 ) that have been exploited in photocatalysts for the degradation of organic contaminants have inherent drawbacks [ 150 ]. For example, WO 3 is a stable photocatalyst for O 2 production within the visible light irradiation range. However, it is not suitable for H 2 evolution because of its low level of conduction band. Additionally, α-Fe 2 O 3 is somewhat stable in acidic solutions, but has the same problems as WO 3 . Moreover, ZnO can be easily corroded under band gap irradiation by photogenerated holes. Ta 2 O 5 photocatalyst with a nanocrystalline mesoporous structure has recently been synthesized for the production of H 2 via a sol–gel process combined with a surfactant-assisted templating mechanism [ 151 ]. Recently, the effect of Fe-doped NiO as a co-catalyst has also been reported [ 152 ]. Additionally, ZnO-CeO 2 nanoparticles synthesized using an efficient and simple low-temperature method have been successfully applied as photocatalysts for the removal of biomedical and environmental contaminants and reported 80.7% and 92.1% degradation for methylene blue and acridine orange within 170 min of irradiation time, respectively [ 130 ]. The Cu x S-TiO 2 composites has shown good efficiency in photo degradation of dyes even under visible light irradiation [ 153 ]. In another report, the photocatalytic activity of high-quality CeO 2 -CdO binary metal oxide nanocomposites was evaluated, showing acceptable growth inhibition of P. aeruginosa ( Figure 7 ) [ 154 ]. Metal Sulfides Metal sulfides have been widely used as visible light responsive photocatalysts. Compared to metal oxides, 3p orbitals of sulfur in their valence band result in a more occupied valence band and a narrower band gap. Recently, among other metal sulfides, ZnS and CdS have attracted great attention. CdS is commonly used for visible light-assisted water splitting due to its suitable band position and band gap (2.4 eV). However, photo-corrosion, which is a common issue in most metal sulfide photocatalysts, occurs when using both CdS and ZnS. Therefore, recent studies have focused on the development of ZnS and CdS photocatalysts, mostly through four different means of improvement: matrixing and supporting the structures of CdS, adding cocatalysts to CdS, preparing porous and one-dimensional CdS, and doping solid solutions of CdS and ZnS [ 155 ]. To synthesize porous CdS, a solvothermal method has been used to synthesize CdS nanowires [ 156 ] and nanorods [ 157 ]. Additionally, mesoporous CdS nanoparticles have been synthesized via ultrasonic and template-free precipitation at room temperature [ 158 ]. Nanoporous CdS have also been prepared by including hollow nanorods and nanosheets with 3 nm diameter of pores through a two-step aqueous method [ 159 ]. Additionally, CdS quantum dots have been recently loaded on porous polysaccharides and applied as highly efficient contrast imaging agents [ 160 ]. On the other hand, due to having an extremely broad band gap to respond to visible light (3.6 eV), solid solutions of ZnS are formed in which the narrow band gap can increase the use of ZnS in visible light. Both CdS and ZnS have the same crystal structures, making it easy to form solid solutions of them [ 161 ]. Magnetic Nanophotocatalysts The incorporation of magnetic nanophotocatalysts in contaminant removal strategies has recently received significant attention due to their improved chemical and physical properties. Therefore, cost-effective, efficient, and environmentally friendly disinfection processes can be achieved due to their easy separation using an external magnetic field, which allows recycling and multiple use of the nanophotocatalyst [ 162 ]. They mostly have a core–shell structure consisting of a magnetic core (e.g., iron, cobalt, nickel, and their oxides like maghemite (α-Fe 2 O 3 ), magnetite (Fe 3 O 4 ), cobalt ferrite (CoFe 2 O 4 )) and a photocatalytic shell (e.g., TiO 2 , ZnO, AgBr, BiOCl) [ 163 ]. Furthermore, some nanoferrites like ZnFe 2 O 4 have shown desirable degradation efficiency of organic target compounds under both visible light and UV irradiation [ 164 ]. Similar studies have reported degradation of different contaminants using Fe 3 O 4 [ 165 , 166 , 167 ], NiFe 2 O 4 [ 168 ], CoFe 2 O 4 [ 169 ], ZnFe 2 O 4 [ 170 ], BaFe 12 O 19 [ 171 ], SrFe 12 O 19 [ 172 ] -doped TiO 2 nanophotocatalysts ( Table 3 ). A schematic of the use of magnetic nanophotocatalysts (MNPCs) in water treatment [ 173 ] is illustrated in Figure 8 . Graphene Graphene (G), due to its one-of-a-kind nanostructure and particular properties has been studied widely from both experimental and theoretical scientific points of view [ 208 ]. It has already shown promising applications in nanocomposites, nanoelectronics, optoelectronics, drug delivery systems, electrochemical super-capacitors, transistors, solar cells, and chemical sensors (e.g., biosensors, gas sensors, pH sensors) [ 209 ]. As shown in Figure 9 , graphene has been employed to enhance photocatalytic efficiency, due to its electron scavenging nature, in the conduction band of metal oxide [ 26 ]. Some of its novel applications include ultrasensitive high-adsorption ability for various types of contaminations, including arsenic in drinking water [ 210 ], brackish water desalination and drinking water purification [ 211 ], metal removal from the contaminated environment [ 212 ], detection of biomarkers [ 213 ], electrochemical sensor for paracetamol [ 214 ], treatment of thrombosis [ 215 ], protection of DNA from cleavage and its effective cellular delivery [ 216 ], MRI and localized photothermal therapy for cancer cell treatment [ 217 ], electrochemical immunosensor for sensitive detection of carbohydrate antigen 1.5-3 (CA 15-3) [ 218 ], and photothermal agents in NIR region [ 219 ]. Modified graphene nanostructures such as P25–G [ 174 , 175 ], TiO 2 –G [ 176 , 177 , 178 , 179 ], SnO 2 –G [ 180 ], Bi 2 WO 6 –G [ 181 ], ZnO–G [ 182 ], ZnFe 2 O 4 –G [ 183 ], BiVO 4 –G [ 184 ], CdS–G [ 185 ] have also been reported to have different photodegradation applications ( Table 3 ). Quantum Dots Quantum dots (QDs), as zero-dimensional semiconductor multifunctional nanomaterials have been receiving significant attention for the degradation of pollutants [ 220 ]. Since QDs have the advantage of the wide band gap of a semiconductor material, they have a promising application as photocatalysts, owing to the swift generation of electron–hole pairs through photoexcitation [ 204 ]. On the other hand, photocatalytic, chemical and optical properties of QDs can be improved by surface modification, which also improves the photostability of QDs, as well as the efficacy of light-induced reactions on the QD surface and the generation of new traps on the QD surface [ 221 ]. For example, in the self-photosensitization pathway of fuchsin dye degradation, photodegradation can be initiated in the presence of graphene quantum dots (GQDs) under visible light irradiation, as demonstrated in Figure 10 [ 222 ]. The application of modified QDs as a photocatalytic agent to degrade pollutants is illustrated in Table 3 . Smart Materials (Self-Cleaning) Smart photocatalytic materials have been developed widely over the past two decades [ 49 ]. Different kinds of applications such as simultaneous self-cleaning and air cleaning have mostly focused on the use of TiO 2 and ZnO due to their low cost, high stability and strong capacity for the photocatalytic decomposition of organic contaminants [ 205 ]. TiO 2 has been used recently to make self-decontaminant textiles that offer high antibacterial activity performance for UV shielding [ 206 ]. Nanophotocatalysts can be merged onto different surfaces of bulk structures (i.e. concrete) [ 223 ] or onto the glass of windows, flat surfaces, or walls [ 224 ]. TiO 2 -coated membranes offer outstanding antifouling/self-cleaning, photoactive, and bactericidal properties that are based on the UV mechanism of TiO 2 photocatalysis ( Figure 11 ) [ 207 ]. The overall applications of different types of nanophotocatalysts mentioned in this review are summarized in Table 3 . 4.1. Laboratory and Hospital 4.1.1. Basics and Fundamentals Hospital waste can be described as a combination of both biological and non-biological waste that is discarded and not intended to be used again. Hospital and laboratory waste can be broadly defined into two categories: hazardous (risk waste) and non-hazardous (non-risk) waste ( Figure 4 ). Risk waste includes sharps, pathological, pharmaceutical, chemical, and radioactive waste, whereas non-risk waste is equivalent to general domestic garbage and introduce no greater risk than normal home waste (e.g., paper, packaging, and food waste). A portion of hospital and laboratory waste management processes includes segregation, handling, transportation, disinfection, mutilation, storage, and final disposal as the foundation for essential progress towards scientifically safe waste management [ 50 ]. There are two main steps involved in the waste disposal process: treatment and final disposal. In the treatment step, depending on the nature of the waste, procedures such as incineration, autoclaving, chemical disinfection, encapsulation, and microwave irradiation are usually applied. The disposed wastes are likely to wind up in landfill, buried inside the premises, discharged into the sewer, etc. Segregation (separation) has a key role in an efficient waste identification and management process. Sorting the waste based on the color of the container is the most proper method for identifying waste categories [ 51 ]. The last treatment applied to the wastes can be performed by autoclave, hydroclave, incineration, or microwave technologies. Waste disposed of in landfill must be properly designated and managed, despite being the cheapest and most readily available way. On the other hand, shredding involves all types of bulk plastic waste, including risk waste, which are first disinfected, and then cut into small pieces and converted into a compact form. This approach is cost effective, and is much cheaper than an incinerator, while causing comparatively no pollution. Autoclaving, however, as the last type of waste disposal is more expensive, but more promising than shredding. People who are exposed to hazardous hospital waste are potentially at higher risk—either those who handle the waste at any stage or are exposed as a consequence of careless management [ 52 ]. The most to least favorable options of waste management are demonstrated in Figure 5 . 4.1.2. Photocatalytic Point of View Photocatalysis is an efficient, affordable strategy for the decontamination and disinfection of hospital waste that applies ultraviolet rays or solar energy to disintegrate antibiotics and disinfect microbes from the waste at the point of origin (see Figure 6 ). In demanding cases such as medical care systems and especially in microbiological labs, frequent and thorough cleaning of surfaces is required to inhibit bacterial transmission and diminish the emergence of microscopic organisms ( Table 1 ). Traditional sterilization strategies are time- and staff-consuming, and do not usually have long-term effectiveness. Additionally, the use of ultraviolet (UV) light instead of aggressive and dangerous chemicals can cause severe occupational health risks [ 53 ]. Titanium dioxide-coated surfaces can perform photocatalytic oxidation as an alternative to the conventional techniques of surface cleansing. The disinfection attributes of anchored titanium thin films on solid surfaces have been investigated in some previous studies [ 53 , 54 ]. The feasibility of this procedure for hygiene has been exhibited using bacteria, for example, Staphylococcus aureus ( S. aureus ), Pseudomonas aeruginosa ( P. aeruginosa ), Enterococcus faecium ( E. faecium ), and Escherichia coli ( E. coli ) [ 53 ]. Another example of self-disinfecting surface application is the inactivation of the deposited E. coli (ATCC8739) cells on membrane filters during fluorescent light irradiation [ 56 ]. In another study, a novel flame-assisted chemical vapor deposition (CVD) technique was used to test the antimicrobial activity of the stainless steel coated with TiO 2 thin films on E. coli [ 57 ]. Due to the extensive applications of the self-sterilizing material stainless steel, because of corrosion resistance, has been used for sterilization of Bacillus pumilus (B. pumilus) and has shown higher photocatalytic activity compared to the glass substrates coated with self-sterilizing materials [ 58 ]. Surfaces coated with CuO-doped Titania photocatalysts were also assessed for their biocidal activity and synergistic impact of photocatalysis and lethality of copper to deactivate bacteriophage T4 and E. coli [ 59 ]. Nitrogen-doped TiO 2 photocatalysts have been used owing to their visible light-induced bactericidal activity against human pathogens [ 60 , 70 ]. Visible light photocatalytic disinfection offers a continuous cleansing of surfaces that are constantly in contact with people, such as push buttons or door handles. This quality has been applied in the inactivation of E. coli using nitrogen/sulfur-co-doped Titania [ 61 , 62 , 63 , 71 ], which presents new disinfectant applications in public places that are consistently exposed to the transmission of pathogens (e.g., hospitals, schools, stations, hotels, airports, public toilets, and public transportation) [ 72 , 73 ]. In another study, prions, an infectious agent of a group of transmissible, fatal neurodegenerative disorders influencing both animals and humans, were inactivated using photocatalysts [ 48 ]. Photocatalytic oxidation, which is the cause of prion inactivation, can decrease the risk of spread, and has shown a significant effect on the pragmatic employment of this approach towards disinfection and cleansing of contaminated objects and surfaces, since these prions can be transmitted by ingestion of contaminated food or during medical treatments using contaminated surgical tools or biological materials. Controlling Legionnaire's disease, which is related to the contamination of hot water distribution systems with Legionella bacteria, is another application of photocatalysts in a laboratory or hospital [ 64 ]. It has also been shown that TiO 2 /UV photocatalytic oxidation has the capacity to mineralize the four strains of L. pneumophilia serogroup 1 cells (strains 977, 1004, 1009, and ATCC 33153) in laboratory-scale surveys, which proves that this could be used as a reasonable procedure for the routine disinfection and sterilization in order to control Legionella bacteria species in the hot water systems of emergency clinics and hospitals (e.g., hyperchlorination and thermal eradication) [ 48 ]. The application of TiO 2 as an effective treatment has also been reported to remove pharmaceutical contaminants from water with high efficiency. Due to their high photolysis sensitivity in aquatic systems, they have been used, for instance, to treat diclofenac [ 74 , 75 ], and triclosan [ 76 ]. 4.1.1. Basics and Fundamentals Hospital waste can be described as a combination of both biological and non-biological waste that is discarded and not intended to be used again. Hospital and laboratory waste can be broadly defined into two categories: hazardous (risk waste) and non-hazardous (non-risk) waste ( Figure 4 ). Risk waste includes sharps, pathological, pharmaceutical, chemical, and radioactive waste, whereas non-risk waste is equivalent to general domestic garbage and introduce no greater risk than normal home waste (e.g., paper, packaging, and food waste). A portion of hospital and laboratory waste management processes includes segregation, handling, transportation, disinfection, mutilation, storage, and final disposal as the foundation for essential progress towards scientifically safe waste management [ 50 ]. There are two main steps involved in the waste disposal process: treatment and final disposal. In the treatment step, depending on the nature of the waste, procedures such as incineration, autoclaving, chemical disinfection, encapsulation, and microwave irradiation are usually applied. The disposed wastes are likely to wind up in landfill, buried inside the premises, discharged into the sewer, etc. Segregation (separation) has a key role in an efficient waste identification and management process. Sorting the waste based on the color of the container is the most proper method for identifying waste categories [ 51 ]. The last treatment applied to the wastes can be performed by autoclave, hydroclave, incineration, or microwave technologies. Waste disposed of in landfill must be properly designated and managed, despite being the cheapest and most readily available way. On the other hand, shredding involves all types of bulk plastic waste, including risk waste, which are first disinfected, and then cut into small pieces and converted into a compact form. This approach is cost effective, and is much cheaper than an incinerator, while causing comparatively no pollution. Autoclaving, however, as the last type of waste disposal is more expensive, but more promising than shredding. People who are exposed to hazardous hospital waste are potentially at higher risk—either those who handle the waste at any stage or are exposed as a consequence of careless management [ 52 ]. The most to least favorable options of waste management are demonstrated in Figure 5 . 4.1.2. Photocatalytic Point of View Photocatalysis is an efficient, affordable strategy for the decontamination and disinfection of hospital waste that applies ultraviolet rays or solar energy to disintegrate antibiotics and disinfect microbes from the waste at the point of origin (see Figure 6 ). In demanding cases such as medical care systems and especially in microbiological labs, frequent and thorough cleaning of surfaces is required to inhibit bacterial transmission and diminish the emergence of microscopic organisms ( Table 1 ). Traditional sterilization strategies are time- and staff-consuming, and do not usually have long-term effectiveness. Additionally, the use of ultraviolet (UV) light instead of aggressive and dangerous chemicals can cause severe occupational health risks [ 53 ]. Titanium dioxide-coated surfaces can perform photocatalytic oxidation as an alternative to the conventional techniques of surface cleansing. The disinfection attributes of anchored titanium thin films on solid surfaces have been investigated in some previous studies [ 53 , 54 ]. The feasibility of this procedure for hygiene has been exhibited using bacteria, for example, Staphylococcus aureus ( S. aureus ), Pseudomonas aeruginosa ( P. aeruginosa ), Enterococcus faecium ( E. faecium ), and Escherichia coli ( E. coli ) [ 53 ]. Another example of self-disinfecting surface application is the inactivation of the deposited E. coli (ATCC8739) cells on membrane filters during fluorescent light irradiation [ 56 ]. In another study, a novel flame-assisted chemical vapor deposition (CVD) technique was used to test the antimicrobial activity of the stainless steel coated with TiO 2 thin films on E. coli [ 57 ]. Due to the extensive applications of the self-sterilizing material stainless steel, because of corrosion resistance, has been used for sterilization of Bacillus pumilus (B. pumilus) and has shown higher photocatalytic activity compared to the glass substrates coated with self-sterilizing materials [ 58 ]. Surfaces coated with CuO-doped Titania photocatalysts were also assessed for their biocidal activity and synergistic impact of photocatalysis and lethality of copper to deactivate bacteriophage T4 and E. coli [ 59 ]. Nitrogen-doped TiO 2 photocatalysts have been used owing to their visible light-induced bactericidal activity against human pathogens [ 60 , 70 ]. Visible light photocatalytic disinfection offers a continuous cleansing of surfaces that are constantly in contact with people, such as push buttons or door handles. This quality has been applied in the inactivation of E. coli using nitrogen/sulfur-co-doped Titania [ 61 , 62 , 63 , 71 ], which presents new disinfectant applications in public places that are consistently exposed to the transmission of pathogens (e.g., hospitals, schools, stations, hotels, airports, public toilets, and public transportation) [ 72 , 73 ]. In another study, prions, an infectious agent of a group of transmissible, fatal neurodegenerative disorders influencing both animals and humans, were inactivated using photocatalysts [ 48 ]. Photocatalytic oxidation, which is the cause of prion inactivation, can decrease the risk of spread, and has shown a significant effect on the pragmatic employment of this approach towards disinfection and cleansing of contaminated objects and surfaces, since these prions can be transmitted by ingestion of contaminated food or during medical treatments using contaminated surgical tools or biological materials. Controlling Legionnaire's disease, which is related to the contamination of hot water distribution systems with Legionella bacteria, is another application of photocatalysts in a laboratory or hospital [ 64 ]. It has also been shown that TiO 2 /UV photocatalytic oxidation has the capacity to mineralize the four strains of L. pneumophilia serogroup 1 cells (strains 977, 1004, 1009, and ATCC 33153) in laboratory-scale surveys, which proves that this could be used as a reasonable procedure for the routine disinfection and sterilization in order to control Legionella bacteria species in the hot water systems of emergency clinics and hospitals (e.g., hyperchlorination and thermal eradication) [ 48 ]. The application of TiO 2 as an effective treatment has also been reported to remove pharmaceutical contaminants from water with high efficiency. Due to their high photolysis sensitivity in aquatic systems, they have been used, for instance, to treat diclofenac [ 74 , 75 ], and triclosan [ 76 ]. 4.2. Biological and Medical 4.2.1. Categories and Conventional Methods Biomedical waste refers to any type of waste generated during treatment, diagnosis, or research activities—testing biological products and the immunization of human beings or animals. Biomedical waste management has become a general medical issue, as it is not only a legal necessity, but also a social responsibility. An improvement of biomedical waste management starts with the reduction of the waste produced [ 77 ]. According to an earlier report, around half of the world's population is in danger of incompatible biomedical waste management, which can impact both work and public places [ 78 ]. About 75% to 90% of total biomedical waste is general non-hazardous waste. The remaining 10% to 25% is dangerous and hazardous (including infectious waste, sharp waste, pathological waste, cytotoxic waste, pharmaceutical waste, liquid infectious waste, chemical waste, radioactive waste, and general health-care waste), which leads to a wide range of issues such as environmental and health risks. The main risks associated with biomedical waste include the microbial, systemic, and local infections caused by exposure to biomedical waste, among which pesticides, disinfectants, and mercury have multiple impacts in different ways, while inappropriate handling of sharps can lead to needle stick injuries and cause infections with blood-borne pathogens such as human immunodeficiency viruses (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) [ 79 ]. The essential needs in biomedical waste management include reducing, waste storage and transport, recycling and reusing, expenses in the annual budget, storage management, separate chemical and pharmaceutical waste segregation, separate storage zones, and documentation related to biomedical waste management [ 80 ]. Waste treatment technologies can be summarized as thermal (autoclaves, steam, microwave, dry heat treatment technologies) [ 80 , 81 ], chemical processes (sodium hypochlorite (NaOCl, 1%–12%), chlorine dioxide, calcium hydroxide, glutaraldehyde and peracetic acid) [ 82 ], incineration [ 83 ], encapsulation and inertization, irradiation technologies, biological processes, membrane bioreactors [ 84 ], disinfection and sterilization, as well as emerging technologies such as alkaline hydrolysis, plasma pyrolysis, superheated steam, ozone and promession [ 85 ]. Other upcoming technologies for the destruction of biomedical waste include base-catalyzed decomposition, gas-phase chemical reduction, sodium reduction, supercritical water oxidation, verification, superheated steam reforming, Fe-tetra-amido macrocyclic ligand (TAML)/peroxide treatment (pharmaceutical waste), biodegradation (using mealworm or bacteria to eat plastics), mechanochemical treatment, sonic technology, electrochemical technologies, phytotechnology, and solvated electron technology [ 86 ]. 4.2.2. Photocatalytic Strategies Photocatalytic processes are used in biomedical applications due to their disinfection abilities. It has been investigated that Staphylococcus aureus ( S. aureus ), a typical pathogenic bacterium in implant-related infection, shows photocatalytic activity using TiO 2 film on titanium substrates and stainless steel [ 87 ]. TiO 2 coatings have been used on bioimplants to apply photocatalysis for antibacterial purposes. The coatings exhibited a bactericidal impact upon UV irradiation, so the implementation of these photocatalytic-coated substrates are a valuable system for controlling infections related to biomedical implants [ 57 , 65 , 66 ]. As recently suggested for air purification systems, photocatalysis can be used to extract dangerous airborne biological risks, such as Anthrax, for example. Therefore, it is a versatile procedure for circumventing the spread of airborne biological threats and preventing bioterror dangers [ 67 ]. Reducing the number of bacteria and preventing their dissemination is essential and can be achieved through disinfection of surfaces in microbiological laboratories, food processing plants, veterinary medicine clinics, and hospitals. Traditional disinfection methods are generally time-consuming and tedious, and not usually sufficient (e.g., cleansing with chemical disinfectants) [ 88 ]. A significant alternative to conventional disinfection methods is the photocatalytic process by coating the surfaces with a thin layer of metal oxide nanostructures [ 89 ]. The high bactericidal property has added value for the practical application of photocatalysis in the treatment of microorganisms (e.g., Pseudomonas aeruginosa , E. coli , Enterococcus faecium , and S. aureus ) [ 68 , 90 ], and plays a crucial role in public health protection [ 69 , 91 , 92 ]. Metal Oxides The development of nanostructured metal oxides, semiconducting oxides, conducting oxides, composites, and polymers have been broadly investigated for quantification and detection of different hazardous biochemicals and chemicals [ 93 ]. It has been shown that metal oxides can photo-oxidize a wide range of organic substances such as alkenes, alkanes, surfactants, aromatics and pesticides [ 94 ]. Several metal oxides, including TiO 2 , ZnO, α-Fe 2 O 3 , MoO 3 , WO 3 , ZrO 2 and SnO 2 can be applied as photocatalysts, which are of great interest in current studies [ 95 ]. Titanium Dioxide (TiO 2 ) TiO 2 is desirable for photocatalysis due to its stability, inertness, and low cost. It is also recyclable and self-regenerating. One of the most important industrial applications of TiO 2 -based photocatalysts is the degradation of expired drugs and pharmaceutical compounds [ 96 ], dyes in textile industries [ 97 ], toxic compounds spills (e.g., pesticides) [ 98 ], natural toxins (e.g., cyanobacterial toxin microcystin-LR) [ 99 ], and a series of parabens as personal care products [ 100 ]. Another application of these nanophotocatalysts is in the treatment of winery wastewater by a photocatalytic reactor [ 101 ]. It has been evidenced in numerous studies that properties such as surface adsorption and photocatalytic reactions of nanocrystalline semiconductor particles are different from those of bulk materials owing to the increased reactive surface area. Application of Nano-Sized TiO 2 has been approved as a component of photocatalytic film covering scalpels [ 102 ], surgical masks [ 103 ], and catheters [ 104 ]. The UV-based disinfection effectiveness of nano-TiO 2 -coated catheters is three times greater than uncoated ones. Another TiO 2 /UV application with similar results has been demonstrated in infected dental implants [ 105 , 106 ]. The TiO 2 /UV process also showed high bactericidal efficiency in orthopedics, and cosmetic surgery in the presence of S. aureus on nano-TiO 2 coated implants, which is a valuable method for decreasing bacterial infection caused by the application of implants in medicine and biomedical fields [ 107 ]. Conventional TiO 2 photocatalysts, however, cannot provide purified and sufficiently safe drinkable water, since water remains toxic following treatment with TiO 2 nanoparticles. Therefore, the novel three-dimensional (3D) structured TiO 2 nanophotocatalyst can be replaced with TiO 2 nanoparticles by which both the safety level and efficiency of purification of the final purified water are preserved. These structures are suitable for environmental and biomedical applications, as they meet the human key safety conditions, according to an in vitro cytotoxicity test of well-purified water by eco-TiO 2 [ 13 , 108 ]. Various types of TiO 2 photocatalytic degradation of organic and inorganic pollutants are summarized in Table 2 . 2. Zinc Oxide (ZnO) ZnO shows efficient activity in photocatalytic degradation of organic contaminants as compared to TiO 2 [ 109 ]. Among different forms of zinc oxides (i.e. ZnO and ZnO 2 ), ZnO can form stable, protective coatings, which act as smart materials. Three different crystalline phases of ZnO include zinc-blende, wurtzite, and rock-salt. Due to these multi-functional qualities, ZnO is extensively used for various applications including photocatalysis [ 110 ], light-emitting diodes [ 111 ], biosensors, solar-cells [ 112 ], field-emission and gas sensing [ 113 ]. Different morphologies of ZnO have been reported, such as nanorods [ 114 ], nanonails, nanopencils [ 115 , 116 ], nanowires [ 117 ], nanotubes, nanobullets [ 118 , 119 ], nanocomb-like structures, nanobelts [ 120 ], nanoribbons [ 121 , 122 , 123 ], nanohelices [ 123 ], nanoneedles [ 124 ], and nanopins [ 125 ]. It has been proved that the properties of these materials are strongly dependent on the size and shape of the ZnO nanoparticles. The effective photocatalytic degradation of acridine orange up to about 90% after 80 min of exposure to UV light by ZnO nanocapsule is reported [ 126 , 127 ]. Photocatalytic degradation of methyl orange using ZnO as the photocatalyst has also been achieved—about 99.7% removal of the azo dye in 180 min [ 128 ]. In another study, almost complete degradation of methylene blue was obtained within 85 min of irradiation time using ZnO nanoparticles synthesized by hydrothermal treatment [ 129 ]. Additionally, ZnO-CeO 2 nanoparticles, which are discussed in the binary metal oxide category, have been applied to remove methylene blue and acridine orange [ 130 ]. 3. Iron Oxide (Fe 2 O 3 ) Iron oxides play an important role in many biological and geological processes. They are increasingly applied as pigments, iron ores, catalysts, and as hemoglobin in the blood. Freely dispersed bulk-, sonic-, and nano-Fe 2 O 3 have been used for photocatalytic oxidation of water under visible and UV irradiation [ 131 ]. Iron oxide (α-Fe 2 O 3 ) exhibits desirable efficiency as an important photocatalyst [ 132 ], with low cost, simple preparation, and n-type semiconducting behavior [ 133 ] with no secondary pollution [ 132 ]. Due to its various applications as sensors, pigments, actuators, and catalysts, it has attracted considerable attention in recent studies [ 134 , 135 , 136 ]. Photocatalytic degradation of organic pollutants via Fe 2 O 3 has been investigated. However, the corresponding photocatalytic mechanism has not been described in detail. The valence electrons of Fe 2 O 3 compared to those of TiO 2 can be excited to the conduction band at wavelengths shorter than 560 nm, which can extensively enhance the efficiency of the sunlight use. The maximum degradation efficiency of 94% for dibutyl phthalate in wastewater (as an excellent plasticizer in different resins, especially nitrocellulose and resins and also an vital additive in special paints and adhesives with about 20 years of hydrolysis half-life) was obtained using Fe 2 O 3 in a photocatalytic process [ 137 ]. Compared to α-Fe 2 O 3 powders, porous α-Fe 2 O 3 films exhibit better photocatalytic activity by water splitting under UV radiation for hydrogen generation [ 138 ]. Photocatalytic oxidation of aniline to azobenzene by Fe 2 O 3 under UV irradiation and natural sunlight in aprotic and protic solvents has also been reported [ 139 ]. 4. Gadolinium Oxide (Gd 2 O 3 ) The global interest in using rare earth metals is increasing, due to their distinctive magnetic and electronic attributes in the fashioning of interfaces and surfaces compared to common bulk materials. Gadolinia (gadolinium (III) oxide) is the most widely available derivative form of gadolinium, and is a potential contrast agent in magnetic resonance imaging (MRI). The Gd 2 O 3 -modified bismuth vanadate (BiVO 4 ) composite, as a photocatalyst, exhibits significantly greater visible-light photocatalytic activity than pure BiVO 4 for methyl orange degradation under visible light irradiation [ 140 ]. Gd 2 O 3 nanorods used to detect ethanol by facile hydrothermal routes demonstrated a lower detection limit with higher sensitivity and shorter response time [ 141 ] compared to the annealed Gd 2 O 3 nanostructures [ 142 ]. Moreover, a moderate photocatalytic activity was evaluated for degradation of methyl orange by uniform Gd 2 O 3 hollow microspheres [ 143 ]. The degradation of about 90% of 4-chlorophenol using modified Gd 2 O 3 photocatalyst prepared by the sol–gel method was measured after 4 h of UV light irradiation [ 144 ]. In another study, a challenging photocatalyst of Gd 2 O 3 nanorods was designed for the degradation of neurotoxic chloramphenicol drugs [ 145 ]. 5. Antimony Oxide (Sb 2 O 4 ) Antimony oxide is classified based on its oxidation states, Sb (III) and Sb (V). Antimony has been applied as a pacifier in enamel, flame retardants, paint and glass art crafts and for making bullets and bullet tracers. It has been used as an alloy for the synthesis of plain bearings, batteries, and solders, as well as as a stabilizer and a catalyst for the preparation of polyethylene terephthalate. The photocatalytic activity of α-Sb 2 O 4 has been demonstrated, with almost 52% degradation of acridine orange in 170 min, with low detection limit, good sensitivity, long linear dynamic range with good linearity in a very short response time [ 146 ], as well as for the removal of heavy metals (e.g., mercury) from waste water [ 147 ], while it has also been reported that synthetic Uranyl Selective Polymeric Membrane sensors based on p-tert-butylbiscalix4arene can be used for the determination of Thorium [ 148 ]. The unique characteristics of nanostructures, such as their large surface area, excellent adsorbing and absorbing activity, bio-friendly nature, and high electron exchange could be reasons for the good sensitivity of these systems [ 149 ]. Binary Metal Oxides In addition to metal oxides, some other metal oxides have also been studied previously for use in the field of photocatalysis, because of their unique benefits and wide range of applications as catalysts, semiconductors, superconductors, ceramics, antifungal agents, adsorbents, and their specific applications in medicines. Many metal oxide semiconductors (e.g., WO 3 , ZrO 2 , ZnO, and Fe 2 O 3 ) that have been exploited in photocatalysts for the degradation of organic contaminants have inherent drawbacks [ 150 ]. For example, WO 3 is a stable photocatalyst for O 2 production within the visible light irradiation range. However, it is not suitable for H 2 evolution because of its low level of conduction band. Additionally, α-Fe 2 O 3 is somewhat stable in acidic solutions, but has the same problems as WO 3 . Moreover, ZnO can be easily corroded under band gap irradiation by photogenerated holes. Ta 2 O 5 photocatalyst with a nanocrystalline mesoporous structure has recently been synthesized for the production of H 2 via a sol–gel process combined with a surfactant-assisted templating mechanism [ 151 ]. Recently, the effect of Fe-doped NiO as a co-catalyst has also been reported [ 152 ]. Additionally, ZnO-CeO 2 nanoparticles synthesized using an efficient and simple low-temperature method have been successfully applied as photocatalysts for the removal of biomedical and environmental contaminants and reported 80.7% and 92.1% degradation for methylene blue and acridine orange within 170 min of irradiation time, respectively [ 130 ]. The Cu x S-TiO 2 composites has shown good efficiency in photo degradation of dyes even under visible light irradiation [ 153 ]. In another report, the photocatalytic activity of high-quality CeO 2 -CdO binary metal oxide nanocomposites was evaluated, showing acceptable growth inhibition of P. aeruginosa ( Figure 7 ) [ 154 ]. Metal Sulfides Metal sulfides have been widely used as visible light responsive photocatalysts. Compared to metal oxides, 3p orbitals of sulfur in their valence band result in a more occupied valence band and a narrower band gap. Recently, among other metal sulfides, ZnS and CdS have attracted great attention. CdS is commonly used for visible light-assisted water splitting due to its suitable band position and band gap (2.4 eV). However, photo-corrosion, which is a common issue in most metal sulfide photocatalysts, occurs when using both CdS and ZnS. Therefore, recent studies have focused on the development of ZnS and CdS photocatalysts, mostly through four different means of improvement: matrixing and supporting the structures of CdS, adding cocatalysts to CdS, preparing porous and one-dimensional CdS, and doping solid solutions of CdS and ZnS [ 155 ]. To synthesize porous CdS, a solvothermal method has been used to synthesize CdS nanowires [ 156 ] and nanorods [ 157 ]. Additionally, mesoporous CdS nanoparticles have been synthesized via ultrasonic and template-free precipitation at room temperature [ 158 ]. Nanoporous CdS have also been prepared by including hollow nanorods and nanosheets with 3 nm diameter of pores through a two-step aqueous method [ 159 ]. Additionally, CdS quantum dots have been recently loaded on porous polysaccharides and applied as highly efficient contrast imaging agents [ 160 ]. On the other hand, due to having an extremely broad band gap to respond to visible light (3.6 eV), solid solutions of ZnS are formed in which the narrow band gap can increase the use of ZnS in visible light. Both CdS and ZnS have the same crystal structures, making it easy to form solid solutions of them [ 161 ]. Magnetic Nanophotocatalysts The incorporation of magnetic nanophotocatalysts in contaminant removal strategies has recently received significant attention due to their improved chemical and physical properties. Therefore, cost-effective, efficient, and environmentally friendly disinfection processes can be achieved due to their easy separation using an external magnetic field, which allows recycling and multiple use of the nanophotocatalyst [ 162 ]. They mostly have a core–shell structure consisting of a magnetic core (e.g., iron, cobalt, nickel, and their oxides like maghemite (α-Fe 2 O 3 ), magnetite (Fe 3 O 4 ), cobalt ferrite (CoFe 2 O 4 )) and a photocatalytic shell (e.g., TiO 2 , ZnO, AgBr, BiOCl) [ 163 ]. Furthermore, some nanoferrites like ZnFe 2 O 4 have shown desirable degradation efficiency of organic target compounds under both visible light and UV irradiation [ 164 ]. Similar studies have reported degradation of different contaminants using Fe 3 O 4 [ 165 , 166 , 167 ], NiFe 2 O 4 [ 168 ], CoFe 2 O 4 [ 169 ], ZnFe 2 O 4 [ 170 ], BaFe 12 O 19 [ 171 ], SrFe 12 O 19 [ 172 ] -doped TiO 2 nanophotocatalysts ( Table 3 ). A schematic of the use of magnetic nanophotocatalysts (MNPCs) in water treatment [ 173 ] is illustrated in Figure 8 . Graphene Graphene (G), due to its one-of-a-kind nanostructure and particular properties has been studied widely from both experimental and theoretical scientific points of view [ 208 ]. It has already shown promising applications in nanocomposites, nanoelectronics, optoelectronics, drug delivery systems, electrochemical super-capacitors, transistors, solar cells, and chemical sensors (e.g., biosensors, gas sensors, pH sensors) [ 209 ]. As shown in Figure 9 , graphene has been employed to enhance photocatalytic efficiency, due to its electron scavenging nature, in the conduction band of metal oxide [ 26 ]. Some of its novel applications include ultrasensitive high-adsorption ability for various types of contaminations, including arsenic in drinking water [ 210 ], brackish water desalination and drinking water purification [ 211 ], metal removal from the contaminated environment [ 212 ], detection of biomarkers [ 213 ], electrochemical sensor for paracetamol [ 214 ], treatment of thrombosis [ 215 ], protection of DNA from cleavage and its effective cellular delivery [ 216 ], MRI and localized photothermal therapy for cancer cell treatment [ 217 ], electrochemical immunosensor for sensitive detection of carbohydrate antigen 1.5-3 (CA 15-3) [ 218 ], and photothermal agents in NIR region [ 219 ]. Modified graphene nanostructures such as P25–G [ 174 , 175 ], TiO 2 –G [ 176 , 177 , 178 , 179 ], SnO 2 –G [ 180 ], Bi 2 WO 6 –G [ 181 ], ZnO–G [ 182 ], ZnFe 2 O 4 –G [ 183 ], BiVO 4 –G [ 184 ], CdS–G [ 185 ] have also been reported to have different photodegradation applications ( Table 3 ). Quantum Dots Quantum dots (QDs), as zero-dimensional semiconductor multifunctional nanomaterials have been receiving significant attention for the degradation of pollutants [ 220 ]. Since QDs have the advantage of the wide band gap of a semiconductor material, they have a promising application as photocatalysts, owing to the swift generation of electron–hole pairs through photoexcitation [ 204 ]. On the other hand, photocatalytic, chemical and optical properties of QDs can be improved by surface modification, which also improves the photostability of QDs, as well as the efficacy of light-induced reactions on the QD surface and the generation of new traps on the QD surface [ 221 ]. For example, in the self-photosensitization pathway of fuchsin dye degradation, photodegradation can be initiated in the presence of graphene quantum dots (GQDs) under visible light irradiation, as demonstrated in Figure 10 [ 222 ]. The application of modified QDs as a photocatalytic agent to degrade pollutants is illustrated in Table 3 . Smart Materials (Self-Cleaning) Smart photocatalytic materials have been developed widely over the past two decades [ 49 ]. Different kinds of applications such as simultaneous self-cleaning and air cleaning have mostly focused on the use of TiO 2 and ZnO due to their low cost, high stability and strong capacity for the photocatalytic decomposition of organic contaminants [ 205 ]. TiO 2 has been used recently to make self-decontaminant textiles that offer high antibacterial activity performance for UV shielding [ 206 ]. Nanophotocatalysts can be merged onto different surfaces of bulk structures (i.e. concrete) [ 223 ] or onto the glass of windows, flat surfaces, or walls [ 224 ]. TiO 2 -coated membranes offer outstanding antifouling/self-cleaning, photoactive, and bactericidal properties that are based on the UV mechanism of TiO 2 photocatalysis ( Figure 11 ) [ 207 ]. The overall applications of different types of nanophotocatalysts mentioned in this review are summarized in Table 3 . 4.2.1. Categories and Conventional Methods Biomedical waste refers to any type of waste generated during treatment, diagnosis, or research activities—testing biological products and the immunization of human beings or animals. Biomedical waste management has become a general medical issue, as it is not only a legal necessity, but also a social responsibility. An improvement of biomedical waste management starts with the reduction of the waste produced [ 77 ]. According to an earlier report, around half of the world's population is in danger of incompatible biomedical waste management, which can impact both work and public places [ 78 ]. About 75% to 90% of total biomedical waste is general non-hazardous waste. The remaining 10% to 25% is dangerous and hazardous (including infectious waste, sharp waste, pathological waste, cytotoxic waste, pharmaceutical waste, liquid infectious waste, chemical waste, radioactive waste, and general health-care waste), which leads to a wide range of issues such as environmental and health risks. The main risks associated with biomedical waste include the microbial, systemic, and local infections caused by exposure to biomedical waste, among which pesticides, disinfectants, and mercury have multiple impacts in different ways, while inappropriate handling of sharps can lead to needle stick injuries and cause infections with blood-borne pathogens such as human immunodeficiency viruses (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) [ 79 ]. The essential needs in biomedical waste management include reducing, waste storage and transport, recycling and reusing, expenses in the annual budget, storage management, separate chemical and pharmaceutical waste segregation, separate storage zones, and documentation related to biomedical waste management [ 80 ]. Waste treatment technologies can be summarized as thermal (autoclaves, steam, microwave, dry heat treatment technologies) [ 80 , 81 ], chemical processes (sodium hypochlorite (NaOCl, 1%–12%), chlorine dioxide, calcium hydroxide, glutaraldehyde and peracetic acid) [ 82 ], incineration [ 83 ], encapsulation and inertization, irradiation technologies, biological processes, membrane bioreactors [ 84 ], disinfection and sterilization, as well as emerging technologies such as alkaline hydrolysis, plasma pyrolysis, superheated steam, ozone and promession [ 85 ]. Other upcoming technologies for the destruction of biomedical waste include base-catalyzed decomposition, gas-phase chemical reduction, sodium reduction, supercritical water oxidation, verification, superheated steam reforming, Fe-tetra-amido macrocyclic ligand (TAML)/peroxide treatment (pharmaceutical waste), biodegradation (using mealworm or bacteria to eat plastics), mechanochemical treatment, sonic technology, electrochemical technologies, phytotechnology, and solvated electron technology [ 86 ]. 4.2.2. Photocatalytic Strategies Photocatalytic processes are used in biomedical applications due to their disinfection abilities. It has been investigated that Staphylococcus aureus ( S. aureus ), a typical pathogenic bacterium in implant-related infection, shows photocatalytic activity using TiO 2 film on titanium substrates and stainless steel [ 87 ]. TiO 2 coatings have been used on bioimplants to apply photocatalysis for antibacterial purposes. The coatings exhibited a bactericidal impact upon UV irradiation, so the implementation of these photocatalytic-coated substrates are a valuable system for controlling infections related to biomedical implants [ 57 , 65 , 66 ]. As recently suggested for air purification systems, photocatalysis can be used to extract dangerous airborne biological risks, such as Anthrax, for example. Therefore, it is a versatile procedure for circumventing the spread of airborne biological threats and preventing bioterror dangers [ 67 ]. Reducing the number of bacteria and preventing their dissemination is essential and can be achieved through disinfection of surfaces in microbiological laboratories, food processing plants, veterinary medicine clinics, and hospitals. Traditional disinfection methods are generally time-consuming and tedious, and not usually sufficient (e.g., cleansing with chemical disinfectants) [ 88 ]. A significant alternative to conventional disinfection methods is the photocatalytic process by coating the surfaces with a thin layer of metal oxide nanostructures [ 89 ]. The high bactericidal property has added value for the practical application of photocatalysis in the treatment of microorganisms (e.g., Pseudomonas aeruginosa , E. coli , Enterococcus faecium , and S. aureus ) [ 68 , 90 ], and plays a crucial role in public health protection [ 69 , 91 , 92 ]. Metal Oxides The development of nanostructured metal oxides, semiconducting oxides, conducting oxides, composites, and polymers have been broadly investigated for quantification and detection of different hazardous biochemicals and chemicals [ 93 ]. It has been shown that metal oxides can photo-oxidize a wide range of organic substances such as alkenes, alkanes, surfactants, aromatics and pesticides [ 94 ]. Several metal oxides, including TiO 2 , ZnO, α-Fe 2 O 3 , MoO 3 , WO 3 , ZrO 2 and SnO 2 can be applied as photocatalysts, which are of great interest in current studies [ 95 ]. Titanium Dioxide (TiO 2 ) TiO 2 is desirable for photocatalysis due to its stability, inertness, and low cost. It is also recyclable and self-regenerating. One of the most important industrial applications of TiO 2 -based photocatalysts is the degradation of expired drugs and pharmaceutical compounds [ 96 ], dyes in textile industries [ 97 ], toxic compounds spills (e.g., pesticides) [ 98 ], natural toxins (e.g., cyanobacterial toxin microcystin-LR) [ 99 ], and a series of parabens as personal care products [ 100 ]. Another application of these nanophotocatalysts is in the treatment of winery wastewater by a photocatalytic reactor [ 101 ]. It has been evidenced in numerous studies that properties such as surface adsorption and photocatalytic reactions of nanocrystalline semiconductor particles are different from those of bulk materials owing to the increased reactive surface area. Application of Nano-Sized TiO 2 has been approved as a component of photocatalytic film covering scalpels [ 102 ], surgical masks [ 103 ], and catheters [ 104 ]. The UV-based disinfection effectiveness of nano-TiO 2 -coated catheters is three times greater than uncoated ones. Another TiO 2 /UV application with similar results has been demonstrated in infected dental implants [ 105 , 106 ]. The TiO 2 /UV process also showed high bactericidal efficiency in orthopedics, and cosmetic surgery in the presence of S. aureus on nano-TiO 2 coated implants, which is a valuable method for decreasing bacterial infection caused by the application of implants in medicine and biomedical fields [ 107 ]. Conventional TiO 2 photocatalysts, however, cannot provide purified and sufficiently safe drinkable water, since water remains toxic following treatment with TiO 2 nanoparticles. Therefore, the novel three-dimensional (3D) structured TiO 2 nanophotocatalyst can be replaced with TiO 2 nanoparticles by which both the safety level and efficiency of purification of the final purified water are preserved. These structures are suitable for environmental and biomedical applications, as they meet the human key safety conditions, according to an in vitro cytotoxicity test of well-purified water by eco-TiO 2 [ 13 , 108 ]. Various types of TiO 2 photocatalytic degradation of organic and inorganic pollutants are summarized in Table 2 . 2. Zinc Oxide (ZnO) ZnO shows efficient activity in photocatalytic degradation of organic contaminants as compared to TiO 2 [ 109 ]. Among different forms of zinc oxides (i.e. ZnO and ZnO 2 ), ZnO can form stable, protective coatings, which act as smart materials. Three different crystalline phases of ZnO include zinc-blende, wurtzite, and rock-salt. Due to these multi-functional qualities, ZnO is extensively used for various applications including photocatalysis [ 110 ], light-emitting diodes [ 111 ], biosensors, solar-cells [ 112 ], field-emission and gas sensing [ 113 ]. Different morphologies of ZnO have been reported, such as nanorods [ 114 ], nanonails, nanopencils [ 115 , 116 ], nanowires [ 117 ], nanotubes, nanobullets [ 118 , 119 ], nanocomb-like structures, nanobelts [ 120 ], nanoribbons [ 121 , 122 , 123 ], nanohelices [ 123 ], nanoneedles [ 124 ], and nanopins [ 125 ]. It has been proved that the properties of these materials are strongly dependent on the size and shape of the ZnO nanoparticles. The effective photocatalytic degradation of acridine orange up to about 90% after 80 min of exposure to UV light by ZnO nanocapsule is reported [ 126 , 127 ]. Photocatalytic degradation of methyl orange using ZnO as the photocatalyst has also been achieved—about 99.7% removal of the azo dye in 180 min [ 128 ]. In another study, almost complete degradation of methylene blue was obtained within 85 min of irradiation time using ZnO nanoparticles synthesized by hydrothermal treatment [ 129 ]. Additionally, ZnO-CeO 2 nanoparticles, which are discussed in the binary metal oxide category, have been applied to remove methylene blue and acridine orange [ 130 ]. 3. Iron Oxide (Fe 2 O 3 ) Iron oxides play an important role in many biological and geological processes. They are increasingly applied as pigments, iron ores, catalysts, and as hemoglobin in the blood. Freely dispersed bulk-, sonic-, and nano-Fe 2 O 3 have been used for photocatalytic oxidation of water under visible and UV irradiation [ 131 ]. Iron oxide (α-Fe 2 O 3 ) exhibits desirable efficiency as an important photocatalyst [ 132 ], with low cost, simple preparation, and n-type semiconducting behavior [ 133 ] with no secondary pollution [ 132 ]. Due to its various applications as sensors, pigments, actuators, and catalysts, it has attracted considerable attention in recent studies [ 134 , 135 , 136 ]. Photocatalytic degradation of organic pollutants via Fe 2 O 3 has been investigated. However, the corresponding photocatalytic mechanism has not been described in detail. The valence electrons of Fe 2 O 3 compared to those of TiO 2 can be excited to the conduction band at wavelengths shorter than 560 nm, which can extensively enhance the efficiency of the sunlight use. The maximum degradation efficiency of 94% for dibutyl phthalate in wastewater (as an excellent plasticizer in different resins, especially nitrocellulose and resins and also an vital additive in special paints and adhesives with about 20 years of hydrolysis half-life) was obtained using Fe 2 O 3 in a photocatalytic process [ 137 ]. Compared to α-Fe 2 O 3 powders, porous α-Fe 2 O 3 films exhibit better photocatalytic activity by water splitting under UV radiation for hydrogen generation [ 138 ]. Photocatalytic oxidation of aniline to azobenzene by Fe 2 O 3 under UV irradiation and natural sunlight in aprotic and protic solvents has also been reported [ 139 ]. 4. Gadolinium Oxide (Gd 2 O 3 ) The global interest in using rare earth metals is increasing, due to their distinctive magnetic and electronic attributes in the fashioning of interfaces and surfaces compared to common bulk materials. Gadolinia (gadolinium (III) oxide) is the most widely available derivative form of gadolinium, and is a potential contrast agent in magnetic resonance imaging (MRI). The Gd 2 O 3 -modified bismuth vanadate (BiVO 4 ) composite, as a photocatalyst, exhibits significantly greater visible-light photocatalytic activity than pure BiVO 4 for methyl orange degradation under visible light irradiation [ 140 ]. Gd 2 O 3 nanorods used to detect ethanol by facile hydrothermal routes demonstrated a lower detection limit with higher sensitivity and shorter response time [ 141 ] compared to the annealed Gd 2 O 3 nanostructures [ 142 ]. Moreover, a moderate photocatalytic activity was evaluated for degradation of methyl orange by uniform Gd 2 O 3 hollow microspheres [ 143 ]. The degradation of about 90% of 4-chlorophenol using modified Gd 2 O 3 photocatalyst prepared by the sol–gel method was measured after 4 h of UV light irradiation [ 144 ]. In another study, a challenging photocatalyst of Gd 2 O 3 nanorods was designed for the degradation of neurotoxic chloramphenicol drugs [ 145 ]. 5. Antimony Oxide (Sb 2 O 4 ) Antimony oxide is classified based on its oxidation states, Sb (III) and Sb (V). Antimony has been applied as a pacifier in enamel, flame retardants, paint and glass art crafts and for making bullets and bullet tracers. It has been used as an alloy for the synthesis of plain bearings, batteries, and solders, as well as as a stabilizer and a catalyst for the preparation of polyethylene terephthalate. The photocatalytic activity of α-Sb 2 O 4 has been demonstrated, with almost 52% degradation of acridine orange in 170 min, with low detection limit, good sensitivity, long linear dynamic range with good linearity in a very short response time [ 146 ], as well as for the removal of heavy metals (e.g., mercury) from waste water [ 147 ], while it has also been reported that synthetic Uranyl Selective Polymeric Membrane sensors based on p-tert-butylbiscalix4arene can be used for the determination of Thorium [ 148 ]. The unique characteristics of nanostructures, such as their large surface area, excellent adsorbing and absorbing activity, bio-friendly nature, and high electron exchange could be reasons for the good sensitivity of these systems [ 149 ]. Binary Metal Oxides In addition to metal oxides, some other metal oxides have also been studied previously for use in the field of photocatalysis, because of their unique benefits and wide range of applications as catalysts, semiconductors, superconductors, ceramics, antifungal agents, adsorbents, and their specific applications in medicines. Many metal oxide semiconductors (e.g., WO 3 , ZrO 2 , ZnO, and Fe 2 O 3 ) that have been exploited in photocatalysts for the degradation of organic contaminants have inherent drawbacks [ 150 ]. For example, WO 3 is a stable photocatalyst for O 2 production within the visible light irradiation range. However, it is not suitable for H 2 evolution because of its low level of conduction band. Additionally, α-Fe 2 O 3 is somewhat stable in acidic solutions, but has the same problems as WO 3 . Moreover, ZnO can be easily corroded under band gap irradiation by photogenerated holes. Ta 2 O 5 photocatalyst with a nanocrystalline mesoporous structure has recently been synthesized for the production of H 2 via a sol–gel process combined with a surfactant-assisted templating mechanism [ 151 ]. Recently, the effect of Fe-doped NiO as a co-catalyst has also been reported [ 152 ]. Additionally, ZnO-CeO 2 nanoparticles synthesized using an efficient and simple low-temperature method have been successfully applied as photocatalysts for the removal of biomedical and environmental contaminants and reported 80.7% and 92.1% degradation for methylene blue and acridine orange within 170 min of irradiation time, respectively [ 130 ]. The Cu x S-TiO 2 composites has shown good efficiency in photo degradation of dyes even under visible light irradiation [ 153 ]. In another report, the photocatalytic activity of high-quality CeO 2 -CdO binary metal oxide nanocomposites was evaluated, showing acceptable growth inhibition of P. aeruginosa ( Figure 7 ) [ 154 ]. Metal Sulfides Metal sulfides have been widely used as visible light responsive photocatalysts. Compared to metal oxides, 3p orbitals of sulfur in their valence band result in a more occupied valence band and a narrower band gap. Recently, among other metal sulfides, ZnS and CdS have attracted great attention. CdS is commonly used for visible light-assisted water splitting due to its suitable band position and band gap (2.4 eV). However, photo-corrosion, which is a common issue in most metal sulfide photocatalysts, occurs when using both CdS and ZnS. Therefore, recent studies have focused on the development of ZnS and CdS photocatalysts, mostly through four different means of improvement: matrixing and supporting the structures of CdS, adding cocatalysts to CdS, preparing porous and one-dimensional CdS, and doping solid solutions of CdS and ZnS [ 155 ]. To synthesize porous CdS, a solvothermal method has been used to synthesize CdS nanowires [ 156 ] and nanorods [ 157 ]. Additionally, mesoporous CdS nanoparticles have been synthesized via ultrasonic and template-free precipitation at room temperature [ 158 ]. Nanoporous CdS have also been prepared by including hollow nanorods and nanosheets with 3 nm diameter of pores through a two-step aqueous method [ 159 ]. Additionally, CdS quantum dots have been recently loaded on porous polysaccharides and applied as highly efficient contrast imaging agents [ 160 ]. On the other hand, due to having an extremely broad band gap to respond to visible light (3.6 eV), solid solutions of ZnS are formed in which the narrow band gap can increase the use of ZnS in visible light. Both CdS and ZnS have the same crystal structures, making it easy to form solid solutions of them [ 161 ]. Magnetic Nanophotocatalysts The incorporation of magnetic nanophotocatalysts in contaminant removal strategies has recently received significant attention due to their improved chemical and physical properties. Therefore, cost-effective, efficient, and environmentally friendly disinfection processes can be achieved due to their easy separation using an external magnetic field, which allows recycling and multiple use of the nanophotocatalyst [ 162 ]. They mostly have a core–shell structure consisting of a magnetic core (e.g., iron, cobalt, nickel, and their oxides like maghemite (α-Fe 2 O 3 ), magnetite (Fe 3 O 4 ), cobalt ferrite (CoFe 2 O 4 )) and a photocatalytic shell (e.g., TiO 2 , ZnO, AgBr, BiOCl) [ 163 ]. Furthermore, some nanoferrites like ZnFe 2 O 4 have shown desirable degradation efficiency of organic target compounds under both visible light and UV irradiation [ 164 ]. Similar studies have reported degradation of different contaminants using Fe 3 O 4 [ 165 , 166 , 167 ], NiFe 2 O 4 [ 168 ], CoFe 2 O 4 [ 169 ], ZnFe 2 O 4 [ 170 ], BaFe 12 O 19 [ 171 ], SrFe 12 O 19 [ 172 ] -doped TiO 2 nanophotocatalysts ( Table 3 ). A schematic of the use of magnetic nanophotocatalysts (MNPCs) in water treatment [ 173 ] is illustrated in Figure 8 . Graphene Graphene (G), due to its one-of-a-kind nanostructure and particular properties has been studied widely from both experimental and theoretical scientific points of view [ 208 ]. It has already shown promising applications in nanocomposites, nanoelectronics, optoelectronics, drug delivery systems, electrochemical super-capacitors, transistors, solar cells, and chemical sensors (e.g., biosensors, gas sensors, pH sensors) [ 209 ]. As shown in Figure 9 , graphene has been employed to enhance photocatalytic efficiency, due to its electron scavenging nature, in the conduction band of metal oxide [ 26 ]. Some of its novel applications include ultrasensitive high-adsorption ability for various types of contaminations, including arsenic in drinking water [ 210 ], brackish water desalination and drinking water purification [ 211 ], metal removal from the contaminated environment [ 212 ], detection of biomarkers [ 213 ], electrochemical sensor for paracetamol [ 214 ], treatment of thrombosis [ 215 ], protection of DNA from cleavage and its effective cellular delivery [ 216 ], MRI and localized photothermal therapy for cancer cell treatment [ 217 ], electrochemical immunosensor for sensitive detection of carbohydrate antigen 1.5-3 (CA 15-3) [ 218 ], and photothermal agents in NIR region [ 219 ]. Modified graphene nanostructures such as P25–G [ 174 , 175 ], TiO 2 –G [ 176 , 177 , 178 , 179 ], SnO 2 –G [ 180 ], Bi 2 WO 6 –G [ 181 ], ZnO–G [ 182 ], ZnFe 2 O 4 –G [ 183 ], BiVO 4 –G [ 184 ], CdS–G [ 185 ] have also been reported to have different photodegradation applications ( Table 3 ). Quantum Dots Quantum dots (QDs), as zero-dimensional semiconductor multifunctional nanomaterials have been receiving significant attention for the degradation of pollutants [ 220 ]. Since QDs have the advantage of the wide band gap of a semiconductor material, they have a promising application as photocatalysts, owing to the swift generation of electron–hole pairs through photoexcitation [ 204 ]. On the other hand, photocatalytic, chemical and optical properties of QDs can be improved by surface modification, which also improves the photostability of QDs, as well as the efficacy of light-induced reactions on the QD surface and the generation of new traps on the QD surface [ 221 ]. For example, in the self-photosensitization pathway of fuchsin dye degradation, photodegradation can be initiated in the presence of graphene quantum dots (GQDs) under visible light irradiation, as demonstrated in Figure 10 [ 222 ]. The application of modified QDs as a photocatalytic agent to degrade pollutants is illustrated in Table 3 . Smart Materials (Self-Cleaning) Smart photocatalytic materials have been developed widely over the past two decades [ 49 ]. Different kinds of applications such as simultaneous self-cleaning and air cleaning have mostly focused on the use of TiO 2 and ZnO due to their low cost, high stability and strong capacity for the photocatalytic decomposition of organic contaminants [ 205 ]. TiO 2 has been used recently to make self-decontaminant textiles that offer high antibacterial activity performance for UV shielding [ 206 ]. Nanophotocatalysts can be merged onto different surfaces of bulk structures (i.e. concrete) [ 223 ] or onto the glass of windows, flat surfaces, or walls [ 224 ]. TiO 2 -coated membranes offer outstanding antifouling/self-cleaning, photoactive, and bactericidal properties that are based on the UV mechanism of TiO 2 photocatalysis ( Figure 11 ) [ 207 ]. The overall applications of different types of nanophotocatalysts mentioned in this review are summarized in Table 3 . Metal Oxides The development of nanostructured metal oxides, semiconducting oxides, conducting oxides, composites, and polymers have been broadly investigated for quantification and detection of different hazardous biochemicals and chemicals [ 93 ]. It has been shown that metal oxides can photo-oxidize a wide range of organic substances such as alkenes, alkanes, surfactants, aromatics and pesticides [ 94 ]. Several metal oxides, including TiO 2 , ZnO, α-Fe 2 O 3 , MoO 3 , WO 3 , ZrO 2 and SnO 2 can be applied as photocatalysts, which are of great interest in current studies [ 95 ]. Titanium Dioxide (TiO 2 ) TiO 2 is desirable for photocatalysis due to its stability, inertness, and low cost. It is also recyclable and self-regenerating. One of the most important industrial applications of TiO 2 -based photocatalysts is the degradation of expired drugs and pharmaceutical compounds [ 96 ], dyes in textile industries [ 97 ], toxic compounds spills (e.g., pesticides) [ 98 ], natural toxins (e.g., cyanobacterial toxin microcystin-LR) [ 99 ], and a series of parabens as personal care products [ 100 ]. Another application of these nanophotocatalysts is in the treatment of winery wastewater by a photocatalytic reactor [ 101 ]. It has been evidenced in numerous studies that properties such as surface adsorption and photocatalytic reactions of nanocrystalline semiconductor particles are different from those of bulk materials owing to the increased reactive surface area. Application of Nano-Sized TiO 2 has been approved as a component of photocatalytic film covering scalpels [ 102 ], surgical masks [ 103 ], and catheters [ 104 ]. The UV-based disinfection effectiveness of nano-TiO 2 -coated catheters is three times greater than uncoated ones. Another TiO 2 /UV application with similar results has been demonstrated in infected dental implants [ 105 , 106 ]. The TiO 2 /UV process also showed high bactericidal efficiency in orthopedics, and cosmetic surgery in the presence of S. aureus on nano-TiO 2 coated implants, which is a valuable method for decreasing bacterial infection caused by the application of implants in medicine and biomedical fields [ 107 ]. Conventional TiO 2 photocatalysts, however, cannot provide purified and sufficiently safe drinkable water, since water remains toxic following treatment with TiO 2 nanoparticles. Therefore, the novel three-dimensional (3D) structured TiO 2 nanophotocatalyst can be replaced with TiO 2 nanoparticles by which both the safety level and efficiency of purification of the final purified water are preserved. These structures are suitable for environmental and biomedical applications, as they meet the human key safety conditions, according to an in vitro cytotoxicity test of well-purified water by eco-TiO 2 [ 13 , 108 ]. Various types of TiO 2 photocatalytic degradation of organic and inorganic pollutants are summarized in Table 2 . 2. Zinc Oxide (ZnO) ZnO shows efficient activity in photocatalytic degradation of organic contaminants as compared to TiO 2 [ 109 ]. Among different forms of zinc oxides (i.e. ZnO and ZnO 2 ), ZnO can form stable, protective coatings, which act as smart materials. Three different crystalline phases of ZnO include zinc-blende, wurtzite, and rock-salt. Due to these multi-functional qualities, ZnO is extensively used for various applications including photocatalysis [ 110 ], light-emitting diodes [ 111 ], biosensors, solar-cells [ 112 ], field-emission and gas sensing [ 113 ]. Different morphologies of ZnO have been reported, such as nanorods [ 114 ], nanonails, nanopencils [ 115 , 116 ], nanowires [ 117 ], nanotubes, nanobullets [ 118 , 119 ], nanocomb-like structures, nanobelts [ 120 ], nanoribbons [ 121 , 122 , 123 ], nanohelices [ 123 ], nanoneedles [ 124 ], and nanopins [ 125 ]. It has been proved that the properties of these materials are strongly dependent on the size and shape of the ZnO nanoparticles. The effective photocatalytic degradation of acridine orange up to about 90% after 80 min of exposure to UV light by ZnO nanocapsule is reported [ 126 , 127 ]. Photocatalytic degradation of methyl orange using ZnO as the photocatalyst has also been achieved—about 99.7% removal of the azo dye in 180 min [ 128 ]. In another study, almost complete degradation of methylene blue was obtained within 85 min of irradiation time using ZnO nanoparticles synthesized by hydrothermal treatment [ 129 ]. Additionally, ZnO-CeO 2 nanoparticles, which are discussed in the binary metal oxide category, have been applied to remove methylene blue and acridine orange [ 130 ]. 3. Iron Oxide (Fe 2 O 3 ) Iron oxides play an important role in many biological and geological processes. They are increasingly applied as pigments, iron ores, catalysts, and as hemoglobin in the blood. Freely dispersed bulk-, sonic-, and nano-Fe 2 O 3 have been used for photocatalytic oxidation of water under visible and UV irradiation [ 131 ]. Iron oxide (α-Fe 2 O 3 ) exhibits desirable efficiency as an important photocatalyst [ 132 ], with low cost, simple preparation, and n-type semiconducting behavior [ 133 ] with no secondary pollution [ 132 ]. Due to its various applications as sensors, pigments, actuators, and catalysts, it has attracted considerable attention in recent studies [ 134 , 135 , 136 ]. Photocatalytic degradation of organic pollutants via Fe 2 O 3 has been investigated. However, the corresponding photocatalytic mechanism has not been described in detail. The valence electrons of Fe 2 O 3 compared to those of TiO 2 can be excited to the conduction band at wavelengths shorter than 560 nm, which can extensively enhance the efficiency of the sunlight use. The maximum degradation efficiency of 94% for dibutyl phthalate in wastewater (as an excellent plasticizer in different resins, especially nitrocellulose and resins and also an vital additive in special paints and adhesives with about 20 years of hydrolysis half-life) was obtained using Fe 2 O 3 in a photocatalytic process [ 137 ]. Compared to α-Fe 2 O 3 powders, porous α-Fe 2 O 3 films exhibit better photocatalytic activity by water splitting under UV radiation for hydrogen generation [ 138 ]. Photocatalytic oxidation of aniline to azobenzene by Fe 2 O 3 under UV irradiation and natural sunlight in aprotic and protic solvents has also been reported [ 139 ]. 4. Gadolinium Oxide (Gd 2 O 3 ) The global interest in using rare earth metals is increasing, due to their distinctive magnetic and electronic attributes in the fashioning of interfaces and surfaces compared to common bulk materials. Gadolinia (gadolinium (III) oxide) is the most widely available derivative form of gadolinium, and is a potential contrast agent in magnetic resonance imaging (MRI). The Gd 2 O 3 -modified bismuth vanadate (BiVO 4 ) composite, as a photocatalyst, exhibits significantly greater visible-light photocatalytic activity than pure BiVO 4 for methyl orange degradation under visible light irradiation [ 140 ]. Gd 2 O 3 nanorods used to detect ethanol by facile hydrothermal routes demonstrated a lower detection limit with higher sensitivity and shorter response time [ 141 ] compared to the annealed Gd 2 O 3 nanostructures [ 142 ]. Moreover, a moderate photocatalytic activity was evaluated for degradation of methyl orange by uniform Gd 2 O 3 hollow microspheres [ 143 ]. The degradation of about 90% of 4-chlorophenol using modified Gd 2 O 3 photocatalyst prepared by the sol–gel method was measured after 4 h of UV light irradiation [ 144 ]. In another study, a challenging photocatalyst of Gd 2 O 3 nanorods was designed for the degradation of neurotoxic chloramphenicol drugs [ 145 ]. 5. Antimony Oxide (Sb 2 O 4 ) Antimony oxide is classified based on its oxidation states, Sb (III) and Sb (V). Antimony has been applied as a pacifier in enamel, flame retardants, paint and glass art crafts and for making bullets and bullet tracers. It has been used as an alloy for the synthesis of plain bearings, batteries, and solders, as well as as a stabilizer and a catalyst for the preparation of polyethylene terephthalate. The photocatalytic activity of α-Sb 2 O 4 has been demonstrated, with almost 52% degradation of acridine orange in 170 min, with low detection limit, good sensitivity, long linear dynamic range with good linearity in a very short response time [ 146 ], as well as for the removal of heavy metals (e.g., mercury) from waste water [ 147 ], while it has also been reported that synthetic Uranyl Selective Polymeric Membrane sensors based on p-tert-butylbiscalix4arene can be used for the determination of Thorium [ 148 ]. The unique characteristics of nanostructures, such as their large surface area, excellent adsorbing and absorbing activity, bio-friendly nature, and high electron exchange could be reasons for the good sensitivity of these systems [ 149 ]. Binary Metal Oxides In addition to metal oxides, some other metal oxides have also been studied previously for use in the field of photocatalysis, because of their unique benefits and wide range of applications as catalysts, semiconductors, superconductors, ceramics, antifungal agents, adsorbents, and their specific applications in medicines. Many metal oxide semiconductors (e.g., WO 3 , ZrO 2 , ZnO, and Fe 2 O 3 ) that have been exploited in photocatalysts for the degradation of organic contaminants have inherent drawbacks [ 150 ]. For example, WO 3 is a stable photocatalyst for O 2 production within the visible light irradiation range. However, it is not suitable for H 2 evolution because of its low level of conduction band. Additionally, α-Fe 2 O 3 is somewhat stable in acidic solutions, but has the same problems as WO 3 . Moreover, ZnO can be easily corroded under band gap irradiation by photogenerated holes. Ta 2 O 5 photocatalyst with a nanocrystalline mesoporous structure has recently been synthesized for the production of H 2 via a sol–gel process combined with a surfactant-assisted templating mechanism [ 151 ]. Recently, the effect of Fe-doped NiO as a co-catalyst has also been reported [ 152 ]. Additionally, ZnO-CeO 2 nanoparticles synthesized using an efficient and simple low-temperature method have been successfully applied as photocatalysts for the removal of biomedical and environmental contaminants and reported 80.7% and 92.1% degradation for methylene blue and acridine orange within 170 min of irradiation time, respectively [ 130 ]. The Cu x S-TiO 2 composites has shown good efficiency in photo degradation of dyes even under visible light irradiation [ 153 ]. In another report, the photocatalytic activity of high-quality CeO 2 -CdO binary metal oxide nanocomposites was evaluated, showing acceptable growth inhibition of P. aeruginosa ( Figure 7 ) [ 154 ]. Metal Sulfides Metal sulfides have been widely used as visible light responsive photocatalysts. Compared to metal oxides, 3p orbitals of sulfur in their valence band result in a more occupied valence band and a narrower band gap. Recently, among other metal sulfides, ZnS and CdS have attracted great attention. CdS is commonly used for visible light-assisted water splitting due to its suitable band position and band gap (2.4 eV). However, photo-corrosion, which is a common issue in most metal sulfide photocatalysts, occurs when using both CdS and ZnS. Therefore, recent studies have focused on the development of ZnS and CdS photocatalysts, mostly through four different means of improvement: matrixing and supporting the structures of CdS, adding cocatalysts to CdS, preparing porous and one-dimensional CdS, and doping solid solutions of CdS and ZnS [ 155 ]. To synthesize porous CdS, a solvothermal method has been used to synthesize CdS nanowires [ 156 ] and nanorods [ 157 ]. Additionally, mesoporous CdS nanoparticles have been synthesized via ultrasonic and template-free precipitation at room temperature [ 158 ]. Nanoporous CdS have also been prepared by including hollow nanorods and nanosheets with 3 nm diameter of pores through a two-step aqueous method [ 159 ]. Additionally, CdS quantum dots have been recently loaded on porous polysaccharides and applied as highly efficient contrast imaging agents [ 160 ]. On the other hand, due to having an extremely broad band gap to respond to visible light (3.6 eV), solid solutions of ZnS are formed in which the narrow band gap can increase the use of ZnS in visible light. Both CdS and ZnS have the same crystal structures, making it easy to form solid solutions of them [ 161 ]. Magnetic Nanophotocatalysts The incorporation of magnetic nanophotocatalysts in contaminant removal strategies has recently received significant attention due to their improved chemical and physical properties. Therefore, cost-effective, efficient, and environmentally friendly disinfection processes can be achieved due to their easy separation using an external magnetic field, which allows recycling and multiple use of the nanophotocatalyst [ 162 ]. They mostly have a core–shell structure consisting of a magnetic core (e.g., iron, cobalt, nickel, and their oxides like maghemite (α-Fe 2 O 3 ), magnetite (Fe 3 O 4 ), cobalt ferrite (CoFe 2 O 4 )) and a photocatalytic shell (e.g., TiO 2 , ZnO, AgBr, BiOCl) [ 163 ]. Furthermore, some nanoferrites like ZnFe 2 O 4 have shown desirable degradation efficiency of organic target compounds under both visible light and UV irradiation [ 164 ]. Similar studies have reported degradation of different contaminants using Fe 3 O 4 [ 165 , 166 , 167 ], NiFe 2 O 4 [ 168 ], CoFe 2 O 4 [ 169 ], ZnFe 2 O 4 [ 170 ], BaFe 12 O 19 [ 171 ], SrFe 12 O 19 [ 172 ] -doped TiO 2 nanophotocatalysts ( Table 3 ). A schematic of the use of magnetic nanophotocatalysts (MNPCs) in water treatment [ 173 ] is illustrated in Figure 8 . Graphene Graphene (G), due to its one-of-a-kind nanostructure and particular properties has been studied widely from both experimental and theoretical scientific points of view [ 208 ]. It has already shown promising applications in nanocomposites, nanoelectronics, optoelectronics, drug delivery systems, electrochemical super-capacitors, transistors, solar cells, and chemical sensors (e.g., biosensors, gas sensors, pH sensors) [ 209 ]. As shown in Figure 9 , graphene has been employed to enhance photocatalytic efficiency, due to its electron scavenging nature, in the conduction band of metal oxide [ 26 ]. Some of its novel applications include ultrasensitive high-adsorption ability for various types of contaminations, including arsenic in drinking water [ 210 ], brackish water desalination and drinking water purification [ 211 ], metal removal from the contaminated environment [ 212 ], detection of biomarkers [ 213 ], electrochemical sensor for paracetamol [ 214 ], treatment of thrombosis [ 215 ], protection of DNA from cleavage and its effective cellular delivery [ 216 ], MRI and localized photothermal therapy for cancer cell treatment [ 217 ], electrochemical immunosensor for sensitive detection of carbohydrate antigen 1.5-3 (CA 15-3) [ 218 ], and photothermal agents in NIR region [ 219 ]. Modified graphene nanostructures such as P25–G [ 174 , 175 ], TiO 2 –G [ 176 , 177 , 178 , 179 ], SnO 2 –G [ 180 ], Bi 2 WO 6 –G [ 181 ], ZnO–G [ 182 ], ZnFe 2 O 4 –G [ 183 ], BiVO 4 –G [ 184 ], CdS–G [ 185 ] have also been reported to have different photodegradation applications ( Table 3 ). Quantum Dots Quantum dots (QDs), as zero-dimensional semiconductor multifunctional nanomaterials have been receiving significant attention for the degradation of pollutants [ 220 ]. Since QDs have the advantage of the wide band gap of a semiconductor material, they have a promising application as photocatalysts, owing to the swift generation of electron–hole pairs through photoexcitation [ 204 ]. On the other hand, photocatalytic, chemical and optical properties of QDs can be improved by surface modification, which also improves the photostability of QDs, as well as the efficacy of light-induced reactions on the QD surface and the generation of new traps on the QD surface [ 221 ]. For example, in the self-photosensitization pathway of fuchsin dye degradation, photodegradation can be initiated in the presence of graphene quantum dots (GQDs) under visible light irradiation, as demonstrated in Figure 10 [ 222 ]. The application of modified QDs as a photocatalytic agent to degrade pollutants is illustrated in Table 3 . Smart Materials (Self-Cleaning) Smart photocatalytic materials have been developed widely over the past two decades [ 49 ]. Different kinds of applications such as simultaneous self-cleaning and air cleaning have mostly focused on the use of TiO 2 and ZnO due to their low cost, high stability and strong capacity for the photocatalytic decomposition of organic contaminants [ 205 ]. TiO 2 has been used recently to make self-decontaminant textiles that offer high antibacterial activity performance for UV shielding [ 206 ]. Nanophotocatalysts can be merged onto different surfaces of bulk structures (i.e. concrete) [ 223 ] or onto the glass of windows, flat surfaces, or walls [ 224 ]. TiO 2 -coated membranes offer outstanding antifouling/self-cleaning, photoactive, and bactericidal properties that are based on the UV mechanism of TiO 2 photocatalysis ( Figure 11 ) [ 207 ]. The overall applications of different types of nanophotocatalysts mentioned in this review are summarized in Table 3 . 5. Future Prospects, and Concluding Remarks Biomedical waste eradication strategies are still very limited due to low photocatalytic efficacy. Therefore, extensive assessments are highly demanded from the practical point of view [ 225 ]. The applications of TiO 2 photocatalysts, for example, is restricted because of poor quantum efficacy as a result of limited absorption within only UV range (4% of sunlight). The recent advancement in the field of photocatalysis technology is investigating novel agents with higher photocatalytic performance to expand their light respond range. To address these issues, coupled semiconductor, noble metal deposition and ion modification, are the proposed methods to improve energy band and photocatalytic efficacy of explicit applications. In addition, due to concurrent photocatalytic and redox reactions, sensible photocatalytic systems can be designed for the simultaneously photocatalytic treatment of two or more contaminants [ 226 ]. However, several issues (e.g., evaluating the immobilization of photocatalysts and suspension systems) should be contemplated for further advancement. Solar-Based photocatalytic methods have shown better performance than conventional methods in the removal of tenacious organic contaminations [ 227 ]. To improve the photodegradation of wastewater, suitable surface modification technique of photocatalysts is an essential need. In addition, development of nanostructures, photoactivity of recycled photocatalysts, mechanisms of degradation, recovery of photocatalyst during treatment, and interactions between the photocatalysts and the pollutants are yet to be further improvements. To predict the kinetics, quantum yield and optimized conditions of the process, further investigations are required to verify the mathematical models for photocatalytic systems. The future improvement of nanophotocatalysts would be made by making them multifunctional and controllable enough to be subsequently transformed into nano-gadgets. To facilitate the accessibility of these innovations, extensive endeavors are expected to defeat the challenges in the future. The implementation of the photocatalytic method for disinfection and cleansing is an adaptable and effective procedure for incapacitating a broad range of adverse microorganisms in different media. This approach is a non-toxic, safe, and cost-effective sterilization technique whose versatility enables it to be used in various purposes. Nanotechnology has shown an incredible potential to improve the effectiveness of biomedical waste treatment [ 228 ]. Therefore, the use of nanophotocatalysts has become of great interest in biomedical waste management as they can provide excellent and practical applicability. Improving the biomedical waste management; however, should start with the reduction of wastes production according to the norms, rules, and standards in each country, with respect to regulation of biomedical waste disposal in various categories. Besides, practicing the optimized models for monitoring the waste produced by hospitals, health-care centers as well as research into eco-friendly sustainable technologies, recycling and PVC-free devices will go in a long way for a safe environment. Globally, more focused research in the field of biomedical waste management required to comprehend its impact on the field of public health better. There is an ongoing research in field of nanomaterials to design and develop nanophotocatalytic reactors. Despot of the advancements in field of nanophotocatalytic materials, further investigations are required to be done to characteristics the nanophotocatalytic materials. The major remaining challenges are strengthening the process, mass transfer limitations and high photons consumption. Therefore, the concept of using nanocomposites is ideal to resolve the issues related to electron pair recombination which can be prolonged by combining the nanocomposites with nanophotocatalytic reactor structures. The recent reactors known as microfluidic reactors open a new opportunity for intense characteristics study in reaction and synthesis phase. Microfluidic reactors are based on micro level reactants. The remarkable features of these reactors are; improved diffusion effect and great mass transfer coefficient factor, large surface-to-volume ratio, highly stable hydrodynamics, less Reynold's flow, and informal handling which make them better candidate compared to the conventional reactors. However, implementing the photocatalysis in a larger scale and actual wastewater is still challenging. Synthesis of structures such as nanorod, nanosphere, nanoflowers, nanoflakes and nanocones with improved functional and structural properties could open a new area of study in this subject. Crucially nanophotocatalysts with excellent efficiency, inexpensive, eco-friendly and high stability are needed to be synthesized [ 229 ]. 6. Importance of Waste Management as a Crucial Public Service during the COVID-19 Outbreak Since the outbreak of the coronavirus (COVID-19), its impact upon human health and the economy has been increasing day by day; administrations are advised to treat waste management, including of medical, household and other harmful waste, as an urgent and crucial public service in order to reduce the possible risk of secondary impacts upon public health and the environment. During such an epidemic, various types of medical and hazardous waste are generated, including infected personal protective equipment (PPE) masks, gloves and other protective equipment, as well as a higher volume of non-infected items of the same nature that can easily become mixed with domestic garbage, but should be treated as hazardous waste and disposed of separately from other household waste streams. Unreliable waste management could lead to unexpected "knock-on" impacts on human health and the environment. During the COVID-19 emergency, safe management of household waste is also expected to be critical. To overcome this enormous and unprecedented challenge, decision-makers are urged to make every effort to ensure that waste management, including that from medical and household sources, is given the attention—indeed priority—it requires in order to minimize its impact upon human health and the environment. The COVID-19 outbreak has made a global demand to effectively diagnose, treat and mitigate the spread of the infection, through comprehensive approaches such as specific alternative antiviral methods and classical disinfection protocols. The physicochemical properties of materials can be engineered to offer distinctive approaches to manage this emergency. Considering the life cycle of the virus, it is envisioned that nanotechnology could be employed to encounter the disease; nanoparticles (NPs), for instance, can be used as an alternative to the conventional disinfection protocols used in the healthcare system, owing to their inherent antipathogenic properties and their ability to inactivate viruses, bacteria, fungi, and yeasts either through photothermal or generation of photocatalysis-induced reactive oxygen species (ROS). In conclusion, nanotechnology can play a critical role in counteracting COVID-19 and preparing for future pandemics [ 230 ].

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8224260/

Nanostructured ZnO and ZnO: Pd with MXene overlayer SPR biosensors

The development of biosensors based on various novel techniques has become highly significant in the context of the outburst of a pandemic like COVID 19. The present work reports the theoretical modeling of two surface plasmon resonance (SPR) based biosensing probe configurations on optical fiber employing Metal/ZnO/MXene and Metal/ZnO: Pd/MXene. Maximum sensitivities of 19,400 nm/RIU and 8350 nm/RIU are calculated with Metal/ZnO/MXene and Metal/ZnO: Pd/MXene, respectively. The sensors are suitable for analytes with refractive index values ranging from 1.354 to 1.422. The refractive index of mucus and serum being in this range, the proposed biosensor, can be a potential tool for COVID-19 detection. Introduction Today we are at a crucial junction where the entire world is being shocked by the fatal epidemic of Corona Virus disease (COVID-19). The corona virus belongs to the viruses that may cause the common cold, influenza-like illnesses, nephritis, respiratory illnesses like pneumonia and bronchitis. Middle East Respiratory Symptom (MERS), Severe Acute Respiratory Symptom (SARS), and the COVID-19 are some of the severe symptoms caused by the corona virus family (Huang et al. 2009 ). Despite all the acquired medical and technological achievements, the epidemic is spreading at lightning speed, raising the death toll to around two lakhs. Hence early, accurate, and fast detection of the human corona virus is essential. Studies confirm that the corona virus gets transmitted through body fluids and social contacts. The endogenous biochemical processes within the human body and the defense mechanism of the tissues against intruders like virus results in the formation of many volatile organic compounds (VOC) as by-products and reflect in the body fluids and the exhaled gases. The constituents in the exhaled air and the body fluids can throw light into the health condition of human beings (Galassetti et al. 2005 ; Ruzsányi and Péter Kalapos 2017 ). It can provide diagnostic information on the early detection of various diseases. Many types of biosensors have been developed so far. The biosensors have become a potential tool in the early detection of diabetes (Gruber et al. 2016 ), abnormalities with lungs (Nardi-Agmon and Peled 2017 ) and blood serum (Liu et al. 2019 ). A wide variety of sensors working on different technologies and materials are available. The quality of sensors has been improved much with hybrid nanostructures (Andre et al. 2018 ). For accurate and quick sensing, the sensors should have better adsorption to various functional groups, quick response characteristics, and chemical stability. Recently, MXenes have revolutionized the sensing arena (Sinha et al. 2018 ; Zhu et al. 2017 ). Among the various optical sensors based on different techniques, such as infrared spectroscopy (Li et al. 1990 ), ultraviolet–visible spectroscopy (Tredgold et al. 1985 ), ellipsometry (Beck et al. 2011 ), and surface plasmon resonance (SPR) (Tabassum and Gupta 2015a , b ), the SPR sensors stand apart as a better choice. The sensors based on the principle of SPR have proved to yield quick and accurate sensing of physical, chemical, and biochemical parameters (Sharma and Gupta 2007 ; Jorgenson and Yee 1993 ; Ghosh et al. 2013 ). SPR sensors make use of the Kretschmann configuration (Homola 1997 ). When a metal is placed in contact with a dielectric and is excited by an optical signal, resonant oscillation called surface plasmons, of the free electrons at the interface between a metal and a dielectric may happen when the wave vectors of the conducting electron cloud and evanescent waves match. SPR sensors operate on the above principle. Surface plasmons are exponentially decaying transverse magnetic (TM) polarized non-radiative electromagnetic signals propagating parallel to the metal–dielectric interface. The principal advantageous feature of SPR is that, even minor variations of refractive index (RI) in the boundary of the metal and dielectric will largely influence the resonance conditions of the surface plasmons. Adsorption of external materials to the surface alter the refractive index, and hence SPR based sensors ensure efficient sensing applications. For harnessing the advantages of optical fibers, the Kretschmann configuration is modeled to be implemented with optical fibers and are called fiber optic SPR sensors (Villuendas and Pelayo 1990 ). When white light is focused on one end of the fiber optic sensor, maximum energy from the transmitted light gets transferred at resonance to the surface plasmons. Hence, a corresponding dip appears at the recorded output waveform at the other end of the fiber. This waveform is the SPR curve. The resonant wavelength ( λ res ) changes with change in the RI of the sensing medium, and hence calibration of the wavelength shift for a specific change in RI can yield the sensitivity. These fiber optic SPR sensors enable compactness, ease of sensing, high sensitivity, fast sensing, and reliability for noninvasive measurements in sensing and biosensing. In the proposed theoretical modeling, fiber optic SPR hybrid nanostructured sensing probes with aluminium (Al), zinc oxide (ZnO), Palladium (Pd), and Ti 3 C 2 T x MXene, suitable for biomedical applications is considered. MXenes are two dimensional materials derived from the M n+1 AX n (MAX) phases, where M is an early transition metal, A is a group 13 or 14 element, and X is carbon or a combination of carbon and nitrogen. They possess high hydrophilicity, large surface area, and are easy to functionalize. The unique physical, chemical, and ion transport properties make them a prominent choice in many applications, especially sensing and biosensing (Sinha et al. 2018 ; Zhu et al. 2017 ; Wu et al. 2018 ). MXenes feature excellent biocompatibility and allow secure immobilization of enzymes and proteins on its surface (Wu et al. 2018 ; Rakhi et al. 2016 ; Wang et al. 2015a ). They possess a large specific surface area and anisotropic conductivity of charge carriers, together with environmental friendliness and excellent chemical stability (Sinha et al. 2018 ; Zhu et al. 2017 ). The presence of surface dangling bonds enables MXenes to attract various functional groups, which make them a unique choice for sensing applications (Sinha et al. 2018 ; Zhu et al. 2017 ; Wang et al. 2015b ; Sudheer et al. 2020 ). ZnO is an II-VI compound semiconductor with a direct and large bandgap. The ionicity resides at the borderline between covalent and ionic semiconductors. ZnO has high thermal and chemical stability, substantial mechanical strength, sizeable piezoelectric coefficient, high exciton binding energy, and optical gain, making it an ideal material to realize excitonic devices at room temperature (Tabassum and Gupta 2015a , b ; Özgür et al. 2005 ; Shukla et al. 2015 ). Palladium is an attractive metal that has been used as a stable Schottky contact of wide bandgap and is a good conductor with superior thermal and chemical stabilities (Tabassum and Gupta 2015a , b ; Abd Rahim et al. 2011 ). Pd nanoparticles possess excellent SPR features, which could lead to applications in colorimetric sensing, plasmonic waveguiding, enhancement of electromagnetic fields and light transmission, and optical sensing of hydrogen (Hamed 2016 ). Pd is very resistant to oxidation and corrosion and possesses excellent catalytic properties. It is also chemically stable. Reverse transcription-polymerase chain reaction (RT-PCR), serological testing, and antibody-based testing were earlier used to detect the coronavirus. Detection based on antigens like N-protein also yielded better results as these proteins can be detected even early as on the first day of infection. Different methods like enzyme-linked immunosorbent assay (ELISA), immunofluorescence assay, and enzyme-linked immunosorbent assay using chemiluminescence (CLEIA) were utilized for detecting N-proteins (Huang et al. 2009 ). Literature shows that studies were also conducted on the refractive index (RI) of the coronavirus (Hamed 2016 ; Pang et al. 2016 ). All of the above test procedures are time-consuming. The proposed fiber-optic SPR sensor has the advantage of the compact size and flexibility of the optical fiber combined with the unique properties of the outer layer of MXene. Hence the sensor can facilitate simple, fast, and accurate invasive and noninvasive sensing of body fluids. We strongly believe that our proposed sensing probe structures can guide the development of sensors for fast and accurate detection of coronavirus infection. Experimental considerations and theoretical modeling Many sensor configurations have already been proposed with ZnO and nanocomposite of ZnO and Pd with enhanced sensitivity. For harnessing the features and properties of ZnO, Pd, and MXene, two different hybrid nanostructure sensor configurations are proposed—Metal/ZnO/ Ti 3 C 2 T x and Metal/ZnO: Pd/ Ti 3 C 2 T x . Simulations are done with aluminium as the metal and Ti 3 C 2 T x as the outer layer interacting with the analyte. Here we propose to remove the cladding for a length of 1.5 cm from a Plastic Clad Silica (PCS) fiber of core diameter 600 μm and numerical aperture of 0.24. This cladding removed portion is coated with the thin film of aluminium. The proposed sensing probe is illustrated in Fig. 1 . In the first configuration, a thin layer of ZnO is sandwiched between the metal and Ti 3 C 2 T x layers, whereas in the second configuration, the ZnO layer is replaced with a nanocomposite of ZnO and Palladium. A beam of collimated white light is made to fall on one face of the fiber with the sensing probe at a suitable angle. The SPR curve recorded at the other end of the fiber gives ample information on the resonance wavelength, change of RI of sensing medium, and sensitivity of the sensor and accuracy of sensing. Usually, the Full-Width Half Maximum (FWHM) of the curve is considered for calculating the spectral width. Figure 2 illustrates two different SPR curves. Depending on parameters like the type and thickness of the different layers of the sensing probe and the RI of the analyte(n s ), the SPR curve need not always be perfect like curve A, shown in Fig. 2 . Calculation of FWHM is possible with the standard SPR curve A. Sometimes, the right-hand shoulder may fall much lower than the left one, as depicted by curve B in Fig. 2 . The depth and sharpness of the SPR curve is critical as it affects the accuracy of determination of the resonant wavelength. The larger the spectral width, the lesser is the accuracy. Hence we have calculated the spectral width at a height corresponding to 0.1 unit above the tip of the SPR dip. For better clarity in representation and interpretation, the inverse of the spectral width is taken and is termed as detection accuracy (DA) (Sudheer et al. 2020 ). Use of DA facilitates easy analysis of the graphs and interpretation of the dependence of resonance wavelength on changes in RI. Fig. 1 Schematic of the proposed sensing probes a Al/ZnO/Ti 3 C 2 T x and b Al/ZnO: Pd/Ti 3 C 2 T x Fig. 2 SPR curves- curve A is a typical SPR curve and curve B illustrates an SPR curve with variation depending on parameters like RI of the analyte and the nature and thickness of the layers of the sensor The proposed SPR sensing probe is based on the Kretschmann configuration with the wavelength interrogation technique. Hence wavelength dependence of the dielectric constants of the core material of the optical fiber, aluminium, ZnO, Palladium, and MXene need to be considered in the theoretical modeling. The dependence of the wavelength on the refractive index of the silica core of the PCS fiber is calculated as defined by the Sellmeier relation as (Ghatak and Thyagarajan 1998 ). 1 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$n\left(\lambda ight)=\sqrt{1+ rac{{A}_{1}{\lambda }^{2}}{{\lambda }^{2}-{B}_{1}^{2}}+ rac{{A}_{2}{\lambda }^{2}}{{\lambda }^{2}-{B}_{2}^{2}}+ rac{{A}_{3}{\lambda }^{2}}{{\lambda }^{2}-{B}_{3}^{2}}}$$\end{document} n λ = 1 + A 1 λ 2 λ 2 - B 1 2 + A 2 λ 2 λ 2 - B 2 2 + A 3 λ 2 λ 2 - B 3 2 where λ is the wavelength in μ m and A 1 , A 2 , A 3 , B 1 , B 2 and B 3 are Sellmeier coefficients with values of A 1 = 0.6961663, A 2 = 0.4079426, A 3 = 0.8974794, B 1 = 0.0684043, B 2 = 0.1162414 and B 3 = 9.896161. The Drude dispersion model quantifies the wavelength dependence of the dielectric constant of the metals with the relation (Sharma and Gupta 2007 ; Sharma et al. 2007 ) 2 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ arepsilon _{m} \left( \lambda ight) = ~ arepsilon _{{{ ext{mr}}}} + i arepsilon _{{{ ext{mi}}}} = 1 - rac{{\lambda ^{2} \lambda _{c} }}{{\lambda _{p}^{2} \left( {\lambda _{c} + i\lambda } ight)}}$$\end{document} ε m λ = ε mr + i ε mi = 1 - λ 2 λ c λ p 2 λ c + i λ where λ p and λ c are respectively the plasma and collision wavelengths of the metal, ε mr and ε mi are respectively the real and imaginary parts of the dielectric constant of the metals and ε ∞ correspond to the dielectric constant for the infinite value of the frequency. For metals, the value of ε ∞ is 1. The values of λ p and λ c for aluminium (A l ) are 1.0657 × 10 –7 and 2.4511 × 10 − 5 , respectively, where λ is in meters (Ordal et al. 1983 ; Lambrecht and Reynaud 2000 ). Similarly, the dielectric constant of ZnO and Pd as a function of the wavelength of light is given as (Tabassum and Gupta 2015a , 2015b ) 3 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${ arepsilon }_{ZnO}\left(\lambda ight)=2.81418- rac{0.87968{\lambda }^{2}}{\left({\lambda }^{2}-{0.3042}^{2} ight)-{0.00711}^{2}}$$\end{document} ε ZnO λ = 2.81418 - 0.87968 λ 2 λ 2 - 0.3042 2 - 0.00711 2 4 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${ arepsilon }_{Pd}\left(\lambda ight)=\left[-24.96364{e}^{\left( rac{-\lambda }{1.09383} ight)}+28.65615 ight]+i\left[15.31843{e}^{\left( rac{\lambda }{0.76839} ight)}-16.97702 ight]$$\end{document} ε Pd λ = - 24.96364 e - λ 1.09383 + 28.65615 + i 15.31843 e λ 0.76839 - 16.97702 For the nanocomposite of ZnO and Pd, particle size is considered to be much smaller than the wavelength of light. Nanoparticles of Pd is considered to be dispersed in the host solution of ZnO and is modeled with the Maxwell–Garnett model. The normalized transmitted power through the fiber measured at one end due to a white light source fed at the other end of the fiber in the SPR based sensor can be expressed as (Singh and Gupta 2010 ) 5 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${P}_{trans}= rac{{\int }_{{ heta }_{cr}}^{\pi /2}{R}_{P}^{{N}_{ref}( heta )} rac{{n}_{1}^{2}\mathrm{sin} heta \mathrm{cos} heta }{{(1-{n}_{1}^{2}{cos}^{2} heta )}^{2}}d heta }{{\int }_{{ heta }_{cr}}^{\pi /2} rac{{n}_{1}^{2}\mathrm{sin} heta \mathrm{cos} heta }{{(1-{n}_{1}^{2}{cos}^{2} heta )}^{2}}d heta }$$\end{document} P trans = ∫ θ cr π / 2 R P N ref ( θ ) n 1 2 sin θ cos θ ( 1 - n 1 2 cos 2 θ ) 2 d θ ∫ θ cr π / 2 n 1 2 sin θ cos θ ( 1 - n 1 2 cos 2 θ ) 2 d θ where R p is the net reflection coefficient of the ray incident at the core metal interface. θ cr is the critical angle of the fiber, n 2 is the refractive index of the fiber cladding, N ref (θ) represents the number of reflections the ray launched at an angle θ undergoes inside the fiber core. Also, N ref (θ) = L / D tan θ, where L is the length of the sensing probe, and D is the diameter of the fiber core. In this work, the sensing probes considered are the multilayered structure, including the fiber core and the sensing medium. Hence, the N-layer matrix method (Hansen 1968 ) has been utilized to determine the value of R p accurately. Accordingly, the characteristic matrix for an N-layer structure can be expressed as 6 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$M=\prod _{k=2}^{N-1}{M}_{k}=\left[egin{array}{cc}{M}_{11}& {M}_{12}\ {M}_{21}& {M}_{22}\end{array} ight]=\left[egin{array}{cc}\mathrm{cos}{eta }_{k}& (-i\mathrm{sin}{eta }_{k})/{q}_{k}\ -i{q}_{k}\mathrm{sin}{eta }_{k}& \mathrm{cos}{eta }_{k}\end{array} ight]$$\end{document} M = ∏ k = 2 N - 1 M k = M 11 M 12 M 21 M 22 = cos β k ( - i sin β k ) / q k - i q k sin β k cos β k where 7 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${q}_{k}={\left( rac{{\mu }_{k}}{{ arepsilon }_{k}} ight)}^{1/2}\mathrm{cos}{ heta }_{k}= rac{{({ arepsilon }_{k}-{n}_{1}^{2}{sin}^{2}{ heta }_{1})}^{1/2}}{{ arepsilon }_{k}}$$\end{document} q k = μ k ε k 1 / 2 cos θ k = ( ε k - n 1 2 sin 2 θ 1 ) 1 / 2 ε k and 8 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${eta }_{k}= rac{2\pi }{\lambda }{n}_{k}\mathrm{cos}{ heta }_{k}\left({z}_{k}-{z}_{k-1} ight)= rac{2\pi {d}_{k}}{\lambda }{({ arepsilon }_{k}-{n}_{1}^{2}{sin}^{2}{ heta }_{1})}^{1/2}$$\end{document} β k = 2 π λ n k cos θ k z k - z k - 1 = 2 π d k λ ( ε k - n 1 2 sin 2 θ 1 ) 1 / 2 The reflection coefficient r p of the incident wave through the N layered structure is given by 9 \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$${r}_{p}= rac{\left({M}_{11}+{M}_{12}{q}_{N} ight){q}_{1}-\left({M}_{21}+{M}_{22}{q}_{N} ight)}{\left({M}_{11}+{M}_{12}{q}_{N} ight){q}_{1}+\left({M}_{21}+{M}_{22}{q}_{N} ight)}$$\end{document} r p = M 11 + M 12 q N q 1 - M 21 + M 22 q N M 11 + M 12 q N q 1 + M 21 + M 22 q N and the corresponding reflectance is \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$R_{p} = ~\left| {r_{p} } ight|^{2}$$\end{document} R p = r p 2 The qualities of any SPR sensors are judged by three main parameters (Wu et al. 2018 ; Ghatak and Thyagarajan 1998 ). (a) Sensitivity, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$S = rac{{{ ext{change}}\;~{ ext{in}}\;~{ ext{resonant wavelenth}}}}{{{ ext{change~in}}~{ ext{refractive}}~{ ext{index}}}} = ~ rac{{\Delta \lambda _{{{ ext{res}}}} }}{{\Delta n_{s} }}$$\end{document} S = change in resonant wavelenth change in refractive index = Δ λ res Δ n s (b) The spectral width of the SPR curve. (c) Range of refractive index values over which efficient sensing is possible. Sensitivity is a major parameter that determines the quality of sensing. The more the shift in resonant wavelength for a unit change in n s , the more accurate will be the sensing for the very small changes in n s . In the same way, the spectral width of the SPR curve is also very important for an SPR sensor in that the accuracy of detection increases with smaller spectral width. Hence, a trade-off between the sensitivity and spectral width needs to be considered. Results and discussion Certain benchmark values for the quality parameters (Sudheer et al. 2020 ; Sharma et al. 2007 ) are fixed by the authors to ensure better sensing—(i) λ res must lie in the visible region (ii) the sensitivity should not be less than 3000 nm/RIU (iii) The DA should not be less than 0.015 nm −1 . Simulation results on the λ res , its spectral width, and depth of the SPR curve are collected for a change in RI of 0.0001 RIU. Al/ZnO/Ti 3 C 2 Tx Figure 3 a illustrates the variation of Sensitivity and DA for different thickness of the aluminium layer, with 10 nm of ZnO layer and a single layer of Ti 3 C 2 T x . No noticeable SPR curve is obtained for 10 nm thickness of aluminium. Up to 20 nm thickness of aluminium, the observed sensitivity is very low. After that, the sensitivity increases with an increase in the thickness of the aluminium layer up to 40 nm. It remains at an almost saturation value of ~ 10,600 nm/RIU up to 50 nm and is observed to decrease with further increase in aluminium layer thickness (up to the considered thickness of 70 nm). The DA is found to increase with an increase in the thickness of the Al layer. That is, even though aluminium layer thickness of 40 nm and 50 nm yielded almost the same sensitivity, the DA is better for 50 nm thickness. Figure 3 b depicts the normalized transmitted power corresponding to the different thickness of the aluminium layer as mentioned above. The resonant wavelength cannot be determined with the SPR curve corresponding to 20 nm thickness. The curve corresponding to 40 nm thickness exhibited larger red shifts and is of comparatively larger spectral width. For 50 nm thickness, the curve made only a lesser red shift, with a comparatively lesser spectral width and very good depth. The curves for 60 nm and above (not shown) exhibits very small depth. Fig. 3 a Variation of sensitivity and detection accuracy as a function of different thickness of Al measured at n s = 1.413 with a single layer of MXene and 10 nm of ZnO. b Normalised transmitted power as a function of the wavelength for different thickness of Al measured at n s = 1.413 with a single layer of MXene and 10 nm of ZnO Hence 50 nm is considered as the optimum thickness of aluminium in all the configurations. The variation of sensitivity with different layers of Ti 3 C 2 T x for 50 nm of aluminium and 10 nm of a ZnO is plotted in Fig. 4 a, which shows that a single layer of Ti 3 C 2 T x yields higher sensitivity of ~ 10,350 nm/RIU. The sensitivity is found to decrease for a further increase in the number of Ti 3 C 2 T x layers. The DA increases with an increase in the number of Ti 3 C 2 T x layers. Also, the λ res experienced a red shift with an increase in Ti 3 C 2 T x layers. Variation in sensitivity and DA as a function of RI, corresponding to 5 nm and 10 nm thickness of ZnO, 50 nm of aluminium and a single layer of Ti 3 C 2 T x are shown in Fig. 5 a. Fig. 4 a Variation of sensitivity and detection accuracy as a function of the number of layers of MXene, measured at n s = 1.413 with 50 nm of Al and 10 nm of ZnO. b Wavelength dependence of the normalized transmitted power at n s = 1.413 for the different number of MXene layers, with 50 nm of Al and 10 nm of ZnO Fig. 5 a Variation of sensitivity and detection accuracy as a function of the refractive index of the analytes, with 50 nm of Al and a single layer of MXene for different thickness of ZnO. b Variation of sensitivity and detection accuracy as a function of the thickness of ZnO measured at n s = 1.413 with 50 nm of Al and a single layer of Mxene. c Normalised transmitted power as a function of wavelength for different thickness of ZnO measured at n s = 1.413 with 50 nm of Al and a single layer of MXene With 5 nm of ZnO, no noticeable SPR dip is observed for values of n s < 1.36, whereas for n s = 1.361 to 1.375, very low sensitivity is observed. The calculated sensitivity is well above the acceptable benchmark value for n s values from 1.38 to 1.416. Beyond n s = 1.416, the λ res is observed to shift to the IR region. The sensitivity is found to increase slightly up to n s = 1.392 and after that increases progressively with n s . A sensitivity of 11,250 nm/RIU is calculated for n s = 1.416. In fact, with 5 nm thickness of ZnO, the effective sensing range is from n s = 1.38 to 1.416. The DA is observed to decrease with an increase in n s , but is within the acceptable benchmark value. Calculations with 10 nm thickness of ZnO observe no noticeable SPR dip up to n s values of 1.39 and thereafter minimal sensitivity value up to n s = 1.39. For n s = 1.4 to 1.42 steady increase in sensitivity with n s is observed. Sensitivity of 14,000 nm/RIU is calculated for n s = 1.419. For n s greater than 1.42, the λ res shifts to the IR region. The sensing range is minimal, ranging from n s = 1.4 to 1.42. The observed SPR curves are so asymmetric that DA could be calculated only for a very narrow range of refractive index and hence not included in the graph. The calculated DA is in the acceptable range. A much higher sensitivity of 19,400 nm/RIU is calculated for n s = 1.422 (not shown in the graph). The DA calculated is slightly higher. The sensing range is constricted to n s = 1.419 to 1.422. Thus it is observed that an increase in the thickness of ZnO increases the sensitivity and the value of RI that can be sensed. But the lower range of sensing range got constricted largely. The variation of sensitivity and DA with different thickness of ZnO, 50 nm of aluminium, a single layer of Ti 3 C 2 T x and n s = 1.412 is shown in Fig. 5 b. Up to around 4.5 nm thickness of ZnO, no noticeable effect on the sensitivity enhancement is seen. Thereafter the sensitivity gradually increases and reaches a maximum value of 11,550 nm/RIU at 11 nm thickness of ZnO, and with further increase in ZnO thickness, shows a reduction in sensitivity. The DA decreases progressively but is within the acceptable range. The plot of the normalized transmitted power for different thicknesses of ZnO with 50 nm of aluminium, a single layer of Ti 3 C 2 T x , and n s = 1.412 depicted in Fig. 5 c shows that all the SPR curves are of almost the same dip. But they experience a blue shift with an increase in ZnO thickness. This configuration can lead to designing SPR biosensors and also sensors for gases and liquids of RI greater than 1.38.Thus 50 nm of aluminium, 11 nm of ZnO, and a single layer of Ti 3 C 2 T x will be the optimum values for efficient sensing. Al/ZnO: Pd/Ti 3 C 2 T x This is a three-layer configuration, just like the Al/ZnO/Ti 3 C 2 T x sensing probe described above, where the ZnO sandwich layer is replaced by a nanocomposite of palladium dispersed in ZnO. Here also the optimum thickness of aluminium is taken as 50 nm and of Ti 3 C 2 T x as a single layer. Calculations have been done with different volume fractions (f) and thickness of the ZnO: Pd nanocomposite. The effect of incorporating the ZnO: Pd nanocomposite can be understood with the illustrative graphs in Fig. 6 . Figure 6 a illustrates the variations in sensitivity and detection accuracy of the sensing probe with 10 nm thick ZnO: Pd nanocomposite layer for f = 0.2, 0.4 and 0.6. Fig. 6 Sensitivity and detection accuracy as a function of the refractive index of the analytes. a for the different volume fraction of ZnO: Pd with 50 nm of Al, a single layer of MXene and 10 nm thickness of ZnO: Pd nanocomposite. b for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.2. c for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.4 d for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.6 For f = 0.2, no SPR dip is observed for n s values less than or equal to 1.3. For n s = 1.3 to 1.35, very low sensitivity is observed. Appreciable sensitivity is seen for n s values from 1.37 to 1.398, beyond which the λ res shifted to the IR region. Maximum Sensitivity of 7250 nm/RIU is calculated for n s = 1.398. Though the DA is observed to be lesser than that of Al/ZnO/Ti 3 C 2 T x configuration, the values are well within the acceptable range. With f = 0.4, SPR dip is observed only after n s = 1.22. An acceptable range of sensitivity is observed from n s = 1.37 to 1.382. Beyond n s = 1.382 the λ res shifts to the IR region. Similarly, for f = 0.6, the simulation does not yield any SPR dip up to n s = 1.109. From n s = 1.1 to 1.35, the observed sensitivity is very low. The acceptable range of sensitivity is calculated for n s = 1.352 to 1.366, above which the λ res shifts to the IR region. Sensitivity of 4100 nm/RIU is calculated for n s = 1.366. The DA observed is lower than that for f = 0.4. The calculated sensitivity is found to increase with the increase in the value of possible n s . For any value of n s in the possible range, the sensitivity is found to increase with an increase in volume fraction. Also, the lowering of the lower and upper limits of the sensing range can be noted. In all the cases, the λ res experience a redshift with an increase in the value of n s . The SPR curves observed for f = 0.8 is much wide that accurate detection of λ res will be difficult. The analysis can be much clearer with the plots of Fig. 6 b, c, and d, which depicts the variation of sensitivity and DA with n s for the selected thickness values of ZnO: Pd nanocomposite for different volume fractions of 0.2, 0.4 and 0.6 respectively. Figure 6 b illustrates the variations in sensitivity and DA with the nanocomposite thickness of 5 nm, 10 nm, and 15 nm for f = 0.2. With 5 nm thickness of the nanocomposite, effective sensing is possible only for values of n s = 1.38 to 1.406. The maximum sensitivity of 8350 nm/RIU is calculated for n s = 1.406. With 10 nm thickness, the sensing range is from n s = 1.37 to 1.398, with a maximum sensitivity of 7250 nm/RIU. For 15 nm, the effective sensing range gets limited to n s = 1.36 to 1.388 with enhanced sensitivity. Sensitivity of 6200 nm/RIU is calculated for n s = 1.388. The DA is also observed to decrease with an increase in thickness of the nanocomposite. Figure 6 c shows the variations in sensitivity and DA with ZnO: Pd thickness of 5 nm, 10 nm and 15 nm for f = 0.4. With ZnO: Pd thickness of 5 nm, the sensing range extended from n s = 1.373 to 1.396 imparting 6150 nm/RIU sensitivity for n s = 1.396. For 10 nm thickness, the sensing range is observed to be from n s = 1.37 to 1.382 with a sensitivity of 5200 nm/RIU at n s = 1.382. Though SPR curve can be obtained for values of n s from 1.11 onwards, effective sensing can be had only for a very narrow span of n s = 1.348 to 1.354, when the thickness of the nanocomposite is 15 nm and f = 0.4. A sensitivity of 3400 nm/RIU is observed for n s = 1.354. The nanocomposite of 15 nm thickness yields higher sensitivity and lesser DA than that of 10 nm. Figure 6 d depicts the variation of sensitivity and DA with 10 nm, 15 nm of the nanocomposite for f = 0.6. The sensing range from n s = 1.37 to 1.392 for a thickness of 5 nm imparting a sensitivity of 5850 nm/RIU at n s = 1.392. With 10 nm thickness, an acceptable range of sensitivity is observed only for n s = 1.352 to 1.366. With 15 nm of nanocomposite and volume fraction of 0.6, though SPR curves are observed for n s = 1.01 to 1.328, the sensitivity is not up to the acceptable benchmarked value. Figure 7 a depicts the SPR curves corresponding to ZnO: Pd thickness of 5 nm, 10 nm, and 15 nm for a volume fraction of 0.2 and n s = 1.355. Similarly, the SPR curves corresponding to volume fractions of 0.2, 0.4, and 0.6 for ZnO: Pd thickness of 10 nm and n s = 1.355 is illustrated in Fig. 7 b. It is observed that with an increase in volume fraction and thickness of the nanocomposite, the SPR curves experience a decrease in depth and DA as well as redshift in λ res . Thus in all the cases analyzed, with an increase in the thickness of the nanocomposite and also the volume fraction, for any value of RI, the λ res experience redshift, and the sensitivity is found to increase. The nanocomposite of 20 nm thickness does not provide acceptable values of sensitivity (not shown). Fig. 7 Normalised transmitted power as a function of wavelength for the different, a thickness of ZnO: Pd nanocomposite at f = 0.2, with 50 nm of Al, a single layer of MXene and n s = 1.355. b volume fraction of the nanocomposite, with 50 nm of Al, a single layer of MXene, 10 nm of ZnO: Pd nanocomposite and n s = 1.355 Both of the proposed sensor configurations are of the same number of layers (five layers, including the fiber core and the analyte). The consolidation of the performance analysis of both the configurations on different analytes, subject to the benchmark conditions of sensitivity (≥ 3000 nm/RIU), detection accuracy(≥ 0.013 nm −1 ), and spectral range of λ res (visible range) is furnished below in Table 1 . The table provides clear information on the possible sensing range and the valid sensitivity for each configuration. The sensor with ZnO as the sandwich layer finds application for analytes of RI 1.388–1.422. Similarly, the second configuration with ZnO: Pd nanocomposite can be made use of for RI values 1.354 to 1.406. Though sensors with ZnO: Pd offers lesser Sensitivity and DA compared to that with ZnO, they possess a broader sensing range of the analytes. These sensors can be made use of for biosensing different body fluids and also in sensing of VOCs like acetone and ethanol. The optimum thickness of aluminium is confirmed to be 50 nm, and of Ti 3 C 2 T x is a single layer. Therefore, the physical realization of the sensor can be done with proper selection of the configuration and thickness and volume fraction of the sandwich layer depending on the application and the required performance parameter values. But of course, since the calculations and consolidations in this work are based on theoretical modeling, slight changes may happen while dealing with the real cases. Table 1 Sensitivity of the sensor configurations for different thickness and volume fractions of ZnO and ZnO: Pd at different RI of analytes. Each column shows the minimum and maximum values of the acceptable sensitivities and the corresponding values of RI. Each row shows the acceptable sensitivity values of each configuration for different RI values RI ZnO (5 nm) ZnO (10 nm) ZnO (15 nm) ZnO:Pd (5 nm) ZnO:Pd (10 nm) ZnO:Pd (15 nm) f 0.2 0.4 0.6 0.2 0.4 0.6 0.2 0.4 0.6 1.354 – – – – – – – – 3 – 3.4 – 1.364 – – – – – – – – 4.1 – – – 1.376 – – – 3 3.3 3.55 3.65 4.35 – 4.35 – – 1.388 4.41 – – 4.2 4.65 5.1 5.15 – – 6.2 – – 1.396 4.85 – – 5.45 6.16 –7.25 – – – – – – 1.400 5.45 – – 6.35 – – – – – – – – 1.406 – – – 8.35 – – – – – – – – 1.41 10 11 – – – – – – – – – – 1.416 11.25 11.95 – – – – – – – – – – 1.41 – 14 – – – – – – – – – – 1.42 – – 19.4 – – – – – – – – – Al/ZnO/Ti 3 C 2 Tx Figure 3 a illustrates the variation of Sensitivity and DA for different thickness of the aluminium layer, with 10 nm of ZnO layer and a single layer of Ti 3 C 2 T x . No noticeable SPR curve is obtained for 10 nm thickness of aluminium. Up to 20 nm thickness of aluminium, the observed sensitivity is very low. After that, the sensitivity increases with an increase in the thickness of the aluminium layer up to 40 nm. It remains at an almost saturation value of ~ 10,600 nm/RIU up to 50 nm and is observed to decrease with further increase in aluminium layer thickness (up to the considered thickness of 70 nm). The DA is found to increase with an increase in the thickness of the Al layer. That is, even though aluminium layer thickness of 40 nm and 50 nm yielded almost the same sensitivity, the DA is better for 50 nm thickness. Figure 3 b depicts the normalized transmitted power corresponding to the different thickness of the aluminium layer as mentioned above. The resonant wavelength cannot be determined with the SPR curve corresponding to 20 nm thickness. The curve corresponding to 40 nm thickness exhibited larger red shifts and is of comparatively larger spectral width. For 50 nm thickness, the curve made only a lesser red shift, with a comparatively lesser spectral width and very good depth. The curves for 60 nm and above (not shown) exhibits very small depth. Fig. 3 a Variation of sensitivity and detection accuracy as a function of different thickness of Al measured at n s = 1.413 with a single layer of MXene and 10 nm of ZnO. b Normalised transmitted power as a function of the wavelength for different thickness of Al measured at n s = 1.413 with a single layer of MXene and 10 nm of ZnO Hence 50 nm is considered as the optimum thickness of aluminium in all the configurations. The variation of sensitivity with different layers of Ti 3 C 2 T x for 50 nm of aluminium and 10 nm of a ZnO is plotted in Fig. 4 a, which shows that a single layer of Ti 3 C 2 T x yields higher sensitivity of ~ 10,350 nm/RIU. The sensitivity is found to decrease for a further increase in the number of Ti 3 C 2 T x layers. The DA increases with an increase in the number of Ti 3 C 2 T x layers. Also, the λ res experienced a red shift with an increase in Ti 3 C 2 T x layers. Variation in sensitivity and DA as a function of RI, corresponding to 5 nm and 10 nm thickness of ZnO, 50 nm of aluminium and a single layer of Ti 3 C 2 T x are shown in Fig. 5 a. Fig. 4 a Variation of sensitivity and detection accuracy as a function of the number of layers of MXene, measured at n s = 1.413 with 50 nm of Al and 10 nm of ZnO. b Wavelength dependence of the normalized transmitted power at n s = 1.413 for the different number of MXene layers, with 50 nm of Al and 10 nm of ZnO Fig. 5 a Variation of sensitivity and detection accuracy as a function of the refractive index of the analytes, with 50 nm of Al and a single layer of MXene for different thickness of ZnO. b Variation of sensitivity and detection accuracy as a function of the thickness of ZnO measured at n s = 1.413 with 50 nm of Al and a single layer of Mxene. c Normalised transmitted power as a function of wavelength for different thickness of ZnO measured at n s = 1.413 with 50 nm of Al and a single layer of MXene With 5 nm of ZnO, no noticeable SPR dip is observed for values of n s < 1.36, whereas for n s = 1.361 to 1.375, very low sensitivity is observed. The calculated sensitivity is well above the acceptable benchmark value for n s values from 1.38 to 1.416. Beyond n s = 1.416, the λ res is observed to shift to the IR region. The sensitivity is found to increase slightly up to n s = 1.392 and after that increases progressively with n s . A sensitivity of 11,250 nm/RIU is calculated for n s = 1.416. In fact, with 5 nm thickness of ZnO, the effective sensing range is from n s = 1.38 to 1.416. The DA is observed to decrease with an increase in n s , but is within the acceptable benchmark value. Calculations with 10 nm thickness of ZnO observe no noticeable SPR dip up to n s values of 1.39 and thereafter minimal sensitivity value up to n s = 1.39. For n s = 1.4 to 1.42 steady increase in sensitivity with n s is observed. Sensitivity of 14,000 nm/RIU is calculated for n s = 1.419. For n s greater than 1.42, the λ res shifts to the IR region. The sensing range is minimal, ranging from n s = 1.4 to 1.42. The observed SPR curves are so asymmetric that DA could be calculated only for a very narrow range of refractive index and hence not included in the graph. The calculated DA is in the acceptable range. A much higher sensitivity of 19,400 nm/RIU is calculated for n s = 1.422 (not shown in the graph). The DA calculated is slightly higher. The sensing range is constricted to n s = 1.419 to 1.422. Thus it is observed that an increase in the thickness of ZnO increases the sensitivity and the value of RI that can be sensed. But the lower range of sensing range got constricted largely. The variation of sensitivity and DA with different thickness of ZnO, 50 nm of aluminium, a single layer of Ti 3 C 2 T x and n s = 1.412 is shown in Fig. 5 b. Up to around 4.5 nm thickness of ZnO, no noticeable effect on the sensitivity enhancement is seen. Thereafter the sensitivity gradually increases and reaches a maximum value of 11,550 nm/RIU at 11 nm thickness of ZnO, and with further increase in ZnO thickness, shows a reduction in sensitivity. The DA decreases progressively but is within the acceptable range. The plot of the normalized transmitted power for different thicknesses of ZnO with 50 nm of aluminium, a single layer of Ti 3 C 2 T x , and n s = 1.412 depicted in Fig. 5 c shows that all the SPR curves are of almost the same dip. But they experience a blue shift with an increase in ZnO thickness. This configuration can lead to designing SPR biosensors and also sensors for gases and liquids of RI greater than 1.38.Thus 50 nm of aluminium, 11 nm of ZnO, and a single layer of Ti 3 C 2 T x will be the optimum values for efficient sensing. Al/ZnO: Pd/Ti 3 C 2 T x This is a three-layer configuration, just like the Al/ZnO/Ti 3 C 2 T x sensing probe described above, where the ZnO sandwich layer is replaced by a nanocomposite of palladium dispersed in ZnO. Here also the optimum thickness of aluminium is taken as 50 nm and of Ti 3 C 2 T x as a single layer. Calculations have been done with different volume fractions (f) and thickness of the ZnO: Pd nanocomposite. The effect of incorporating the ZnO: Pd nanocomposite can be understood with the illustrative graphs in Fig. 6 . Figure 6 a illustrates the variations in sensitivity and detection accuracy of the sensing probe with 10 nm thick ZnO: Pd nanocomposite layer for f = 0.2, 0.4 and 0.6. Fig. 6 Sensitivity and detection accuracy as a function of the refractive index of the analytes. a for the different volume fraction of ZnO: Pd with 50 nm of Al, a single layer of MXene and 10 nm thickness of ZnO: Pd nanocomposite. b for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.2. c for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.4 d for different thickness of ZnO: Pd nanocomposite with 50 nm of Al, a single layer of MXene and f = 0.6 For f = 0.2, no SPR dip is observed for n s values less than or equal to 1.3. For n s = 1.3 to 1.35, very low sensitivity is observed. Appreciable sensitivity is seen for n s values from 1.37 to 1.398, beyond which the λ res shifted to the IR region. Maximum Sensitivity of 7250 nm/RIU is calculated for n s = 1.398. Though the DA is observed to be lesser than that of Al/ZnO/Ti 3 C 2 T x configuration, the values are well within the acceptable range. With f = 0.4, SPR dip is observed only after n s = 1.22. An acceptable range of sensitivity is observed from n s = 1.37 to 1.382. Beyond n s = 1.382 the λ res shifts to the IR region. Similarly, for f = 0.6, the simulation does not yield any SPR dip up to n s = 1.109. From n s = 1.1 to 1.35, the observed sensitivity is very low. The acceptable range of sensitivity is calculated for n s = 1.352 to 1.366, above which the λ res shifts to the IR region. Sensitivity of 4100 nm/RIU is calculated for n s = 1.366. The DA observed is lower than that for f = 0.4. The calculated sensitivity is found to increase with the increase in the value of possible n s . For any value of n s in the possible range, the sensitivity is found to increase with an increase in volume fraction. Also, the lowering of the lower and upper limits of the sensing range can be noted. In all the cases, the λ res experience a redshift with an increase in the value of n s . The SPR curves observed for f = 0.8 is much wide that accurate detection of λ res will be difficult. The analysis can be much clearer with the plots of Fig. 6 b, c, and d, which depicts the variation of sensitivity and DA with n s for the selected thickness values of ZnO: Pd nanocomposite for different volume fractions of 0.2, 0.4 and 0.6 respectively. Figure 6 b illustrates the variations in sensitivity and DA with the nanocomposite thickness of 5 nm, 10 nm, and 15 nm for f = 0.2. With 5 nm thickness of the nanocomposite, effective sensing is possible only for values of n s = 1.38 to 1.406. The maximum sensitivity of 8350 nm/RIU is calculated for n s = 1.406. With 10 nm thickness, the sensing range is from n s = 1.37 to 1.398, with a maximum sensitivity of 7250 nm/RIU. For 15 nm, the effective sensing range gets limited to n s = 1.36 to 1.388 with enhanced sensitivity. Sensitivity of 6200 nm/RIU is calculated for n s = 1.388. The DA is also observed to decrease with an increase in thickness of the nanocomposite. Figure 6 c shows the variations in sensitivity and DA with ZnO: Pd thickness of 5 nm, 10 nm and 15 nm for f = 0.4. With ZnO: Pd thickness of 5 nm, the sensing range extended from n s = 1.373 to 1.396 imparting 6150 nm/RIU sensitivity for n s = 1.396. For 10 nm thickness, the sensing range is observed to be from n s = 1.37 to 1.382 with a sensitivity of 5200 nm/RIU at n s = 1.382. Though SPR curve can be obtained for values of n s from 1.11 onwards, effective sensing can be had only for a very narrow span of n s = 1.348 to 1.354, when the thickness of the nanocomposite is 15 nm and f = 0.4. A sensitivity of 3400 nm/RIU is observed for n s = 1.354. The nanocomposite of 15 nm thickness yields higher sensitivity and lesser DA than that of 10 nm. Figure 6 d depicts the variation of sensitivity and DA with 10 nm, 15 nm of the nanocomposite for f = 0.6. The sensing range from n s = 1.37 to 1.392 for a thickness of 5 nm imparting a sensitivity of 5850 nm/RIU at n s = 1.392. With 10 nm thickness, an acceptable range of sensitivity is observed only for n s = 1.352 to 1.366. With 15 nm of nanocomposite and volume fraction of 0.6, though SPR curves are observed for n s = 1.01 to 1.328, the sensitivity is not up to the acceptable benchmarked value. Figure 7 a depicts the SPR curves corresponding to ZnO: Pd thickness of 5 nm, 10 nm, and 15 nm for a volume fraction of 0.2 and n s = 1.355. Similarly, the SPR curves corresponding to volume fractions of 0.2, 0.4, and 0.6 for ZnO: Pd thickness of 10 nm and n s = 1.355 is illustrated in Fig. 7 b. It is observed that with an increase in volume fraction and thickness of the nanocomposite, the SPR curves experience a decrease in depth and DA as well as redshift in λ res . Thus in all the cases analyzed, with an increase in the thickness of the nanocomposite and also the volume fraction, for any value of RI, the λ res experience redshift, and the sensitivity is found to increase. The nanocomposite of 20 nm thickness does not provide acceptable values of sensitivity (not shown). Fig. 7 Normalised transmitted power as a function of wavelength for the different, a thickness of ZnO: Pd nanocomposite at f = 0.2, with 50 nm of Al, a single layer of MXene and n s = 1.355. b volume fraction of the nanocomposite, with 50 nm of Al, a single layer of MXene, 10 nm of ZnO: Pd nanocomposite and n s = 1.355 Both of the proposed sensor configurations are of the same number of layers (five layers, including the fiber core and the analyte). The consolidation of the performance analysis of both the configurations on different analytes, subject to the benchmark conditions of sensitivity (≥ 3000 nm/RIU), detection accuracy(≥ 0.013 nm −1 ), and spectral range of λ res (visible range) is furnished below in Table 1 . The table provides clear information on the possible sensing range and the valid sensitivity for each configuration. The sensor with ZnO as the sandwich layer finds application for analytes of RI 1.388–1.422. Similarly, the second configuration with ZnO: Pd nanocomposite can be made use of for RI values 1.354 to 1.406. Though sensors with ZnO: Pd offers lesser Sensitivity and DA compared to that with ZnO, they possess a broader sensing range of the analytes. These sensors can be made use of for biosensing different body fluids and also in sensing of VOCs like acetone and ethanol. The optimum thickness of aluminium is confirmed to be 50 nm, and of Ti 3 C 2 T x is a single layer. Therefore, the physical realization of the sensor can be done with proper selection of the configuration and thickness and volume fraction of the sandwich layer depending on the application and the required performance parameter values. But of course, since the calculations and consolidations in this work are based on theoretical modeling, slight changes may happen while dealing with the real cases. Table 1 Sensitivity of the sensor configurations for different thickness and volume fractions of ZnO and ZnO: Pd at different RI of analytes. Each column shows the minimum and maximum values of the acceptable sensitivities and the corresponding values of RI. Each row shows the acceptable sensitivity values of each configuration for different RI values RI ZnO (5 nm) ZnO (10 nm) ZnO (15 nm) ZnO:Pd (5 nm) ZnO:Pd (10 nm) ZnO:Pd (15 nm) f 0.2 0.4 0.6 0.2 0.4 0.6 0.2 0.4 0.6 1.354 – – – – – – – – 3 – 3.4 – 1.364 – – – – – – – – 4.1 – – – 1.376 – – – 3 3.3 3.55 3.65 4.35 – 4.35 – – 1.388 4.41 – – 4.2 4.65 5.1 5.15 – – 6.2 – – 1.396 4.85 – – 5.45 6.16 –7.25 – – – – – – 1.400 5.45 – – 6.35 – – – – – – – – 1.406 – – – 8.35 – – – – – – – – 1.41 10 11 – – – – – – – – – – 1.416 11.25 11.95 – – – – – – – – – – 1.41 – 14 – – – – – – – – – – 1.42 – – 19.4 – – – – – – – – – Conclusion In the present work, two different fiber optic SPR sensor configurations are proposed. Both the configurations utilize Ti 3 C 2 T x as the outermost layer interacting with the analytes. The unique characteristics of Ti 3 C 2 T x offer several advantages to find application in invasive and noninvasive biosensing of body fluids and VOCs. The Al/ZnO/Ti 3 C 2 T x configuration works effectively for RI values of 1.388–1.422 and the Al/ZnO: Pd/Ti 3 C 2 T x configuration with 1.354–1.406. Proper choice of the thickness and volume fractions of the sandwich layer will yield better biosensors. Since the calculations and consolidations in this work are based on theoretical modeling, the chances of slight deviation from the calculated values cannot be neglected while dealing with the physical implementation of the sensor. The proposed biosensor can be an excellent tool for COVID-19 detection as the refractive index of mucus and serum falls in this range, which is to be explored using real samples.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4562184/

Extracellular Vesicles: Role in Inflammatory Responses and Potential Uses in Vaccination in Cancer and Infectious Diseases

Almost all cells and organisms release membrane structures containing proteins, lipids, and nucleic acids called extracellular vesicles (EVs), which have a wide range of functions concerning intercellular communication and signaling events. Recently, the characterization and understanding of their biological role have become a main research area due to their potential role in vaccination, as biomarkers antigens, early diagnostic tools, and therapeutic applications. Here, we will overview the recent advances and studies of Evs shed by tumor cells, bacteria, parasites, and fungi, focusing on their inflammatory role and their potential use in vaccination and diagnostic of cancer and infectious diseases. 1. Introduction Extracellular vesicles (EVs) are particles of 20 nm up to 5 μ m in diameter composed of proteins, nucleic acid, and lipids that are found in body fluids such as plasma, serum, saliva, urine, breast milk, ascites, and cerebrospinal fluids [ 1 ]. These particles are involved in intercellular communication, modulating a wide range of signaling events during innate and acquired immune responses ( Figure 1 and Table 1 ) [ 2 – 4 ]. EVs are secreted during health conditions or upon inflammation during the course of diseases by all mammalian cells types [ 2 , 3 , 5 ]. EVs include different types of particles and may be named or classified depending on the cell type or function. They can be derived from dendritic cells (dexosomes), prostate tissue (prostasomes), bone, cartilage and atherosclerotic plaques (matrix vesicles), neurons (synaptic vesicles), apoptotic blebs or apoptotic bodies (microparticles, exosomes, and apoptotic vesicles), shed vesicles, shedding microvesicles or microparticles (ectosomes or microvesicles), and membrane fragments of virus infected cells, protozoa, fungi, and bacteria outer membrane vesicles [ 1 , 2 , 4 , 6 – 10 ]. The vesicles derived from mammalian cells contain a family of integral membrane proteins that cross four times the lipid bilayer and are called tetraspanins [ 11 ], including the surface markers of lymphocytes and antigen-presenting cells such as CD37, CD9, CD53, CD63, CD81, and CD82. EVs also contain molecules of the major histocompatibility complex (MHC classes I and II) ( http://www.exocarta.org/ ) [ 11 , 12 ]. EVs derived from normal cells cause either suppression or activation of the immune response by modulating the production of inflammatory mediators. For example, T-cells and monocytes secrete vesicles that contain FasL on the surface that modulate apoptosis of the other cells ( Figure 2 ) [ 13 ]. Vesicles isolated from monocytes deliver proinflammatory mediators that activate endothelial cells [ 14 , 15 ]. Tumor cells secrete EVs that are able to downregulate the immune system, allowing the escape from the immune system. Furthermore, these vesicles can control tumor development and growth, by decreasing the expression and release of IL-2 reducing the proliferation of natural killer (NK) cells [ 14 , 15 ]. Therefore, EVs are potential biomarkers and antigens for vaccination, with potential uses for early diagnostic, and therapeutic applications in several diseases. The purpose of this review is to provide an updated overview of the vesicles released by distinct pathogens and mammalian tissues, highlighting their potential use in vaccination and diagnostic of cancer and infectious diseases. 2. Extracellular Vesicles in Cancer EVs derived from tumors may be involved in tumor growth control and in the communication events between tumor and normal cells by delivering oncogenic proteins and growth factors [ 16 , 17 ]. In some cases, EVs suppress tumor growth by exposing dendritic cells MHC classes I or II molecules, peptides, and costimulatory molecules for the immune system. This amplifies the immunological response, preventing tumor growth [ 18 ]. EVs can also stimulate the resistance to chemotherapeutic agents. Moreover, EVs contain proteins and genetic material from the originating tumor cells that can be used as diagnostic biomarkers. In this regard, recent efforts to elucidate different roles and signaling pathways of EVs have been conducted. A pivotal role of EVs during cancer cell migration and invasion has been reported in different cell types. For instance, EVs derived from 786-0 renal tumor cells enhance their migration and invasion properties [ 23 ]. This occurs through induction of type 4-chemokine receptor (CXCR4) and matrix metalloproteinase-9 (MMP-9) expression by EVs. In addition, adhesion and invasion of the gastrointestinal interstitial stroma are enhanced by the oncogenic protein tyrosine kinase (KIT) present in tumor cell EVs [ 46 ]. More importantly, those structures have also been involved in drug resistance. Tamoxifen-resistant breast tumor cells release exosomes that contain microRNAs (miR221/222) and promote drug resistance in naive cells [ 19 ]. Similarly, resistance to docetaxel in breast tumors and prostate cancer, as well as cisplatin in human lung cancer line (A549 cells), was associated with the content of vesicular microRNAs transferred to susceptible cells [ 47 – 49 ]. Moreover, EVs from A549 cells containing TrkB, EGFR, and sortilin receptors (TES complex) were related to angiogenesis induction through endothelial cells [ 24 ]. In hepatocellular carcinoma (HCC), one of the most lethal cancers, the tumor becomes more resistant to TGF β -dependent chemotherapy through long noncoding RNAs (lncRNAs) obtained from EVs [ 50 ]. Therefore, the extracellular communication through EVs is an important mechanism to activate/deactivate certain crucial events in tumor cell biology. Other studies have evidenced the role of microRNAs present in EVs in cancer establishment. For instance, miR-105, detected in EVs from breast tumor, is associated with metastasis formation via destruction of endothelial monolayers. Interestingly, it is possible to detect miR-105 in the blood circulation before the metastasis establishment reinforcing its potential role as a diagnostic biomarker [ 20 ]. Likewise, gastric cancer stromal cells deliver exosomes to gastric tumor cells. Expression of miR-214, miR-221, and miR-222 present in these EVs is related to lymph node metastasis, venous invasion, and tumor development [ 21 ]. In some cases, miR-containing EVs repress proangiogenic events and impair tumor development on a bone cancer model [ 22 ]. The study of biogenesis of stress-induced vesicles also becomes crucial to understand the development of metastasis. For example, the elevated expression of RAB22A gene in breast tumor cells induced by hypoxia, common in advanced tumors, increases the shedding of vesicles, and the Rab protein colocalizes with the sites of budding EVs. Moreover, the knockdown of RAB22A prevents metastasis, supporting the idea that Rab is involved in the generation of EVs [ 25 ]. EVs released from heat-stressed tumors in a mouse model can induce antitumor immunity [ 51 ]. These vesicles showed chemotactic effects on CD4+ and CD8+ T-cells, efficiently activating dendritic cells (DC). Another study showed that EVs derived from breast cancers can alter the tumor microenvironment and promote tumorigenesis of normal cells via induction of autophagy, response to DNA damage repair (DDR), and induction of reactive oxygen species (ROS) in normal breast epithelial cells [ 52 ]. EVs also carry potential cancer biomarker molecules, as reported by several groups. This includes the polyadenylate-binding protein 1 (Pabp1), predominant in EVs from metastatic duodenal tumor cell lines [ 26 ], prostate-specific membrane antigen (PSA) related to prostate cancer progression, angiogenesis, and metastasis [ 27 ], miR-21 and miR-146a in cervical cancer [ 53 ], and finally lncRNAs in skin cancer (secreted into the blood or urine through EVs) [ 54 ]. All the above-mentioned microRNAs are proposed as potential biomarkers for cancer noninvasive diagnosis. It was also shown that EVs from pancreatic tumor cells contain fragments of double-stranded genomic DNA (dsDNA), suggesting that mutations may be identified in this dsDNA as predictors of cancer and streamline therapeutics [ 55 ]. Based on these findings, it is clear that new biomarkers, once optimized, could be used in therapeutic conducts, offering great advantage over other established methods. In cancer therapy, EVs can also be employed as vehicles to deliver drugs. EVs from tumor cells are able to associate better with their recipient cells than liposomes (>10-fold), due to their lipid and protein composition [ 56 ]. In addition, microRNAs can be delivered to tumor cells and interfere with cancer progression and metastasis. In this logic, synthetic miR-143 was introduced into mesenchymal stem cells, and the secreted exosomes containing miR-143 was transferred to osteosarcoma cells to reduce the migration of the latter cell [ 57 ]. Interestingly, a feedback regulatory mechanism for controlling exosome release was suggested, in which exosomes derived from normal human mammary epithelial cells could impair the release of exosomes from breast tumor cells [ 58 ]. These authors suggest that this may be used as a novel therapeutic approach, attenuating carcinogenic effects of tumor exosomes. Another interesting strategy is to use a synthetic structure based on tumor-derived exosomes and staphylococcal enterotoxin B to induce apoptosis in breast tumor cells [ 59 , 60 ]. The vesicles could be used as a diagnostic, because tumor cells release vesicles in biologic fluids like urine, blood, ascites, and pleural fluids. For example, patients with ovarian cancer shed vesicles derived from tumor cells in the circulation. These vesicles are enriched up to 4-fold more in patients with cancer than healthy controls. Therefore, they can be used as biomarkers to identify early cancers in asymptomatic patients that will potentially develop malignancy. In addition, specific miRNAs are found in extracellular vesicles from patients with lung cancers [ 61 ]. DC have been widely used in the research of therapeutic cancer vaccines. For example, DC were primed with interferon-gamma (IFN- γ ) to induce the expression CD40, CD80, CD86, and CD54 in exosomes, endowing a potent CD8+ T-cell-triggering potential in vitro and in vivo [ 28 ]. Yao et al. [ 62 ] compared the antitumor immunities between EG7 tumor cell-derived exosomes [EXO (EG7)] and EXO- (EG7-) targeted dendritic cells [DC (EXO)]. They showed that the latter DC (EXO) was more effective in inducing antitumor immunity, and this was independent from the host DC, emphasizing the role of the host DC in tumor cell-derived exosomes (TEX) vaccines. In contrast, CD8+ T-cell responses could be induced in vivo when mice were immunized with protein-loaded instead of peptide-loaded dexosomes. Recently, protein-loaded dexosomes were used to protect against tumor growth, whereby CD8+ T-cell responses occurred in vivo [ 63 , 64 ]. Purified MHC classes I and II inserted in exosomes and delivered to melanoma were recognized by specific T-cells. This was used to transfer functional MHC/peptide complexes to antigen-presenting cells [ 65 ]. In this way, antitumor response could be elicited as these complexes may stimulate CD8+ and CD4+ T-cell responses in an "acellular" immunotherapy approach. In another study, exosomes from Rab27a overexpressing cells increased significantly CD4+ T-cell proliferation in vitro because these exosomes upregulated MHC class II, CD80, and CD86 molecules in DC. Moreover, exosomes containing a small GTPase protein involved in secretion of exosomes also were capable of retaining tumor growth in vivo [ 29 ]. Plasmid DNA vaccines encoding EV-associated antigens were recently used as vaccines in mice in order to produce ovalbumin containing-EV antigens in vivo , either exposed on the surface of vesicles or incorporated inside membrane-enclosed virus-like particles [ 66 ]. In both cases, these vaccines were able to induce specific T-cell responses and efficiently prevent the growth of ovalbumin-expressing tumors in vivo , showing that immunotherapy based on EVs may be a valuable method to promote tumor control and other diseases. 3. Bacterial Vesicles Bacteria release vesicles sizing from 20 to 250 nm [ 33 , 67 ] are named outer membrane vesicles (OMVs) for Gram-negative and membrane vesicles, or blebs, for Gram-positive bacteria [ 68 , 69 ]. EVs are required for the exchange of genetic information between bacteria such as Bacillus anthracis , Staphylococcus aureus , Mycobacterium ulcerans , Bacillus spp., Escherichia coli , Pseudomonas aeruginosa , and Helicobacter pylori . Additionally, EVs contain toxins and deliver virulence factors to host cells [ 32 , 70 – 80 ]. Bacterial EVs are composed of cytosolic and membrane proteins, lipoproteins, phospholipids, glycolipids, and nucleic acids [ 31 , 32 , 81 – 84 ]. Detailed composition analysis and biogenesis of OMVs from different Gram-negative bacteria are available [ 33 ]. For example, OMVs from Bordetella parapertussis contain surface immunogenic molecules, porin, outer membrane protein OmpQ, and pertactin that were used in a murine model to assess the protection against infection [ 30 ]. In the same way, OMVs of Pseudomonas putida KT2440 have outer membrane proteins such as OprC, OprD, OprE, OprF, OprH, OprG, and OprW which can serve as adjuvants or vaccine [ 85 ]. Vibrio cholerae OMVs contain several proteins that contribute for the virulence and are essential for cell growth and colonization in vivo [ 86 ]. Another interesting aspect of OMVs is their role in delivering endotoxins to host cells as demonstrated for enterogenic and uropathogenic Escherichia coli ((ETEC) and (UPEC)), the causative agents of traveler's diarrhea and human urinary tract infections. Both ETEC and UPEC strains are able to produce many virulence factors including the heat-labile enterotoxin (LT), homologous to cholera toxin, and cytotoxic necrotizing factor type 1 (CNF1). These toxins are released from bacteria in OMVs and delivered to host cells through vesicle internalization [ 74 , 87 ]. In particular, LT also acts as a ligand for vesicle binding, which is internalized via lipid rafts. Once inside the cell, the toxin is trafficked via retrograde transport through the Golgi and the endoplasmic reticulum [ 74 ]. An outstanding role of OMVs in biotechnology is their use as general vehicles to deliver human, heterologous, or viral antigens [ 33 , 88 ]. Neisseria meningitidis serogroup B OMVs showed remarkable adjuvant properties for anti-HIV-1 antigens and induced a production of IFN- γ and IL-4 [ 89 ]. Vesicles isolated from DC infected with Mycobacterium tuberculosis were able to induce a protective host immunity response [ 90 , 91 ]. There are also potential uses of these EVs as cancer vaccines through immune stimulation [ 92 ]. OMVs from different species of Gram-negative bacteria contain lipopolysaccharide (LPS), proteins, and nucleic acids, which are strong agonists in the modulation of inflammatory reactions through the activation of Toll-like receptors (TLRs). These activations require the action of LPS, which is sensed by Toll-like receptor 4 (TLR4) on host cells, and induce an innate immune response to Gram-negative bacteria leading to inflammatory cytokine production [ 93 – 96 ]. In the case of Pseudomonas aeruginosa , OMVs appear to deliver virulence factors to distant locations by fusing with lipid rafts of several host cell membranes [ 97 ]. Proteins present in secreted vesicles released from P. aeruginosa also seem to play important roles in pathogenesis. This is the case of the inhibitory factor of the cystic fibrosis transmembrane conductance regulator, which promotes changes in the epithelium, allowing reduced clearance of P. aeruginosa toxin A that hijacks the host ubiquitin proteolytic system [ 97 ]. Therefore, P. aeruginosa EVs have the potential to protect the immunized host against subsequent infection and for this reason they have been proposed as vaccines candidates against infection. Another interesting example is OMVs isolated from Haemophilus influenzae , which increases the expression of CD69 and CD86 and activating of the humoral response. In addition, they induce TLR9 signaling through bacterial DNA, which causes a significant proliferative response of inflammatory cells [ 98 ]. Vesicles from Gram-negative bacteria are released naturally as blebs of the outer membrane through bulging and "pinching off." Alternatively, vesicles can be prepared from the detergent-treated bacteria either from normal or from bacteria carrying genetic modifications such as the generalized modules for membrane antigens (GMMA) to induce a strong immune response [ 99 ]. All these vesicles are called OMVs, but it is important to note that they have different composition and properties. Naturally shed blebs are almost free of cytoplasmic and inner membrane components and maintain lipophilic proteins, unlike detergent extracted OMVs derived from bacteria. These differences are relevant when considering the use of vesicles for immunization or diagnostic purposes [ 67 ]. Several vaccines are prepared based on OMVs isolated from Gram-negative bacteria [ 100 ]. One example is the case of Neisseria meningitidis -OMVs vaccine, named Bexsero (Novartis) [ 67 ]. These particles activate the immune response and protection against a challenge with bacteria in murine models [ 82 , 101 – 107 ]. There are, however, several cases that vaccination with OMVs requires further developments to improve better antigenicity, manufacturability, and reduction of pyrogenicity, detergent extract, and toxicity via LPS detoxification [ 82 ]. The mechanism of how Gram-negative bacteria-derived OMVs elicit a vaccination effect, for example E. coli used as a model to study the effect of the adaptive immune response decrease against bacteria-induced lethality. However, with high doses these OMVs induced systemic inflammatory, characterized by hypothermia, tachypnea, and leukopenia (sepsis) [ 108 ]. Because of the thick cell wall of Gram-positive bacteria, extracellular vesicle secretion has been less studied in these bacteria. Nevertheless, it has been reported that S. aureus and Bacillus subtilis secrete membrane vesicles to the extracellular milieu. Proteomic analysis revealed that vesicles derived from S. aureus harbor several pathogenic components [ 109 ]. Furthermore, S. aureus extracellular vesicles induce atopic dermatitis-like skin inflammation in mice. These observations provided hints on the possible roles of Gram-positive secreted vesicles. Recently, a study on the immune activating role of Gram-positive bacteria-derived EV has been published [ 110 ]. The Gram-positive Bacillus anthracis , the agent of the Anthrax disease, also shed membrane-derived vesicles. These EVs are formed by a double membrane and have a spherical shape sizing from 50 to 150 nm [ 83 ]. They are enriched by molecular chaperons and molecules of the cell wall involved in the cellular architecture and include the lethal toxin (LeTx) and the antholysin (ALO). BALB/c mice immunized with these EVs were able to produce more protective IgM to the toxin in comparison with the isolated toxin, prompting to further use these preparations to elaborate vaccines. The protection induced by vesicles obtained from Gram-positive bacteria was not as effective when compared to Gram-negative bacteria OMVs indicating that further work might be necessary to improve their potential. In summary, OMVs include multiple virulence factors, overcoming the limitation of a single antigen immunization. Furthermore, OMVs can act as adjuvant and antigen carrier. 4. Parasite Vesicles Cultured protozoan parasites release EVs that contain several molecules that might affect the host ( Figure 1 ). They are composed of membrane fragments and cytosolic components, including proteins, lipids, and nucleic acids that accumulate in the supernatant of the protozoan cultivated in the presence or absence of host cells [ 40 , 41 ]. When injected in animal models or added to in vitro systems, these EVs were found to affect the course of infection and alter the disease progression caused by the parasite, through the modulation of the host innate and acquired immune response. EVs are described in many protozoa such as Leishmania spp. [ 37 , 38 , 90 , 111 , 112 ], Trypanosoma cruzi [ 40 , 41 , 113 – 115 ], Trypanosoma brucei [ 116 ], Plasmodium spp. [ 117 , 118 ], Trichomonas vaginalis [ 119 ], Toxoplasma gondii [ 120 – 122 ], and Eimeria parasites [ 123 ]. Helminthes have also released EVs in Dicrocoelium dendriticum [ 124 ]. Trypanosoma cruzi is a flagellate protozoan that causes Chagas disease. It is acquired by humans either by the insect vector, blood transfusion, or through maternal transmission during new born delivery [ 125 ]. When T. cruzi enters the host, the first line of defense is the innate immune response, which initiates when receptors that recognize microbial products are activated [ 126 ]. This occurs through Toll-like receptors (TLR) signaling and macrophage activation by mucin-like glycoproteins, which corresponds to 60–80% of the parasite surface molecules [ 35 , 127 ], resulting in the increased production of IL-12, IFN- γ , and nitric oxide (NO) [ 128 ]. The production of these proinflammatory cytokines leads to the activation of several kinds of cells such as natural killer (NK) typical of the acute phase of Chagas disease [ 129 ]. Very little is known about how mucin-like glycoproteins and other surface components are presented to the host. Infective parasites obtained from cultured mammalian cells shed large amounts of EVs that are rich in these surface molecules [ 40 ]. EVs isolated from infective T. cruzi forms promote macrophage activation with an increase in parasitemia levels and amastigotes nests in the heart tissue [ 41 ]. These effects are caused by parasite surface glycoproteins present in the vesicles that attenuate the host immune system. T. cruzi EVs are enriched in α -gal containing glycoconjugates, found preferentially in the mucin-like molecules [ 35 ], and several surface glycoproteins, known as members of a trans -sialidase (TS) family that participate in adhesion and invasion of host cells [ 40 , 41 , 114 , 130 ]. Mucins containing α -gal residues elicit high titers of IgG antibodies decreasing parasitemia during the chronic phase [ 131 ]. Therefore, the production and release of EVs might have a key role in the establishment of infection and may be considered a platform to develop preventive or prophylactic vaccines for Chagas disease [ 132 ]. Leishmania genus encloses protozoan species that cause visceral, cutaneous, and mucocutaneous leishmaniasis in humans. The disease is transmitted by sandfly vectors ( Lutzomyia and Phlebotomus ), which inject parasites into the host during the insect blood meal [ 39 ]. In culture, the insect stages of several Leishmania species release EVs containing parasite antigens, such as the surface glycoprotein of 63 kDa (gp63) that has a strong suppressive effect on host macrophages [ 37 , 38 ]. However, a missing step in Leishmania EVs biogenesis is whether those structures also contain the major surface lipophosphoglycan (LPG), a multivirulence factor involved in the interaction with the vertebrate and invertebrate host [ 39 ]. EVs derived from Leishmania donovani are involved in immune response evasion mechanisms, enabling parasite survival in the host [ 37 , 133 , 134 ]. In contrast, EVs derived from macrophages infected with Leishmania amazonensis induce proinflammatory response in vitro by stimulating the production of proinflammatory cytokines TNF- α , IL-12, and IL-1 β [ 112 ]. These host-derived vesicles have been characterized and contain both parasite and host components [ 37 , 38 ], which indicates that a cross talk of signaling events occurs during infection. Indeed, the immunization of mice with dexosomes derived from DC pulsed with Leishmania major antigen was able to provide protection against the parasite [ 90 ]. This finding could help to improve the available canine vaccines, used to stop transmission [ 135 ], and eventually develop a preventive prophylactic human therapy for leishmaniasis. Trichomonas vaginalis , a flagellated protozoan that colonizes human vaginal and urethral epithelia, also secretes vesicles that act at the host-parasite interface. T. vaginalis EVs stimulate the immune response by increasing the production of IL-6 and IL-8 [ 15 ] and promote greater adherence of less adherent strains of the parasite to the epithelium [ 15 ]. T. vaginalis EVs fuse with and deliver their contents to host cells [ 136 ] and are clearly involved in the colonization of the genital host's tract. It is also possible that EVs from this parasite could provide a more suitable environment to other sexually transmitted diseases such as HIV or HPV. There are several studies about EVs of Plasmodium spp., Apicomplexa parasites that cause human and animal malaria focusing mainly in the immunization alternatives. For example, EVs derived from reticulocytes infected with Plasmodium yoelii , a rodent malaria, induce protection to infection in mice [ 137 ]. Plasmodium berghei , another rodent malaria, secretes microparticles in the plasma of infected mice that induce an intense macrophage activation, which results in inflammatory reaction [ 138 ] via TLR4 and MyD88 [ 12 ]. Therefore, these EVs are key components in the modulation and communication between the parasite and the host [ 118 ]. However, one of the main difficulties in working with human Plasmodium , especially Plasmodium vivax , is the availability to have enough amounts of EVs. Toxoplasma gondii is another intracellular Apicomplexa protozoan that causes Toxoplasmosis. The disease is usually transmitted by eating contaminated meat, accidental ingestion of cat feces with oocytes, and congenital contact. It may cause abortion in pregnant women [ 139 ]. The infection is severe in immune-compromised individuals. EVs derived from DC incubated with T. gondii antigens induce an intense immune response, increasing the levels of MHC class II and the specific production of T-cells and cytokines [ 140 ]. Studies of immunization with these DC are promising alternatives in promoting protection against T. gondii [ 141 , 142 ]. Eimeria tenella , Eimeria maxima , and Eimeria acervulina are also coccidian parasite of chickens that also release EVs, which confer protective immune response against the parasite [ 123 , 143 , 144 ]. In summary, EVs isolated from several parasites or from infected cells have major effects on the immune response and are also potential candidates for immunoprevention of parasitic diseases. 5. Fungal Vesicles Fungi have the capacity to cause devastating human diseases, some of them with high mortality rates, in both immunocompetent and immunocompromised individuals [ 145 ]. Pathogenic fungi exhibit a singular genetic flexibility that facilitates rapid adaptation to the host or environment [ 146 ]. However, there are several open questions of how these pathogens colonize and cause morbidity. As other eukaryotic organisms, fungi use membrane trafficking to connect intracellular and extracellular compartments allowing sorting of protein and lipids to their final cellular sites [ 147 ]. For a variety of proteins, the extracellular milieu is the final destination of the cell wall components, digestive enzymes, and, in the pathogenic species, virulence factors [ 148 ]. In fungi, the cell wall represents the final step of secretion, an event that brings additional complexity to the secretory mechanisms used by these cells [ 147 ]. The cell wall is a complex and rigid structure basically composed of chitin, chitosan, β -1,3-glucan, β -1,6-glucan, mixed β -1,3-/ β -1,4-glucan, α -1,3-glucan, melanin, and glycoproteins as major constituents [ 149 ]. EVs are now recognized as important structures for transcell transport of virulence factors that modulate host immune responses [ 147 , 148 , 150 , 151 ], suggesting the importance of these structures in the pathogenesis of many fungal diseases. The production of fungal EVs was initially characterized in the pathogenic yeast Cryptococcus neoformans [ 152 ]. Currently, EVs were identified in several pathogenic fungi such as Histoplasma capsulatum , Paracoccidioides brasiliensis , Sporothrix schenckii , Candida albicans , Candida parapsilosis , Malassezia sympodialis [ 42 , 78 , 152 , 153 ], and nonpathogenic yeast Saccharomyces cerevisiae [ 154 ]. Different proteins, sterols, phospholipids, polysaccharides, and pigments have been characterized in these fungal EVs isolated from culture supernatants [ 42 , 78 , 150 – 158 ]. Many of these molecules have been identified as known virulence factors or inducers of host humoral responses. For example, in C. neoformans the most important virulence factor and immunomodulator, the glucuronoxylomannan (GXM) [ 43 ], was detected in vesicles released during in vitro macrophage infection [ 150 ]. In P. brasiliensis , similar GXM that interacts with α 1,3-glucans was detected in EVs [ 159 ]. GXM acts differently on the host immune response, depending on its specific molecular characteristics [ 44 , 45 ] making it a possible target for antifungal therapy or vaccination [ 45 ]. Another key molecule in fungal infection is glucosylceramide (GlcCer), a glycolipid component of the fungal cell wall [ 160 ], which has been detected in EVs of C. neoformans [ 152 , 158 ], P. brasiliensis [ 155 ], and C. albicans [ 151 ]. Fungal GlcCer is an antigenic glycosphingolipid that elicits antibody responses in experimental infection models [ 161 ] and in patients affected by some mycoses, such as cryptococcosis [ 162 ]. GlcCer is described as a virulence regulator of C. albicans and C. neoformans [ 163 , 164 ]. Furthermore, GlcCer from P. brasiliensis , Aspergillus fumigatus , and S. schenckii inhibited T-cell proliferation in vitro [ 165 ]. The GlcCer from A. fumigatus was able to activate in vitro mouse and human NK cells and to induce airway hyperreactivity in mice [ 166 ]. These findings indicate that fungal GlcCer may influence both humoral and cellular responses and that inhibition or blocking the GlcCer action can be a therapeutic approach [ 160 ]. Other studies have evidenced that vesicles isolated from C. neoformans culture supernatant were able to melanize after incubation with L-DOPA [ 158 ], a substrate for melanization [ 167 ]. Melanin has been identified in several pathogenic fungi [ 168 ]. Although it is immunologically active, little is known about its role in the immune response. Melanin protects fungal cells from phagocytosis by macrophages, a key step in the host defense against these pathogens [ 169 ]. It also reduces proinflammatory cytokines [ 170 ] and decreases their susceptibility to antifungal drugs [ 148 ], mainly to amphotericin B and caspofungin, and is less evident or absent in ketoconazole, fluconazole, or itraconazole [ 171 , 172 ]. Therefore, it seems that melanization is a distinguished feature observed in EVs released during fungal infections and its role should be further explored in the fungal pathogenesis. Many studies indicated that acquired immunity against EVs is observed during fungal infections. Vesicular components reacted with immune serum from patients with cryptococcosis, histoplasmosis, and paracoccidioidomycosis (PCM) [ 42 , 78 , 153 ] or with serum from C. albicans -infected mice [ 151 ]. Particularly, EVs of P. brasiliensis transport components carrying α -galactopyranosyl ( α -gal) epitopes, a highly immunogenic molecule, which were efficiently recognized by anti- α -gal antibodies from patient with PCM [ 42 ]. These data showed that the fungal vesicular products might be important serological markers produced during this disease. The immunomodulatory activity of fungal EVs is still poorly understood. In vitro studies have demonstrated that mammalian macrophages can incorporate fungal EVs, resulting in increased levels of both pro- and anti-inflammatory cytokines [ 150 , 151 ]. Specifically, in C. neoformans , the exposure of macrophages to EVs resulted in their internalization and production of IL-10, TGF- β , and TNF- α , while for C. albicans , the production of IL-10, IL-12, and TGF- β was observed. In both studies, fungal EVs stimulated murine macrophages to produce higher levels of NO [ 150 , 151 ]. This effect probably occurred due to the fungal EVs preparations, which were composed of heterogeneous populations of different size and probably content [ 148 , 150 ]. M. sympodialis releases EVs carrying allergen, which induce high levels of TNF- α and IL-4, suggesting that vesicles have multiple immunoregulatory functions in atopic eczema. Despite this controversy in host immune response, fungal EVs were capable of stimulating a protective response against infection. Recently, Vargas et al. [ 151 ] showed that inoculation of Galleria mellonella , a larvae model, with EVs followed by challenge with C. albicans reduced the number of recovered viable yeasts in comparison to infected larvae control. Moreover, these authors also observed immunomodulation of DC after internalization of EVs from C. albicans . The synthesis of IL-12, IL-10, TGF- β , and TNF- α was also significantly increased in comparison to nonstimulated DC [ 151 ]. Proteomic-based approaches have been used to characterize C. neoformans , P. brasiliensis , H. capsulatum , and C. albicans and S. cerevisiae EVs [ 78 , 150 , 151 , 153 , 156 ]. Interestingly, most of the identified proteins in P. brasiliensis and C. neoformans lacked the characteristic signal peptide required for conventional secretion [ 78 , 156 ], suggesting that fungal vesicles can also be derived from unconventional secretory mechanisms, as observed in mammalian cells [ 173 ]. These proteomic analyses also revealed a large complexity of proteins with diverse biological functions in fungi EVs. Remarkably, we notice the presence of four proteins repeated in all EVs analyzed as follows: glyceraldehyde-3-phosphate dehydrogenase (GADPH), phosphoglycerate kinase, elongation factor 1-alpha, and 6-phosphogluconate dehydrogenase. Thus, it is possible to consider the potential of these molecules as biomarkers of fungal EVs. 6. Concluding Remarks EVs are remarkable structures found in all biological fluids in mammals. The major reported functions of EVs are highlighted in Figure 1 . In normal and tumor cells, they affect the following: antigen presentation, immune suppression, intercellular communication, inflammation, cellular homeostasis, and coagulation. In pathogens, they are considered virulence factors and are involved in the following: cell adhesion and invasion, evasion and modulation of the immune response, and drug resistance. There are many molecules in EVs. The EVs from mammalian cells contain molecules such as MHC classes I and II, mRNA, miRNA, caspase 3, signaling factors, structural proteins, and cytokines. The EVs isolated from tumor cells express, for example, FasL, MHC classes I and II, mRNA, miRNA, FADD, P-glycoprotein, MMPs, PS, and TF. In protozoan, EVs are formed by key membrane components involved in host-parasite interaction. OMV or EVs from bacteria have antigenic material providing gene transference of resistance to antibiotics and adaptation factors. Fungal EVs are structures for transcell transport of virulence factors, immunomodulatory molecules, and serological markers. Therefore, EVs extend the cell-to-cell communication between host and pathogens. By preventing this communication, EVs can be used as targets for vaccination. In addition, the presence of EVs and the characterization of their composition can provide new diagnostic information on several diseases. Furthermore, studies on EVs in the different situations can be useful to understand the intimate mechanisms of pathogenesis. In conclusion, EVs represent a rich and challenging subject for basic and applied research enabling the understanding of a plethora of different mechanisms and opening new tools to combat diseases ( Figure 2 ). Conflict of Interests The authors declare that there is no conflict of interests regarding the publication of this paper.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10162409/

Cholesterol- and actin-centered view of the plasma membrane: updating the Singer–Nicolson fluid mosaic model to commemorate its 50th anniversary †

Two very polarized views exist for understanding the cellular plasma membrane (PM). For some, it is the simple fluid described by the original Singer–Nicolson fluid mosaic model. For others, due to the presence of thousands of molecular species that extensively interact with each other, the PM forms various clusters and domains that are constantly changing and therefore, no simple rules exist that can explain the structure and molecular dynamics of the PM. In this article, we propose that viewing the PM from its two predominant components, cholesterol and actin filaments, provides an excellent and transparent perspective of PM organization, dynamics, and mechanisms for its functions. We focus on the actin-induced membrane compartmentalization and lipid raft domains coexisting in the PM and how they interact with each other to perform PM functions. This view provides an important update of the fluid mosaic model. INTRODUCTION When we consider plasma membrane (PM) structure, dynamics, and functions, we are often struck by the feeling that we are engaged in military intelligence, facing the so-called VUCA problems: volatility, uncertainty, complexity, and ambiguity ( Bennis and Nanus, 1986 ). The PM is basically in a fluid liquid state ( Singer and Nicolson, 1972 ) and probably contains more than 10,000 protein and lipid species ( Holthuis and Menon, 2014 ). They interact with each other in complex ways ( Simons, 2016 ), often forming a variety of nano- to micronscale molecular complexes and domains with very broad ranges of lifetimes, from milliseconds to perhaps weeks (see Figure 5 of Kusumi et al. , 2012b ) ( Simons and Sampaio, 2011 ), and sometimes with quite ambiguous boundaries with other domains/complexes, as exemplified by the presence of so-called lipid raft domains. The thermodynamic state of the PM has been proposed to be close to miscibility critical points exhibiting immense local compositional fluctuations, which might be essential for turning on cooperative switches to trigger important signaling processes and PM structural changes ( Veatch and Keller, 2003 ; Honerkamp-Smith et al. , 2008 ). This complicates the analyses of the system, because subtle variations in experimental conditions employed by different research groups could lead to quite dissimilar results. Despite the VUCA of the PM as a system, in terms of its molecular composition two molecules stand out among all molecular species because of their abundance and essential roles in PM functions: cholesterol (CHOL) and actin. CHOL is the most common PM molecular species, representing ∼35–45 mol% of all the PM lipids ( Subczynski et al. , 1991 ; Meer et al. , 2008 ; Simons and Sampaio, 2011 ). The actin "membrane skeleton," which is the actin meshwork closely apposed to the PM cytoplasmic surface, and the cortical actin filament layers, with a thickness of ∼25 nm (three to five layers of actin filaments) from the PM, are found virtually everywhere throughout the PM in eukaryotic cell lines ( Morone et al. , 2006 ; Shirai et al. , 2017 ). In this review, we advance the arguments that the CHOL- and actin-centered view of the PM can provide excellent perspectives for understanding a variety of key PM properties and functions, including those of "lipid rafts." This year is the 50th anniversary of the publication of the fluid mosaic model by Singer and Nicolson (1972 ), which is undoubtedly still the most fundamental model for the PM structure and molecular dynamics. Even now, the fluid mosaic model is believed to represent the basic structure of the PMs of virtually all cells existing on Earth. Such universality is comparable to that of the double helical structure of DNA, although this point is often missed even in classrooms and textbooks. Just as many DNA functions are based on the DNA's double helical structure, the various PM functions are enabled by the PM's fluid mosaic structure. However, to develop clearer synthetic understandings of how the PM performs its various functions, perhaps based on the PM VUCA, scientists have been modifying the fluid mosaic model to higher and finer levels by incorporating spatiotemporal heterogeneity ( Manley et al. , 2008 ; Schuberth and Wedlich-Söldner, 2015 ; Sezgin et al. , 2017 ; Jacobson et al. , 2019 ; Kalappurakkal et al. , 2020 ). For this purpose, in this article we propose that one of the useful ways to reach a synthetic understanding of the PM structure, dynamics, and functions is to regard the PM VUCA from the perspectives of CHOL and actin. We will explain that two major causes of PM spatiotemporal heterogeneity are brought about by CHOL and the parts of the cortical actin filaments directly bound to the PM (actin-based membrane skeleton). CHOL induces "raft domains" (defined later in this review), and the actin-based membrane skeleton induces the compartmentalization of the entire PM into 30–230-nm domains. Each raft domain basically exists within an actin-induced compartment. CHOL-CENTERED VIEW OF THE PM Strikingly nonlinear dependence of many signaling functions on the CHOL concentration in the PM Mild CHOL depletion is a prevalent method to study the involvement of CHOL and so-called "lipid rafts" in biological processes of interest. However, the following key point is rarely considered. The CHOL concentration in the PM of mammalian cells is generally around 35–45 mol% of total lipids. In normal CHOL-depletion experiments, the CHOL concentration is typically decreased by only 40% to 20–27 mol% ( Green et al. , 1999 ; Surviladze et al. , 2001 ; Suzuki et al. , 2007b ). For conciseness, in the following we state 40 and 25% to refer to the CHOL mole fractions in the PM before and after CHOL depletion, respectively. However, despite the presence of high concentrations (25 mol%) of CHOL remaining in the "CHOL-depleted" PM, many signaling reactions were shut down (we mostly consider events occurring at 37°C), as exemplified in the following reports: Green et al. , 1999 ; Sheets et al. , 1999 ; Abrami et al. , 2003 ; Seveau et al. , 2004 ; Monastyrskaya et al. , 2005 ; Vial and Evans, 2005 ; Hunter and Nixon, 2006 ; Suzuki et al. , 2007a , b ; McGraw et al. , 2012 ; Korinek et al. , 2015 ; Ridone et al. , 2020 ; Wang et al. , 2021 . Therefore, the dependence of these signaling processes on the CHOL content is strikingly nonlinear: full reaction at 40 mol% CHOL and no reaction at 25 mol% CHOL. This clearly shows that these signaling reactions do not depend on simple interactions with single CHOL molecules but instead are enabled by the domains/structures in the PM that are formed by cooperative interactions with CHOL in the presence of overall CHOL concentrations greater than 25 mol%. Previous and currently prevalent "definitions" of raft domains The special domains induced by CHOL are enriched in sphingomyelins. Accordingly, the idea that CHOL and sphingomyelins together form special lipid-based domains that stand out among the PM VUCA was born, and these domains were named "lipid rafts" about a quarter of century ago ( Simons and Meer, 1988 ; Simons and Ikonen, 1997 ; Simons and Toomre, 2000 ; Harder and Simons, 1997 ; Rietveld and Simons, 1998 ). Although many researchers erroneously assumed that generally mild CHOL-depletion protocols removed most of the CHOL from the PM (rather than just ≈40%),due to the experimental ease CHOL depletion became a very useful and popular method to examine the involvement of raft domains in various cellular events and molecular functions ( Moran and Miceli, 1998 ; Fessler and Parks, 2011 ). Furthermore, this protocol was often coupled with biochemical purification methods to prepare lipid raft domains as aggregates of some fractions of lipid raft domains and raft-associated molecules, after extraction from various cells and tissues using nonionic detergents at 4°C (called detergent-resistant membranes or DRMs) ( Brown and Rose, 1992 ). However, the interesting observations obtained by this approach failed to provide a clear definition of raft domains in the PM. The lipid raft domain research field has long been plagued by the lack of a clear definition. If the raft domains had been visualized by microscopic methods, then the definition would have been easier. For example, the micronscale structures on the PM with lifetimes longer than several tens of seconds, such as synapses, adherens junctions, and clathrin-coated pits, and containing many (>10) copies of each key protein species could be defined simply, because they can be imaged by immunofluorescence and immunoelectron microscopy. The raft domains that could be visualized are limited to those that were enlarged and stabilized by external stimulation (often by non–physiologically strong stimulation) and by artificial cross-linking of raft-associated molecules. Therefore, the focus of interest in raft research has shifted to revealing whether the precursor raft domains that are triggered to become large, stabilized raft domains exist in quiescent cells before stimulation. If they do exist, what are the precursor domains, how do they behave, and how are they triggered to become enlarged, stabilized raft domains to perform cellular signaling? For such investigations, microscopic visualizations of precursor domains in the quiescent steady state must be made, but these observations have turned out to be enormously difficult. At the 2006 Keystone Symposium on "Lipid Rafts and Cell Function," lipid rafts were "defined" as "small (10–200 nm), heterogeneous, highly dynamic, sterol- and sphingolipid-enriched domains that compartmentalize cellular processes. Small rafts can sometimes be stabilized to form larger platforms through protein–protein interactions." However, although this "definition" summarizes the general properties of raft domains in the PM quite well, it might not be very useful for research examining whether the observed molecules and events occur in raft domains and whether any raft domains are involved in the observed phenomena. As such, the older definitions of raft domains were often mixtures of definitions and properties, and usually difficult to apply to actual research. Furthermore, the definition of raft domains often depended on the context in which they are considered. For example, in the biochemical context, the raft domains have long been defined as the DRM fraction. This definition has been useful for identifying the molecular candidates that might be involved in raft domains in the PM. However, the largest problem is that DRMs could not be readily related to the actual raft domains in the PM, particularly those before stimulation (for example, in terms of their sizes, lifetimes, and distributions). Therefore, a general definition of raft domains existing in the PM must be developed for the robust development of this research field. It must be useful for research in the future and in broad fields of biomedicine. In addition, the definition must include the cooperative interaction of CHOL with other molecules for the formation of raft domains, as explained in the preceding subsection. Another key observation for defining raft domains in the PM As described, many signaling reactions that depend on CHOL do not rely on simple interactions with single CHOL molecules, but instead are enabled by the CHOL-enriched domains/structures in the PM that are formed by cooperative interactions of CHOL with other constituent molecules. Such domains could be generated in the presence of overall CHOL concentrations greater than 25 mol% in the PM. Another critically important observation was made using giant plasma membrane vesicles (GPMVs) or PM spheres, formed by PM blebbing after the dissociation of the actin-based membrane skeleton from the PM cytoplasmic surface. Upon cooling to ∼10°C, GPMVs generally have two coexisting and complementary micronscale liquid domains ( Baumgart et al. , 2007 ; Lingwood et al. , 2008 ; Levental et al. , 2009 , 2010 ; Johnson et al. , 2010 ). These domains exhibit the characteristics of the liquid-ordered and liquid-disordered phases (Lo- and Ld-phases, respectively) found in artificial giant unilamellar vesicles (GUVs) with the proper lipid compositions, typically a ternary 1:1:1 mixture of long, saturated phosphatidylcholine (PC) or sphingomyelin, CHOL, and unsaturated PC, at temperatures below 37°C (for the Lo- and Ld-phases, see the next subsection, Detailed explanation 1 ; Figure 1, A–C ). FIGURE 1: Key molecular interactions for the formation of various raft domains. (A) All trans conformations and two gauche conformations ([+] gauche and [–] gauche ) of the alkyl chains (–CH 2 –CH 2 –) n . The gauche conformation is induced by the rotation around a single bond between two adjacent CH 2 groups. The gauche conformation induces 60° bend in the acyl chain. Because the inclusion of a single gauche conformation in a chain is quite unfavorable for the acyl chain packing in the membrane (and thus energy), the second gauche conformation occurs frequently near the first gauche conformation. Such double gauche conformations are called kinks. The formation of gauche conformations (rotation around single bonds) and the rapid interconversions between the trans and gauche conformations are the fundamental causes for inducing fluidity (liquid-like property) of the PM. (B) (a) Chemical structures of CHOL and representative saturated and unsaturated phospholipids frequently used in the studies of GUVs. Saturated lipid: l -α-distearoylphosphatidylcholine (DSPC). Unsaturated lipid: L-α-dioleoylphosphatidylcholine (DOPC). When saturated lipids, such as DSPC, are located next to a CHOL molecule, due to the rigid tetracyclic skeletal structure of CHOL, the gauche conformations of the acyl chain between the carbonyl carbon C1 (-COO) and the 12th carbon C12 are suppressed, inducing more trans conformations and increasing the acyl chain packing. However, unsaturated lipids, such as DOPC, have at least a single cis bond between C9 and C10, which induces a mandatory 60° bend there. The mandatory 60° bend of the unsaturated acyl chain causes steric problems when it is next to a CHOL molecule. (b) (i) Schematic figure showing three possible configurations for placing CHOL and an unsaturated chain next to each other, indicating that good packing of these two structures is difficult. This is due to the mismatch (nonconformability) between the mandatory 60° bend at the double bond in the unsaturated acyl chain and the rigid tetracyclic skeletal structure of CHOL (magenta rectangles). Therefore, CHOL and unsaturated acyl chains tend to exclude each other, which is called lateral nonconformability, and CHOL tends to associate with saturated lipids if given the choice. Such accommodating and nonconformable interactions of CHOL with saturated and unsaturated lipids, respectively, are the main cause of phase separation in the ternary mixtures of CHOL, saturated lipids, and unsaturated lipids. (ii) The rigid ring structure of CHOL, located adjacent to the rough surface of the TM domain of the TM protein (induced by the protruding hydrophobic side chains of the amino acids in the membrane), would generate vacant pockets (packing defects). In this manner, CHOL is excluded from the first annulus of most TM proteins. Therefore, the CHOL-based raft domains are excluded from the PM compartment boundaries due to their exclusion from the rows of the TM picket proteins anchored to and aligned along the actin-based membrane skeleton fence. The TM proteins with raft affinities are expected to be well miscible with CHOL molecules. (C) Schematic snapshot drawings of the phase-separated GUV membrane, made of saturated lipids, such as DSPC, and unsaturated lipids, such as DOPC, and CHOL (top view). Open circles represent acyl chains (saturated chain in green and unsaturated chain in cyan; two chains form a lipid molecule), and solid magenta structures indicate CHOL molecules. The regions enriched in saturated lipids and CHOL represent the Lo-phase domains, and those enriched in unsaturated lipids represent the Ld-phase domains. Note that Lo-phase domains additionally contain smaller amounts of unsaturated lipids and Ld-phase domains additionally contain smaller amounts of saturated lipids and CHOL. Fluorescence microscopy observations might suggest the presence of submicronscale–micronscale Lo-phase domains, but they might actually represent the membrane regions enriched in nanoscale (several to several tens of nanometers) Lo-phase domains. (D) (a, left) Schematic model of a GPI-AP. The protein moiety is linked to phosphatidylinositol (PI), a phospholipid, via a short glycochain. The glycosylated PI (GPI) anchors the protein moiety to the PM outer leaflet. The monomeric GPI-AP can associate with raft domains, but its affinity is low (left). (a, right) Schematic model of the metastable GPI-AP homodimer rafts, which are formed by specific homoprotein interactions (P-P) and stabilized by raft-lipid interactions (which prolong the lifetime of homodimers). All GPI-APs examined thus far form metastable homodimer rafts with lifetimes on the order of ∼200 ms. (b) GPI-AP homodimer rafts merge to form GPI-AP tetramer rafts by raft-lipid interactions (green ripples), rather than protein–protein interactions. The tetramer raft lifetimes are shorter (∼100 ms) than the homodimer raft lifetimes. When GPI-AP homodimer rafts of the same GPI-AP species merge, they form homotetramer rafts, and when GPI-AP homodimer rafts of different GPI-AP species merge, they form heterotetramer rafts. Such flexibility of protein species in tetramer rafts is possible because the merging is mediated by raft-lipid interactions, rather than protein–protein interactions. These results suggest that the formation of metastable GPI-AP homodimer rafts represents the first step in the formation of raft domains and that the GPI-AP homodimer rafts are one of the basic units for generating greater raft domains. The formation of GPI-AP tetramer rafts would be the second step for the formation of greater raft domains. The next step required for the formation of greater raft domains is the stabilization of greater raft domains, which would be triggered by signaling, including the pathways of Src-family kinases and trimeric G proteins, and probably by interactions with active actomyosin systems. Importantly, these results showed that the PM is capable of undergoing large-scale phase separations into Lo-phase-like and Ld-phase-like domains, when the actin-based membrane skeleton is removed and the temperature is lowered to ∼10°C. Interestingly, after mild partial CHOL depletion, the Lo-phase-like micronscale domains could not be induced upon cooling to 10°C ( Johnson et al. , 2010 ), suggesting that the presence of overall CHOL concentrations of 25 mol% or more in the PM is essential for producing Lo-phase-like domains in GPMVs at ∼10°C. This requirement is essentially the same as that for many PM signaling functions, as described in preceding paragraphs, as well as for the induction of Lo–Ld-phase separation in GUVs (see Section 4 of Kusumi et al. , 2020 ). Detailed explanation 1: Lo- and Ld-phases in GUVs and biological membranes Readers familiar with Lo- and Ld-phases should skip this subsection. They can do so without losing track of the main flow of this article. The liquid properties of biological membranes are induced by the thermally driven conformational changes of individual single bonds in the fatty acyl chains of phospholipids and sphingolipids (–CH 2 –CH 2 –), from the more stable trans conformation to either one of the two gauche conformations ([+] gauche and [–] gauche ) ( Figure 1A ). The gauche conformations induce a 60° bend in the acyl chain, which prevents the lipid molecules from being packed closely in the membrane. The acyl chain order, which is often defined by the spread (distribution) of the angles of the C–C bonds in each –CH 2 –CH 2 – group relative to its average orientation (averaged over all the molecules and also over the characteristic time span for the observation method used), is generally coupled with how well the acyl chains are packed in the membrane: higher packing generally means higher acyl chain order (fewer gauche conformations). The double bond in a fatty acyl chain works like a fixed gauche bond; that is, a mandatory 60° bend in the acyl chain (although their exact conformations are different) ( Figure 1B ). Therefore, the membranes containing lipids with one or two unsaturated fatty acyl chains (unsaturated lipids) are less packed, blocking the solidification of the membrane, which is important for organisms living in cold environments. In artificial membranes, like GUVs, in the absence of CHOL the acyl chain packing and order are low at higher temperatures and thus constituent lipid molecules can undergo rapid translational diffusion. Thermodynamically, such membranes are in the liquid-disordered phase (Ld-phase), which was so named because lipid molecules undergo Brownian diffusion in the membrane (a characteristic of the liquid) and the acyl chains are disordered. The effects of CHOL on the membrane structure and dynamics in GUVs can be understood in terms of three types of CHOL–acyl chain interactions. When an acyl chain is located adjacent to a CHOL molecule, due to the rigid tetracyclic skeletal structure of CHOL, the gauche conformation is suppressed, thus inducing more trans conformations and increasing the acyl chain packing ( Figure 1B ). Therefore, the presence of CHOL generally enhances the acyl chain order in membranes in the Ld-phase. Upon lowering the temperature, because the presence of CHOL induces defects in the alignment of acyl chains, the packing in the solid phase is lower. Namely, CHOL enhances alkyl chain disorder in membranes in the solid phase (where acyl chains are quite well aligned and no translational diffusion of lipids occurs). CHOL and unsaturated acyl chains tend to exclude each other because the forced 60° bend at the double bond prevents the acyl chain from following along the rigid tetracyclic skeletal structure of CHOL (called "lateral non-conformability"; Kusumi et al. , 2005 ) ( Figure 1B ). At sufficiently high temperatures, GUVs composed of ternary mixtures of >25 mol% CHOL and saturated and unsaturated lipids form the Ld-phase membrane. Upon cooling, unsaturated lipids are excluded from the CHOL's rigid tetracyclic backbone surface (Interaction 3), where saturated lipids must come in and interact. Under certain conditions, such CHOL effects induce phase-separated membranes consisting of the Ld-phase domain containing a high content of unsaturated lipids and lower amounts of saturated lipids with smaller amounts of CHOL and the other phase domain mostly consisting of CHOL and saturated lipids with small amounts of unsaturated lipids ( Figure 1C ). The latter membrane domain exhibits higher acyl chain order compared with the Ld-phase domain, due to the interaction of CHOL with saturated fatty acyl chains (Interaction 1), but the packing is less dense than that in the solid phase due to the lower alignment of saturated lipids next to CHOL (Interaction 2), allowing lateral diffusion of the constituent molecules (liquid property). Accordingly, the latter domain tends to be in the liquid-ordered phase (Lo-phase domain). The thermodynamic phase can be strictly defined for GUVs comprising a few defined lipid species. However, for membranes composed of many molecular species, like biological membranes, because the phase cannot be defined readily we call the disordered and ordered liquid states "Ld-phase-like" and "Lo-phase-like" or simply "Ld-like" and "Lo-like" states, respectively. Some authors call these states of the biological membranes the Ld- and Lo-phases, but readers should know that this would be a loose terminology. In artificial membranes consisting of only CHOL and unsaturated lipids, CHOL molecules form transient clusters (oligomers), probably due to exclusion from the space around unsaturated acyl chains. Such CHOL clusters form and disperse continually with a lifetime of the order of 0.1–1 ns ( Subczynski et al. , 1990 ). The occurrence of Interaction 1 increases the number of trans configurations in the acyl chains and enhances the lipid packing, which would make the Lo-phase domains thicker than the Ld-phase domains. Cryoelectron microscopy/tomography directly visualized the membrane thickness in the Lo-phase domain to be greater than that in the Ld-phase domain in GUVs ( Cornell et al. , 2020 ; Heberle et al. , 2020 ). The thickness distribution of GPMVs suggested the nanoscale lateral heterogeneities ( Heberle et al. , 2020 ). Revised definition of raft domains On the basis of the argument developed in the preceding section (Lo-phase-like and Ld-phase-like domains coexist in the PM of living cells), as the definition of raft domains in the PM, we propose the following: "Raft domains in the PM are Lo-phase-like, CHOL-centered molecular complexes/domains formed by cooperative interactions of CHOL with saturated acyl chains as well as unsaturated acyl chains and transmembrane (TM) proteins ." This definition represents the fundamental mechanism for the formation of raft domains in the PM: the cooperative interactions due to saturated acyl chains' weak multiple accommodating interactions with CHOL and, importantly, the low miscibility of CHOL with unsaturated acyl chains and TM proteins (collectively called "raft-lipid interactions"), which occur in the presence of >25 mol% CHOL. The latter half of this definition is quite different from previous definitions in that it includes CHOL exclusion from the sites adjacent to unsaturated lipids and TM domains. For details about the interactions of CHOL with saturated and unsaturated acyl chains and TM proteins, see the preceding subsection, Detailed explanation 1 , and Figure 1, A–C . In developing this new definition, we wanted to make it very general but also very strong. Here, the critical molecular species to form raft domains are only CHOL and molecules containing saturated and unsaturated acyl chains. The definition is based on CHOL's cooperative interactions with saturated acyl chains (accommodating) as well as with unsaturated acyl chains and proteins' TM domains (nonconformable). These accommodating and nonconformable interactions together induce cooperative interactions. Previously, the "lateral nonconformability" between CHOL and unsaturated acyl chains was often neglected, making the meaning of the cooperativity unclear. This concept is close to the mechanism for the formation of the Lo-phase in the artificial ternary mixtures of CHOL, lipids with long saturated acyl chains, and lipids with unsaturated acyl chains, but in the definition of raft domains the Lo-phase formation is unnecessary. The key is not the formation of the phase, but the cooperative interactions. The cooperativity might start with only a few molecules that are included in raft domains (such as homodimer rafts of glycosylphosphatidylinositol-anchored proteins [GPI-APs] and CHOL; Suzuki et al. , 2012 ) and the surrounding unsaturated chains (although the Lo-phase might be defined for a molecular assembly of ≈10 molecules surrounded by unsaturated lipids; Kusumi et al. , 2020 ). Because this definition is based on the fundamental mechanism for the formation of raft domains, it should be applicable and useful for testing molecular complexes in the PM found in diverse biomedical fields. The application of this definition to experimental research is described in the next subsection, Practical guide for studying raft domains in the PM based on the new raft definition . It is critically important to realize that the interactions of individual molecules in and near the CHOL-centered raft domains are transient and weak. The sum of many weak accommodating and nonconformable interactions is the key for raft formation and for explaining the physical properties of raft domains. Therefore, all of the involved molecules in and around the raft domains should dynamically undergo thermal diffusion, readily moving in and out of raft domains ( Kenworthy et al. , 2004 ; Suzuki et al. , 2007a , b , 2012 ). This is consistent with the concept that the thermodynamic state of the PM is close to miscibility critical points exhibiting immense local compositional fluctuations ( Veatch and Keller, 2003 ; Honerkamp-Smith et al. , 2008 ). Although the space scales of the fluctuation were considered to be quite long, ≈20 nm ( Veatch and Keller, 2003 ; Honerkamp-Smith et al. , 2008 ), the experimental data about GPI-AP homodimer rafts suggest that the sizes of raft domains can be as small as a few nanometers ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ). A cautionary remark is necessary here. Lo-phase-like domains that do not involve CHOL might exist in the PM. Fluorescence lifetime measurements, using environmentally sensitive membrane dyes that report the degree of lipid packing, suggested that ∼76% of the liquid is in the "ordered" state, although the exact meaning of the ordered state must be further clarified ( Owen et al. , 2012 ). Practical guide for studying raft domains in the PM based on the new raft definition This definition of raft domains would be convenient to use for future research in a variety of biomedical fields. On the basis of the definition, we propose the following three criteria for determining whether the molecule of interest preferentially partitions into and functions in raft domains in the PM. Preferential partitioning into Lo-like domains in GPMVs at 10°C; Preferential partitioning into detergent-resistant membranes using Triton X-100 at 4°C; Greatly reduced/altered function after mild CHOL depletion, and recovery of the original function soon after CHOL replenishment. Basically, all three criteria should be satisfied for a molecule to be categorized as a raft-associated molecule. This is because each of these criteria is not based on the all-or-none law, but the preference addresses a quantitative difference and thus higher probability. Furthermore, these criteria are based on molecular behaviors in very different contexts/conditions. Therefore, we think that satisfying all three of these criteria is important. Previously, we proposed another criterion, "Preferential partitioning into the Lo-phase domains in Lo–Ld-phase-separated GUVs after reconstitution at 25°C," but we replaced it with the third criterion described here because the use of GUVs requires extensive experiments and might not be practical in various research projects (for details, see Sections 10 and 13 of Kusumi et al. , 2020 ). We believe that this third criterion effectively represents the definition and would be more practical for applications to molecular cell biological studies. Note that CHOL repletion experiments must be performed to confirm that the effect of CHOL depletion was directly induced by the loss of raft domains, rather than artifactual side effects of CHOL depletion ( Kenworthy et al. , 2004 ; Suzuki et al. , 2007a , b , 2012 ), although CHOL repletion experiments are rarely performed. For example, because acute CHOL depletion could induce actin reorganization and reduce the mobilities of PM molecules, its effect can be strong and broad in various assays ( Kwik et al. , 2003 ). However, because this effect does not disappear for 12–24 h after CHOL repletion ( Kwik et al. , 2003 ), if CHOL repletion quickly restores the molecular functions and behaviors before CHOL depletion, then the CHO-depletion effect is likely induced by the loss of raft domains. To assess the involvement of raft domains in biological events/phenomena of interest in the live-cell PM, we recommend examinations of the recruitment/involvement of GPI-APs ( Suzuki et al ., 2012 ) and the recently developed fluorescent analogues of sphingomyelin and gangliosides, GM1, GM2, GM3, and GD1b ( Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ); for an extensive summary, see Table 4 of Kusumi et al. (2020) . Note that virtually all fluorescent ganglioside and sphingomyelin analogues used before these publications preferentially partitioned into Ld domains ( Sezgin et al. , 2012a ) and thus could not and will not serve as raft markers (these studies tended to erroneously arrive at wrong conclusions). Protein toxin–based CHOL and sphingomyelin probes could also be used, but due caution is required because their affinities tend to be low and affected by their environment (readers interested in the protein probes for CHOL and sphingomyelins, see the next subsection, Detailed explanation 2 ). Because the sizes of PM raft domains in quiescent cells might be on the order of 2–20 nm ( Kusumi et al. , 2020 ), the recruitment and colocalizations would be better observed by employing single-molecule imaging methods ( Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ). When single-molecule localization microscopy, such as PALM/STORM, is utilized, note that simply gathering images would not be very useful. Instead, the spatial correlations and colocalizations of individual molecules of interest (fluorescent spots) with individual molecules of the raftophilic lipid probes (fluorescent spots) must be obtained ( Stone and Veatch, 2014 ; Pageon et al. , 2016 ; Simoncelli et al. , 2020 ); for the method to circumvent the overcounting problem, see Arnold et al. (2020 ). However, because lipid probes cannot be chemically fixed ( Tanaka et al. , 2010 ), observations in living cells, using fast data acquisition frame rates, would be preferable ( Fujiwara et al. , 2021b ). For further understanding of raft domains, excellent recent reviews, which have not been cited thus far in this article, are available. For example, refer to Sevcsik and Schütz (2016) , Levental et al. (2020) , and Regen (2020) . Recently, using protein toxin–based CHOL probes, the existence of three distinct CHOL pools in the PM outer leaflet has been found ( Das et al. , 2014 ; Endapally et al. , 2019 ). Their relationships with raft domains are not clear, but the 2–20-nm-scale raft domains could be classified into those that do or do not contain sphingomyelin. These three CHOL pools are summarized in Detailed explanation 3 . Detailed explanation 2: Protein-based CHOL probes Readers who are not interested in this subject matter can skip this subsection without losing track of the main flow of this article. CHOL probes based on bacterial and mushroom toxins (proteins) have recently been developed. The most popular probes employ the carboxy-terminal ∼13-kDa domain, referred to as domain 4 (D4), derived from perfringolysin O (PFO) ( Ohno-Iwashita et al. , 1990 ; Nakamura et al. , 1995 ; Shimada et al. , 2002 ) or anthrolysin O (ALO) ( Bourdeau et al. , 2009 ; Farrand et al. , 2010 ). D4H, the D434S mutant, exhibited higher affinity to CHOL ( Johnson et al. , 2012 ; Maekawa and Fairn, 2015 ). Liu et al. (2017 ) developed a ratiometric fluorescence imaging method that allows simultaneous in situ CHOL quantifications in both PM leaflets, using four orthogonal CHOL sensors, which facilitate the selection of probes with proper affinities for each experiment ( Cho et al. , 2022 ). Although these probes turned out to be extremely useful ( Zhang et al. , 2018 ; Cho et al. , 2022 ), the results obtained with them should be interpreted carefully because their binding might be sensitive to particular lipid-lipid interactions, phase states, and masking by the binding of CHOL's physiological interaction partner proteins ( Maxfield and Wüstner, 2012 ; Courtney et al. , 2018 ). Probes based on mushroom toxins, ostreolysin A (OlyA) and nakanori, which bind only to the sphingomyelin-CHOL complex, have also been developed ( Ota et al. , 2013 ; Das et al. , 2014 ; Skocˇaj et al. , 2014 ; Makino et al. , 2017 ; Endapally et al. , 2019 ). Structural and biochemical analyses showed that the conformation of CHOL-bound sphingomyelin is different from that of free sphingomyelin ( Endapally et al. , 2019 ). The E69A and E69S mutants of OlyA can bind to both free sphingomyelin and sphingomyelin complexed with CHOL and thus could be used as sphingomyelin probes (another powerful sphingomyelin probe is lysenin [ Yamaji et al. , 1998 ; Tomishige et al. , 2021 ]). Interestingly, the E69G mutant exhibited specificity to the sphingomyelin-CHOL complex, whereas the E69N and E69D mutants failed to bind either form of sphingomyelin. The D4 domain of PFO (and Y181A, called PFO*, used at 4°C; at 37°C, like PFO, PFO* forms pores in the PM) is very useful. It binds to free CHOL, but not the sphingomyelin-CHOL complex. Together with OlyA, which binds only to the sphingomyelin-CHOL complex, three distinct CHOL pools in the PM outer leaflet were identified ( Das et al. , 2014 ; Endapally et al. , 2019 ). They are the D4-accessible (∼16 mol% CHOL vs. total lipid), sphingomyelin-sequestered (∼15 mol%), and essential (∼12 mol%) CHOL pools (see Detailed explanation 3 , the next subsection). The essential CHOL pool is thus termed because its depletion causes cells to round up and dissociate from the substrate. This pool is protected from D4 and OlyA binding and could be removed only by a treatment with 2-hydroxypropyl-β-cyclodextrin (and methyl-β-cyclodextrin) after the D4-accessible and sphingomyelin-sequestered pools were first depleted ( Das et al. , 2014 ). Detailed explanation 3: Three distinct CHOL pools exist in the PM outer leaflet Readers who are not interested in this subject matter should skip this subsection and directly move to the next subsection. Three distinct CHOL pools in the PM outer leaflet were identified ( Das et al. , 2014 ; Endapally et al. , 2019 ) using toxins that bind to CHOL in different states. The carboxy-terminal ∼13-kDa domain of the bacterial toxin PFO, referred to as domain 4 (D4), binds only to free CHOL, and the mushroom toxin OlyA binds only to the sphingomyelin-CHOL complex (see Detailed explanation 2 for the protein-based CHOL probes). The three distinct CHOL pools are called the D4-accessible (∼16 mol% CHOL vs. total lipid), sphingomyelin-sequestered (∼15 mol%; OlyA binding), and essential (∼12 mol%) CHOL pools. The essential CHOL pool is thus termed because its depletion causes cells to round up and dissociate from the Petri dish. This pool is protected from D4 and OlyA binding and could be removed only by a treatment with 2-hydroxypropyl-β-cyclodextrin (and methyl-β-cyclodextrin) after the accessible and sphingomyelin-sequestered pools are first depleted ( Das et al. , 2014 ). The D4-accessible and sphingomyelin-sequestered CHOL pools appear only when the CHOL contents in the PM are greater than 35 and 25 mol%, respectively. Clarification of the relationships of these three CHOL pools with the Lo–Ld-phases in GUVs and GPMVs is the next key issue. The D4-accessible pool is involved in many cellular processes, including hedgehog signaling ( Kinnebrew et al. , 2019 ), resistance to cytolysin formation ( Zhou et al. , 2020 ), and blocking bacterial infection ( Abrams et al. , 2020 ). The biological functions of the sphingomyelin-sequestered CHOL pools are barely known, but recently their function during the initial stages of the influenza virus (IFV) entry via clathrin-coated structures was reported ( Tang et al. , 2022 ). The sphingomyelin-CHOL complex nanodomain is recruited to the IFV-containing clathrin-coated structure, and then formin-binding protein 17 (FBP17), a membrane-bending protein that activates actin nucleation, is recruited to this sphingomyelin-CHOL complex, facilitating the neck constriction of the IFV-containing clathrin-coated structure. Detailed explanation 4: GPI-APs occupy an important position in raft research Readers who are familiar with the research fields of raft domains including GPI-APs can skip this subsection without losing track of the main flow of this article. In the human genome, more than 150 protein species have been identified as GPI-APs, in which the protein moieties located at the PM extracellular surface are anchored to the PM by way of GPI, a phospholipid ( Kinoshita and Fujita, 2015 ; Figure 1Da ). Most of the GPIs of GPI-APs contain two saturated fatty acyl chains of C18 and longer. Very small fractions of GPIs contain an unsaturated C24:1 acyl chain, but interestingly, the double bond exists near the terminal methyl group ( Kusumi et al. , 2004 , 2020 ). Therefore, their interactions with CHOL would be close to those of saturated acyl chains. This suggests that GPI-APs might be involved in the formation and functions of raft domains. Indeed, in raft research, GPI-APs occupy an important position. First, all GPI-APs examined thus far appear to satisfy all three criteria described in the subsection Practical guide for studying raft domains in the PM based on the new raft definition to be categorized as raft-associated molecules. For example, for "1) preferential partitioning into Lo-like domains in GPMVs at 10°C," see Sengupta et al. (2008) and Sezgin et al. (2012b) ; for "2) preferential partitioninginto detergent-resistant membranes using Triton X-100 at 4°C," see Brown and Rose (1992) and Suzuki et al. (2012) ; and for "3) greatly reduced/altered function after mild CHOL depletion, and recovery of the original function upon CHOL replenishment," see Suzuki et al. (2007a , b , 2012 ). Second, GPI-APs have played historically important roles in raft research. GPI-APs were key molecules for developing the use of DRMs in raft research ( Brown and Rose, 1992 ), examining possible raft involvement in polarized sorting and transport ( Dotti et al. , 1991 ; Fiedler et al. , 1993 ; Zurzolo and Simons, 2016 ), and investigating raft involvement in receptor signaling ( Stulnig et al. , 1997 ; Moran and Miceli, 1998 ). Gangliosides and sphingomyelins might be extensively associated with raft domains, but the raft functions involved in signaling could be better studied using GPI-APs. As such, GPI-APs have been extensively used as representative molecular species located and functioning in raft domains in the PM and have facilitated important conceptual advances about raft domains in the PM. One of the key characteristics of GPI-APs is that by replacing the GPI-anchoring chain with the transmembrane domain of a nonraftophilic single-pass transmembrane protein, without CHOL depletion, their raft involvement can be totally suppressed, and thus this property allows us to investigate the raft involvement of GPI-AP–related functions without CHOL depletion ( Varma and Mayor, 1998 ; Suzuki et al. , 2012 ). Third, many GPI-APs are receptors and trigger signaling pathways involving tyrosine kinases, phospholipase C, and Ca 2+ ( Štefanová et al. , 1991 ; Green et al. , 1999 ; Sheets et al. , 1999 ; Abrami et al. , 2003 ; Seveau et al. , 2004 ; Monastyrskaya et al. , 2005 ; Vial and Evans, 2005 ; Hunter and Nixon, 2006 ; Suzuki et al. , 2007a , b ; McGraw et al. , 2012 ; Korinek et al. , 2015 ; Ridone et al. , 2020 ; Wang et al. , 2021 ). Because GPI-APs are associated with raft domains, they are often chosen to study the mechanisms by which raft domains work for signal transduction. Meanwhile, the GPI-anchored structure is almost paradoxical for receptors because, although it relays the signal from the outside environment to the inside of the cell, it spans only halfway through the membrane ( Figure 1Da ). The involvement of "raft domains" has been implied in the transbilayer signaling by the GPI-AP receptors, but exactly how raft domains or raft-based lipid interactions participate in the transbilayer signal transduction of GPI-AP receptors remains unknown, despite extensive research (in addition to the references cited in the preceding paragraph, we suggest the following references for further reading: Omidvar et al. , 2006 ; Paulick and Bertozzi, 2008 ; Lingwood and Simons, 2010 ; Eisenberg et al. , 2011 ; Fessler and Parks, 2011 ; Kusumi et al. , 2014 ; Raghupathy et al. , 2015 ). Fourth, as described, all GPI-APs examined thus far form metastable homodimer rafts ( Suzuki et al. , 2012 ). These metastable homodimer rafts are considered to be one of the basic raft "units" that will form greater raft domains ( Suzuki et al. , 2012 ), and thus the generation of such homodimer rafts would constitute the first fundamental step in raft domain formation ( Figure 1Db ). As such, GPI-APs occupy a critical and unique position in raft research. Therefore, in this review we will extensively discuss the behaviors and functions of GPI-APs. Metastable homodimer rafts of GPI-APs are one of the basic units for raft formation and function As summarized in the preceding subsection, GPI-APs have occupied very special positions in raft research ( Figure 1D, a and b ). Using GPI-APs, Suzuki et al. (2012) identified one of the most fundamental steps, and probably the first step, for the formation of raft domains. This is based on the finding that essentially all molecular species of GPI-APs, including CD59, DAF, Thy-1, and Prion Protein, form dynamic, metastable homodimer rafts , with lifetimes on the order of 200 ms ( Figure 1Db ) ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ). The formation of GPI-AP homodimer rafts requires ectodomain protein homointeractions, which are stabilized by raft-lipid interactions. In the context of raft studies, attention was typically paid only to raft-lipid interactions, but here, specific protein–protein interactions were found to be critical for the formation of homodimer rafts (because GPI-APs form homodimers rather than heterodimers, the protein interaction is very specific). Because all GPI-APs examined thus far form metastable homodimer rafts, these structures are considered to be one of the common basic characteristics of GPI-APs. These observations were made possible by advanced single-fluorescent-molecule imaging. Meanwhile, the crystal structure of a GPI-AP dimer, human urokinase-type plasminogen activator receptor (uPAR), was recently reported and revealed enormous conformational changes of the dimer as compared with the monomeric structure ( Yu et al. , 2022 ). Such extensive conformational changes might also occur to induce metastable homodimers of other GPI-APs. The formation of metastable GPI-AP homodimer rafts might be the first step for the formation of any raft domains containing GPI-APs, because greater raft domains containing GPI-APs, such as GPI-AP trimer rafts and tetramer rafts, are generated by the merging of GPI-AP homodimer rafts with a GPI-AP monomer or another GPI-AP homodimer raft, indicating that GPI-AP homodimer rafts are the major building blocks for the production of greater raft domains ( Figure 1Db ) ( Suzuki et al. , 2012 ). Furthermore, the merging is likely induced by raft-lipid interactions, rather than protein–protein interactions, because GPI-AP hetero -trimer and tetramer rafts are formed as readily as homo -trimer and tetramer rafts and their lifetimes are about the same (not much involvement of protein–protein interactions for the merging; Figure 1Db ) and also because the merging is CHOL dependent ( Suzuki et al. , 2012 ). Because these greater GPI-AP–containing rafts are formed by raft-lipid interactions, they can have flexible GPI-AP compositions. Therefore, we consider the formation of GPI-AP homodimer rafts as the first step for the formation of raft domains containing GPI-APs, and the merging of GPI-AP homodimer rafts by raft-lipid interactions is the second step. This further means that the results of many studies of GPI-APs might have to be reinterpreted, because what the authors assumed to be GPI-AP monomers might have actually been GPI-AP homodimer rafts. Using CD59, Suzuki et al. (2012 ) also found that the homodimer rafts are the basic units for inducing cytoplasmic signals after the membrane attack complex binds to CD59. Suppression of the homodimer raft formation, using a mutant CD59 in which CD59's GPI anchor was replaced by the TM domain of the nonraft LDL receptor, made the cytoplasmic Ca 2+ mobilization triggered by binding of the membrane attack complex to CD59 much smaller and slower. These results further indicate that the GPI-AP's metastable homodimer rafts are one of the basic building blocks for the formation of greater functional raft domains containing GPI-APs. We raise the possibility that this process, in which the homodimers induced by specific molecular interactions form the cores for producing basic metastable homodimer raft domains, might also be true for other key raft-forming molecules, such as glycosphingolipids and sphingomyelins. We further consider that these various homodimer rafts are the basic fundamental units for building most of the greater raft domains in the PM and that the greater raft domains are generated by the merging of these basic raft units by raft-lipid interactions. Submicronscale–micronscale GPI-AP-raft–enriched domains In single-molecule imaging studies, the number density of the observed (fluorescently labeled) GPI-APs in the PM is ∼0.6 copies/µm 2 . This is the condition where single molecules and the interactions of single molecules in the PM could readily be observed. However, such expression levels are not suitable to observe greater raft domains. Mayor's group employed cells expressing GPI-APs with a total number density of ≈400 copies/µm 2 , expression levels ≈700x higher than those used by Suzuki et al. (2012) (still within the physiological range), and observed GPI-AP oligomers and assemblies by using time-resolved förster resonance energy transfer (FRET) imaging. Sharma et al. (2004 ) showed that 20–40% of GPI-APs exist in clusters smaller than hetero -pentamers (nondiscrimination of the protein species), with the remaining 60–80% being monomers (which might actually be homodimer rafts, because the time-resolved FRET method is not sensitive enough to detect homodimers). Mayor and colleagues further found that the submicronscale–micronscale rafts containing heteromixtures of GPI-APs form in a manner depending on ATP and actin filaments ( Goswami et al. , 2008 ; Zanten et al. , 2009 , 2010 ; Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ). Because the formation of GPI-AP homodimer, trimer, and tetramer rafts does not depend on actin filaments ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ), these results together indicate that active processes involving actin filaments may be important for the formation of much greater submicronscale–micronscale rafts containing heteromixtures of GPI-APs ( Plowman et al. , 2005 ; Goswami et al. , 2008 ; Gowrishankar et al. , 2012 ; Raghupathy et al. , 2015 ). We propose that these greater rafts would consist of GPI-AP homodimer rafts that are weakly glued together by raft-lipid interactions, which might associate with the actin filaments located on the PM cytoplasmic surface and thus become stabilized. Nevertheless, the GPI-AP homodimer rafts in such actin-induced submicronscale–micronscale rafts would be continually (all the time) exchanging with those existing as GPI-AP homodimer rafts and monomers in the bulk PM. The submicronscale–micronscale rafts might not be continuous entities ( Heberle et al. , 2010 , 2013 , 2020 ; Pathak and London, 2011 ; Bhatia et al. , 2014 , 2016 ; Cornell et al. , 2020 ). They are detected by optical microscopy, and, considering their limited spatial resolutions, these "rafts" might simply be the PM regions where true smaller raft domains are concentrated by the actomyosin mechanical system. However, when the raft domains are concentrated, they would fuse readily. However, the decomposition rates should not be affected by the concentration. Therefore, in these raft-enriched domains, extensive and rapid fusing and splitting of raft domains must be occurring continually. We think that these raft-enriched domains are extremely interesting in terms of their functions and the dynamic mechanisms by which they are formed. Strikingly nonlinear dependence of many signaling functions on the CHOL concentration in the PM Mild CHOL depletion is a prevalent method to study the involvement of CHOL and so-called "lipid rafts" in biological processes of interest. However, the following key point is rarely considered. The CHOL concentration in the PM of mammalian cells is generally around 35–45 mol% of total lipids. In normal CHOL-depletion experiments, the CHOL concentration is typically decreased by only 40% to 20–27 mol% ( Green et al. , 1999 ; Surviladze et al. , 2001 ; Suzuki et al. , 2007b ). For conciseness, in the following we state 40 and 25% to refer to the CHOL mole fractions in the PM before and after CHOL depletion, respectively. However, despite the presence of high concentrations (25 mol%) of CHOL remaining in the "CHOL-depleted" PM, many signaling reactions were shut down (we mostly consider events occurring at 37°C), as exemplified in the following reports: Green et al. , 1999 ; Sheets et al. , 1999 ; Abrami et al. , 2003 ; Seveau et al. , 2004 ; Monastyrskaya et al. , 2005 ; Vial and Evans, 2005 ; Hunter and Nixon, 2006 ; Suzuki et al. , 2007a , b ; McGraw et al. , 2012 ; Korinek et al. , 2015 ; Ridone et al. , 2020 ; Wang et al. , 2021 . Therefore, the dependence of these signaling processes on the CHOL content is strikingly nonlinear: full reaction at 40 mol% CHOL and no reaction at 25 mol% CHOL. This clearly shows that these signaling reactions do not depend on simple interactions with single CHOL molecules but instead are enabled by the domains/structures in the PM that are formed by cooperative interactions with CHOL in the presence of overall CHOL concentrations greater than 25 mol%. Previous and currently prevalent "definitions" of raft domains The special domains induced by CHOL are enriched in sphingomyelins. Accordingly, the idea that CHOL and sphingomyelins together form special lipid-based domains that stand out among the PM VUCA was born, and these domains were named "lipid rafts" about a quarter of century ago ( Simons and Meer, 1988 ; Simons and Ikonen, 1997 ; Simons and Toomre, 2000 ; Harder and Simons, 1997 ; Rietveld and Simons, 1998 ). Although many researchers erroneously assumed that generally mild CHOL-depletion protocols removed most of the CHOL from the PM (rather than just ≈40%),due to the experimental ease CHOL depletion became a very useful and popular method to examine the involvement of raft domains in various cellular events and molecular functions ( Moran and Miceli, 1998 ; Fessler and Parks, 2011 ). Furthermore, this protocol was often coupled with biochemical purification methods to prepare lipid raft domains as aggregates of some fractions of lipid raft domains and raft-associated molecules, after extraction from various cells and tissues using nonionic detergents at 4°C (called detergent-resistant membranes or DRMs) ( Brown and Rose, 1992 ). However, the interesting observations obtained by this approach failed to provide a clear definition of raft domains in the PM. The lipid raft domain research field has long been plagued by the lack of a clear definition. If the raft domains had been visualized by microscopic methods, then the definition would have been easier. For example, the micronscale structures on the PM with lifetimes longer than several tens of seconds, such as synapses, adherens junctions, and clathrin-coated pits, and containing many (>10) copies of each key protein species could be defined simply, because they can be imaged by immunofluorescence and immunoelectron microscopy. The raft domains that could be visualized are limited to those that were enlarged and stabilized by external stimulation (often by non–physiologically strong stimulation) and by artificial cross-linking of raft-associated molecules. Therefore, the focus of interest in raft research has shifted to revealing whether the precursor raft domains that are triggered to become large, stabilized raft domains exist in quiescent cells before stimulation. If they do exist, what are the precursor domains, how do they behave, and how are they triggered to become enlarged, stabilized raft domains to perform cellular signaling? For such investigations, microscopic visualizations of precursor domains in the quiescent steady state must be made, but these observations have turned out to be enormously difficult. At the 2006 Keystone Symposium on "Lipid Rafts and Cell Function," lipid rafts were "defined" as "small (10–200 nm), heterogeneous, highly dynamic, sterol- and sphingolipid-enriched domains that compartmentalize cellular processes. Small rafts can sometimes be stabilized to form larger platforms through protein–protein interactions." However, although this "definition" summarizes the general properties of raft domains in the PM quite well, it might not be very useful for research examining whether the observed molecules and events occur in raft domains and whether any raft domains are involved in the observed phenomena. As such, the older definitions of raft domains were often mixtures of definitions and properties, and usually difficult to apply to actual research. Furthermore, the definition of raft domains often depended on the context in which they are considered. For example, in the biochemical context, the raft domains have long been defined as the DRM fraction. This definition has been useful for identifying the molecular candidates that might be involved in raft domains in the PM. However, the largest problem is that DRMs could not be readily related to the actual raft domains in the PM, particularly those before stimulation (for example, in terms of their sizes, lifetimes, and distributions). Therefore, a general definition of raft domains existing in the PM must be developed for the robust development of this research field. It must be useful for research in the future and in broad fields of biomedicine. In addition, the definition must include the cooperative interaction of CHOL with other molecules for the formation of raft domains, as explained in the preceding subsection. Another key observation for defining raft domains in the PM As described, many signaling reactions that depend on CHOL do not rely on simple interactions with single CHOL molecules, but instead are enabled by the CHOL-enriched domains/structures in the PM that are formed by cooperative interactions of CHOL with other constituent molecules. Such domains could be generated in the presence of overall CHOL concentrations greater than 25 mol% in the PM. Another critically important observation was made using giant plasma membrane vesicles (GPMVs) or PM spheres, formed by PM blebbing after the dissociation of the actin-based membrane skeleton from the PM cytoplasmic surface. Upon cooling to ∼10°C, GPMVs generally have two coexisting and complementary micronscale liquid domains ( Baumgart et al. , 2007 ; Lingwood et al. , 2008 ; Levental et al. , 2009 , 2010 ; Johnson et al. , 2010 ). These domains exhibit the characteristics of the liquid-ordered and liquid-disordered phases (Lo- and Ld-phases, respectively) found in artificial giant unilamellar vesicles (GUVs) with the proper lipid compositions, typically a ternary 1:1:1 mixture of long, saturated phosphatidylcholine (PC) or sphingomyelin, CHOL, and unsaturated PC, at temperatures below 37°C (for the Lo- and Ld-phases, see the next subsection, Detailed explanation 1 ; Figure 1, A–C ). FIGURE 1: Key molecular interactions for the formation of various raft domains. (A) All trans conformations and two gauche conformations ([+] gauche and [–] gauche ) of the alkyl chains (–CH 2 –CH 2 –) n . The gauche conformation is induced by the rotation around a single bond between two adjacent CH 2 groups. The gauche conformation induces 60° bend in the acyl chain. Because the inclusion of a single gauche conformation in a chain is quite unfavorable for the acyl chain packing in the membrane (and thus energy), the second gauche conformation occurs frequently near the first gauche conformation. Such double gauche conformations are called kinks. The formation of gauche conformations (rotation around single bonds) and the rapid interconversions between the trans and gauche conformations are the fundamental causes for inducing fluidity (liquid-like property) of the PM. (B) (a) Chemical structures of CHOL and representative saturated and unsaturated phospholipids frequently used in the studies of GUVs. Saturated lipid: l -α-distearoylphosphatidylcholine (DSPC). Unsaturated lipid: L-α-dioleoylphosphatidylcholine (DOPC). When saturated lipids, such as DSPC, are located next to a CHOL molecule, due to the rigid tetracyclic skeletal structure of CHOL, the gauche conformations of the acyl chain between the carbonyl carbon C1 (-COO) and the 12th carbon C12 are suppressed, inducing more trans conformations and increasing the acyl chain packing. However, unsaturated lipids, such as DOPC, have at least a single cis bond between C9 and C10, which induces a mandatory 60° bend there. The mandatory 60° bend of the unsaturated acyl chain causes steric problems when it is next to a CHOL molecule. (b) (i) Schematic figure showing three possible configurations for placing CHOL and an unsaturated chain next to each other, indicating that good packing of these two structures is difficult. This is due to the mismatch (nonconformability) between the mandatory 60° bend at the double bond in the unsaturated acyl chain and the rigid tetracyclic skeletal structure of CHOL (magenta rectangles). Therefore, CHOL and unsaturated acyl chains tend to exclude each other, which is called lateral nonconformability, and CHOL tends to associate with saturated lipids if given the choice. Such accommodating and nonconformable interactions of CHOL with saturated and unsaturated lipids, respectively, are the main cause of phase separation in the ternary mixtures of CHOL, saturated lipids, and unsaturated lipids. (ii) The rigid ring structure of CHOL, located adjacent to the rough surface of the TM domain of the TM protein (induced by the protruding hydrophobic side chains of the amino acids in the membrane), would generate vacant pockets (packing defects). In this manner, CHOL is excluded from the first annulus of most TM proteins. Therefore, the CHOL-based raft domains are excluded from the PM compartment boundaries due to their exclusion from the rows of the TM picket proteins anchored to and aligned along the actin-based membrane skeleton fence. The TM proteins with raft affinities are expected to be well miscible with CHOL molecules. (C) Schematic snapshot drawings of the phase-separated GUV membrane, made of saturated lipids, such as DSPC, and unsaturated lipids, such as DOPC, and CHOL (top view). Open circles represent acyl chains (saturated chain in green and unsaturated chain in cyan; two chains form a lipid molecule), and solid magenta structures indicate CHOL molecules. The regions enriched in saturated lipids and CHOL represent the Lo-phase domains, and those enriched in unsaturated lipids represent the Ld-phase domains. Note that Lo-phase domains additionally contain smaller amounts of unsaturated lipids and Ld-phase domains additionally contain smaller amounts of saturated lipids and CHOL. Fluorescence microscopy observations might suggest the presence of submicronscale–micronscale Lo-phase domains, but they might actually represent the membrane regions enriched in nanoscale (several to several tens of nanometers) Lo-phase domains. (D) (a, left) Schematic model of a GPI-AP. The protein moiety is linked to phosphatidylinositol (PI), a phospholipid, via a short glycochain. The glycosylated PI (GPI) anchors the protein moiety to the PM outer leaflet. The monomeric GPI-AP can associate with raft domains, but its affinity is low (left). (a, right) Schematic model of the metastable GPI-AP homodimer rafts, which are formed by specific homoprotein interactions (P-P) and stabilized by raft-lipid interactions (which prolong the lifetime of homodimers). All GPI-APs examined thus far form metastable homodimer rafts with lifetimes on the order of ∼200 ms. (b) GPI-AP homodimer rafts merge to form GPI-AP tetramer rafts by raft-lipid interactions (green ripples), rather than protein–protein interactions. The tetramer raft lifetimes are shorter (∼100 ms) than the homodimer raft lifetimes. When GPI-AP homodimer rafts of the same GPI-AP species merge, they form homotetramer rafts, and when GPI-AP homodimer rafts of different GPI-AP species merge, they form heterotetramer rafts. Such flexibility of protein species in tetramer rafts is possible because the merging is mediated by raft-lipid interactions, rather than protein–protein interactions. These results suggest that the formation of metastable GPI-AP homodimer rafts represents the first step in the formation of raft domains and that the GPI-AP homodimer rafts are one of the basic units for generating greater raft domains. The formation of GPI-AP tetramer rafts would be the second step for the formation of greater raft domains. The next step required for the formation of greater raft domains is the stabilization of greater raft domains, which would be triggered by signaling, including the pathways of Src-family kinases and trimeric G proteins, and probably by interactions with active actomyosin systems. Importantly, these results showed that the PM is capable of undergoing large-scale phase separations into Lo-phase-like and Ld-phase-like domains, when the actin-based membrane skeleton is removed and the temperature is lowered to ∼10°C. Interestingly, after mild partial CHOL depletion, the Lo-phase-like micronscale domains could not be induced upon cooling to 10°C ( Johnson et al. , 2010 ), suggesting that the presence of overall CHOL concentrations of 25 mol% or more in the PM is essential for producing Lo-phase-like domains in GPMVs at ∼10°C. This requirement is essentially the same as that for many PM signaling functions, as described in preceding paragraphs, as well as for the induction of Lo–Ld-phase separation in GUVs (see Section 4 of Kusumi et al. , 2020 ). Detailed explanation 1: Lo- and Ld-phases in GUVs and biological membranes Readers familiar with Lo- and Ld-phases should skip this subsection. They can do so without losing track of the main flow of this article. The liquid properties of biological membranes are induced by the thermally driven conformational changes of individual single bonds in the fatty acyl chains of phospholipids and sphingolipids (–CH 2 –CH 2 –), from the more stable trans conformation to either one of the two gauche conformations ([+] gauche and [–] gauche ) ( Figure 1A ). The gauche conformations induce a 60° bend in the acyl chain, which prevents the lipid molecules from being packed closely in the membrane. The acyl chain order, which is often defined by the spread (distribution) of the angles of the C–C bonds in each –CH 2 –CH 2 – group relative to its average orientation (averaged over all the molecules and also over the characteristic time span for the observation method used), is generally coupled with how well the acyl chains are packed in the membrane: higher packing generally means higher acyl chain order (fewer gauche conformations). The double bond in a fatty acyl chain works like a fixed gauche bond; that is, a mandatory 60° bend in the acyl chain (although their exact conformations are different) ( Figure 1B ). Therefore, the membranes containing lipids with one or two unsaturated fatty acyl chains (unsaturated lipids) are less packed, blocking the solidification of the membrane, which is important for organisms living in cold environments. In artificial membranes, like GUVs, in the absence of CHOL the acyl chain packing and order are low at higher temperatures and thus constituent lipid molecules can undergo rapid translational diffusion. Thermodynamically, such membranes are in the liquid-disordered phase (Ld-phase), which was so named because lipid molecules undergo Brownian diffusion in the membrane (a characteristic of the liquid) and the acyl chains are disordered. The effects of CHOL on the membrane structure and dynamics in GUVs can be understood in terms of three types of CHOL–acyl chain interactions. When an acyl chain is located adjacent to a CHOL molecule, due to the rigid tetracyclic skeletal structure of CHOL, the gauche conformation is suppressed, thus inducing more trans conformations and increasing the acyl chain packing ( Figure 1B ). Therefore, the presence of CHOL generally enhances the acyl chain order in membranes in the Ld-phase. Upon lowering the temperature, because the presence of CHOL induces defects in the alignment of acyl chains, the packing in the solid phase is lower. Namely, CHOL enhances alkyl chain disorder in membranes in the solid phase (where acyl chains are quite well aligned and no translational diffusion of lipids occurs). CHOL and unsaturated acyl chains tend to exclude each other because the forced 60° bend at the double bond prevents the acyl chain from following along the rigid tetracyclic skeletal structure of CHOL (called "lateral non-conformability"; Kusumi et al. , 2005 ) ( Figure 1B ). At sufficiently high temperatures, GUVs composed of ternary mixtures of >25 mol% CHOL and saturated and unsaturated lipids form the Ld-phase membrane. Upon cooling, unsaturated lipids are excluded from the CHOL's rigid tetracyclic backbone surface (Interaction 3), where saturated lipids must come in and interact. Under certain conditions, such CHOL effects induce phase-separated membranes consisting of the Ld-phase domain containing a high content of unsaturated lipids and lower amounts of saturated lipids with smaller amounts of CHOL and the other phase domain mostly consisting of CHOL and saturated lipids with small amounts of unsaturated lipids ( Figure 1C ). The latter membrane domain exhibits higher acyl chain order compared with the Ld-phase domain, due to the interaction of CHOL with saturated fatty acyl chains (Interaction 1), but the packing is less dense than that in the solid phase due to the lower alignment of saturated lipids next to CHOL (Interaction 2), allowing lateral diffusion of the constituent molecules (liquid property). Accordingly, the latter domain tends to be in the liquid-ordered phase (Lo-phase domain). The thermodynamic phase can be strictly defined for GUVs comprising a few defined lipid species. However, for membranes composed of many molecular species, like biological membranes, because the phase cannot be defined readily we call the disordered and ordered liquid states "Ld-phase-like" and "Lo-phase-like" or simply "Ld-like" and "Lo-like" states, respectively. Some authors call these states of the biological membranes the Ld- and Lo-phases, but readers should know that this would be a loose terminology. In artificial membranes consisting of only CHOL and unsaturated lipids, CHOL molecules form transient clusters (oligomers), probably due to exclusion from the space around unsaturated acyl chains. Such CHOL clusters form and disperse continually with a lifetime of the order of 0.1–1 ns ( Subczynski et al. , 1990 ). The occurrence of Interaction 1 increases the number of trans configurations in the acyl chains and enhances the lipid packing, which would make the Lo-phase domains thicker than the Ld-phase domains. Cryoelectron microscopy/tomography directly visualized the membrane thickness in the Lo-phase domain to be greater than that in the Ld-phase domain in GUVs ( Cornell et al. , 2020 ; Heberle et al. , 2020 ). The thickness distribution of GPMVs suggested the nanoscale lateral heterogeneities ( Heberle et al. , 2020 ). Revised definition of raft domains On the basis of the argument developed in the preceding section (Lo-phase-like and Ld-phase-like domains coexist in the PM of living cells), as the definition of raft domains in the PM, we propose the following: "Raft domains in the PM are Lo-phase-like, CHOL-centered molecular complexes/domains formed by cooperative interactions of CHOL with saturated acyl chains as well as unsaturated acyl chains and transmembrane (TM) proteins ." This definition represents the fundamental mechanism for the formation of raft domains in the PM: the cooperative interactions due to saturated acyl chains' weak multiple accommodating interactions with CHOL and, importantly, the low miscibility of CHOL with unsaturated acyl chains and TM proteins (collectively called "raft-lipid interactions"), which occur in the presence of >25 mol% CHOL. The latter half of this definition is quite different from previous definitions in that it includes CHOL exclusion from the sites adjacent to unsaturated lipids and TM domains. For details about the interactions of CHOL with saturated and unsaturated acyl chains and TM proteins, see the preceding subsection, Detailed explanation 1 , and Figure 1, A–C . In developing this new definition, we wanted to make it very general but also very strong. Here, the critical molecular species to form raft domains are only CHOL and molecules containing saturated and unsaturated acyl chains. The definition is based on CHOL's cooperative interactions with saturated acyl chains (accommodating) as well as with unsaturated acyl chains and proteins' TM domains (nonconformable). These accommodating and nonconformable interactions together induce cooperative interactions. Previously, the "lateral nonconformability" between CHOL and unsaturated acyl chains was often neglected, making the meaning of the cooperativity unclear. This concept is close to the mechanism for the formation of the Lo-phase in the artificial ternary mixtures of CHOL, lipids with long saturated acyl chains, and lipids with unsaturated acyl chains, but in the definition of raft domains the Lo-phase formation is unnecessary. The key is not the formation of the phase, but the cooperative interactions. The cooperativity might start with only a few molecules that are included in raft domains (such as homodimer rafts of glycosylphosphatidylinositol-anchored proteins [GPI-APs] and CHOL; Suzuki et al. , 2012 ) and the surrounding unsaturated chains (although the Lo-phase might be defined for a molecular assembly of ≈10 molecules surrounded by unsaturated lipids; Kusumi et al. , 2020 ). Because this definition is based on the fundamental mechanism for the formation of raft domains, it should be applicable and useful for testing molecular complexes in the PM found in diverse biomedical fields. The application of this definition to experimental research is described in the next subsection, Practical guide for studying raft domains in the PM based on the new raft definition . It is critically important to realize that the interactions of individual molecules in and near the CHOL-centered raft domains are transient and weak. The sum of many weak accommodating and nonconformable interactions is the key for raft formation and for explaining the physical properties of raft domains. Therefore, all of the involved molecules in and around the raft domains should dynamically undergo thermal diffusion, readily moving in and out of raft domains ( Kenworthy et al. , 2004 ; Suzuki et al. , 2007a , b , 2012 ). This is consistent with the concept that the thermodynamic state of the PM is close to miscibility critical points exhibiting immense local compositional fluctuations ( Veatch and Keller, 2003 ; Honerkamp-Smith et al. , 2008 ). Although the space scales of the fluctuation were considered to be quite long, ≈20 nm ( Veatch and Keller, 2003 ; Honerkamp-Smith et al. , 2008 ), the experimental data about GPI-AP homodimer rafts suggest that the sizes of raft domains can be as small as a few nanometers ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ). A cautionary remark is necessary here. Lo-phase-like domains that do not involve CHOL might exist in the PM. Fluorescence lifetime measurements, using environmentally sensitive membrane dyes that report the degree of lipid packing, suggested that ∼76% of the liquid is in the "ordered" state, although the exact meaning of the ordered state must be further clarified ( Owen et al. , 2012 ). Practical guide for studying raft domains in the PM based on the new raft definition This definition of raft domains would be convenient to use for future research in a variety of biomedical fields. On the basis of the definition, we propose the following three criteria for determining whether the molecule of interest preferentially partitions into and functions in raft domains in the PM. Preferential partitioning into Lo-like domains in GPMVs at 10°C; Preferential partitioning into detergent-resistant membranes using Triton X-100 at 4°C; Greatly reduced/altered function after mild CHOL depletion, and recovery of the original function soon after CHOL replenishment. Basically, all three criteria should be satisfied for a molecule to be categorized as a raft-associated molecule. This is because each of these criteria is not based on the all-or-none law, but the preference addresses a quantitative difference and thus higher probability. Furthermore, these criteria are based on molecular behaviors in very different contexts/conditions. Therefore, we think that satisfying all three of these criteria is important. Previously, we proposed another criterion, "Preferential partitioning into the Lo-phase domains in Lo–Ld-phase-separated GUVs after reconstitution at 25°C," but we replaced it with the third criterion described here because the use of GUVs requires extensive experiments and might not be practical in various research projects (for details, see Sections 10 and 13 of Kusumi et al. , 2020 ). We believe that this third criterion effectively represents the definition and would be more practical for applications to molecular cell biological studies. Note that CHOL repletion experiments must be performed to confirm that the effect of CHOL depletion was directly induced by the loss of raft domains, rather than artifactual side effects of CHOL depletion ( Kenworthy et al. , 2004 ; Suzuki et al. , 2007a , b , 2012 ), although CHOL repletion experiments are rarely performed. For example, because acute CHOL depletion could induce actin reorganization and reduce the mobilities of PM molecules, its effect can be strong and broad in various assays ( Kwik et al. , 2003 ). However, because this effect does not disappear for 12–24 h after CHOL repletion ( Kwik et al. , 2003 ), if CHOL repletion quickly restores the molecular functions and behaviors before CHOL depletion, then the CHO-depletion effect is likely induced by the loss of raft domains. To assess the involvement of raft domains in biological events/phenomena of interest in the live-cell PM, we recommend examinations of the recruitment/involvement of GPI-APs ( Suzuki et al ., 2012 ) and the recently developed fluorescent analogues of sphingomyelin and gangliosides, GM1, GM2, GM3, and GD1b ( Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ); for an extensive summary, see Table 4 of Kusumi et al. (2020) . Note that virtually all fluorescent ganglioside and sphingomyelin analogues used before these publications preferentially partitioned into Ld domains ( Sezgin et al. , 2012a ) and thus could not and will not serve as raft markers (these studies tended to erroneously arrive at wrong conclusions). Protein toxin–based CHOL and sphingomyelin probes could also be used, but due caution is required because their affinities tend to be low and affected by their environment (readers interested in the protein probes for CHOL and sphingomyelins, see the next subsection, Detailed explanation 2 ). Because the sizes of PM raft domains in quiescent cells might be on the order of 2–20 nm ( Kusumi et al. , 2020 ), the recruitment and colocalizations would be better observed by employing single-molecule imaging methods ( Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ). When single-molecule localization microscopy, such as PALM/STORM, is utilized, note that simply gathering images would not be very useful. Instead, the spatial correlations and colocalizations of individual molecules of interest (fluorescent spots) with individual molecules of the raftophilic lipid probes (fluorescent spots) must be obtained ( Stone and Veatch, 2014 ; Pageon et al. , 2016 ; Simoncelli et al. , 2020 ); for the method to circumvent the overcounting problem, see Arnold et al. (2020 ). However, because lipid probes cannot be chemically fixed ( Tanaka et al. , 2010 ), observations in living cells, using fast data acquisition frame rates, would be preferable ( Fujiwara et al. , 2021b ). For further understanding of raft domains, excellent recent reviews, which have not been cited thus far in this article, are available. For example, refer to Sevcsik and Schütz (2016) , Levental et al. (2020) , and Regen (2020) . Recently, using protein toxin–based CHOL probes, the existence of three distinct CHOL pools in the PM outer leaflet has been found ( Das et al. , 2014 ; Endapally et al. , 2019 ). Their relationships with raft domains are not clear, but the 2–20-nm-scale raft domains could be classified into those that do or do not contain sphingomyelin. These three CHOL pools are summarized in Detailed explanation 3 . Detailed explanation 2: Protein-based CHOL probes Readers who are not interested in this subject matter can skip this subsection without losing track of the main flow of this article. CHOL probes based on bacterial and mushroom toxins (proteins) have recently been developed. The most popular probes employ the carboxy-terminal ∼13-kDa domain, referred to as domain 4 (D4), derived from perfringolysin O (PFO) ( Ohno-Iwashita et al. , 1990 ; Nakamura et al. , 1995 ; Shimada et al. , 2002 ) or anthrolysin O (ALO) ( Bourdeau et al. , 2009 ; Farrand et al. , 2010 ). D4H, the D434S mutant, exhibited higher affinity to CHOL ( Johnson et al. , 2012 ; Maekawa and Fairn, 2015 ). Liu et al. (2017 ) developed a ratiometric fluorescence imaging method that allows simultaneous in situ CHOL quantifications in both PM leaflets, using four orthogonal CHOL sensors, which facilitate the selection of probes with proper affinities for each experiment ( Cho et al. , 2022 ). Although these probes turned out to be extremely useful ( Zhang et al. , 2018 ; Cho et al. , 2022 ), the results obtained with them should be interpreted carefully because their binding might be sensitive to particular lipid-lipid interactions, phase states, and masking by the binding of CHOL's physiological interaction partner proteins ( Maxfield and Wüstner, 2012 ; Courtney et al. , 2018 ). Probes based on mushroom toxins, ostreolysin A (OlyA) and nakanori, which bind only to the sphingomyelin-CHOL complex, have also been developed ( Ota et al. , 2013 ; Das et al. , 2014 ; Skocˇaj et al. , 2014 ; Makino et al. , 2017 ; Endapally et al. , 2019 ). Structural and biochemical analyses showed that the conformation of CHOL-bound sphingomyelin is different from that of free sphingomyelin ( Endapally et al. , 2019 ). The E69A and E69S mutants of OlyA can bind to both free sphingomyelin and sphingomyelin complexed with CHOL and thus could be used as sphingomyelin probes (another powerful sphingomyelin probe is lysenin [ Yamaji et al. , 1998 ; Tomishige et al. , 2021 ]). Interestingly, the E69G mutant exhibited specificity to the sphingomyelin-CHOL complex, whereas the E69N and E69D mutants failed to bind either form of sphingomyelin. The D4 domain of PFO (and Y181A, called PFO*, used at 4°C; at 37°C, like PFO, PFO* forms pores in the PM) is very useful. It binds to free CHOL, but not the sphingomyelin-CHOL complex. Together with OlyA, which binds only to the sphingomyelin-CHOL complex, three distinct CHOL pools in the PM outer leaflet were identified ( Das et al. , 2014 ; Endapally et al. , 2019 ). They are the D4-accessible (∼16 mol% CHOL vs. total lipid), sphingomyelin-sequestered (∼15 mol%), and essential (∼12 mol%) CHOL pools (see Detailed explanation 3 , the next subsection). The essential CHOL pool is thus termed because its depletion causes cells to round up and dissociate from the substrate. This pool is protected from D4 and OlyA binding and could be removed only by a treatment with 2-hydroxypropyl-β-cyclodextrin (and methyl-β-cyclodextrin) after the D4-accessible and sphingomyelin-sequestered pools were first depleted ( Das et al. , 2014 ). Detailed explanation 3: Three distinct CHOL pools exist in the PM outer leaflet Readers who are not interested in this subject matter should skip this subsection and directly move to the next subsection. Three distinct CHOL pools in the PM outer leaflet were identified ( Das et al. , 2014 ; Endapally et al. , 2019 ) using toxins that bind to CHOL in different states. The carboxy-terminal ∼13-kDa domain of the bacterial toxin PFO, referred to as domain 4 (D4), binds only to free CHOL, and the mushroom toxin OlyA binds only to the sphingomyelin-CHOL complex (see Detailed explanation 2 for the protein-based CHOL probes). The three distinct CHOL pools are called the D4-accessible (∼16 mol% CHOL vs. total lipid), sphingomyelin-sequestered (∼15 mol%; OlyA binding), and essential (∼12 mol%) CHOL pools. The essential CHOL pool is thus termed because its depletion causes cells to round up and dissociate from the Petri dish. This pool is protected from D4 and OlyA binding and could be removed only by a treatment with 2-hydroxypropyl-β-cyclodextrin (and methyl-β-cyclodextrin) after the accessible and sphingomyelin-sequestered pools are first depleted ( Das et al. , 2014 ). The D4-accessible and sphingomyelin-sequestered CHOL pools appear only when the CHOL contents in the PM are greater than 35 and 25 mol%, respectively. Clarification of the relationships of these three CHOL pools with the Lo–Ld-phases in GUVs and GPMVs is the next key issue. The D4-accessible pool is involved in many cellular processes, including hedgehog signaling ( Kinnebrew et al. , 2019 ), resistance to cytolysin formation ( Zhou et al. , 2020 ), and blocking bacterial infection ( Abrams et al. , 2020 ). The biological functions of the sphingomyelin-sequestered CHOL pools are barely known, but recently their function during the initial stages of the influenza virus (IFV) entry via clathrin-coated structures was reported ( Tang et al. , 2022 ). The sphingomyelin-CHOL complex nanodomain is recruited to the IFV-containing clathrin-coated structure, and then formin-binding protein 17 (FBP17), a membrane-bending protein that activates actin nucleation, is recruited to this sphingomyelin-CHOL complex, facilitating the neck constriction of the IFV-containing clathrin-coated structure. Detailed explanation 4: GPI-APs occupy an important position in raft research Readers who are familiar with the research fields of raft domains including GPI-APs can skip this subsection without losing track of the main flow of this article. In the human genome, more than 150 protein species have been identified as GPI-APs, in which the protein moieties located at the PM extracellular surface are anchored to the PM by way of GPI, a phospholipid ( Kinoshita and Fujita, 2015 ; Figure 1Da ). Most of the GPIs of GPI-APs contain two saturated fatty acyl chains of C18 and longer. Very small fractions of GPIs contain an unsaturated C24:1 acyl chain, but interestingly, the double bond exists near the terminal methyl group ( Kusumi et al. , 2004 , 2020 ). Therefore, their interactions with CHOL would be close to those of saturated acyl chains. This suggests that GPI-APs might be involved in the formation and functions of raft domains. Indeed, in raft research, GPI-APs occupy an important position. First, all GPI-APs examined thus far appear to satisfy all three criteria described in the subsection Practical guide for studying raft domains in the PM based on the new raft definition to be categorized as raft-associated molecules. For example, for "1) preferential partitioning into Lo-like domains in GPMVs at 10°C," see Sengupta et al. (2008) and Sezgin et al. (2012b) ; for "2) preferential partitioninginto detergent-resistant membranes using Triton X-100 at 4°C," see Brown and Rose (1992) and Suzuki et al. (2012) ; and for "3) greatly reduced/altered function after mild CHOL depletion, and recovery of the original function upon CHOL replenishment," see Suzuki et al. (2007a , b , 2012 ). Second, GPI-APs have played historically important roles in raft research. GPI-APs were key molecules for developing the use of DRMs in raft research ( Brown and Rose, 1992 ), examining possible raft involvement in polarized sorting and transport ( Dotti et al. , 1991 ; Fiedler et al. , 1993 ; Zurzolo and Simons, 2016 ), and investigating raft involvement in receptor signaling ( Stulnig et al. , 1997 ; Moran and Miceli, 1998 ). Gangliosides and sphingomyelins might be extensively associated with raft domains, but the raft functions involved in signaling could be better studied using GPI-APs. As such, GPI-APs have been extensively used as representative molecular species located and functioning in raft domains in the PM and have facilitated important conceptual advances about raft domains in the PM. One of the key characteristics of GPI-APs is that by replacing the GPI-anchoring chain with the transmembrane domain of a nonraftophilic single-pass transmembrane protein, without CHOL depletion, their raft involvement can be totally suppressed, and thus this property allows us to investigate the raft involvement of GPI-AP–related functions without CHOL depletion ( Varma and Mayor, 1998 ; Suzuki et al. , 2012 ). Third, many GPI-APs are receptors and trigger signaling pathways involving tyrosine kinases, phospholipase C, and Ca 2+ ( Štefanová et al. , 1991 ; Green et al. , 1999 ; Sheets et al. , 1999 ; Abrami et al. , 2003 ; Seveau et al. , 2004 ; Monastyrskaya et al. , 2005 ; Vial and Evans, 2005 ; Hunter and Nixon, 2006 ; Suzuki et al. , 2007a , b ; McGraw et al. , 2012 ; Korinek et al. , 2015 ; Ridone et al. , 2020 ; Wang et al. , 2021 ). Because GPI-APs are associated with raft domains, they are often chosen to study the mechanisms by which raft domains work for signal transduction. Meanwhile, the GPI-anchored structure is almost paradoxical for receptors because, although it relays the signal from the outside environment to the inside of the cell, it spans only halfway through the membrane ( Figure 1Da ). The involvement of "raft domains" has been implied in the transbilayer signaling by the GPI-AP receptors, but exactly how raft domains or raft-based lipid interactions participate in the transbilayer signal transduction of GPI-AP receptors remains unknown, despite extensive research (in addition to the references cited in the preceding paragraph, we suggest the following references for further reading: Omidvar et al. , 2006 ; Paulick and Bertozzi, 2008 ; Lingwood and Simons, 2010 ; Eisenberg et al. , 2011 ; Fessler and Parks, 2011 ; Kusumi et al. , 2014 ; Raghupathy et al. , 2015 ). Fourth, as described, all GPI-APs examined thus far form metastable homodimer rafts ( Suzuki et al. , 2012 ). These metastable homodimer rafts are considered to be one of the basic raft "units" that will form greater raft domains ( Suzuki et al. , 2012 ), and thus the generation of such homodimer rafts would constitute the first fundamental step in raft domain formation ( Figure 1Db ). As such, GPI-APs occupy a critical and unique position in raft research. Therefore, in this review we will extensively discuss the behaviors and functions of GPI-APs. Metastable homodimer rafts of GPI-APs are one of the basic units for raft formation and function As summarized in the preceding subsection, GPI-APs have occupied very special positions in raft research ( Figure 1D, a and b ). Using GPI-APs, Suzuki et al. (2012) identified one of the most fundamental steps, and probably the first step, for the formation of raft domains. This is based on the finding that essentially all molecular species of GPI-APs, including CD59, DAF, Thy-1, and Prion Protein, form dynamic, metastable homodimer rafts , with lifetimes on the order of 200 ms ( Figure 1Db ) ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ). The formation of GPI-AP homodimer rafts requires ectodomain protein homointeractions, which are stabilized by raft-lipid interactions. In the context of raft studies, attention was typically paid only to raft-lipid interactions, but here, specific protein–protein interactions were found to be critical for the formation of homodimer rafts (because GPI-APs form homodimers rather than heterodimers, the protein interaction is very specific). Because all GPI-APs examined thus far form metastable homodimer rafts, these structures are considered to be one of the common basic characteristics of GPI-APs. These observations were made possible by advanced single-fluorescent-molecule imaging. Meanwhile, the crystal structure of a GPI-AP dimer, human urokinase-type plasminogen activator receptor (uPAR), was recently reported and revealed enormous conformational changes of the dimer as compared with the monomeric structure ( Yu et al. , 2022 ). Such extensive conformational changes might also occur to induce metastable homodimers of other GPI-APs. The formation of metastable GPI-AP homodimer rafts might be the first step for the formation of any raft domains containing GPI-APs, because greater raft domains containing GPI-APs, such as GPI-AP trimer rafts and tetramer rafts, are generated by the merging of GPI-AP homodimer rafts with a GPI-AP monomer or another GPI-AP homodimer raft, indicating that GPI-AP homodimer rafts are the major building blocks for the production of greater raft domains ( Figure 1Db ) ( Suzuki et al. , 2012 ). Furthermore, the merging is likely induced by raft-lipid interactions, rather than protein–protein interactions, because GPI-AP hetero -trimer and tetramer rafts are formed as readily as homo -trimer and tetramer rafts and their lifetimes are about the same (not much involvement of protein–protein interactions for the merging; Figure 1Db ) and also because the merging is CHOL dependent ( Suzuki et al. , 2012 ). Because these greater GPI-AP–containing rafts are formed by raft-lipid interactions, they can have flexible GPI-AP compositions. Therefore, we consider the formation of GPI-AP homodimer rafts as the first step for the formation of raft domains containing GPI-APs, and the merging of GPI-AP homodimer rafts by raft-lipid interactions is the second step. This further means that the results of many studies of GPI-APs might have to be reinterpreted, because what the authors assumed to be GPI-AP monomers might have actually been GPI-AP homodimer rafts. Using CD59, Suzuki et al. (2012 ) also found that the homodimer rafts are the basic units for inducing cytoplasmic signals after the membrane attack complex binds to CD59. Suppression of the homodimer raft formation, using a mutant CD59 in which CD59's GPI anchor was replaced by the TM domain of the nonraft LDL receptor, made the cytoplasmic Ca 2+ mobilization triggered by binding of the membrane attack complex to CD59 much smaller and slower. These results further indicate that the GPI-AP's metastable homodimer rafts are one of the basic building blocks for the formation of greater functional raft domains containing GPI-APs. We raise the possibility that this process, in which the homodimers induced by specific molecular interactions form the cores for producing basic metastable homodimer raft domains, might also be true for other key raft-forming molecules, such as glycosphingolipids and sphingomyelins. We further consider that these various homodimer rafts are the basic fundamental units for building most of the greater raft domains in the PM and that the greater raft domains are generated by the merging of these basic raft units by raft-lipid interactions. Submicronscale–micronscale GPI-AP-raft–enriched domains In single-molecule imaging studies, the number density of the observed (fluorescently labeled) GPI-APs in the PM is ∼0.6 copies/µm 2 . This is the condition where single molecules and the interactions of single molecules in the PM could readily be observed. However, such expression levels are not suitable to observe greater raft domains. Mayor's group employed cells expressing GPI-APs with a total number density of ≈400 copies/µm 2 , expression levels ≈700x higher than those used by Suzuki et al. (2012) (still within the physiological range), and observed GPI-AP oligomers and assemblies by using time-resolved förster resonance energy transfer (FRET) imaging. Sharma et al. (2004 ) showed that 20–40% of GPI-APs exist in clusters smaller than hetero -pentamers (nondiscrimination of the protein species), with the remaining 60–80% being monomers (which might actually be homodimer rafts, because the time-resolved FRET method is not sensitive enough to detect homodimers). Mayor and colleagues further found that the submicronscale–micronscale rafts containing heteromixtures of GPI-APs form in a manner depending on ATP and actin filaments ( Goswami et al. , 2008 ; Zanten et al. , 2009 , 2010 ; Komura et al. , 2016 ; Kinoshita et al. , 2017 ; Arumugam et al. , 2021 ). Because the formation of GPI-AP homodimer, trimer, and tetramer rafts does not depend on actin filaments ( Suzuki et al. , 2012 ; Tiwari et al. , 2018 ), these results together indicate that active processes involving actin filaments may be important for the formation of much greater submicronscale–micronscale rafts containing heteromixtures of GPI-APs ( Plowman et al. , 2005 ; Goswami et al. , 2008 ; Gowrishankar et al. , 2012 ; Raghupathy et al. , 2015 ). We propose that these greater rafts would consist of GPI-AP homodimer rafts that are weakly glued together by raft-lipid interactions, which might associate with the actin filaments located on the PM cytoplasmic surface and thus become stabilized. Nevertheless, the GPI-AP homodimer rafts in such actin-induced submicronscale–micronscale rafts would be continually (all the time) exchanging with those existing as GPI-AP homodimer rafts and monomers in the bulk PM. The submicronscale–micronscale rafts might not be continuous entities ( Heberle et al. , 2010 , 2013 , 2020 ; Pathak and London, 2011 ; Bhatia et al. , 2014 , 2016 ; Cornell et al. , 2020 ). They are detected by optical microscopy, and, considering their limited spatial resolutions, these "rafts" might simply be the PM regions where true smaller raft domains are concentrated by the actomyosin mechanical system. However, when the raft domains are concentrated, they would fuse readily. However, the decomposition rates should not be affected by the concentration. Therefore, in these raft-enriched domains, extensive and rapid fusing and splitting of raft domains must be occurring continually. We think that these raft-enriched domains are extremely interesting in terms of their functions and the dynamic mechanisms by which they are formed. ACTIN-CENTERED VIEW OF THE PM Dynamic actin meshwork partitions the PM into ≈100-nm compartments Cortical actin filament layers, with a thickness of ≈25 nm from the PM cytoplasmic surface and representing three to five overlaid actin filaments, are found virtually everywhere throughout the PM cytoplasmic surface in eukaryotic cells, except for newly forming thin lamellipodial and filopodial regions ( Morone et al. , 2006 ). Together with the stress fibers and other cytoskeletal elements, they play important roles in morphological changes and maintenance of the cell shape (which is the cell's PM shape), motility, endocytosis, and exocytosis. Among the cortical actin filaments, those closely apposed or bound to the PM cytoplasmic surface crisscross throughout this surface, forming a mesh-like structure, and thus are called the "actin membrane skeleton" ( Bennett, 1990 ; Morone et al. , 2006 , 2008 ). The actin membrane skeleton, which in some cases includes general cortical actin filaments, is a key constituent of the PM. Owing to its prevalence on the PM and ubiquity among cells, we will advance the actin-centered view of the PM here, in addition to the CHOL-centered view. The entire bulk PM is compartmentalized (∼100 nm and in the range of 30–230 nm, depending on the cell line) by the actin-based membrane-skeleton "fences" and rows of TM-protein "pickets" anchored to and aligned along these fences ( Figure 2Aa ) ( Fujiwara et al. , 2002 , 2016 ; Murase et al. , 2004 ; Trimble and Grinstein, 2015 ; Ostrowski et al. , 2016 ; Freeman et al. , 2018 ). In the compartmentalized PM, both TM proteins and lipids undergo short-term confined diffusion within a compartment plus occasional hop movements to an adjacent compartment, which is termed hop diffusion ( Figure 2Aa ) ( Fujiwara et al. , 2002 , 2016 , 2021a,b ; Murase et al. , 2004 ; Kusumi et al. , 2005 , 2012a ,b; Jaqaman and Grinstein, 2012 ; Xia et al. , 2019 ). As a consequence, each molecule in the PM exhibits two diffusion coefficients: the microscopic diffusion coefficient ( D micro : 3–10 µm 2 /s), describing the unhindered diffusion within a compartment, and the macroscopic diffusion coefficient ( D MACRO : 0.1–1 µm 2 /s), determined by the compartment size and the hop frequency across intercompartmental boundaries composed of the picket fence. FIGURE 2: Actin-induced PM compartmentalization and its coupling with raft domains for signal transduction of GPI-AP receptors. (A) (a) Schematic model showing the actin-induced PM compartmentalization and the raft domains coexisting in the PM. Virtually the entire PM is compartmentalized by the actin-based membrane-skeleton meshes (fences) and rows of TM-protein pickets anchored to and aligned along the actin fence (blue molecules), with sizes between 40 and 230 nm (for different cell types). Both TM proteins (magenta) and lipids (black) undergo short-term confined diffusion within a compartment plus occasional hop movements to an adjacent compartment, termed hop diffusion. The phospholipids located in the PM outer leaflet also undergo hop diffusion due to the steric hindrance of TM picket proteins and the presence of the annular belt zone around the picket proteins, where the diffusion of membrane molecules is slowed due to the hydrodynamic friction–like effect of immobile picket proteins in the viscous liquid (red circular area surrounding blue TM picket proteins in the bottom figure; also see b). Raft domains would mostly be confined within the compartment (purple domains; see b for the mechanism). Meanwhile, submicronscale hetero GPI-AP complex rafts (pink with red slashes in the bottom figure) might interact with the actin membrane skeleton mesh and actin filaments associated with the membrane skeleton mesh, perhaps forming asters. The organization of the actin-based membrane skeleton dynamically changes, thus varying the compartment shapes all the time, partly due to myosin activities. (b) Proposed mechanisms showing how the rows of TM picket proteins anchored to and aligned along the actin-based membrane skeleton (fence) contribute to the formation of compartment boundaries and suppress the growth of raft domains across the compartment boundaries. For the diffusion of all PM-associated molecules, the TM picket proteins generate steric hindrance effects and, more importantly, the hydrodynamic friction–like effects of immobile molecules in the viscous liquid, which are propagated around two diameters of the immobilized TM domain cross-section, by approximately 0.7–3 nm (propagating 1.4–6 nm away from the TM picket protein surface; magenta gradation surrounding blue TM picket proteins). Monte Carlo simulations revealed that when ∼20% of the compartment boundary is occupied by the TM picket proteins, the experimentally observed temporary confinement of phospholipids was observed ( Fujiwara et al. , 2002 ). This is consistent with the presence of TM picket proteins 3–12 nm away from each other. The large variations in these numbers are due to the different sizes of the TM picket proteins. Raft growth across the compartment boundaries would be suppressed, because the compartment boundaries are full of CHOL-excluding zones formed by the first annular lipids or boundary lipids around the TM picket protein ( Figure 1Bb-ii ), with a belt thickness of ∼0.5 nm (the areas shown by blue diagonal lines surrounding the blue TM picket proteins). Together with the steric hindrance effect of the TM picket proteins, they form CHOL-excluding zones of 1.7–5 nm in diameter (surrounding the TM picket proteins) located 2–11 nm away from each other. Therefore, for the growth of raft domains across the compartment boundaries, they might initially have to use these small gaps between the CHOL-excluding zones. (B) Schematic figure showing the process of raft-based signaling, using the case of CD59, a GPI-AP, as an example. See the main text for details. The TM picket proteins would not have to be packed on the actin filament meshes ( Fujiwara et al. , 2002 ). Owing to the hydrodynamic friction–like effect of immobile molecules in viscous liquid, the diffusion coefficients of molecules within ∼5 nm from the TM picket proteins would be greatly decreased ( Figure 2Ab ). Therefore, ∼20% picket coverage of the fence area would be sufficient to generate the diffusion barriers. Only several microseconds are needed for a membrane molecule to pass the compartment boundary on an actin filament (assuming 5 µm 2 /s for the phospholipid diffusion coefficient within the compartment and 10 nm for the width of the compartment boundary). Therefore, if a TM protein binds to an actin filament for >10 µs, it would block the passage of membrane molecules across the compartment boundary. Consequently, although the TM protein pickets might be dynamically exchanging with diffusing molecules all the time, the TM proteins transiently bound to an actin filament still work well as diffusion obstacles and creators of annular areas within which diffusion is slowed. Membrane molecules probably move from one compartment to an adjacent one due to the temporary breakdown of the fence or pickets. Short-term severing and spatial fluctuations of the actin filament mesh, as well as temporary loss of the TM picket proteins, would provide space for the passage of membrane molecules. For example, in the erythrocyte membrane, by pulling the spectrin meshwork with laser tweezers, Tomishige et al. (1998 ) found that spectrin tetramer dissociation to dimers (severing of the fence) is a key step for the hop movement of band 3, a TM protein ( spe ctrin tetramer-dimer eq uilibrium [SPEQ] gate model) ( Tsuji et al. , 1988 ; Bennett, 1990 ; Tomishige et al. , 1998 ). Therefore, it is wrong to assume that such picket fences are static structures, as mentioned by some authors ( Gowrishankar et al. , 2012 ; Freeman et al. , 2018 ). In longer timescales, laser tweezer experiments in which membrane molecules were laterally dragged in the PM plane revealed that the compartment boundaries shift and break within a minute ( Ritchie and Kusumi, 2002 ). Actin membrane skeleton is dynamically reorganized and continually modified by ATP- and myosin II–dependent processes The hop movements of membrane molecules and the laser tweezer data suggest temporary severing and spatial fluctuations of the actin filament mesh, as well as dynamic large-scale modifications of the actin membrane skeleton. Superresolution microscopy of cortical actin filaments labeled with Lifeact-mGFP in live cells, conducted at time resolutions of 0.5–2.3 s and spatial resolutions of ∼120 nm, detected actin clusters linking two or more actin filaments in the fine actin meshwork (confirmed by phalloidin staining using fixed cells), acting as nodes of the meshwork ( Shirai et al. , 2017 ). About two-thirds of the actin nodes are located within 3.5 nm from the PM cytoplasmic surface, as found by using Lifeact linked to a TM peptide (Lifeact-TM), indicating that the majority of the observed actin nodes are on the actin membrane skeleton ( Shirai et al. , 2017 ). Interestingly, the use of Lifeact-TM revealed that 89% of the stress fibers are also in the proximity (within 3.5 nm) of the PM cytoplasmic surface. The node formation depended on the Arp2/3 actin nucleation and filament branching activities. The actin nodes dynamically moved on/along the actin meshwork in a myosin II–dependent manner. However, despite these myosin II–dependent movements of the actin nodes in the actin membrane skeleton, myosin II is not present in the actin nodes. Myosin II is probably located near the nodes, because it induces their directed and jittering diffusion-like movements. The fluctuating movements of the actin node are caused by several nearby myosin II filaments, which undergo a tug of war over the actin node. These ATP-dependent active processes might be involved in the formation and dynamics of the actin membrane skeleton meshwork. As described, scanning optical trap experiments demonstrated that the PM compartments drift and change shapes in less than a minute ( Ritchie and Kusumi, 2002 ). Dynamic actin meshwork partitions the PM into ≈100-nm compartments Cortical actin filament layers, with a thickness of ≈25 nm from the PM cytoplasmic surface and representing three to five overlaid actin filaments, are found virtually everywhere throughout the PM cytoplasmic surface in eukaryotic cells, except for newly forming thin lamellipodial and filopodial regions ( Morone et al. , 2006 ). Together with the stress fibers and other cytoskeletal elements, they play important roles in morphological changes and maintenance of the cell shape (which is the cell's PM shape), motility, endocytosis, and exocytosis. Among the cortical actin filaments, those closely apposed or bound to the PM cytoplasmic surface crisscross throughout this surface, forming a mesh-like structure, and thus are called the "actin membrane skeleton" ( Bennett, 1990 ; Morone et al. , 2006 , 2008 ). The actin membrane skeleton, which in some cases includes general cortical actin filaments, is a key constituent of the PM. Owing to its prevalence on the PM and ubiquity among cells, we will advance the actin-centered view of the PM here, in addition to the CHOL-centered view. The entire bulk PM is compartmentalized (∼100 nm and in the range of 30–230 nm, depending on the cell line) by the actin-based membrane-skeleton "fences" and rows of TM-protein "pickets" anchored to and aligned along these fences ( Figure 2Aa ) ( Fujiwara et al. , 2002 , 2016 ; Murase et al. , 2004 ; Trimble and Grinstein, 2015 ; Ostrowski et al. , 2016 ; Freeman et al. , 2018 ). In the compartmentalized PM, both TM proteins and lipids undergo short-term confined diffusion within a compartment plus occasional hop movements to an adjacent compartment, which is termed hop diffusion ( Figure 2Aa ) ( Fujiwara et al. , 2002 , 2016 , 2021a,b ; Murase et al. , 2004 ; Kusumi et al. , 2005 , 2012a ,b; Jaqaman and Grinstein, 2012 ; Xia et al. , 2019 ). As a consequence, each molecule in the PM exhibits two diffusion coefficients: the microscopic diffusion coefficient ( D micro : 3–10 µm 2 /s), describing the unhindered diffusion within a compartment, and the macroscopic diffusion coefficient ( D MACRO : 0.1–1 µm 2 /s), determined by the compartment size and the hop frequency across intercompartmental boundaries composed of the picket fence. FIGURE 2: Actin-induced PM compartmentalization and its coupling with raft domains for signal transduction of GPI-AP receptors. (A) (a) Schematic model showing the actin-induced PM compartmentalization and the raft domains coexisting in the PM. Virtually the entire PM is compartmentalized by the actin-based membrane-skeleton meshes (fences) and rows of TM-protein pickets anchored to and aligned along the actin fence (blue molecules), with sizes between 40 and 230 nm (for different cell types). Both TM proteins (magenta) and lipids (black) undergo short-term confined diffusion within a compartment plus occasional hop movements to an adjacent compartment, termed hop diffusion. The phospholipids located in the PM outer leaflet also undergo hop diffusion due to the steric hindrance of TM picket proteins and the presence of the annular belt zone around the picket proteins, where the diffusion of membrane molecules is slowed due to the hydrodynamic friction–like effect of immobile picket proteins in the viscous liquid (red circular area surrounding blue TM picket proteins in the bottom figure; also see b). Raft domains would mostly be confined within the compartment (purple domains; see b for the mechanism). Meanwhile, submicronscale hetero GPI-AP complex rafts (pink with red slashes in the bottom figure) might interact with the actin membrane skeleton mesh and actin filaments associated with the membrane skeleton mesh, perhaps forming asters. The organization of the actin-based membrane skeleton dynamically changes, thus varying the compartment shapes all the time, partly due to myosin activities. (b) Proposed mechanisms showing how the rows of TM picket proteins anchored to and aligned along the actin-based membrane skeleton (fence) contribute to the formation of compartment boundaries and suppress the growth of raft domains across the compartment boundaries. For the diffusion of all PM-associated molecules, the TM picket proteins generate steric hindrance effects and, more importantly, the hydrodynamic friction–like effects of immobile molecules in the viscous liquid, which are propagated around two diameters of the immobilized TM domain cross-section, by approximately 0.7–3 nm (propagating 1.4–6 nm away from the TM picket protein surface; magenta gradation surrounding blue TM picket proteins). Monte Carlo simulations revealed that when ∼20% of the compartment boundary is occupied by the TM picket proteins, the experimentally observed temporary confinement of phospholipids was observed ( Fujiwara et al. , 2002 ). This is consistent with the presence of TM picket proteins 3–12 nm away from each other. The large variations in these numbers are due to the different sizes of the TM picket proteins. Raft growth across the compartment boundaries would be suppressed, because the compartment boundaries are full of CHOL-excluding zones formed by the first annular lipids or boundary lipids around the TM picket protein ( Figure 1Bb-ii ), with a belt thickness of ∼0.5 nm (the areas shown by blue diagonal lines surrounding the blue TM picket proteins). Together with the steric hindrance effect of the TM picket proteins, they form CHOL-excluding zones of 1.7–5 nm in diameter (surrounding the TM picket proteins) located 2–11 nm away from each other. Therefore, for the growth of raft domains across the compartment boundaries, they might initially have to use these small gaps between the CHOL-excluding zones. (B) Schematic figure showing the process of raft-based signaling, using the case of CD59, a GPI-AP, as an example. See the main text for details. The TM picket proteins would not have to be packed on the actin filament meshes ( Fujiwara et al. , 2002 ). Owing to the hydrodynamic friction–like effect of immobile molecules in viscous liquid, the diffusion coefficients of molecules within ∼5 nm from the TM picket proteins would be greatly decreased ( Figure 2Ab ). Therefore, ∼20% picket coverage of the fence area would be sufficient to generate the diffusion barriers. Only several microseconds are needed for a membrane molecule to pass the compartment boundary on an actin filament (assuming 5 µm 2 /s for the phospholipid diffusion coefficient within the compartment and 10 nm for the width of the compartment boundary). Therefore, if a TM protein binds to an actin filament for >10 µs, it would block the passage of membrane molecules across the compartment boundary. Consequently, although the TM protein pickets might be dynamically exchanging with diffusing molecules all the time, the TM proteins transiently bound to an actin filament still work well as diffusion obstacles and creators of annular areas within which diffusion is slowed. Membrane molecules probably move from one compartment to an adjacent one due to the temporary breakdown of the fence or pickets. Short-term severing and spatial fluctuations of the actin filament mesh, as well as temporary loss of the TM picket proteins, would provide space for the passage of membrane molecules. For example, in the erythrocyte membrane, by pulling the spectrin meshwork with laser tweezers, Tomishige et al. (1998 ) found that spectrin tetramer dissociation to dimers (severing of the fence) is a key step for the hop movement of band 3, a TM protein ( spe ctrin tetramer-dimer eq uilibrium [SPEQ] gate model) ( Tsuji et al. , 1988 ; Bennett, 1990 ; Tomishige et al. , 1998 ). Therefore, it is wrong to assume that such picket fences are static structures, as mentioned by some authors ( Gowrishankar et al. , 2012 ; Freeman et al. , 2018 ). In longer timescales, laser tweezer experiments in which membrane molecules were laterally dragged in the PM plane revealed that the compartment boundaries shift and break within a minute ( Ritchie and Kusumi, 2002 ). Actin membrane skeleton is dynamically reorganized and continually modified by ATP- and myosin II–dependent processes The hop movements of membrane molecules and the laser tweezer data suggest temporary severing and spatial fluctuations of the actin filament mesh, as well as dynamic large-scale modifications of the actin membrane skeleton. Superresolution microscopy of cortical actin filaments labeled with Lifeact-mGFP in live cells, conducted at time resolutions of 0.5–2.3 s and spatial resolutions of ∼120 nm, detected actin clusters linking two or more actin filaments in the fine actin meshwork (confirmed by phalloidin staining using fixed cells), acting as nodes of the meshwork ( Shirai et al. , 2017 ). About two-thirds of the actin nodes are located within 3.5 nm from the PM cytoplasmic surface, as found by using Lifeact linked to a TM peptide (Lifeact-TM), indicating that the majority of the observed actin nodes are on the actin membrane skeleton ( Shirai et al. , 2017 ). Interestingly, the use of Lifeact-TM revealed that 89% of the stress fibers are also in the proximity (within 3.5 nm) of the PM cytoplasmic surface. The node formation depended on the Arp2/3 actin nucleation and filament branching activities. The actin nodes dynamically moved on/along the actin meshwork in a myosin II–dependent manner. However, despite these myosin II–dependent movements of the actin nodes in the actin membrane skeleton, myosin II is not present in the actin nodes. Myosin II is probably located near the nodes, because it induces their directed and jittering diffusion-like movements. The fluctuating movements of the actin node are caused by several nearby myosin II filaments, which undergo a tug of war over the actin node. These ATP-dependent active processes might be involved in the formation and dynamics of the actin membrane skeleton meshwork. As described, scanning optical trap experiments demonstrated that the PM compartments drift and change shapes in less than a minute ( Ritchie and Kusumi, 2002 ). INTERACTION OF THE ACTIN-BASED MEMBRANE SKELETON AND CORTICAL ACTIN FILAMENTS WITH RAFT DOMAINS In this section, we describe the relationship between CHOL-centered raft domains and the actin-based membrane skeleton. Raft exclusion from PM compartment boundaries by rows of TM picket proteins aligned along the actin membrane skeleton fence In general, TM proteins and CHOL are sterically nonconformable with each other, due to the rigid α-helical structure and the rugged surface (due to the amino acid side chains protruding from the α-helix) of the TM domain and the rigid, bulky tetracyclic CHOL structure ( Figure 1Bb ) (however, see the paragraph after the next one). Therefore, raft domains, whether they exist in only the PM outer leaflet or in both the outer and inner leaflets, would hardly grow across the compartment boundaries where the TM picket proteins are abundant. Consequently, the sizes of raft domains in the PM would be limited by the meshes of the actin membrane skeleton. This conclusion is supported by in vitro experiments in which actin was placed on the supported bilayer or GUVs and by simulations ( Honigmann et al. , 2014 ; Arumugam et al. , 2015 ). In addition, the GPMV experiments revealed that cooling-induced large micronscale phase separation into Lo-phase-like and Ld-phase-like domains can occur only after the removal of the actin membrane skeleton ( Baumgart et al. , 2007 ; Lingwood et al. , 2008 ; Levental et al. , 2009 , 2010 ; Johnson et al. , 2010 ). These results clearly showed that the actin membrane skeleton blocks the formation of micronscale raft domains. Meanwhile, some TM proteins have been proposed to partition into raft domains ( Levental et al. , 2010 ). Furthermore, immune TM receptors reportedly take advantage of raft domains in their signaling. These results are summarized in Detailed explanation 5 , which is placed right before Concluding remarks . We sometimes hear arguments about whether the PM is organized by actin-based compartmentalization or raft domains, but it is important to realize that both occur, as clarified by the preceding discussions; also see Kusumi et al. (2012a,b ). The actin membrane skeleton compartmentalizes the entire PM, and the nanoscale (2–200 nm) rafts exist within the compartments ( Figure 2A ). Because both actin-induced compartments and raft domains coexist in the PM and the sizes and densities of raft domains as well as the sizes of the actin-induced compartments vary greatly among different types of cells, applications of the so-called diffusion law for interpreting the results of fluorescence correlation spectroscopy obtained by confocal and stimulated emission depletion (STED) microscopy must be done carefully. Depending on the details of the raft domains and actin-induced compartments, the dependence of the diffusion coefficients on the observation area size would vary in very complex ways ( Wawrezinieck et al. , 2005 ; Andrade et al. , 2015 ; Schneider et al. , 2017 ; Veerapathiran and Wohland, 2018 ; Sezgin et al. , 2019 ). Submicronscale–micronscale rafts containing heteromixtures of GPI-APs are coupled with cytoplasmic actin filaments by way of TM proteins and long, saturated phosphatidylserine bound to actin filaments As described, Mayor's group found submicronscale–micronscale rafts containing heteromixtures of GPI-APs in the PM outer leaflet, whose formation is actively driven by actomyosin, located on the PM inner surface. These submicronscale–micronscale rafts are considered to be surrounded by radially spreading actin filaments. Such structures are termed "asters," and were visualized in vitro ( Gowrishankar et al. , 2012 ) but apparently not in vivo. The asters are colocalized with myosin II and filamin. For beautiful and clear presentations of asters and the actin meshwork on the PM, see Figure 1b of Honigmann and Pralle (2016 ). How can these actin filaments located on the PM inner surface become coupled to the hetero-GPI-AP complex rafts in the PM outer leaflet? Mayor's group proposed that clusters of phosphatidylserine (PS), with long, saturated acyl chains located in the inner leaflet, are bound to actin filaments and also coupled to the submicronscale–micronscale rafts containing heteromixtures of GPI-APs in the outer leaflet by interdigitation in the middle of the bilayer ( Raghupathy et al. , 2015 ). Because CHOL is much shorter than the long, saturated chains of GPI-APs and PS ( Figure 1Ba ), the inner surfaces of both the PS and GPI-AP clusters in the central part of the bilayer are rough, and thus once they come together, they might not readily break apart from each other. The transbilayer coupling is enhanced upon the activation of β1 integrin, which triggers actin nucleation via formins and myosin, and the involvement of talin and the mechanotransducer vinculin in the cluster ( Kalappurakkal et al. , 2019 ). Interestingly, GPI-AP clustering was required for integrin-induced cell spreading and migration. Some of the actin nodes detected by superresolution microscopy might be the same as actin asters, but some may differ because myosin II and filamin A are not appreciably colocalized with the actin nodes ( Shirai et al. , 2017 ). Previously, such actin clusters (nodes and asters) were found in vitro ( Köster et al. , 2016 ) or after drug-induced partial actin depolymerization in live cells ( Luo et al. , 2013 ). However, superresolution microscopy of live cells clarified that actin nodes constitutively exist in the actin membrane skeleton within 3.5 nm from the PM inner surface and are dynamically driven by myosin II in an ATP-dependent manner ( Shirai et al. , 2017 ). Involvement of actin membrane skeleton in GPI-AP receptors' raft-based signaling Reports about the involvement of the actin membrane skeleton (or cortical actin filaments) in the signal transduction of GPI-AP receptors have been quite scarce. Here, we summarize the results reported previously ( Suzuki et al. , 2007a , b ), which are rare publications addressing this issue. When CD59, a GPI-AP working as the receptor for the membrane attack complex, is engaged, it forms clusters containing an average of four to five CD59 molecules, which in turn form nano-sized stabilized CD59-cluster rafts ( Figure 2B 1). Experimentally, this process could be mimicked by closely cross-linking the CD59 molecules by the addition of 40-nm gold particles coated with an anti-CD59 monoclonal antibody. Both the natural ligation and artificial cross-linking can trigger very similar intracellular signals, such as PLCγ recruitment to CD59-cluster rafts, leading to IP 3 production and then to intracellular calcium responses ( Figure 2B ). Engaged CD59-cluster rafts exhibited very peculiar behaviors. They repeatedly and continually undergo temporary immobilizations lasting for 0.56 s (exponential lifetime) that are induced by the binding of CD59-cluster rafts to the actin-based membrane skeleton (called stimulation-induced arrest of lateral diffusion = STALL; Figure 2B 2). This binding is induced by Src-family kinases, such as Lyn, which is activated by Giα, the alpha subunit of the inhibitory trimeric G protein, after they are both recruited to CD59-cluster rafts ( Figure 2B 3). The protein(s) phosphorylated by Lyn is unknown, but it (or one of them) could be an as-yet-unknown TM protein X, which might mediate the PLCγ recruitment to the site on the PM inner surface linked to the CD59-cluster raft located in the PM outer leaflet ( Figure 2B 4). Meanwhile, a cytoplasmic protein Y might also be involved in this process ( Figure 2B 5). Very interestingly, PLCγ is recruited to the CD59-cluster only when the CD59-cluster raft is undergoing temporary immobilization due to its binding to the actin-based membrane skeleton ( Figure 2B 6). These results suggest the possibility that signal transduction platforms for GPI-APs might abundantly exist on the actin-based membrane skeleton. When the CD59-cluster raft meets Giα and Lyn on the platform, it becomes temporarily immobilized on the signaling platform by binding to proteins X and Y, which are in turn phosphorylated by Lyn, producing the binding sites for PLCγ. PLCγ recruited to CD59-cluster rafts produces IP 3 and diacylglycerol, leading to intracellular Ca 2+ mobilization and PKC activation ( Figure 2B 7). Detailed explanation 5: TM proteins partitioning into raft domains and TM receptors taking advantage of raft domains for their signaling Note that some TM proteins have been proposed to partition into raft domains ( Levental et al. , 2010 ). Lorent et al. (2017 ) identified three physical features of the TM domains that independently affect the raft partitioning of TM proteins: smaller surface area, longer TM domain, and palmitoylation. More specifically, they represent smaller protruding amino acid side chains from the α-helical TM domain, matching of the hydrophobic length of the TM domain with the extended saturated acyl chains (due to the association with CHOL), and smoothing of the rugged TM domain surface by the palmitoyl chains, respectively. These features might be coupled with the dimerization–oligomerization of TM proteins linked by palmitoyl chains. Rhodopsin, a G protein–coupled receptor with seven TM domains, has two covalently linked palmitoyl chains, but becomes associated with raft domains only upon dimer formation ( Seno and Hayashi, 2017 ; Hayashi et al. , 2019 ). This result suggests that the rugged parts of the TM domain surface of the dimer are more easily covered with four palmitate chains than those of the monomer with two palmitate chains. Immune receptors, including high-affinity Fcε receptors (FcεRI) in mast cells and B-cell receptors, take advantage of the raft domains that the ligand-induced receptor oligomers generate in/around them to enhance their downstream signaling. The involvement of raft domains in immune receptor signaling had long been suspected, but it was difficult to show unequivocally. However, this difficulty was solved by developing/improving microscopic imaging-analysis methods. By improving imaging fluorescence correlation spectroscopy, Baird and colleagues revealed that stabilized Lo-like nanoscale raft domains are created around clustered FcεRI in immune mast cells ( Bag et al. , 2021 ). These domains strongly augment the Lyn recruitment and suppress the recruitment of the TM phosphatase PTPα. Proof of the Lyn recruitment to clustered FcεRI had previously been elusive, but extensive examinations of very many molecular trajectories (several tens of thousands), made possible by the improvement and application of imaging fluorescence correlation spectroscopy, unequivocally revealed Lyn recruitment to clustered FcεRI. This result is consistent with the enhanced recruitment of sphingomyelin, a prototypical raftophilic phospholipid, to antigen-induced clusters of FcεRI but not to FcεRI in resting cells, found by single-molecule imaging tracking of a newly developed fluorescently labeled sphingomyelin ( Kinoshita et al. , 2017 ). In immune B cells, by applying a pair cross-correlation analysis method to superresolution single-molecule localization microscopy data ( Stone and Veatch, 2014 ), Veatch and colleagues demonstrated that stimulation-induced B-cell receptor (BCR) clusters induce Lo-phase-like domains around them, which are capable of sorting key regulators of BCR activation ( Stone et al. , 2017 ). Taken together, these results provide evidence for the role of the receptor clustering–induced membrane domains and a plausible mechanism for recruiting downstream signaling molecules to the receptor clusters. They agree well with the signaling raft domain formation induced by the clustering of a GPI-AP, CD59 ( Suzuki et al. , 2007a , b , 2012 ) and suggest that similar mechanisms are employed in many other receptor signaling pathways ( Kusumi et al. , 2004 ; Koyama-Honda et al. , 2020 ). For other examples in which raft-based interactions might be involved in the signal transduction by TM receptors, refer to the following papers: Sohn et al. (2008) ; Coskun et al. (2011) ; Chung et al. (2016) ; Shelby et al. (2016) . Raft exclusion from PM compartment boundaries by rows of TM picket proteins aligned along the actin membrane skeleton fence In general, TM proteins and CHOL are sterically nonconformable with each other, due to the rigid α-helical structure and the rugged surface (due to the amino acid side chains protruding from the α-helix) of the TM domain and the rigid, bulky tetracyclic CHOL structure ( Figure 1Bb ) (however, see the paragraph after the next one). Therefore, raft domains, whether they exist in only the PM outer leaflet or in both the outer and inner leaflets, would hardly grow across the compartment boundaries where the TM picket proteins are abundant. Consequently, the sizes of raft domains in the PM would be limited by the meshes of the actin membrane skeleton. This conclusion is supported by in vitro experiments in which actin was placed on the supported bilayer or GUVs and by simulations ( Honigmann et al. , 2014 ; Arumugam et al. , 2015 ). In addition, the GPMV experiments revealed that cooling-induced large micronscale phase separation into Lo-phase-like and Ld-phase-like domains can occur only after the removal of the actin membrane skeleton ( Baumgart et al. , 2007 ; Lingwood et al. , 2008 ; Levental et al. , 2009 , 2010 ; Johnson et al. , 2010 ). These results clearly showed that the actin membrane skeleton blocks the formation of micronscale raft domains. Meanwhile, some TM proteins have been proposed to partition into raft domains ( Levental et al. , 2010 ). Furthermore, immune TM receptors reportedly take advantage of raft domains in their signaling. These results are summarized in Detailed explanation 5 , which is placed right before Concluding remarks . We sometimes hear arguments about whether the PM is organized by actin-based compartmentalization or raft domains, but it is important to realize that both occur, as clarified by the preceding discussions; also see Kusumi et al. (2012a,b ). The actin membrane skeleton compartmentalizes the entire PM, and the nanoscale (2–200 nm) rafts exist within the compartments ( Figure 2A ). Because both actin-induced compartments and raft domains coexist in the PM and the sizes and densities of raft domains as well as the sizes of the actin-induced compartments vary greatly among different types of cells, applications of the so-called diffusion law for interpreting the results of fluorescence correlation spectroscopy obtained by confocal and stimulated emission depletion (STED) microscopy must be done carefully. Depending on the details of the raft domains and actin-induced compartments, the dependence of the diffusion coefficients on the observation area size would vary in very complex ways ( Wawrezinieck et al. , 2005 ; Andrade et al. , 2015 ; Schneider et al. , 2017 ; Veerapathiran and Wohland, 2018 ; Sezgin et al. , 2019 ). Submicronscale–micronscale rafts containing heteromixtures of GPI-APs are coupled with cytoplasmic actin filaments by way of TM proteins and long, saturated phosphatidylserine bound to actin filaments As described, Mayor's group found submicronscale–micronscale rafts containing heteromixtures of GPI-APs in the PM outer leaflet, whose formation is actively driven by actomyosin, located on the PM inner surface. These submicronscale–micronscale rafts are considered to be surrounded by radially spreading actin filaments. Such structures are termed "asters," and were visualized in vitro ( Gowrishankar et al. , 2012 ) but apparently not in vivo. The asters are colocalized with myosin II and filamin. For beautiful and clear presentations of asters and the actin meshwork on the PM, see Figure 1b of Honigmann and Pralle (2016 ). How can these actin filaments located on the PM inner surface become coupled to the hetero-GPI-AP complex rafts in the PM outer leaflet? Mayor's group proposed that clusters of phosphatidylserine (PS), with long, saturated acyl chains located in the inner leaflet, are bound to actin filaments and also coupled to the submicronscale–micronscale rafts containing heteromixtures of GPI-APs in the outer leaflet by interdigitation in the middle of the bilayer ( Raghupathy et al. , 2015 ). Because CHOL is much shorter than the long, saturated chains of GPI-APs and PS ( Figure 1Ba ), the inner surfaces of both the PS and GPI-AP clusters in the central part of the bilayer are rough, and thus once they come together, they might not readily break apart from each other. The transbilayer coupling is enhanced upon the activation of β1 integrin, which triggers actin nucleation via formins and myosin, and the involvement of talin and the mechanotransducer vinculin in the cluster ( Kalappurakkal et al. , 2019 ). Interestingly, GPI-AP clustering was required for integrin-induced cell spreading and migration. Some of the actin nodes detected by superresolution microscopy might be the same as actin asters, but some may differ because myosin II and filamin A are not appreciably colocalized with the actin nodes ( Shirai et al. , 2017 ). Previously, such actin clusters (nodes and asters) were found in vitro ( Köster et al. , 2016 ) or after drug-induced partial actin depolymerization in live cells ( Luo et al. , 2013 ). However, superresolution microscopy of live cells clarified that actin nodes constitutively exist in the actin membrane skeleton within 3.5 nm from the PM inner surface and are dynamically driven by myosin II in an ATP-dependent manner ( Shirai et al. , 2017 ). Involvement of actin membrane skeleton in GPI-AP receptors' raft-based signaling Reports about the involvement of the actin membrane skeleton (or cortical actin filaments) in the signal transduction of GPI-AP receptors have been quite scarce. Here, we summarize the results reported previously ( Suzuki et al. , 2007a , b ), which are rare publications addressing this issue. When CD59, a GPI-AP working as the receptor for the membrane attack complex, is engaged, it forms clusters containing an average of four to five CD59 molecules, which in turn form nano-sized stabilized CD59-cluster rafts ( Figure 2B 1). Experimentally, this process could be mimicked by closely cross-linking the CD59 molecules by the addition of 40-nm gold particles coated with an anti-CD59 monoclonal antibody. Both the natural ligation and artificial cross-linking can trigger very similar intracellular signals, such as PLCγ recruitment to CD59-cluster rafts, leading to IP 3 production and then to intracellular calcium responses ( Figure 2B ). Engaged CD59-cluster rafts exhibited very peculiar behaviors. They repeatedly and continually undergo temporary immobilizations lasting for 0.56 s (exponential lifetime) that are induced by the binding of CD59-cluster rafts to the actin-based membrane skeleton (called stimulation-induced arrest of lateral diffusion = STALL; Figure 2B 2). This binding is induced by Src-family kinases, such as Lyn, which is activated by Giα, the alpha subunit of the inhibitory trimeric G protein, after they are both recruited to CD59-cluster rafts ( Figure 2B 3). The protein(s) phosphorylated by Lyn is unknown, but it (or one of them) could be an as-yet-unknown TM protein X, which might mediate the PLCγ recruitment to the site on the PM inner surface linked to the CD59-cluster raft located in the PM outer leaflet ( Figure 2B 4). Meanwhile, a cytoplasmic protein Y might also be involved in this process ( Figure 2B 5). Very interestingly, PLCγ is recruited to the CD59-cluster only when the CD59-cluster raft is undergoing temporary immobilization due to its binding to the actin-based membrane skeleton ( Figure 2B 6). These results suggest the possibility that signal transduction platforms for GPI-APs might abundantly exist on the actin-based membrane skeleton. When the CD59-cluster raft meets Giα and Lyn on the platform, it becomes temporarily immobilized on the signaling platform by binding to proteins X and Y, which are in turn phosphorylated by Lyn, producing the binding sites for PLCγ. PLCγ recruited to CD59-cluster rafts produces IP 3 and diacylglycerol, leading to intracellular Ca 2+ mobilization and PKC activation ( Figure 2B 7). Detailed explanation 5: TM proteins partitioning into raft domains and TM receptors taking advantage of raft domains for their signaling Note that some TM proteins have been proposed to partition into raft domains ( Levental et al. , 2010 ). Lorent et al. (2017 ) identified three physical features of the TM domains that independently affect the raft partitioning of TM proteins: smaller surface area, longer TM domain, and palmitoylation. More specifically, they represent smaller protruding amino acid side chains from the α-helical TM domain, matching of the hydrophobic length of the TM domain with the extended saturated acyl chains (due to the association with CHOL), and smoothing of the rugged TM domain surface by the palmitoyl chains, respectively. These features might be coupled with the dimerization–oligomerization of TM proteins linked by palmitoyl chains. Rhodopsin, a G protein–coupled receptor with seven TM domains, has two covalently linked palmitoyl chains, but becomes associated with raft domains only upon dimer formation ( Seno and Hayashi, 2017 ; Hayashi et al. , 2019 ). This result suggests that the rugged parts of the TM domain surface of the dimer are more easily covered with four palmitate chains than those of the monomer with two palmitate chains. Immune receptors, including high-affinity Fcε receptors (FcεRI) in mast cells and B-cell receptors, take advantage of the raft domains that the ligand-induced receptor oligomers generate in/around them to enhance their downstream signaling. The involvement of raft domains in immune receptor signaling had long been suspected, but it was difficult to show unequivocally. However, this difficulty was solved by developing/improving microscopic imaging-analysis methods. By improving imaging fluorescence correlation spectroscopy, Baird and colleagues revealed that stabilized Lo-like nanoscale raft domains are created around clustered FcεRI in immune mast cells ( Bag et al. , 2021 ). These domains strongly augment the Lyn recruitment and suppress the recruitment of the TM phosphatase PTPα. Proof of the Lyn recruitment to clustered FcεRI had previously been elusive, but extensive examinations of very many molecular trajectories (several tens of thousands), made possible by the improvement and application of imaging fluorescence correlation spectroscopy, unequivocally revealed Lyn recruitment to clustered FcεRI. This result is consistent with the enhanced recruitment of sphingomyelin, a prototypical raftophilic phospholipid, to antigen-induced clusters of FcεRI but not to FcεRI in resting cells, found by single-molecule imaging tracking of a newly developed fluorescently labeled sphingomyelin ( Kinoshita et al. , 2017 ). In immune B cells, by applying a pair cross-correlation analysis method to superresolution single-molecule localization microscopy data ( Stone and Veatch, 2014 ), Veatch and colleagues demonstrated that stimulation-induced B-cell receptor (BCR) clusters induce Lo-phase-like domains around them, which are capable of sorting key regulators of BCR activation ( Stone et al. , 2017 ). Taken together, these results provide evidence for the role of the receptor clustering–induced membrane domains and a plausible mechanism for recruiting downstream signaling molecules to the receptor clusters. They agree well with the signaling raft domain formation induced by the clustering of a GPI-AP, CD59 ( Suzuki et al. , 2007a , b , 2012 ) and suggest that similar mechanisms are employed in many other receptor signaling pathways ( Kusumi et al. , 2004 ; Koyama-Honda et al. , 2020 ). For other examples in which raft-based interactions might be involved in the signal transduction by TM receptors, refer to the following papers: Sohn et al. (2008) ; Coskun et al. (2011) ; Chung et al. (2016) ; Shelby et al. (2016) . CONCLUDING REMARKS We propose two major updates of the Singer–Nicolson fluid mosaic model. One is the CHOL-induced cooperativity-based assembly of CHOL and molecules containing saturated acyl chains, which represents the transition of the fluid mosaic model from a simple liquid model to a model of the liquid containing various metastable liquid-like clusters of 2–20 nm in diameter, called raft domains. The formation of such lower-nanoscale liquid rafts is enhanced by homophilic protein–protein interactions in the case of GPI-APs, leading to the formation of GPI-AP homodimer rafts, which are one of the basic unit rafts involved in building greater raft domains containing GPI-APs. Upon extracellular stimulation, GPI-AP receptors form ligand-induced clusters, which induce greater and stabilized raft domains with the help of raft-lipid interactions, and these enlarged, stabilized raft domains become responsible for the downstream signaling. Note that the cooperativity for the raft domain formation is due not only to the accommodating interactions between CHOL and saturated acyl chains, but also to the nonconformable interactions between CHOL and unsaturated acyl chains + proteins' TM domains. The other update is the compartmentalization of the entire PM by the actin-based membrane skeleton (fence), as well as TM proteins anchored to and aligned along the membrane skeleton (pickets). This compartmentalization occurs throughout the entire PM with compartment sizes between 30 and 230 nm, which depends on the cell type. However, within each compartment, the fluid mosaic model with the upgrade of the presence of many nanoscale liquid raft domains is perfectly correct. Raft domains are excluded from the picket-fence compartmental boundary areas due to their nonconformability with the TM pickets. Therefore, the raft domains tend to be localized within a compartment and thus their sizes are smaller than the compartment sizes. However, the stabilized raft domains induced by the engaged GPI-AP receptors might be able to indirectly associate with the actin-based membrane skeleton by way of the signal transduction platform bound to the actin-based membrane skeleton. This was deduced by the finding that the stabilized raft domains induced by the engaged GPI-AP receptor clusters undergo actin-dependent temporary immobilizations and signal transduction during immobilization. In this review, we have advanced the argument that the CHOL- and actin-centered views of the PM provide excellent perspectives for understanding PM structure, molecular dynamics, and functions, while coping with the VUCA of the PM. We hope that readers have been convinced of the usefulness of these views and that they will start considering the PM accordingly.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4285945/

Pharmacokinetics and pharmacodynamics of levofloxacin injection in healthy Chinese volunteers and dosing regimen optimization

What is known and objective The pharmacokinetics (PK) and pharmacodynamics (PD) of levofloxacin were investigated following administration of levofloxacin injection in healthy Chinese volunteers for optimizing dosing regimen. Methods The PK study included single-dose (750 mg/150 mL) and multiple-dose (750 mg/150 mL once daily for 7 days) phases. The concentration of levofloxacin in blood and urine was determined using HPLC method. Both non-compartmental and compartmental analyses were performed to estimate PK parameters. Taking f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 as a target, the cumulative fraction of response (CFR) of levofloxacin 750 mg for treatment of community-acquired pneumonia (CAP) was calculated using Monte Carlo simulation. The probability of target attainment (PTA) of levofloxacin at various minimal inhibitory concentrations (MICs) was also evaluated. Results and discussion The results of PK study showed that the C max and AUC 0–∞ of levofloxacin were 14·94 μg/mL and 80·14 μg h/mL following single-dose infusion of levofloxacin. The half-life and average cumulative urine excretion ratio within 72 h post-dosing were 7·75 h and 86·95%, respectively. The mean C ss,max , C ss,min and AUC 0–τ of levofloxacin at steady state following multiple doses were 13·31 μg/mL, 0·031 μg/mL and 103·7 μg h/mL, respectively. The accumulation coefficient was 1·22. PK/PD analysis revealed that the CFR value of levofloxacin 750-mg regimen against Streptococcus pneumoniae was 96·2% and 95·4%, respectively, in terms of f C max /MIC and f AUC/MIC targets. What is new and conclusion The regimen of 750-mg levofloxacin once daily provides a satisfactory PK/PD profile against the main pathogenic bacteria of CAP, which implies promising clinical and bacteriological efficacy for patients with CAP. A large-scale clinical study is warranted to confirm these results. What is known and objective The pharmacokinetics (PK) and pharmacodynamics (PD) of levofloxacin were investigated following administration of levofloxacin injection in healthy Chinese volunteers for optimizing dosing regimen. Methods The PK study included single-dose (750 mg/150 mL) and multiple-dose (750 mg/150 mL once daily for 7 days) phases. The concentration of levofloxacin in blood and urine was determined using HPLC method. Both non-compartmental and compartmental analyses were performed to estimate PK parameters. Taking f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 as a target, the cumulative fraction of response (CFR) of levofloxacin 750 mg for treatment of community-acquired pneumonia (CAP) was calculated using Monte Carlo simulation. The probability of target attainment (PTA) of levofloxacin at various minimal inhibitory concentrations (MICs) was also evaluated. Results and discussion The results of PK study showed that the C max and AUC 0–∞ of levofloxacin were 14·94 μg/mL and 80·14 μg h/mL following single-dose infusion of levofloxacin. The half-life and average cumulative urine excretion ratio within 72 h post-dosing were 7·75 h and 86·95%, respectively. The mean C ss,max , C ss,min and AUC 0–τ of levofloxacin at steady state following multiple doses were 13·31 μg/mL, 0·031 μg/mL and 103·7 μg h/mL, respectively. The accumulation coefficient was 1·22. PK/PD analysis revealed that the CFR value of levofloxacin 750-mg regimen against Streptococcus pneumoniae was 96·2% and 95·4%, respectively, in terms of f C max /MIC and f AUC/MIC targets. What is new and conclusion The regimen of 750-mg levofloxacin once daily provides a satisfactory PK/PD profile against the main pathogenic bacteria of CAP, which implies promising clinical and bacteriological efficacy for patients with CAP. A large-scale clinical study is warranted to confirm these results. What is Known and Objective Levofloxacin, a levoisomer of ofloxacin, belongs to fluoroquinolone antimicrobial agents with broad-spectrum antibacterial activity. It is highly active against Gram-negative aerobic bacteria including most of the Enterobacteriaceae species. Levofloxacin also provides good activity for Gram-positive aerobic bacteria. Moreover, some atypical pathogens are susceptible to levofloxacin. 1 Levofloxacin is increasingly used in the treatment for community-acquired pneumonia (CAP) due to its potent bactericidal effect, higher blood drug concentration, extensive tissue distribution and higher bioavailability. 2 – 5 The pharmacokinetic/pharmacodynamic (PK/PD) studies in recent years have proved that levofloxacin is a concentration-dependent antimicrobial agent. Its bactericidal activity increases with drug concentration over a certain range. It is proposed that the maximal bactericidal effect may be achieved by delivering fewer or single but high-level dose per day to maintain higher C max /MIC and AUC 0–24 /MIC ratios. Such a regimen may also be helpful to avoid the emergence of antibiotic resistance. For this reason, PK studies have been conducted for levofloxacin 750-mg injection (150 mL) in the United States and European countries. 6 , 7 Intravenous administration of levofloxacin 750-mg injection (150 mL) once daily for 5 days achieved the same efficacy in severe CAP as the regimen of 500 mg once daily for 10 days. This new regimen has a shorter duration of treatment and lower exposure, less medical cost and may be associated with lower risk of developing resistance. 8 , 9 However, no multiple-dose PK study of levofloxacin 750-mg IV infusion has been reported in Chinese subjects yet. It is necessary to clarify the PK/PD and safety profile of levofloxacin 750-mg IV infusion in Chinese volunteers for optimizing the dosing regimen. This study was designed to explore the PK profiles of levofloxacin following single-dose and multiple-dose intravenous infusion of levofloxacin 750 mg for the first time in Chinese healthy volunteers. Meanwhile, considering the PD data of in vitro antimicrobial susceptibility testing for major CAP pathogens such as Streptococcus pneumoniae and Haemophilus influenzae , 10 Monte Carlo simulation (MCS) method was used to perform PK/PD analysis. The results were compared with the PK/PD profile of levofloxacin 500-mg injection. An optimized dosing regimen, which is expected to achieve the maximal bactericidal effect in vivo , is recommended for patients with severe CAP. Methods Subjects In the open-label single-dose study, nine enrolled healthy subjects received levofloxacin 750 mg/150 mL via intravenous infusion. In the randomized, double-blind, multiple-dose phase, 12 healthy subjects received levofloxacin 750 mg/150 mL or placebo (ratio of 3 : 1) via intravenous infusion over 90 min, once daily for seven consecutive days. The study protocol and informed consent were approved by the Institutional Ethics Committee. All subjects signed the written informed consent before participation in this study. The nine male subjects in single-dose study were 23·51 ± 2·00 years of age and 65·67 ± 4·60 kg of body weight at average. Their creatinine clearance (CL cr ) was 138·0 ± 13·28 mL/min before study administration. The nine male subjects receiving levofloxacin in multiple-dose study were 24·56 ± 2·41 years of age and 60·39 ± 4·61 kg of body weight at average. Their CL cr was 125·9 ± 16·5 mL/min before study initiation. The three male subjects receiving placebo were 26·24 ± 0·98 years of age and 62·00 ± 4·27 kg of body weight. Their CL cr was 116·6 ± 7·6 mL/min. Antimicrobial agent Levofloxacin 750-mg injection (Cravit, Lot: EXH0701, 150 mL/bottle containing levofloxacin 0·75 g and NaCl 1·35 g) was the product of Daiichi Pharmaceutical (Beijing, China). Placebo was 0·9% sodium chloride injection [Shanghai Huayuan Changfu Pharmaceutical (Group) Company, Lot: 07071104–1, 250 mL/bottle]. Single-dose study Single dose of levofloxacin 750 mg was administered intravenously with an infusion pump at constant rate over 1·5 h. Blood sample (5 mL) was drawn before infusion, 0·75 h after infusion start and immediately, 0·25, 0·5, 1, 1·5, 2, 3, 4, 8, 12, 24, 36, 48, 60 and 72 h after the end of infusion. All urine was collected according to the following intervals: before infusion, 0–2, 2–4, 4–8, 8–12, 12–24, 24–48 and 48–72 h after the end of infusion. Multiple-dose study Levofloxacin 750 mg was administered intravenously to the volunteers once daily with an infusion pump at constant rate over 1·5 h for seven consecutive days. Blood sample (5 mL) was drawn before infusion, 0·75 h after infusion start and immediately, 0·25, 0·5, 1, 1·5, 2, 3, 4, 8 and 12 h after the end of infusion on Day 1, immediately before infusion (0 h) and immediately after the end of infusion from Day 2 to Day 6. The sampling time points on Day 7 were the same as that in single-dose study. All urine was collected on Day 1 and Day 7 according to the following intervals: before infusion, 0–2, 2–4, 4–8, 8–12 and 12–24 h. Urine volume was measured and recorded. Sample processing and data collection All the blood samples were put on ice bath and subjected to centrifugation within 30 min at 4 °C (1500 × g, 10 min) to separate plasma. The plasma was aliquoted to two tubes and stored at −40 ± 5 °C in a refrigerator for later analysis. The volunteers were observed for any adverse event since the start of infusion, including clinical observation, vital signs (temperature, blood pressure and pulse rate), laboratory tests (haematology, urinalysis and serum biochemical tests), 12-lead electrocardiogram for safety evaluation of levofloxacin. Levofloxacin assay High-performance liquid chromatography (HPLC) (Waters 2695 HPLC System, Waters Company, Milford, MA, USA) was used to determine levofloxacin concentration in blood and urine samples. The solid phase was TSK-gel ODS-80TM gel column (4·6 mm × 150 mm, 5 μm). The mobile phase contained 50 mmol/L KH 2 PO 4 (pH 2·0), THF and 1 m CH 3 COONH 4 (92/7/1, V/V/V ), respectively. The wavelength for fluorescence detection E x /E m was 296 nm/504 nm, column temperature was 35 °C, flow rate was 1·0 mL/min, and injection volume for plasma and urine samples was 10 μL and 20 μL, respectively. The retention time was about 6·5 min for levofloxacin and 8·5 min for the internal standard (DL-8493, lot number 900131, provided by Daiichi Pharmaceutical). The range of standard curve was from 0·010 to 5·000 μg/mL for plasma and 0·100 to 100·0 μg/mL for urine. The curve linearity was 0·9989 and 0·9998, respectively. The assay method and its validation were previously described by Zhang J et al . 11 Pharmacokinetic calculation Phoenix WinNonlin 6.0 software (Pharsight Company, Sunnyvale, CA, USA) was used to calculate the non-compartmental PK parameters of levofloxacin following administration of levofloxacin. Peak concentration (C max ) and C 24 h were expressed as the observed values. AUC 0–t was calculated based on the trapezoidal rule. Linear regression analysis was conducted using the terminal logarithmic concentration–time data to calculate the elimination rate constant (λ). The terminal elimination half-life (T 1/2 ) was 0·693/λ. AUC 0–∞ was calculated as AUC 0–t + terminal concentration/λ. The total clearance (CL t ) was calculated as the ratio of dose to AUC 0–∞ . Renal clearance (CL r ) was the ratio of levofloxacin cumulative urinary excretion to AUC 0–24 h . The apparent volume of distribution (V d ) is the ratio of CL t to λ. The mean residence time (MRT 0–∞ ) was based on AUMC/AUC ratio. Two-compartment model was employed to analyse the PK profile of levofloxacin in healthy volunteers using WinNonlin software. The removal of levofloxacin from central compartment to peripheral compartment and the transport between central compartment and peripheral compartment were all consistent with first-order kinetics. The elimination rate and clearance were represented by K e and CL. The corresponding rate constant was expressed as K 12 and K 21 , respectively. The clearance between compartments was expressed as Q. The distribution volume of central and peripheral compartments was represented by V 1 and V 2 . The rate of distribution phase and elimination phase was represented by α and β, whereas the corresponding half-life was expressed as T 1/2 , α and T 1/2, β . The weight of 1/C was used in the calculation. Statistical analysis The non-compartmental parameters of levofloxacin following multiple doses were compared between Day 1 and Day 7 to evaluate the PK behaviours over time following multiple doses, for example, the change in T 1/2 and drug accumulation. The compartmental parameters of levofloxacin following single- or multiple-dose administration were compared to analyse the details of levofloxacin PK profile. The correlation between compartmental parameters of levofloxacin and the relevant demographic covariates was also evaluated to preliminarily examine the potential factors influencing the disposition of levofloxacin in healthy adults. The data sets were examined by F -test for homogeneity of variance. The homogeneous data were compared by t -test between groups. Statistical significance was assumed at the level of P < 0·05. The above statistical analyses were completed with spss software package (Version 13.0, ibm spss Software, Armonk, NY, USA). The results were expressed as Mean ± SD. Safety evaluation The Criteria for Evaluation of Adverse Drug Reactions and Laboratory Abnormalities due to Antimicrobial Agents in Clinical Study issued by Japanese Society of Chemotherapy were referred to in this study to confirm the adverse events and the severity (mild, moderate and severe). The relationship between an adverse event and the study drug was judged as definitely, probably, possibly related, or possibly unrelated, unrelated or indeterminate. The events definitely, probably, possibly related to the study drug or indeterminate were classified as adverse drug reactions. In this study, the adverse events leading to study discontinuation, or requiring treatment, or resulting in QTc prolongation (≥60 msec longer or QTc interval ≥500 msec) were taken as significant adverse events. PK/PD analysis and dosing regimen evaluation PK/PD analysis was performed and evaluated for levofloxacin 750-mg dosing regimen based on the PK parameters of levofloxacin at steady state following dosing of 750 mg daily for seven consecutive days in this study, and the pharmacodynamic data from two clinical trials were as follows: levofloxacin tablet 500 mg in 881 patients with CAP or other lower respiratory tract infections during 2006–2007, and randomized controlled study of levofloxacin injection 750 mg vs. 500 mg in CAP during 2007–2008. 10 Monte Carlo simulation was used to conduct the PK/PD analysis. The simulation was performed in 5000 patients with MATLAB software version 7.0.1 (Mathworks Company, Natick, MA, USA). It is expected to obtain a target PK/PD value indicating satisfactory clinical and bacteriological efficacy if f C max /MIC ≥5 and/or f AUC 24 h /MIC ≥30. 12 The protein-binding rate of levofloxacin is 30%; hence, the fraction of free drug ( f ) is 0·7. The cumulative fraction of response (CFR) was calculated for 750-mg levofloxacin regimen in attaining the above target PK/PD value specific for major CAP pathogens. The susceptibility break point of levofloxacin against Streptococcus pneumoniae is ≤2 μg/mL. 13 Therefore, this calculation is the pharmacodynamic probability of target attainment (PTA) in terms of f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 when MIC = 0·25, 0·5, 1 or 2 μg/mL. In Monte Carlo simulation, the simulated steady-state AUC 0–24 h and C max data were generated based on logarithmic normal distribution. The simulated MIC data were generated based on discrete distribution according to specified probability at each MIC level. Subjects In the open-label single-dose study, nine enrolled healthy subjects received levofloxacin 750 mg/150 mL via intravenous infusion. In the randomized, double-blind, multiple-dose phase, 12 healthy subjects received levofloxacin 750 mg/150 mL or placebo (ratio of 3 : 1) via intravenous infusion over 90 min, once daily for seven consecutive days. The study protocol and informed consent were approved by the Institutional Ethics Committee. All subjects signed the written informed consent before participation in this study. The nine male subjects in single-dose study were 23·51 ± 2·00 years of age and 65·67 ± 4·60 kg of body weight at average. Their creatinine clearance (CL cr ) was 138·0 ± 13·28 mL/min before study administration. The nine male subjects receiving levofloxacin in multiple-dose study were 24·56 ± 2·41 years of age and 60·39 ± 4·61 kg of body weight at average. Their CL cr was 125·9 ± 16·5 mL/min before study initiation. The three male subjects receiving placebo were 26·24 ± 0·98 years of age and 62·00 ± 4·27 kg of body weight. Their CL cr was 116·6 ± 7·6 mL/min. Antimicrobial agent Levofloxacin 750-mg injection (Cravit, Lot: EXH0701, 150 mL/bottle containing levofloxacin 0·75 g and NaCl 1·35 g) was the product of Daiichi Pharmaceutical (Beijing, China). Placebo was 0·9% sodium chloride injection [Shanghai Huayuan Changfu Pharmaceutical (Group) Company, Lot: 07071104–1, 250 mL/bottle]. Single-dose study Single dose of levofloxacin 750 mg was administered intravenously with an infusion pump at constant rate over 1·5 h. Blood sample (5 mL) was drawn before infusion, 0·75 h after infusion start and immediately, 0·25, 0·5, 1, 1·5, 2, 3, 4, 8, 12, 24, 36, 48, 60 and 72 h after the end of infusion. All urine was collected according to the following intervals: before infusion, 0–2, 2–4, 4–8, 8–12, 12–24, 24–48 and 48–72 h after the end of infusion. Multiple-dose study Levofloxacin 750 mg was administered intravenously to the volunteers once daily with an infusion pump at constant rate over 1·5 h for seven consecutive days. Blood sample (5 mL) was drawn before infusion, 0·75 h after infusion start and immediately, 0·25, 0·5, 1, 1·5, 2, 3, 4, 8 and 12 h after the end of infusion on Day 1, immediately before infusion (0 h) and immediately after the end of infusion from Day 2 to Day 6. The sampling time points on Day 7 were the same as that in single-dose study. All urine was collected on Day 1 and Day 7 according to the following intervals: before infusion, 0–2, 2–4, 4–8, 8–12 and 12–24 h. Urine volume was measured and recorded. Sample processing and data collection All the blood samples were put on ice bath and subjected to centrifugation within 30 min at 4 °C (1500 × g, 10 min) to separate plasma. The plasma was aliquoted to two tubes and stored at −40 ± 5 °C in a refrigerator for later analysis. The volunteers were observed for any adverse event since the start of infusion, including clinical observation, vital signs (temperature, blood pressure and pulse rate), laboratory tests (haematology, urinalysis and serum biochemical tests), 12-lead electrocardiogram for safety evaluation of levofloxacin. Levofloxacin assay High-performance liquid chromatography (HPLC) (Waters 2695 HPLC System, Waters Company, Milford, MA, USA) was used to determine levofloxacin concentration in blood and urine samples. The solid phase was TSK-gel ODS-80TM gel column (4·6 mm × 150 mm, 5 μm). The mobile phase contained 50 mmol/L KH 2 PO 4 (pH 2·0), THF and 1 m CH 3 COONH 4 (92/7/1, V/V/V ), respectively. The wavelength for fluorescence detection E x /E m was 296 nm/504 nm, column temperature was 35 °C, flow rate was 1·0 mL/min, and injection volume for plasma and urine samples was 10 μL and 20 μL, respectively. The retention time was about 6·5 min for levofloxacin and 8·5 min for the internal standard (DL-8493, lot number 900131, provided by Daiichi Pharmaceutical). The range of standard curve was from 0·010 to 5·000 μg/mL for plasma and 0·100 to 100·0 μg/mL for urine. The curve linearity was 0·9989 and 0·9998, respectively. The assay method and its validation were previously described by Zhang J et al . 11 Pharmacokinetic calculation Phoenix WinNonlin 6.0 software (Pharsight Company, Sunnyvale, CA, USA) was used to calculate the non-compartmental PK parameters of levofloxacin following administration of levofloxacin. Peak concentration (C max ) and C 24 h were expressed as the observed values. AUC 0–t was calculated based on the trapezoidal rule. Linear regression analysis was conducted using the terminal logarithmic concentration–time data to calculate the elimination rate constant (λ). The terminal elimination half-life (T 1/2 ) was 0·693/λ. AUC 0–∞ was calculated as AUC 0–t + terminal concentration/λ. The total clearance (CL t ) was calculated as the ratio of dose to AUC 0–∞ . Renal clearance (CL r ) was the ratio of levofloxacin cumulative urinary excretion to AUC 0–24 h . The apparent volume of distribution (V d ) is the ratio of CL t to λ. The mean residence time (MRT 0–∞ ) was based on AUMC/AUC ratio. Two-compartment model was employed to analyse the PK profile of levofloxacin in healthy volunteers using WinNonlin software. The removal of levofloxacin from central compartment to peripheral compartment and the transport between central compartment and peripheral compartment were all consistent with first-order kinetics. The elimination rate and clearance were represented by K e and CL. The corresponding rate constant was expressed as K 12 and K 21 , respectively. The clearance between compartments was expressed as Q. The distribution volume of central and peripheral compartments was represented by V 1 and V 2 . The rate of distribution phase and elimination phase was represented by α and β, whereas the corresponding half-life was expressed as T 1/2 , α and T 1/2, β . The weight of 1/C was used in the calculation. Statistical analysis The non-compartmental parameters of levofloxacin following multiple doses were compared between Day 1 and Day 7 to evaluate the PK behaviours over time following multiple doses, for example, the change in T 1/2 and drug accumulation. The compartmental parameters of levofloxacin following single- or multiple-dose administration were compared to analyse the details of levofloxacin PK profile. The correlation between compartmental parameters of levofloxacin and the relevant demographic covariates was also evaluated to preliminarily examine the potential factors influencing the disposition of levofloxacin in healthy adults. The data sets were examined by F -test for homogeneity of variance. The homogeneous data were compared by t -test between groups. Statistical significance was assumed at the level of P < 0·05. The above statistical analyses were completed with spss software package (Version 13.0, ibm spss Software, Armonk, NY, USA). The results were expressed as Mean ± SD. Safety evaluation The Criteria for Evaluation of Adverse Drug Reactions and Laboratory Abnormalities due to Antimicrobial Agents in Clinical Study issued by Japanese Society of Chemotherapy were referred to in this study to confirm the adverse events and the severity (mild, moderate and severe). The relationship between an adverse event and the study drug was judged as definitely, probably, possibly related, or possibly unrelated, unrelated or indeterminate. The events definitely, probably, possibly related to the study drug or indeterminate were classified as adverse drug reactions. In this study, the adverse events leading to study discontinuation, or requiring treatment, or resulting in QTc prolongation (≥60 msec longer or QTc interval ≥500 msec) were taken as significant adverse events. PK/PD analysis and dosing regimen evaluation PK/PD analysis was performed and evaluated for levofloxacin 750-mg dosing regimen based on the PK parameters of levofloxacin at steady state following dosing of 750 mg daily for seven consecutive days in this study, and the pharmacodynamic data from two clinical trials were as follows: levofloxacin tablet 500 mg in 881 patients with CAP or other lower respiratory tract infections during 2006–2007, and randomized controlled study of levofloxacin injection 750 mg vs. 500 mg in CAP during 2007–2008. 10 Monte Carlo simulation was used to conduct the PK/PD analysis. The simulation was performed in 5000 patients with MATLAB software version 7.0.1 (Mathworks Company, Natick, MA, USA). It is expected to obtain a target PK/PD value indicating satisfactory clinical and bacteriological efficacy if f C max /MIC ≥5 and/or f AUC 24 h /MIC ≥30. 12 The protein-binding rate of levofloxacin is 30%; hence, the fraction of free drug ( f ) is 0·7. The cumulative fraction of response (CFR) was calculated for 750-mg levofloxacin regimen in attaining the above target PK/PD value specific for major CAP pathogens. The susceptibility break point of levofloxacin against Streptococcus pneumoniae is ≤2 μg/mL. 13 Therefore, this calculation is the pharmacodynamic probability of target attainment (PTA) in terms of f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 when MIC = 0·25, 0·5, 1 or 2 μg/mL. In Monte Carlo simulation, the simulated steady-state AUC 0–24 h and C max data were generated based on logarithmic normal distribution. The simulated MIC data were generated based on discrete distribution according to specified probability at each MIC level. Results Levofloxacin pharmacokinetics The PK parameters in the nine subjects following single intravenous dose of levofloxacin 750 mg showed mean C max (14·94 ± 1·48) μg/mL, AUC 0–72 h (79·97 ± 12·53) μg h/mL and AUC 0–∞ (80·14 ± 12·55) μg h/mL. The terminal elimination half-life (t 1/2 ) was (7·75 ± 0·59) h. Renal clearance (CL r ) and CL t were (8·17 ± 1·00) L/h and (9·58 ± 1·62) L/h, respectively. The distribution volume (V d ) was (107·30 ± 20·73) L. The cumulative urinary excretion of levofloxacin was (86·95 ± 5·04)% of the administered dose 72 h post-dose (Table 1 ). Table 1 Non-compartmental and compartmental parameters of levofloxacin following intravenous infusion of levofloxacin 750 mg in healthy Chinese volunteers ( n = 9, Mean ± SD) Non-compartmental Compartmental Parameter Single-dose PK Multiple-dose PK Parameter Single-dose PK Multiple-dose PK First Dose Last Dose C max (μg/mL) 14·9 ± 1·48 17·7 ± 4·22 13·3 ± 2·77 c T 1/2,α (h) 1·02 ± 0·323 0·305 ± 0·0473 e AUC 0–24 h (μg·h/mL) 73·3 ± 10·7 90·1 ± 11·9 b 94·1 ± 15·2 b T 1/2,β (h) 7·19 ± 0·503 7·22 ± 0·783 AUC 0–72 h (μg·h/mL) 80·0 ± 12·5 – 103·7 ± 17·9 K 12 (1/h) 0·220 ± 0·120 1·26 ± 0·336 e AUC 0–∞ (μg·h/mL) 80·1 ± 12·5 99·3 ± 13·7 b 104 ± 18·0 b K 21 (1/h) 0·465 ± 0·119 0·897 ± 0·237 e T 1/2 (h) 7·75 ± 0·591 7·52 ± 0·656 6·91 ± 0·806 a , d T 1/2,Ke (h) 4·62 ± 0·655 2·82 ± 0·854 e CL t (L/h) 9·58 ± 1·62 8·46 ± 1·10 8·27 ± 1·31 CL (L/h) 10·2 ± 1·89 7·84 ± 1·10 e CL r (L/h) 8·17 ± 1·28 6·40 ± 0·972 b 6·85 ± 1·06 a Q (L/h) 13·8 ± 5·60 37·2 ± 6·05 e V d (L) 107·3 ± 20·7 91·3 ± 9·32 81·7 ± 10·5 b , d V 1 (L) 67·8 ± 13·2 31·6 ± 9·68 e MRT 0–∞ (h) 9·18 ± 0·786 9·72 ± 1·04 10·0 ± 0·97 V 2 (L) 29·2 ± 7·61 43·4 ± 9·97 e a P < 0·05. b P < 0·01 vs. single-dose PK (non-compartmental parameter). c P < 0·05. d P < 0·01 vs. multiple-dose PK (first dose, non-compartmental parameter). e P < 0·01 vs. single-dose PK (compartmental parameter); the duration of levofloxacin infusion was 1·5 h. The dosing regimen for the multiple-dose PK study was q24 h × 7 days. In the nine subjects receiving multiple-dose intravenous infusion of levofloxacin 750 mg once daily for seven consecutive days, the C max was (17·69 ± 4·23) μg/mL on Day 1 and (13·31 ± 2·77) μg/mL on Day 7. The plasma concentration of levofloxacin at 24 h was (0·87 ± 0·22) μg/mL on Day 1 and (0·803 ± 0·258) μg/mL on Day 7. The corresponding AUC 0–24 h was (90·05 ± 11·92) μg·h/mL and (94·1 ± 15·2) μg h/mL. The mean 24-hour cumulative urinary excretion rate was (75·69 ± 6·34)% and (84·16 ± 5·14)% of the administered dose, respectively. When steady state was reached following consecutive dosing, the mean C ss,max of levofloxacin was 13·31 μg/mL, mean C ss,min 0·031 μg/mL, average plasma concentration (C avg ) (3·87 ± 0·62) μg/mL, mean AUC 0–τ 94·12 μg h/mL (τ = 24 h) and accumulation factor (1·22 ± 0·14) (Table 1 ). Further analysis indicated that the ratio of geometric mean C max , C 24 h , AUC 0–24 h (Day 7/Day 1) and the corresponding 95% confidence interval were 0·76 (0·61–0·95), 0·91 (0·78–1·06) and 1·04 (0·97–1·11), suggesting no apparent accumulation in the body when steady state was reached following once-daily administration of levofloxacin for 7 days. Compartmental model analysis Two-compartment model well described the PK profile of levofloxacin in healthy volunteers following single- or multiple-dose administration (Fig. 1 ). The results showed that the distribution half-life after single dose of levofloxacin was (1·02 ± 0·32) h. The distribution volume of central and peripheral compartments was (67·8 ± 13·2) L and (29·2 ± 7·61) L, respectively. The clearance from central compartment was (10·2 ± 1·89) L/h and (13·8 ± 5·6) L/h between central and peripheral compartments. Fig 1 Concentration–time profiles of levofloxacin following single- or multiple-dose infusion of 750-mg levofloxacin in healthy Chinese volunteers (Mean ± SD, n = 9). The observed values and fittings obtained from two-compartment model were represented by dots and lines , whereas red and blue symbols represent single-dose PK group and multiple-dose PK group, respectively. The infusion time of single-dose levofloxacin was 1·5 h. The dosing regimen for multiple PK study was q24 h × 7 days. Safety of levofloxacin Six ADRs were reported in 6 of the 9 subjects receiving single dose of levofloxacin. The ADR was local adverse reaction in four subjects, that is, mild injection site hyperaemia, which started within 16–45 min since intravenous infusion, and resolved 30 min after the end of intravenous infusion. Both the increased lactate dehydrogenase in one subject and the conjugated bilirubin elevation in another subject developed 23 h after the end of intravenous infusion and normalized at Day 7 post-dose visit. All these ADRs were mild, transient and well tolerated. Nine adverse events occurred in 8 of the 9 subjects receiving multiple doses of levofloxacin. Eight of these events in seven subjects were ADRs. Six ADRs were intermittent mild injection site hyperaemia during the consecutive 7-day intravenous infusion and resolved within 10–25 min after the end of intravenous infusion the same day. Two cases of laboratory abnormalities were transient creatine kinase elevation and increased total bilirubin. The increase was within two times the upper limit of normal in both cases. These ADRs were mild in severity and well tolerated. One of the three subjects receiving placebo developed two mild adverse events (decreased white blood cell count and increased uric acid), which were unrelated to the study drug. Other laboratory tests, vital signs and 12-lead ECG parameters were all within the normal range in both single-dose and multiple-dose studies. No fatal or other significant adverse events were found. There was no significant adverse event leading to study discontinuation or abnormal QTc interval. PK/PD analysis and dosing regimen evaluation Figure 2 displays the CFR value of levofloxacin targeting 245 strains of Streptococcus pneumoniae , Haemophilus influenzae , Klebsiella pneumoniae , methicillin-susceptible Staphylococcus aureus (MSSA) or methicillin-resistant Staphylococcus aureus (MRSA). The CFR of levofloxacin 750-mg once-daily regimen was 96·2%, 98·5%, 95·3% and 100·0% for Streptococcus pneumoniae , Haemophilus influenzae , Klebsiella pneumoniae and MSSA in terms of f C max /MIC = 5, significantly better than that for MRSA (0·4%). The CFR was 95·4%, 98·0%, 95·2% and 100·0% in terms of f AUC 24 h /MIC = 30, whereas for MRSA, the CFR was 0. Fig 2 Cumulative fraction of response of f C max /MIC (a) and f AUC 24 h /MIC (b) for levofloxacin. The dosing regimen of levofloxacin was 500 mg or 750 mg (q.d.) for seven consecutive days, and f indicates the unbound fraction, the value of which was 0·7. From Fig. 3 , when MIC ≤1 μg/mL, the PTA of levofloxacin 750-mg dosing regimen was kept at 100% in terms of the target value of f C max /MIC or f AUC 24 h /MIC, suggesting that levofloxacin 750-mg dosing regimen can promise good clinical and microbiological efficacy targeting the bacteria for which levofloxacin MIC ≤1 μg/mL. Fig 3 Distribution of CAP pathogens in terms of MIC level and the PTA of f C max /MIC (a) and f AUC 24 h /MIC (b) for levofloxacin following multiple dosing. The dosing regimen of levofloxacin was 500 mg or 750 mg (q.d.) for 7 consecutive days, and f indicates the unbound fraction, the value of which was 0·7. Histograms and lines represent the distribution frequency of MIC and PTA values, respectively. In the graph, the distribution data of MIC for CAP pathogens were obtained from the literature. 10 Levofloxacin pharmacokinetics The PK parameters in the nine subjects following single intravenous dose of levofloxacin 750 mg showed mean C max (14·94 ± 1·48) μg/mL, AUC 0–72 h (79·97 ± 12·53) μg h/mL and AUC 0–∞ (80·14 ± 12·55) μg h/mL. The terminal elimination half-life (t 1/2 ) was (7·75 ± 0·59) h. Renal clearance (CL r ) and CL t were (8·17 ± 1·00) L/h and (9·58 ± 1·62) L/h, respectively. The distribution volume (V d ) was (107·30 ± 20·73) L. The cumulative urinary excretion of levofloxacin was (86·95 ± 5·04)% of the administered dose 72 h post-dose (Table 1 ). Table 1 Non-compartmental and compartmental parameters of levofloxacin following intravenous infusion of levofloxacin 750 mg in healthy Chinese volunteers ( n = 9, Mean ± SD) Non-compartmental Compartmental Parameter Single-dose PK Multiple-dose PK Parameter Single-dose PK Multiple-dose PK First Dose Last Dose C max (μg/mL) 14·9 ± 1·48 17·7 ± 4·22 13·3 ± 2·77 c T 1/2,α (h) 1·02 ± 0·323 0·305 ± 0·0473 e AUC 0–24 h (μg·h/mL) 73·3 ± 10·7 90·1 ± 11·9 b 94·1 ± 15·2 b T 1/2,β (h) 7·19 ± 0·503 7·22 ± 0·783 AUC 0–72 h (μg·h/mL) 80·0 ± 12·5 – 103·7 ± 17·9 K 12 (1/h) 0·220 ± 0·120 1·26 ± 0·336 e AUC 0–∞ (μg·h/mL) 80·1 ± 12·5 99·3 ± 13·7 b 104 ± 18·0 b K 21 (1/h) 0·465 ± 0·119 0·897 ± 0·237 e T 1/2 (h) 7·75 ± 0·591 7·52 ± 0·656 6·91 ± 0·806 a , d T 1/2,Ke (h) 4·62 ± 0·655 2·82 ± 0·854 e CL t (L/h) 9·58 ± 1·62 8·46 ± 1·10 8·27 ± 1·31 CL (L/h) 10·2 ± 1·89 7·84 ± 1·10 e CL r (L/h) 8·17 ± 1·28 6·40 ± 0·972 b 6·85 ± 1·06 a Q (L/h) 13·8 ± 5·60 37·2 ± 6·05 e V d (L) 107·3 ± 20·7 91·3 ± 9·32 81·7 ± 10·5 b , d V 1 (L) 67·8 ± 13·2 31·6 ± 9·68 e MRT 0–∞ (h) 9·18 ± 0·786 9·72 ± 1·04 10·0 ± 0·97 V 2 (L) 29·2 ± 7·61 43·4 ± 9·97 e a P < 0·05. b P < 0·01 vs. single-dose PK (non-compartmental parameter). c P < 0·05. d P < 0·01 vs. multiple-dose PK (first dose, non-compartmental parameter). e P < 0·01 vs. single-dose PK (compartmental parameter); the duration of levofloxacin infusion was 1·5 h. The dosing regimen for the multiple-dose PK study was q24 h × 7 days. In the nine subjects receiving multiple-dose intravenous infusion of levofloxacin 750 mg once daily for seven consecutive days, the C max was (17·69 ± 4·23) μg/mL on Day 1 and (13·31 ± 2·77) μg/mL on Day 7. The plasma concentration of levofloxacin at 24 h was (0·87 ± 0·22) μg/mL on Day 1 and (0·803 ± 0·258) μg/mL on Day 7. The corresponding AUC 0–24 h was (90·05 ± 11·92) μg·h/mL and (94·1 ± 15·2) μg h/mL. The mean 24-hour cumulative urinary excretion rate was (75·69 ± 6·34)% and (84·16 ± 5·14)% of the administered dose, respectively. When steady state was reached following consecutive dosing, the mean C ss,max of levofloxacin was 13·31 μg/mL, mean C ss,min 0·031 μg/mL, average plasma concentration (C avg ) (3·87 ± 0·62) μg/mL, mean AUC 0–τ 94·12 μg h/mL (τ = 24 h) and accumulation factor (1·22 ± 0·14) (Table 1 ). Further analysis indicated that the ratio of geometric mean C max , C 24 h , AUC 0–24 h (Day 7/Day 1) and the corresponding 95% confidence interval were 0·76 (0·61–0·95), 0·91 (0·78–1·06) and 1·04 (0·97–1·11), suggesting no apparent accumulation in the body when steady state was reached following once-daily administration of levofloxacin for 7 days. Compartmental model analysis Two-compartment model well described the PK profile of levofloxacin in healthy volunteers following single- or multiple-dose administration (Fig. 1 ). The results showed that the distribution half-life after single dose of levofloxacin was (1·02 ± 0·32) h. The distribution volume of central and peripheral compartments was (67·8 ± 13·2) L and (29·2 ± 7·61) L, respectively. The clearance from central compartment was (10·2 ± 1·89) L/h and (13·8 ± 5·6) L/h between central and peripheral compartments. Fig 1 Concentration–time profiles of levofloxacin following single- or multiple-dose infusion of 750-mg levofloxacin in healthy Chinese volunteers (Mean ± SD, n = 9). The observed values and fittings obtained from two-compartment model were represented by dots and lines , whereas red and blue symbols represent single-dose PK group and multiple-dose PK group, respectively. The infusion time of single-dose levofloxacin was 1·5 h. The dosing regimen for multiple PK study was q24 h × 7 days. Safety of levofloxacin Six ADRs were reported in 6 of the 9 subjects receiving single dose of levofloxacin. The ADR was local adverse reaction in four subjects, that is, mild injection site hyperaemia, which started within 16–45 min since intravenous infusion, and resolved 30 min after the end of intravenous infusion. Both the increased lactate dehydrogenase in one subject and the conjugated bilirubin elevation in another subject developed 23 h after the end of intravenous infusion and normalized at Day 7 post-dose visit. All these ADRs were mild, transient and well tolerated. Nine adverse events occurred in 8 of the 9 subjects receiving multiple doses of levofloxacin. Eight of these events in seven subjects were ADRs. Six ADRs were intermittent mild injection site hyperaemia during the consecutive 7-day intravenous infusion and resolved within 10–25 min after the end of intravenous infusion the same day. Two cases of laboratory abnormalities were transient creatine kinase elevation and increased total bilirubin. The increase was within two times the upper limit of normal in both cases. These ADRs were mild in severity and well tolerated. One of the three subjects receiving placebo developed two mild adverse events (decreased white blood cell count and increased uric acid), which were unrelated to the study drug. Other laboratory tests, vital signs and 12-lead ECG parameters were all within the normal range in both single-dose and multiple-dose studies. No fatal or other significant adverse events were found. There was no significant adverse event leading to study discontinuation or abnormal QTc interval. PK/PD analysis and dosing regimen evaluation Figure 2 displays the CFR value of levofloxacin targeting 245 strains of Streptococcus pneumoniae , Haemophilus influenzae , Klebsiella pneumoniae , methicillin-susceptible Staphylococcus aureus (MSSA) or methicillin-resistant Staphylococcus aureus (MRSA). The CFR of levofloxacin 750-mg once-daily regimen was 96·2%, 98·5%, 95·3% and 100·0% for Streptococcus pneumoniae , Haemophilus influenzae , Klebsiella pneumoniae and MSSA in terms of f C max /MIC = 5, significantly better than that for MRSA (0·4%). The CFR was 95·4%, 98·0%, 95·2% and 100·0% in terms of f AUC 24 h /MIC = 30, whereas for MRSA, the CFR was 0. Fig 2 Cumulative fraction of response of f C max /MIC (a) and f AUC 24 h /MIC (b) for levofloxacin. The dosing regimen of levofloxacin was 500 mg or 750 mg (q.d.) for seven consecutive days, and f indicates the unbound fraction, the value of which was 0·7. From Fig. 3 , when MIC ≤1 μg/mL, the PTA of levofloxacin 750-mg dosing regimen was kept at 100% in terms of the target value of f C max /MIC or f AUC 24 h /MIC, suggesting that levofloxacin 750-mg dosing regimen can promise good clinical and microbiological efficacy targeting the bacteria for which levofloxacin MIC ≤1 μg/mL. Fig 3 Distribution of CAP pathogens in terms of MIC level and the PTA of f C max /MIC (a) and f AUC 24 h /MIC (b) for levofloxacin following multiple dosing. The dosing regimen of levofloxacin was 500 mg or 750 mg (q.d.) for 7 consecutive days, and f indicates the unbound fraction, the value of which was 0·7. Histograms and lines represent the distribution frequency of MIC and PTA values, respectively. In the graph, the distribution data of MIC for CAP pathogens were obtained from the literature. 10 Discussion Following single- or multiple-dose intravenous infusion of levofloxacin 750 mg in healthy Chinese male volunteers, the C max of levofloxacin was 14·94 μg/mL and 13·31 μg/mL; AUC 0–24 73·26 μg h/mL and 94·12 μg h/mL; T 1/2 7·75 h and 6·91 h. These findings were comparable with the results reported by Chow et al ., 6 which indicated that after single- or consecutive 10-day multiple-dose intravenous infusion of levofloxacin 750 mg, the C max was 11·3 mg/L and 12·4 mg/L, AUC 0–24 90·9 mg h/L and 103 mg h/L, and T 1/2 7·51 h. The multiple-dose PK profile of levofloxacin showed more rapid terminal elimination and lower apparent volume of distribution compared with that after single-dose intravenous infusion (Table 1 ). This is reflected in the parameters of compartmental model, that is, both the clearance from central compartment and distribution volume decreased significantly ( P < 0·01). Furthermore, the distribution rate of levofloxacin was significantly increased; meanwhile, the clearance between compartments and the distribution volume of peripheral compartment were also significantly higher following multiple-dose infusion. These results indicate that after levofloxacin multiple-dose infusion, the central-to-peripheral compartment distribution is faster. A fraction of levofloxacin was transported from central-to-peripheral compartment, reflecting the high tissue penetration and extensive, high-level distribution to peripheral tissues. This is helpful for levofloxacin to reach effective bactericidal concentration at the site of infection. The distribution volume of levofloxacin is quite large, 1·64 L/kg on average, following single dose of 750 mg. This indicates that levofloxacin is distributed widely in tissues and body fluids. It is reported that levofloxacin can be extensively distributed to the skeletal muscle intercellular fluid, inflammatory blister fluid, subcutaneous soft tissue and pulmonary tissue; especially in lungs and bronchi, the concentration of levofloxacin is two times higher than that in blood. The subjects in this study were all healthy volunteers without infectious respiratory disease. Several studies reported that the levofloxacin concentration in the pulmonary tissue and alveolar fluid was higher than the corresponding concentration in blood in the patients with respiratory tract infection, and the levofloxacin concentration in the blood of the patients with infection was higher than that in healthy volunteers. Therefore, levofloxacin can reach higher concentration in the tissues and body fluids of the patients with infection compared with healthy volunteers. 14 – 21 Correlation analysis showed that the endogenous creatinine clearance (CL cr ) was well correlated with levofloxacin clearance (CL) in healthy subjects. Pearson's correlation coefficient ( R ) was 0·529 ( P < 0·05). The body weight of volunteers was also correlated significantly with CL ( R = 0·573, P < 0·01). The body weight of volunteers was positively correlated with the distribution volume of central compartment ( R = 0·443). The distribution volume of peripheral compartment showed a trend of increase with age ( R = 0·525, P < 0·01). This implies that age of volunteers may be an important factor influencing the distribution volume of levofloxacin in peripheral compartment. The endogenous creatinine clearance and body weight are covariates affecting the external clearance. This finding is consistent with the results of levofloxacin population PK study. 11 In our previous study of levofloxacin 500 mg, 22 after intravenous infusion of levofloxacin 500 mg, the mean C max and AUC 0–∞ were (7·6 ± 1·1) μg/mL and (38·3 ± 4·9) μg h/mL, respectively. The average terminal elimination half-life (T 1/2 ) was (6·5 ± 0·6) h. The CL r and CL t were (8·8 ± 1·4) L/h and (13·3 ± 1·7) L/h, respectively. V d was (109·8 ± 10·8) L. At 24 h after dosing, the cumulative urinary excretion was (65 ± 4)% of the administered dose. According to the PK parameters of levofloxacin 750-mg and 500-mg studies, as well as the PD data (results of susceptibility testing of 245 strains of bacteria) from two clinical trials previously conducted in our institute, 10 f C max /MIC 90 and f AUC 24 h /MIC 90 were calculated. The results demonstrated that except MRSA, both levofloxacin 750-mg and 500-mg dosing regimens can attain or exceed the target of f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 for the common bacterial pathogens of CAP. Levofloxacin 750 mg provided higher PTA values than 500 mg (Table 2 ). Table 2 Susceptibility testing and PK/PD parameters of levofloxacin against 245 clinical isolates from community-acquired pneumonia Bacteria (No. of strains) MIC 90 500 mg 750 mg (μg/mL) f C max /MIC 90 f AUC 24 h /MIC 90 f C max /MIC 90 f AUC 24 h /MIC 90 Streptococcus pneumoniae (45) 1 5·3 26·8 9·3 65·9 Haemophilus influenzae (51) 0·5 10·6 53·6 18·6 131·7 Klebsiella pneumoniae (108) 0·5 10·6 53·6 18·6 131·7 Methicillin-susceptible Staphylococcus aureus (36) 0·25 21·3 107·2 37·2 263·5 Methicillin-resistant Staphylococcus aureus (5) 32 0·2 0·8 0·3 2·06 The f means the unbound fraction of levofloxacin, the value of which was 0·7. The C max and AUC 0–24 were the non-compartmental paramters following last dose of levofloxacin in the multiple PK study. Monte Carlo simulation indicated that the CFR of levofloxacin 750-mg and 500-mg dosing regimens was 96·2% and 79·5% for Streptococcus pneumoniae , a common pathogen of CAP, in terms of f C max /MIC ≥5. Similarly, the CFR of these two regimens was 95·4% and 52·9% in terms of f AUC 24 h /MIC ≥30. When MIC ≤0·5 μg/mL, both dosing regimens offered best PTA (100%) in terms of f C max /MIC ≥5 and f AUC 24 h /MIC ≥30 for the common pathogens of CAP. When MIC = 1 μg/mL, the PTA for levofloxacin 750-mg dosing regimen was 99·9% vs. 73·7% for 500-mg dosing regimen in terms of f C max /MIC ≥5, and 99·9% vs. 18·6% in terms of f AUC 24 h /MIC ≥30. This suggests 750-mg dosing regimen of levofloxacin has higher probability in reaching the PD targets. What is New and Conclusion Considering the concentration-dependent PK/PD property of levofloxacin, which means higher blood concentration predicts better bactericidal activity, it is expected that the regimen of levofloxacin 750-mg q.d. will provide more potent bactericidal activity in human body and maintain effective bactericidal concentration at infection site. Such a dosing regimen can shorten treatment duration and decrease the exposure to sub-MIC level of levofloxacin, which can potentially reduce the emergence of resistant bacteria. The results of this study provide positive evidence to support the regimen of levofloxacin 750-mg injection (150 mL) once daily via intravenous infusion for 5 days in the treatment for severe CAP. It is necessary to design a large-scale clinical trial in patients to confirm these results. Conflict of Interest None of the authors has any conflict of interest to declare.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11013133/

Gut Microbial Dysbiosis Differs in Two Distinct Cachectic Tumor-Bearing Models Consuming the Same Diet

The impact of cancer cachexia on the colonic microbiota is poorly characterized. This study assessed the effect of two cachectic-producing tumor types on the gut microbiota to determine if a similar dysbiosis could be found. In addition, it was determined if a diet containing an immunonutrient-rich food (walnuts) known to promote the growth of probiotic bacteria in the colon could alter the dysbiosis and slow cachexia. Male Fisher 344 rats were randomly assigned to a semi-purified diet with or without walnuts. Then, within each diet group, rats were further assigned randomly to a treatment group: tumor-bearing ad libitum fed (TB), non-tumor-bearing ad libitum fed (NTB-AL), and non-tumor-bearing group pair-fed to the TB (NTB-PF). The TB group was implanted either with the Ward colon carcinoma or MCA-induced sarcoma, both transplantable tumor lines. Fecal samples were collected after the development of cachexia, and bacteria species were identified using 16S rRNA gene analysis. Both TB groups developed cachexia but had a differently altered gut microbiome. Beta diversity was unaffected by treatment (NTB-AL, TB, and NTB-PF) regardless of tumor type but was affected by diet. Also, diet consistently changed the relative abundance of several bacteria taxa, while treatment and tumor type did not. The control diet increased the abundance of A. Anaeroplasma , while the walnut diet increased the genus Ruminococcus . There were no common fecal bacterial changes characteristic of cachexia found. Diet consistently changed the gut microbiota, but these changes were insufficient to slow the progression of cachexia, suggesting cancer cachexia is more complex than a few gut microbiota shifts. 1. Introduction Recent studies have demonstrated a synergistic relationship between the microbes in our gut and many physiological processes within our body [ 1 , 2 ]. Sometimes, those physiological processes go awry, resulting in tumor growth [ 3 ]. The effect of a tumor on our gut microbes has not been clearly established. The growth of a tumor is known to impact many physiological processes, including promoting the unexplained loss of body tissues or cachexia [ 4 , 5 ]. This study investigated the impact of tumor-driven cachexia on the gut microbiome and whether a diet rich in walnuts can alter the observed dysbiosis. Cachexia occurs in many terminal disease states, such as cancer, heart disease, etc., and many studies have described the physiological processes perturbed by tumor-driven cachexia. More than 50% of cancer patients experience unexplained weight loss or cachexia [ 6 ], which can be devastating, affecting the patient's quality and length of life as well as response to treatment [ 7 ]. Ultimately, preventing cancer cachexia would be best, but despite decades of research, there are no known cures. The etiology of cancer cachexia remains a mystery. Genomic techniques have allowed us to identify the microbes in our environment, including those on and in our body. Our gastrointestinal tract contains thousands of bacterial species, with the largest population located in the colon. Many correlative studies have demonstrated a profound communication network between the colonic bacterial communities and the host's cells [ 2 , 8 ]. For example, gut bacteria are critical for developing and training the baby's immune system, which continues throughout one's lifespan [ 9 , 10 ]. In response, the host's immune system secretes molecules that target particular bacterial groups in the colon and regulate their growth [ 10 , 11 ]. Both the gut microbiome and the host's immune system work in concert. The gut microbiota plays an essential role in maintaining the homeostasis of the host [ 2 , 10 ]. Bearing a tumor changes the host's immune system and potentially its symbiotic relationship with the gut microbiota [ 3 , 12 , 13 ]. It has been proposed that gut dysbiosis is one of the factors that contribute to the development and progression of cancer cachexia. Dysbiosis can lead to increased intestinal permeability, impaired immune function, and chronic inflammation, all of which can affect the metabolism and muscle function of the host [ 14 ]. Many of these changes are hallmarks of cancer cachexia as well [ 15 ]. Several studies have shown that cancer cachexia is associated with changes in the composition and diversity of the gut microbiota, with a decrease in beneficial bacteria, such as Ruminococcaceae [ 15 ], Lachnospiraceae [ 15 , 16 ], and Lactobacillus [ 17 ], and an increase in harmful bacteria, such as Bacteroidetes [ 18 ], Enterobacteriaceae [ 16 , 18 , 19 ], and Parabacteroides [ 18 ]. These changes may influence the production of metabolites, cytokines, and hormones that modulate appetite, energy expenditure, and muscle and fat mass [ 20 ]. Most studies to date are limited to murine models with colon cancer, neuroblastoma, or leukemia. Lactobacillus reuteri and Lactobacillus gasseri were low in cachectic mice with leukemia [ 17 ]. Potgens et al. linked cachexia, induced by colon carcinoma 26, with Klebsiella oxytoca , a specific gut bacterial species that altered gut barrier function in cachectic mice with colon carcinoma [ 15 ]. A particular strain, Faecalibacterium prausnitzii A2–165 (DSM 17677), was unsuccessful at reversing cancer cachexia in the same mouse model [ 21 ]. One study in cachectic human cancer patients found an unknown genus from the Enterobacteriaceae family ( p < 0.01) and that Proteobacteria ( p < 0.001) and Veillonella ( p < 0.001) were more abundant [ 22 ]. Thus, a clear pattern of gut microbiota dysbiosis has not emerged. The primary goal of this study was to determine if tumor-driven cachexia produces a typical pattern of dysbiosis. While changes in a few individual gut microbes have been reported, no consistent pattern of changes has been identified. This study used two transplantable tumor models, carcinoma and sarcoma, to determine if cachexia driven by different tumor types is associated with similar changes in colonic microbiota's relative abundance. The carcinoma we selected grows more slowly than the sarcoma, but both models have been used extensively to study cancer cachexia. In addition, rats were selected since previous studies have used mice models. Using a rodent model allows the investigator to control extraneous microbiota influencers, such as diet. In humans, daily lifestyle choices, such as diet, sleep, and physical activity, are known to change the relative abundance of bacteria within the colon [ 23 ]. Since diet can alter the gut microbiota, it may be one way to influence the tumor's effect on the host's gut microbiota and the cachectic process. Thus, the second goal of our study was to determine whether diet could remediate the gut dysbiosis observed in cachectic tumor-bearing rats. Walnuts were selected to add to the diet because they have been shown to promote probiotic gut bacteria in non-tumor-bearing rats [ 24 ]. Also, walnuts are an excellent source of two dietary constituents with known anti-cachectic properties [ 25 , 26 ]: omega-3 fatty acid (particularly α -linolenic acid) and antioxidants. Finally, several studies have shown that walnuts can slow or prevent breast and prostate tumor growth in genetically programmed mice [ 25 ] and xenografts [ 26 ]. For the studies reported here, walnuts were added to the animal's diet without compromising nutritional quality. 2. Materials and Methods 2.1. Study Design This study was approved by the Institutional Care and Use Committee at the Louisiana State University Health Sciences Center (LSUHSC) in New Orleans, LA, USA. One cohort of animals consisted of thirty male Fischer 344 rats, and the second cohort was thirty-six male Fischer 344 rats. The animals were housed in the LSUHSC vivarium under controlled conditions, constant temperature, and a 12 h light/dark cycle. The animals were maintained on rat chow for one week, and at the end of the week, they were weighed and randomly assigned to one of two diet groups: (1) control and (2) walnut. Each animal was singly housed and fed their assigned diet for the remainder of the study. The animals were allowed to adjust to their single housing and diet for three weeks before tumor implantation. The study's design is shown in Figure 1 ; the diets are described under Section 2.2 . The day before tumor implantation, the animals were weighed and randomly assigned to one of three treatment groups: (1) a tumor-bearing (TB) group that was implanted with the tumor and fed ad libitum; (2) a non-tumor-bearing (NTB-AL) group that was sham-operated and fed ad libitum; and, (3) a pair-fed (NTB-PF) group that was sham-operated and given the amount of food the TB animals ate the previous 24 h. This grouping is referred to as treatment (NTB-AL, TB, and NTB-PF). NTB-PF animals were assigned by weight to a TB animal, so there was no more than a two-gram weight difference between the TB and NTB-PF animals. The NTB-AL animals were also weight-matched (±5 g) to NTB-PF and TB animals. After assignment to their treatment group, each animal was anesthetized using isoflurane. A 2 × 2 × 2 mm chunk of the Ward colon carcinoma (carcinoma) or MCA-induced sarcoma (sarcoma), referred to as tumor type, was obtained from a donor tumor-bearing animal and implanted subcutaneously on the left hind flank. Cells for the Ward colon carcinoma tumor line were graciously supplied by Dr. Vickie Baracos at the University of Alberta, Canada. The MCA sarcoma cells were obtained from Dr. Lauri Byerley's laboratory. NTB-PF and NTB-AL animals received the same operation (sham) as the TB animals but did not receive the tumor cells. Animals were weighed and fed daily for 21 days (sarcoma) or 49 days (carcinoma) and then euthanized. Twelve hours prior to euthanasia, food was removed from the animals' cages to ensure they were in a similar metabolic state. At euthanasia, the animal was anesthetized using isoflurane, blood was collected by cardiac puncture, and the abdominal artery was cut to ensure death. Fecal samples were collected aseptically from the descending colon, frozen in liquid nitrogen, and stored at −80° C until DNA isolation. 2.2. Diets The diet was reported previously and identical to the one used by Hardman et al. [ 25 ]. Briefly, the diet was based on the AIN-76 diet. The protein (walnut: 15.6 g/100 g; control: 15.5 g/100 g), fat (walnut: 4.3 g/100 g, control: 5.8 g/100 g), carbohydrate (walnut: 61.7 g/100 g; control: 60.9 g/100 g), and crude fiber (walnut: 3.67 g/100 g; control: 2.7 g/100 g) were adjusted in the control diet for the walnuts, so both had a similar macronutrient composition. Each diet was made in small batches. The walnuts were ground to a fine state and mixed with the other ingredients. When the diet was the consistency of cookie dough, it was rolled, vacuum-sealed in small batches, and frozen at −20 °C until fed to the animals. The diet was thawed at the time of feeding, and a weighed cube was given to the animal. A fresh diet was provided every two days. Pair-feeding started ten days after surgery to allow the animals to recover from the surgery. Previous studies by our group have shown that food intake between TB and NTB-AL animals was not different until ten days after tumor implant. On the eleventh day after tumor implantation, the NTB-PF received the amount of food their matched TB animals consumed 24 h earlier. 2.3. DNA Isolation and PCR Amplification A protocol developed by the LSUHSC School of Medicine Microbial Genomics Resource Group ( http://metagenomics.lsuhsc.edu/mgrg (access on 1 January 2024)) was used to extract total DNA from approximately 0.25 g of feces. This method has been previously published [ 17 ]. The QIAamp DNA Stool Kit (Qiagen, Germantown, MD, USA) was modified to include bead-beating and RNAase treatment steps. 2.4. Sequencing The procedure was previously published [ 17 ]. Briefly, the 16S rRNA gene (V3-V4 hypervariable region) was PCR amplified using V3F = CCTACGGGAGGCAGCAG and V4R = GGACTACHVGGGTWTCTAAT primers, Illumina adaptors, and molecular barcodes [ 18 ]. Each sample was ligated with Illumina indexes and multiplexed for sequencing on a single Illumina MiSeq run using the Illumina V3 600-cycle sequencing kit (Illumina, San Diego, CA, USA) in paired-end mode. Microbial Mock Community HM-276D (BEI Resources, Manassas, VA, USA) was used as a positive control. 2.5. Quality Filtering/Picking Forward read files were processed through the UPARSE pipeline [ 19 ]. Reverse reads were discarded due to persistent read quality issues with the reverse sequencing reads from Illumina V3 sequencing kits. Reads were truncated to a uniform length of 280 bp and reads with quality scores less than 16 were filtered out. The UPARSE pipeline steps described by Edgar [ 27 ] were performed in sequence, and OTU clusters were formed at 97% with chimeric OTUs removed from the data. After quality filtering, reads were analyzed using QIIME 1.9.0 (Quantitative Insights Into Microbial Ecology) with the DADA2 plugin [ 20 ]. Forward and reverse reads were truncated to a uniform length of 240 bp, and 20 bp were trimmed off the front of each read to remove the primer. DADA2-identified amplicon sequence variants (ASVs) were merged, and any that ranged outside the expected 250–255 bp amplicon length were discarded. Any ASVs that appeared in only one sample were removed using contingency-based filtering, and chimeric ASVs were removed using the consensus method. ASVs were aligned using MAFFT [ 28 ] and FastTree [ 29 ], and a phylogenetic tree for diversity analysis was built. Greengenes v13.8 was used for taxonomic classification [ 30 ]. After primary data analysis, the remaining reads were analyzed using QIIME2 [ 31 ]. 2.6. Microbial Community Analysis Sixty-six samples (30 sarcoma and 36 carcinoma) were included in the QIIME analysis with read counts ranging from 11,619 to 147,455 with an average read count per sample of 91,143 (sarcoma) and 92,036 (carcinoma). Alpha rarefaction was performed at a level of 11,619 reads to include all samples. Alpha rarefaction plots were produced by plotting the number of sequences in a sample against several different diversity metrics, for example, Shannon, Simpson, and Chao1. Beta diversity was determined by principal coordinate analysis using both unweighted and weighted UniFrac metrics. Emperor 3D viewer was used to visualize the plots [ 32 , 33 ]. 2.7. Predicted Functional Pathways Potential microbial functions were identified from the 16S sequencing data. The raw data were formatted and imported into QIIME2. Closed-reference clustering against the Greengenes 13_5 97% OTUs reference database was used to develop a de-replicated feature table and representative sequences. The closed-reference OTU table was used as input into the PICRUSt [ 22 ] pipeline, and the resulting PICRUSt metagenome data were further analyzed by using STAMP (Statistical Analysis of Metagenomic Profiles) [ 23 ]. Pathways were labeled at Level 2 since several pathways were not classified at Level 1. From this data, KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways were compared between NTB-AL, TB, and NTB-PF groups within each tumor type. 2.8. Statistical Analysis Data are expressed as mean ± standard error of the mean (SEM). SAS ( https://www.sas.com/en_us/home.html (access on 1 January 2024), SAS Institute Inc., Cary, NC, USA), SPSS ( https://www.ibm.com/products/spss-statistics (access on 1 January 2024), IBM Corp. Released 2020. IBM SPSS Statistics for Windows, Version 27.0. IBM Corp: Armonk, NY, USA), and R ( https://www.r-project.org/ (access on 1 January 2024), R Statistical Software (v4.1.2; R Core Team 2021)) software were used to analyze data statistically. Descriptive data such as mean and SEM were determined using SPSS. A p -value less than 0.05 was considered significant. LEFSE was used to select the bacterial species to determine statistical differences between the groups to reduce the number of comparisons [ 34 ]. Differences among the two diet groups (control and walnut) and three treatment groups (NTB-AL, TB, and NTB-PF) for the selected bacterial species were determined using a two-way analysis of variance (SAS). Since multiple analyses were run, the Benjamini–Hochberg procedure was used to control for the false discovery rate. Briefly, the p -values were put in order from the smallest to largest, were ranked (rank of i = 1, i = 2, etc.), and a critical value (CV) was calculated as ( i / m ) Q , where m is the total number of tests and Q is the false discovery rate (0.05). Those taxa with a p -value less than the CV ( P < (( i /m) Q )) were considered significant. If there was a significant effect, differences among the groups were determined using the Newman–Keuls. All taxa that were selected by LEFSE and their Benjamini–Hochberg values are shown in Table A1 and Table A2 . STAMP was used to determine statistical differences in functional pathways between the groups and generate post hoc (Tukey–Kramer) plots for each KEGG pathway significantly different between NTB-AL, TB, and NTB-PF animals. Bonferroni was used to correct for multiple analyses. Figures were created using GraphPad Prism v10 ( https://www.graphpad.com/ (access on 1 January 2024), GraphPad Software, San Diego, CA, USA) and BioRender 2023 ( https://www.biorender.com/ (access on 1 January 2024)). 2.1. Study Design This study was approved by the Institutional Care and Use Committee at the Louisiana State University Health Sciences Center (LSUHSC) in New Orleans, LA, USA. One cohort of animals consisted of thirty male Fischer 344 rats, and the second cohort was thirty-six male Fischer 344 rats. The animals were housed in the LSUHSC vivarium under controlled conditions, constant temperature, and a 12 h light/dark cycle. The animals were maintained on rat chow for one week, and at the end of the week, they were weighed and randomly assigned to one of two diet groups: (1) control and (2) walnut. Each animal was singly housed and fed their assigned diet for the remainder of the study. The animals were allowed to adjust to their single housing and diet for three weeks before tumor implantation. The study's design is shown in Figure 1 ; the diets are described under Section 2.2 . The day before tumor implantation, the animals were weighed and randomly assigned to one of three treatment groups: (1) a tumor-bearing (TB) group that was implanted with the tumor and fed ad libitum; (2) a non-tumor-bearing (NTB-AL) group that was sham-operated and fed ad libitum; and, (3) a pair-fed (NTB-PF) group that was sham-operated and given the amount of food the TB animals ate the previous 24 h. This grouping is referred to as treatment (NTB-AL, TB, and NTB-PF). NTB-PF animals were assigned by weight to a TB animal, so there was no more than a two-gram weight difference between the TB and NTB-PF animals. The NTB-AL animals were also weight-matched (±5 g) to NTB-PF and TB animals. After assignment to their treatment group, each animal was anesthetized using isoflurane. A 2 × 2 × 2 mm chunk of the Ward colon carcinoma (carcinoma) or MCA-induced sarcoma (sarcoma), referred to as tumor type, was obtained from a donor tumor-bearing animal and implanted subcutaneously on the left hind flank. Cells for the Ward colon carcinoma tumor line were graciously supplied by Dr. Vickie Baracos at the University of Alberta, Canada. The MCA sarcoma cells were obtained from Dr. Lauri Byerley's laboratory. NTB-PF and NTB-AL animals received the same operation (sham) as the TB animals but did not receive the tumor cells. Animals were weighed and fed daily for 21 days (sarcoma) or 49 days (carcinoma) and then euthanized. Twelve hours prior to euthanasia, food was removed from the animals' cages to ensure they were in a similar metabolic state. At euthanasia, the animal was anesthetized using isoflurane, blood was collected by cardiac puncture, and the abdominal artery was cut to ensure death. Fecal samples were collected aseptically from the descending colon, frozen in liquid nitrogen, and stored at −80° C until DNA isolation. 2.2. Diets The diet was reported previously and identical to the one used by Hardman et al. [ 25 ]. Briefly, the diet was based on the AIN-76 diet. The protein (walnut: 15.6 g/100 g; control: 15.5 g/100 g), fat (walnut: 4.3 g/100 g, control: 5.8 g/100 g), carbohydrate (walnut: 61.7 g/100 g; control: 60.9 g/100 g), and crude fiber (walnut: 3.67 g/100 g; control: 2.7 g/100 g) were adjusted in the control diet for the walnuts, so both had a similar macronutrient composition. Each diet was made in small batches. The walnuts were ground to a fine state and mixed with the other ingredients. When the diet was the consistency of cookie dough, it was rolled, vacuum-sealed in small batches, and frozen at −20 °C until fed to the animals. The diet was thawed at the time of feeding, and a weighed cube was given to the animal. A fresh diet was provided every two days. Pair-feeding started ten days after surgery to allow the animals to recover from the surgery. Previous studies by our group have shown that food intake between TB and NTB-AL animals was not different until ten days after tumor implant. On the eleventh day after tumor implantation, the NTB-PF received the amount of food their matched TB animals consumed 24 h earlier. 2.3. DNA Isolation and PCR Amplification A protocol developed by the LSUHSC School of Medicine Microbial Genomics Resource Group ( http://metagenomics.lsuhsc.edu/mgrg (access on 1 January 2024)) was used to extract total DNA from approximately 0.25 g of feces. This method has been previously published [ 17 ]. The QIAamp DNA Stool Kit (Qiagen, Germantown, MD, USA) was modified to include bead-beating and RNAase treatment steps. 2.4. Sequencing The procedure was previously published [ 17 ]. Briefly, the 16S rRNA gene (V3-V4 hypervariable region) was PCR amplified using V3F = CCTACGGGAGGCAGCAG and V4R = GGACTACHVGGGTWTCTAAT primers, Illumina adaptors, and molecular barcodes [ 18 ]. Each sample was ligated with Illumina indexes and multiplexed for sequencing on a single Illumina MiSeq run using the Illumina V3 600-cycle sequencing kit (Illumina, San Diego, CA, USA) in paired-end mode. Microbial Mock Community HM-276D (BEI Resources, Manassas, VA, USA) was used as a positive control. 2.5. Quality Filtering/Picking Forward read files were processed through the UPARSE pipeline [ 19 ]. Reverse reads were discarded due to persistent read quality issues with the reverse sequencing reads from Illumina V3 sequencing kits. Reads were truncated to a uniform length of 280 bp and reads with quality scores less than 16 were filtered out. The UPARSE pipeline steps described by Edgar [ 27 ] were performed in sequence, and OTU clusters were formed at 97% with chimeric OTUs removed from the data. After quality filtering, reads were analyzed using QIIME 1.9.0 (Quantitative Insights Into Microbial Ecology) with the DADA2 plugin [ 20 ]. Forward and reverse reads were truncated to a uniform length of 240 bp, and 20 bp were trimmed off the front of each read to remove the primer. DADA2-identified amplicon sequence variants (ASVs) were merged, and any that ranged outside the expected 250–255 bp amplicon length were discarded. Any ASVs that appeared in only one sample were removed using contingency-based filtering, and chimeric ASVs were removed using the consensus method. ASVs were aligned using MAFFT [ 28 ] and FastTree [ 29 ], and a phylogenetic tree for diversity analysis was built. Greengenes v13.8 was used for taxonomic classification [ 30 ]. After primary data analysis, the remaining reads were analyzed using QIIME2 [ 31 ]. 2.6. Microbial Community Analysis Sixty-six samples (30 sarcoma and 36 carcinoma) were included in the QIIME analysis with read counts ranging from 11,619 to 147,455 with an average read count per sample of 91,143 (sarcoma) and 92,036 (carcinoma). Alpha rarefaction was performed at a level of 11,619 reads to include all samples. Alpha rarefaction plots were produced by plotting the number of sequences in a sample against several different diversity metrics, for example, Shannon, Simpson, and Chao1. Beta diversity was determined by principal coordinate analysis using both unweighted and weighted UniFrac metrics. Emperor 3D viewer was used to visualize the plots [ 32 , 33 ]. 2.7. Predicted Functional Pathways Potential microbial functions were identified from the 16S sequencing data. The raw data were formatted and imported into QIIME2. Closed-reference clustering against the Greengenes 13_5 97% OTUs reference database was used to develop a de-replicated feature table and representative sequences. The closed-reference OTU table was used as input into the PICRUSt [ 22 ] pipeline, and the resulting PICRUSt metagenome data were further analyzed by using STAMP (Statistical Analysis of Metagenomic Profiles) [ 23 ]. Pathways were labeled at Level 2 since several pathways were not classified at Level 1. From this data, KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways were compared between NTB-AL, TB, and NTB-PF groups within each tumor type. 2.8. Statistical Analysis Data are expressed as mean ± standard error of the mean (SEM). SAS ( https://www.sas.com/en_us/home.html (access on 1 January 2024), SAS Institute Inc., Cary, NC, USA), SPSS ( https://www.ibm.com/products/spss-statistics (access on 1 January 2024), IBM Corp. Released 2020. IBM SPSS Statistics for Windows, Version 27.0. IBM Corp: Armonk, NY, USA), and R ( https://www.r-project.org/ (access on 1 January 2024), R Statistical Software (v4.1.2; R Core Team 2021)) software were used to analyze data statistically. Descriptive data such as mean and SEM were determined using SPSS. A p -value less than 0.05 was considered significant. LEFSE was used to select the bacterial species to determine statistical differences between the groups to reduce the number of comparisons [ 34 ]. Differences among the two diet groups (control and walnut) and three treatment groups (NTB-AL, TB, and NTB-PF) for the selected bacterial species were determined using a two-way analysis of variance (SAS). Since multiple analyses were run, the Benjamini–Hochberg procedure was used to control for the false discovery rate. Briefly, the p -values were put in order from the smallest to largest, were ranked (rank of i = 1, i = 2, etc.), and a critical value (CV) was calculated as ( i / m ) Q , where m is the total number of tests and Q is the false discovery rate (0.05). Those taxa with a p -value less than the CV ( P < (( i /m) Q )) were considered significant. If there was a significant effect, differences among the groups were determined using the Newman–Keuls. All taxa that were selected by LEFSE and their Benjamini–Hochberg values are shown in Table A1 and Table A2 . STAMP was used to determine statistical differences in functional pathways between the groups and generate post hoc (Tukey–Kramer) plots for each KEGG pathway significantly different between NTB-AL, TB, and NTB-PF animals. Bonferroni was used to correct for multiple analyses. Figures were created using GraphPad Prism v10 ( https://www.graphpad.com/ (access on 1 January 2024), GraphPad Software, San Diego, CA, USA) and BioRender 2023 ( https://www.biorender.com/ (access on 1 January 2024)). 3. Results Body weight did not differ significantly among the NTB-AL, TB, and NTB-PF groups (both tumor types) before the tumor or sham operation occurred (all: 330 ± 2; walnut: 330 ± 3; control: 330 ± 3). At the time of euthanasia, tumor weight was not significantly different between the control and walnut TB groups (both tumor types, Figure 2 A,B). Host body weight (body weight minus tumor weight) at the time of euthanasia is shown in Figure 2 C,D. At the time of euthanasia, host weight (total body weight minus tumor weight) for both the sarcoma- and carcinoma-bearing animals were significantly less than their matched NTB-AL animal regardless of diet, indicating they were cachectic. Total caloric intake (from implant to euthanasia) was not significantly altered between the walnut and control diets regardless of treatment (NTB-AL, TB, and NTB-PF) for either tumor type ( Figure 2 E,F). No differences in alpha diversity (within community diversity) using several different measures (Simpson, Shannon, Chao, observed taxa, and phylogenetic diversity) were found. Beta microbial diversity (differences between communities) is shown in Figure 3 , and both diet and tumor type altered diversity. The walnut and control diets were clearly different from each other for both the sarcoma and carcinoma tumor types. However, the NTB-AL, TB, and NTB-PF overlapped within these four communities, so no differences could be determined except for the carcinoma walnut TB group, which differed from the carcinoma walnut NTB-PF and NTB-AL. Figure 4 A shows nine phyla for the carcinoma and sarcoma animals on each diet and treatment. Together, Firmicutes and Bacteroidetes phyla comprised approximately 90% of the colonic microbiota, with 61% of the microbes from the Firmicutes phylum. The walnut diet consistently produced a similar relative abundance for the Firmicutes phyla (carcinoma: 67 ± 5% (NTB-AL), 63 ± 4% (TB), and 65 ± 4% (NTB-PF); sarcoma: 65 ± 4% (NTB-AL), 62 ± 5% (TB), and 67 ± 9% (NTB-PF)) regardless of tumor type and treatment (NTB-AL, TB, and NTB-PF). That was not the case for the control diet. The relative abundance of the Firmicutes phylum was lower in the NTB-AL animals for both tumor types (carcinoma: 52 ± 2%; sarcoma: 53 ± 2%), but higher in the sarcoma TB (carcinoma: 56 ± 4%; sarcoma: 72 ± 12%), and more in the carcinoma NTB-PF (carcinoma: 65 ± 2%; sarcoma: 60 ± 3%). The sarcoma-TB had a higher relative abundance compared to the carcinoma TB. The same pattern was observed for Bacteroidetes, but there was less variability in the relative abundance of the animals. The Firmicutes-to-Bacteroidetes (F/B) ratio has been proposed as a marker of gut dysbiosis. The Firmicutes-to-Bacteroidetes ratio from our study was not significantly different between treatment and diet for either tumor type ( Figure 4 B). Overall, Firmicutes dominated the OTU-level diversity by approximately 3-fold over the Bacteroidetes. The sarcoma TB consuming the control diet had the most variability and highest ratio. This group also had the lowest relative abundance of Bacteroidetes. We also looked at the relative abundance of all species present in the stool sample. Several microbes were different for either the carcinoma or sarcoma treatment. Not corrected for multiple comparisons, these are shown in Table A1 and Table A2 . We aimed to identify specific bacteria that were consistently elevated or reduced for both tumor types and, for this, corrected for multiple comparisons. Treatment did not significantly and consistently affect the relative abundance of any microbes. Only diet consistently and significantly altered the relative abundance of a few microbes shown in Table 1 . For both tumor types, microbes from the Tenericutes phylum, order Anaeroplasmatales, had a significantly higher relative abundance in animals consuming the control diet regardless of treatment. The Tenericutes phylum's relative abundance was low (0.84 ± 0.25%) compared to the Firmicutes and Bacerodiodetes phyla. The walnut diet significantly increased the relative abundance of several microbes from the Firmicutes phyla, particularly the Bacilli and Clostridia classes. Differences in four functional pathways (KEGG Level 2) were predicted from the gene data ( Figure 5 A,B): cellular processing, genetic information processing, human diseases, and metabolism. Only genetic information processing had significantly different pathways at Level 3: DNA repair and recombination proteins and translation factors. No other pathways at Level 3 were significantly different. The same pathways were affected within each tumor type. From 16s RNA, metabolic pathways that are up- or down-regulated can be predicted. KEGG is a hierarchical collection of pathway maps. Metabolism is one of these, which has seven broad categories. At Level 2, four predicted metabolic pathways emerged as different in the two tumor types. Their percentage difference is shown in Figure 5 A,B. Each tumor type had a different percentage. At Level 3, we found two predicted metabolic pathways altered in carcinoma and sarcoma-bearing rats ( Figure 5 C–F). Pathways in DNA repair and recombination proteins and translation factors were elevated in the TB rats compared to the NTB-AL or the NTB-PF. 4. Discussion Cancer cachexia, which is characterized by weight loss, muscle wasting, anorexia, and systemic inflammation, is a complex syndrome that affects many patients with advanced cancer [ 36 ]. These symptoms impair the quality of life, response to treatment, and the survival of cancer patients [ 37 ]. One of the factors that may contribute to the development and progression of cancer cachexia is the alteration of the gut microbiota [ 20 ]. The gut microbiota plays a vital role in maintaining the homeostasis of the host, but various factors, such as diet, infection, medication, or cancer itself, can promote dysbiosis [ 23 ]. This study used two distinctly different tumor types: carcinoma and sarcoma. Carcinomas account for 80 to 90% of all human cancers, while sarcomas are rare (<1% of adult human tumors). Each arises from different tissue types. Animal models for both tumor types have been developed, and the Ward colon carcinoma and the MCA-induced sarcoma have been used extensively to study cancer cachexia, so these were selected to compare their gut microbiota. The MCA-induced sarcoma is faster growing than the Ward colon carcinoma, but both produce cachexia unrelated to a reduced food intake. Diet has been studied extensively as a tool to improve the health and well-being of cachectic cancer patients. Several nutritional therapies (including prebiotics and probiotics) that target the gut microbiota have been tried in the last several decades. For example, Bindels et al. [ 17 ] administered lactobacilli to cachectic leukemia mice and found it decreased muscle atrophy. This same observation was confirmed in the colon carcinoma 26 mouse model that also develops cachexia [ 38 ]. We showed that walnuts increase probiotic bacteria, Lactobacillus, Ruminococcaceae, and g. Roseburia [ 24 ], so we, therefore, investigated if adding walnuts to the diet could improve the cachectic condition. We reported earlier that a diet with walnuts added does not slow muscle atrophy [ 39 ]. Several studies have shown that gut microbiota diversity and composition are altered in cachectic tumor-bearing animals and humans, thus supporting the notion that dysbiosis may be involved in the pathogenesis of this syndrome. Colonic dysbiosis has been reported in tumor-bearing mice, but the dysbiosis has not been compared to two distinctly different tumor types in a different species, rat. Alpha diversity represents a single sample's richness and community diversity, such as the tumor-bearing animals. There are a variety of different measures that can be used to compare the richness and diversity between samples. Published results for these measures in cachectic tumor-bearing mice and humans are inconsistent. Jeong et al. [ 40 ] found that cachectic mice bearing Lewis lung cancer cell allografts had lower alpha diversity than non-tumor-bearing mice. We found neither community richness nor diversity was different regardless of tumor type (sarcoma vs. carcinoma), the diet consumed (walnut vs. control), and treatment (NTB-AL, TB, and NTB-PF). Ni et al. [ 41 ] also found no differences in alpha diversity in cachectic lung cancer patients compared to non-cachectic lung cancer patients. Beta diversity analysis quantifies the similarity or dissimilarity between microbiome pairs between samples, such as the walnut and control diets. Jeong et al. [ 40 ] found that cachectic mice bearing Lewis lung cancer cell allografts had distinct beta diversity compared to the non-tumor-bearing mice. We found a noticeable difference in beta diversity; tumor type and diet caused a significant separation in the composition of the gut microbiome, while treatment had no effect. These results suggest that tumor type and diet have a greater influence on beta diversity than bearing a tumor and developing cachexia. While we did not find significant shifts in alpha diversity or the F/B ratio (a marker of gut dysbiosis), we observed a few changes in specific bacterial species. For this study, we corrected for multiple comparisons, which drastically reduced the number of significant species. Only diet significantly reduced or increased the relative abundance of several microbes for both tumor types. Diet is known to change the relative abundance of gut microbial communities. The control diet significantly increased the presence of two genera from the Anaeroplasmatales order for both the sarcoma and carcinoma animals. Anaeroplasma is an obligate anaerobe and resides in the gut at relatively low levels. There is minimal information on Anaeroplasma in human diseases, but it has been observed in an aging mouse model [ 42 ]. It is a member of the Tenericutes phylum, which has a low relative abundance compared to other members of the phylum level. De Maria, Y et al. [ 19 ] characterized the gut microbiome of mice bearing Lewis lung carcinoma. They found dysbiosis-involved representatives from seven phyla (Proteobacteria, Cyanobacteria, TM7, Actinobacteria, Bacteroidetes, Firmicutes, and Tenericutes), demonstrating a complex pattern. For the Tenericutes phylum, the F16 order was expanded, not Anaeroplamatales. For the walnut diet, the class Bacilli and genus Ruminococcus had a significantly higher relative abundance for both tumor types. Bacilli are Gram-positive and often rod-shaped bacteria, widely distributed in nature, particularly soil. This class contains several well-known pathogens, including the bacteria that cause anthrax and B. cereus , a known food pathogen [ 43 ]. Although a relatively minor proportion of the gut microbiome, Bacilli class bacteria secrete a wide range of compounds [ 43 ]. Ruminococcus are butyrate-forming anaerobic Gram-positive bacteria that degrade and convert complex polysaccharides, like cellulose, into various nutrients, like glucose, for their hosts [ 44 ]. Byerley et al. reported that a walnut-rich diet increased the relative abundance of this bacteria in healthy, non-tumor-bearing rats [ 24 ]. Several other studies have reported increased [ 45 ] and decreased [ 46 ] Ruminococcus when walnuts are added to the human diet. We are unaware of any studies of cachectic animals or humans that have reported an increase in this particular bacterium. Ni et al. [ 41 ] used shotgun metagenomics to interrogate the gut microbiome of cachectic lung cancer patients. They reported that the catabolic pathways of certain complex carbohydrates and sugar derivatives and the anabolic pathways for several amino acid groups were significantly lower, while the polysaccharide pathways were enriched in the cachectic patients. Our 16s RNA analysis identified two pathways from the KEGG Level 2 genetic information processing pathway that significantly differed in both the sarcoma and carcinoma groups. These two pathways were related to DNA repair and translation factors. 5. Conclusions In summary, we found that cachexia, as a result of bearing a tumor, perturbed the gut microbiome, but the changes were not consistent across the two distinctly different tumor models examined. Therefore, we did not find a unique gut microbiome dysbiosis pattern that could be associated with cachexia. This suggests that gut microbiota changes are a consequence of cachexia and are unique to tumor type. Diet consistently altered the gut microbiome in both tumor types, but it was not enough to slow the progression of cachexia.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6853890/

Editorial: Nanoparticle Vaccines Against Infectious Diseases

Author Contributions All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10029730/

One Health activities to reinforce intersectoral coordination at local levels in India

India's dense human and animal populations, agricultural economy, changing environment, and social dynamics support conditions for emergence/re-emergence of zoonotic diseases that necessitate a One Health (OH) approach for control. In addition to OH national level frameworks, effective OH driven strategies that promote local intersectoral coordination and collaboration are needed to truly address zoonotic diseases in India. We conducted a literature review to assess the landscape of OH activities at local levels in India that featured intersectoral coordination and collaboration and supplemented it with our own experience conducting OH related activities with local partners. We identified key themes and examples in local OH activities. Our landscape assessment demonstrated that intersectoral collaboration primarily occurs through specific research activities and during outbreaks, however, there is limited formal coordination among veterinary, medical, and environmental professionals on the day-to-day prevention and detection of zoonotic diseases at district/sub-district levels in India. Examples of local OH driven intersectoral coordination include the essential role of veterinarians in COVID-19 diagnostics, testing of human samples in veterinary labs for Brucella and leptospirosis in Punjab and Tamil Nadu, respectively, and implementation of OH education targeted to school children and farmers in rural communities. There is an opportunity to strengthen local intersectoral coordination between animal, human and environmental health sectors by building on these activities and formalizing the existing collaborative networks. As India moves forward with broad OH initiatives, OH networks and experience at the local level from previous or ongoing activities can support implementation from the ground up. Introduction One Health (OH) is a framework for a transdisciplinary and cross-sectoral approach to address complex problems at the intersection of the environment, people, plants, and animals. The OH approach saves human and animal lives and is economical due to the efficient use of resources, for example, infrastructure, finances, personnel, and timely action ( 1 , 2 ). In addition to critical global health concerns like food safety and antimicrobial resistance (AMR), the OH approach is essential to address zoonotic diseases, as the interactions at the human-animal-environment interface directly influence the epidemiology of these diseases. The recent COVID-19 pandemic and global monkeypox public health emergency highlight the ongoing global threat of emerging and re-emerging zoonotic diseases, with calls for using a OH approach to prepare for and address the next infectious disease threat ( 3 , 4 ). With 75% of emerging infections having a non-human animal origin ( 5 ), the next global infectious disease emergency will likely require a OH approach. The drivers of infectious diseases' emergence/re-emergence include increasing human population, urbanization, climate change, land use change, intensive farming practices, deforestation, and exploitation of wildlife. Human population density, land area and the human development index at the country level are associated with human, emerging and zoonotic pathogen diversity ( 6 ). A One Health approach is important for India, as it is considered one of the probable global "hotspots" for emerging and re-emerging zoonoses ( 7 ) due to its enormous human and animal populations and complex agricultural economy combined with rapid socio-ecological, environmental and climactic changes. It accounts for 35% of the global burden of rabies ( 8 ), with ~20,000 human deaths annually. Economic loss due to the brucellosis burden in livestock in India was estimated to be 3.4 billion USD ( 9 ). Along with four nations, India contributes to 83% of cases of Kala-azar (visceral leishmaniasis) globally ( 10 ). Moreover, highly pathogenic zoonotic diseases like Nipah virus and Crimean-Congo Hemorrhagic Fever (CCHF) virus have emerged in India in the past two decades. A key component of OH is enhancing interactions and collaborations among diverse stakeholders, not only at the international and national levels, but also at the local level. It is particularly important in India, where most of the human population continues to live in rural areas and have daily interactions with livestock and wildlife. In many areas, animals, including cattle, dogs, and pigs, roam freely in the streets. Environmental and climate factors, such as temperature, humidity, population density, deforestation, pollutants, change in land use etc., may lead to the emergence/re-emergence and increased transmission of zoonotic pathogens. In these areas and communities with high human-animal-environmental interaction, effective OH driven strategies that promote local intersectoral coordination and collaboration are needed to truly address zoonotic diseases in India. While there is support for OH frameworks in India, particularly at the national level, OH-related efforts at local levels (ex. district, sub-district, city, Tehsil or taluk, and village levels) are critical, especially for addressing zoonotic and emerging diseases at the level at which they start. To assess the landscape of OH activities at local levels in India featuring intersectoral coordination and collaboration, we conducted a landscape review of the literature and supplemented it with our own experience conducting OH-related activities with local partners ( Figure 1 ). We conducted a PubMed search including keywords "One Health[Title]" and "India[Title]" on 8 July 2022, which returned 31 relevant published papers. Two co-authors (JT and ND) reviewed 31 abstracts, excluded two that were corrections to already included papers, and shortlisted 22 full-length papers for in-depth reading. Eight co-authors were asked to read these papers and note the following information: (a) Was there intersectoral coordination? If yes, which sectors were involved? (b) What were the findings and overall paper focus? (c) Who was involved at the local level? (d) Was there community involvement? This information enabled us to identify key overarching themes and examples in local OH activities in India. Figure 1 Overview of landscape analysis approach to evaluate OH activities at local levels in India featuring intersectoral coordination and collaboration. A literature review using PubMed search produced 31 abstracts based on search terms. These abstracts were reviewed, and 22 of them were shortlisted for in-depth reading and evaluation of these publications by co-authors. In parallel, examples of local intersectoral coordination and collaboration on OH activities were identified through personal experience conducting and knowledge of relevant activities with local partners. Findings Our review of OH in India literature returned publications falling under the following categories: (1) Studies describing public health, epidemiological, and laboratory research, including reviews, (2) Policy or social sciences research or analysis of Indian OH frameworks, and (3) Case studies of OH relevant programs or activities ( Table 1 ). Our landscape assessment demonstrated that OH intersectoral collaboration primarily occurs through specific research activities, during outbreaks, and for community outreach efforts. However, there is limited formal coordination among veterinary, medical, and environmental professionals on the day-to-day prevention and detection of zoonotic diseases at district/sub-district levels in India. Here we describe examples of OH activities featuring local coordination and collaboration from the published literature and our own experiences implementing them. Table 1 Summary of One Health in India literature review. References Year Category Author-specified keywords Key descriptors Chaudhari et al. ( 11 ) 2021 Laboratory and/or epidemiological research Brucellosis; One Health; Listeriosis; Scrub typhus; Surveillance; Zoonotic diseases Research collaboration, Zoonotic disease surveillance Das et al. ( 12 ) 2020 Laboratory and/or epidemiological research India; One Health; Antibiotics resistance; Bacterial infection; Drug prescriptions; Sentinel surveillance Surveillance; AMR; Research collaboration Sack et al. ( 13 ) 2021 Laboratory and/or epidemiological research Not specified Epidemiological investigation; Research collaboration; Environmental sampling Asaaga et al. ( 14 ) 2021 Policy/Social science Cross-sectoral convergence; Emerging infectious disease; Health system; India; One Health; Zoonoses Operationalisation/implementation facilitators and barriers; Zoonosis control Perez et al. ( 15 ) 2020 Case study; Policy Avian influenza; Flood management; One Health action; Rabies Operationalisation/implementation; Zoonosis control Yasobant et al. ( 16 ) 2021 Case study; Policy/Social science ASHA; CHW; India; Motivation; OHA; One Health Implementation; Community engagement Jani et al. ( 17 ) 2021 Laboratory and/or epidemiological research Antimicrobial resistance, Mass gatherings, AMR containment policy, One Health approach AMR; Review Chattu et al. ( 18 ) 2018 Case study; Policy Global health security; Kerala; Nipah virus; One Health; Pteropus bat species; Paramyxovirus Zoonotic disease outbreak; Research collaboration Rajagopal et al. ( 19 ) 2021 Laboratory and/or epidemiological research Escherichia coli; India; One Health; Antimicrobial resistance; Community-acquired AMR; Review Yasobant et al. ( 20 ) 2021 Public health research; Social science Health system contact; India; One Health; Community awareness; Zoonotic diseases Community awareness investigation Dasgupta et al. ( 21 ) 2021 Policy Intersectoral approach; One Health committee; Leadership; Strategic goals; Zoonoses Operationalisation/implementation Yasobant et al. ( 22 ) 2020 Case study; Policy/Social science India; One Health; Actors; Health system; Intersectoral collaboration Operationalisation; Outbreak Reddy et al. ( 23 ) 2021 Policy Not specified Operationalisation; Outbreak Yasobant et al. ( 24 ) 2019 Policy Health system; India; Initiatives; One Health collaboration; Strategies Zoonoses; Vaccination; Education Yasobant et al. ( 25 ) 2020 Policy Coronavirus disease 2019; India; One Health; One Health surveillance Surveillance; Outbreak Lindahl et al. ( 26 ) 2020 Policy Brucella; Brucellosis; India; Livestock; Public health Zoonotic disease control Fitzpatrick et al. ( 27 ) 2016 Laboratory and/or epidemiological research Cost-effectiveness; Mathematical modeling; Rabies; Sterilization; Vaccination Zoonotic disease control; Cost-effectiveness; Economic analysis Yasobant et al. ( 28 ) 2018 Policy/Social science One Health; Systems thinking; Health systems; Prevention and control; Zoonotic diseases Zoonotic disease control Yasobant et al. ( 29 ) 2019 Policy/Social science Not specified Disease prioritization; Zoonotic diseases; Stakeholder engagement Weiss et al. ( 30 ) 2021 Policy Not specified Perspective piece; One Health in India Chatterjee et al. ( 31 ) 2016 Policy Emerging infectious diseases; Health policy; Intersectoral coordination; Multidisciplinary; One Health; Transdisciplinary; Zoonoses Perspective piece; One Health in India Yasobant et al. ( 32 ) 2021 Case study, Policy Intersectoral collaboration; One Health; Operationalization; Health system; India Zoonotic diseases; Operationalization Mansingh et al. ( 33 ) 2021 Policy; Public health research Anthrax; FDG; One Health; Endemic regions; Surveillance; Zoonotic disease Zoonotic diseases Mourya et al. ( 34 ) 2021 Case study Crimean-Congo haemorrhagic fever; India; Kyasanur forest disease; One Heath; Tick-borne; Zoonotic disease Outbreak Stålsby et al. ( 35 ) 2015 Laboratory and/or epidemiological research One Health; Health seeking behavior; Antibiotic prescribing; Formal and informal health care providers; Escherichia coli in stools of children and water; Antibiotic resistance; Molecular basis of resistance AMR Murhekar et al. ( 36 ) 2021 Case study, Laboratory and/or epidemiological research Acute encephalitis syndrome; One Health; Acute febrile illness; Scrub typhus; Vector Zoonotic disease surveillance; Zoonotic disease control; Outbreak Prejit et al. ( 37 ) 2022 Case study; Public health research Evaluation; Integrated surveillance; KFD virus; One Health; Wayanad; Zoonoses Zoonotic disease surveillance; Zoonotic disease control; Outbreak Gibson et al. ( 38 ) 2022 Case study; Public health research Developing world; Epidemiology; Viral epidemiology; Viral genetics Zoonotic disease surveillance; Zoonotic disease control; Economic analysis Abbas et al. ( 39 ) 2011 Case study; Public health research Rabies, Zoonosis, India, One Health, Health policy, Communicable disease Zoonotic disease control Twenty-nine papers from a search on One Health in India are described by their year of publication, category, author-specified keywords, and other key descriptors. Author-specified keywords were provided by each publication's authors. Key descriptors were derived in our abstract and/or full paper review of each publication. Research and laboratory collaborations Many OH activities in India are public health research-based, largely focusing on the etiology and surveillance of zoonotic diseases and sometimes AMR. Intersectoral collaboration happens primarily between human health and animal health scientists, though AMR studies may also involve environmental sampling and/or participation from the environmental sector. In a systematic review of studies from 18 states examining antimicrobial-resistant E. coli across India, only 10% (4 out of 38) conducted interdisciplinary sampling—defined as sampling from a combination of human, animal, or environmental sources ( 19 ). The research studies often feature a multidisciplinary approach, combining social/epidemiological surveys with laboratory-based and/or economic investigations ( Table 1 ). For example, a transdisciplinary team of human and animal health researchers conducted several studies to identify Orientia tsutsugamushi , the causative agent of scrub typhus, as the major etiology of Acute Encephalitis Syndrome (AES) outbreaks in Gorakhpur region, India ( 36 ). These studies led to a better understanding of AES transmission in the region and recommendations for its control. Veterinary institutes are essential partners in OH research and laboratory diagnostics. In early 2020, Tamil Nadu Veterinary and Animal Sciences University (TANUVAS) signed a MoU with The Tamil Nadu Dr MGR Medical University (TNMGRMU) for conducting joint academic and research activities, including an animal trial for a SARS-CoV2 vaccine candidate developed by TNMGRMU. TANUVAS has also performed in vitro testing for some plant-derived SARS-CoV-2 therapeutic compounds (A.B.R. Parthiban, personal communication, August 2022). Guru Angad Dev Veterinary and Animal Sciences University (GADVASU) receives samples from Dayanand Medical College and Hospital, Ludhiana for testing of Brucella spp. Likewise, TANUVAS initiated diagnostic testing for human leptospirosis about two decades ago using a microscopic agglutination test and dark field microscopy, gold standard confirmatory tests for leptospirosis. TANUVAS maintains leptospirosis reference cultures (serovars) for these testing. Even today, TANUVAS remains the only laboratory to offer these tests in tandem in the southern states of India. In contrast, commercial testing laboratories offer only IgM/ELISA-based screening tests. The testing results on leptospirosis are periodically shared with the Tamil Nadu State Public Health Department (A.B.R. Parthiban, personal communication, August 2022). Moreover, the Ministry of Health and Family Welfare (MoHFW) designated two veterinary labs—National Research Center on Equines, Hissar and Central Military Veterinary Laboratory), Meerut—as reference laboratories for the diagnosis of glanders in humans ( 40 ). Disease outbreaks Intersectoral interactions at the village/sub-district level have occurred most frequently during disease outbreaks and have been observed across the country ( 22 , 36 , 41 , 42 ). These interactions often involve OH coordination and response between stakeholders at national, state, and district levels, usually with a top-down approach that directs, initiates, and supports response from lower-level stakeholders. For example, the Indian Council of Medical Research (ICMR)—National Institute of Virology, Pune acted at the local level to build laboratory capacity and develop laboratory networks for quick diagnosis of emerging infectious diseases through their activities managing Kyasanur Forest disease and CCHF outbreaks in India ( 34 ). Similarly, an intersectoral collaboration featuring the MoHFW, Directorate of Health Research, Indian Council of Agricultural Research (ICAR), State Health Department, State Animal Husbandry, and District Administration, during a 2018 Nipah virus outbreak in Kerala led to zero spread and no mortality in a subsequent outbreak the following year ( 42 ). Currently, a OH approach at multiple levels is being utilized to control anthrax in several villages of a tribal district of Odisha ( 41 ); several stakeholders are involved: clinical service providers, program managers and health workers (health care sector), veterinary doctors, livestock inspectors, forest guards (animals care sector) and service utilizer clients, local governance members, non-governmental organizations (NGOs), self-help groups, cattle owners/gatherers, and village residents (community) ( 41 ). Animal health professionals recently supported their human health counterparts during the COVID-19 pandemic. In the Haryana state, veterinarians were deployed in the isolation wards of village hospitals ( 43 ). Veterinarians were not only engaged from the animal husbandry departments for helping medical staff during this pandemic but also from other departments, such as veterinary corps personnel from the Armed forces. They were deployed through the operation "CO-JEET (Victory over COVID; Jeet means Victory)" to support overstretched medical personnel ( 44 ) and were involved in several activities, including setting up of COVID facilities ( 45 ). Enhanced intersectoral coordination in laboratory diagnostics also featured prominently during the COVID pandemic. For example, scientists from several veterinary schools and agencies supported their medical counterparts during the beginning of the pandemic in India ( 46 ). In April 2020, a team of 10 researchers from GADVASU joined the team of medical doctors at the Government Medical colleges, Amritsar and Patiala, providing two Real-Time PCR testing machines and training on how to use these machines for diagnostic testing ( 47 ). Veterinarians in many parts of the country also stepped in to perform temperature screens, COVID tests, and collect patient samples. Even more, in August 2020, a COVID-19 Testing laboratory was established at GADVASU, playing a critical role in diagnosing COVID-19 in Punjab. Within 6 months, the lab tested more than 100,000 samples ( 48 ). Another example of such intersectoral coordination is when TANUVAS partnered with Kings Institute of Preventive Medicine and Council of Scientific and Industrial Research-Center for Cellular and Molecular Biology from April-May 2020 to sequence 21 complete genomes of SARS-CoV2. These were the first complete genome sequences of SARS-CoV2 circulating in Tamil Nadu to be submitted to the Global Initiative on Sharing Avian Influenza Data's Epicov (GISAID's EpiCoV) database (EPI_SET_220907pk, https://doi.org/10.55876/gis8.220907pk ). Community outreach and engagement The OH-related intersectoral coordination and collaboration must involve communities to improve and sustain the prevention, detection and control of zoonotic and emerging pathogens, particularly in high-risk areas. Beyond medical and veterinary personnel, local leaders, community health workers, and NGOs have been instrumental in outbreak response. Collaboration between several stakeholders, for example, the village sarpanch (head of village-level constitutional body), community health officers, auxiliary nursing midwives, and accredited social health activists (ASHA), enabled overcoming the COVID vaccine hesitancy in a village of Punjab ( 49 ). Similarly, grassroots NGOs that work closely with tribal communities in the Nilgiri district of Tamil Nadu were recruited to address vaccine hesitancy in these communities ( 50 ). Another way the OH approach is implemented in communities is through community outreach and education. GADVASU has established the Center for OH that engages with the community to educate them about zoonotic diseases, AMR, food safety, and biosecurity measures in various training and education programs. Under ICAR's Farmers FIRST Programme, a multi-disciplinary team comprising of animal scientists, an agronomist, and a vegetable expert from GADVASU and Punjab Agricultural University worked closely with sarpanches of five villages to educate and help them in quality crop, vegetable, and fodder production, clean milk production, vector control (flies and ticks), mastitis detection, and in the construction of foot baths at farm gates for stronger biosecurity (J. Singh, personal communication, August 2022). In 2016, GADVASU and the State health department organized an awareness camp at the livestock farmers' fair to inform the visitors about symptoms and preventive measures for prevalent diseases such as brucellosis. They also provided free brucellosis testing for farmers, testing 125 samples ( 51 ). Community outreach has also been extended to school children, educating them about rabies, food safety, hygienic practices and AMR ( 52 – 55 ). Aside from in-person events and training, the university disseminates knowledge on OH topics through print (books, magazines), digital tools (apps, YouTube channels, Facebook, etc.), and mass media (newspaper, radio, and television) (J. Singh, personal communication, August 2022). Research and laboratory collaborations Many OH activities in India are public health research-based, largely focusing on the etiology and surveillance of zoonotic diseases and sometimes AMR. Intersectoral collaboration happens primarily between human health and animal health scientists, though AMR studies may also involve environmental sampling and/or participation from the environmental sector. In a systematic review of studies from 18 states examining antimicrobial-resistant E. coli across India, only 10% (4 out of 38) conducted interdisciplinary sampling—defined as sampling from a combination of human, animal, or environmental sources ( 19 ). The research studies often feature a multidisciplinary approach, combining social/epidemiological surveys with laboratory-based and/or economic investigations ( Table 1 ). For example, a transdisciplinary team of human and animal health researchers conducted several studies to identify Orientia tsutsugamushi , the causative agent of scrub typhus, as the major etiology of Acute Encephalitis Syndrome (AES) outbreaks in Gorakhpur region, India ( 36 ). These studies led to a better understanding of AES transmission in the region and recommendations for its control. Veterinary institutes are essential partners in OH research and laboratory diagnostics. In early 2020, Tamil Nadu Veterinary and Animal Sciences University (TANUVAS) signed a MoU with The Tamil Nadu Dr MGR Medical University (TNMGRMU) for conducting joint academic and research activities, including an animal trial for a SARS-CoV2 vaccine candidate developed by TNMGRMU. TANUVAS has also performed in vitro testing for some plant-derived SARS-CoV-2 therapeutic compounds (A.B.R. Parthiban, personal communication, August 2022). Guru Angad Dev Veterinary and Animal Sciences University (GADVASU) receives samples from Dayanand Medical College and Hospital, Ludhiana for testing of Brucella spp. Likewise, TANUVAS initiated diagnostic testing for human leptospirosis about two decades ago using a microscopic agglutination test and dark field microscopy, gold standard confirmatory tests for leptospirosis. TANUVAS maintains leptospirosis reference cultures (serovars) for these testing. Even today, TANUVAS remains the only laboratory to offer these tests in tandem in the southern states of India. In contrast, commercial testing laboratories offer only IgM/ELISA-based screening tests. The testing results on leptospirosis are periodically shared with the Tamil Nadu State Public Health Department (A.B.R. Parthiban, personal communication, August 2022). Moreover, the Ministry of Health and Family Welfare (MoHFW) designated two veterinary labs—National Research Center on Equines, Hissar and Central Military Veterinary Laboratory), Meerut—as reference laboratories for the diagnosis of glanders in humans ( 40 ). Disease outbreaks Intersectoral interactions at the village/sub-district level have occurred most frequently during disease outbreaks and have been observed across the country ( 22 , 36 , 41 , 42 ). These interactions often involve OH coordination and response between stakeholders at national, state, and district levels, usually with a top-down approach that directs, initiates, and supports response from lower-level stakeholders. For example, the Indian Council of Medical Research (ICMR)—National Institute of Virology, Pune acted at the local level to build laboratory capacity and develop laboratory networks for quick diagnosis of emerging infectious diseases through their activities managing Kyasanur Forest disease and CCHF outbreaks in India ( 34 ). Similarly, an intersectoral collaboration featuring the MoHFW, Directorate of Health Research, Indian Council of Agricultural Research (ICAR), State Health Department, State Animal Husbandry, and District Administration, during a 2018 Nipah virus outbreak in Kerala led to zero spread and no mortality in a subsequent outbreak the following year ( 42 ). Currently, a OH approach at multiple levels is being utilized to control anthrax in several villages of a tribal district of Odisha ( 41 ); several stakeholders are involved: clinical service providers, program managers and health workers (health care sector), veterinary doctors, livestock inspectors, forest guards (animals care sector) and service utilizer clients, local governance members, non-governmental organizations (NGOs), self-help groups, cattle owners/gatherers, and village residents (community) ( 41 ). Animal health professionals recently supported their human health counterparts during the COVID-19 pandemic. In the Haryana state, veterinarians were deployed in the isolation wards of village hospitals ( 43 ). Veterinarians were not only engaged from the animal husbandry departments for helping medical staff during this pandemic but also from other departments, such as veterinary corps personnel from the Armed forces. They were deployed through the operation "CO-JEET (Victory over COVID; Jeet means Victory)" to support overstretched medical personnel ( 44 ) and were involved in several activities, including setting up of COVID facilities ( 45 ). Enhanced intersectoral coordination in laboratory diagnostics also featured prominently during the COVID pandemic. For example, scientists from several veterinary schools and agencies supported their medical counterparts during the beginning of the pandemic in India ( 46 ). In April 2020, a team of 10 researchers from GADVASU joined the team of medical doctors at the Government Medical colleges, Amritsar and Patiala, providing two Real-Time PCR testing machines and training on how to use these machines for diagnostic testing ( 47 ). Veterinarians in many parts of the country also stepped in to perform temperature screens, COVID tests, and collect patient samples. Even more, in August 2020, a COVID-19 Testing laboratory was established at GADVASU, playing a critical role in diagnosing COVID-19 in Punjab. Within 6 months, the lab tested more than 100,000 samples ( 48 ). Another example of such intersectoral coordination is when TANUVAS partnered with Kings Institute of Preventive Medicine and Council of Scientific and Industrial Research-Center for Cellular and Molecular Biology from April-May 2020 to sequence 21 complete genomes of SARS-CoV2. These were the first complete genome sequences of SARS-CoV2 circulating in Tamil Nadu to be submitted to the Global Initiative on Sharing Avian Influenza Data's Epicov (GISAID's EpiCoV) database (EPI_SET_220907pk, https://doi.org/10.55876/gis8.220907pk ). Community outreach and engagement The OH-related intersectoral coordination and collaboration must involve communities to improve and sustain the prevention, detection and control of zoonotic and emerging pathogens, particularly in high-risk areas. Beyond medical and veterinary personnel, local leaders, community health workers, and NGOs have been instrumental in outbreak response. Collaboration between several stakeholders, for example, the village sarpanch (head of village-level constitutional body), community health officers, auxiliary nursing midwives, and accredited social health activists (ASHA), enabled overcoming the COVID vaccine hesitancy in a village of Punjab ( 49 ). Similarly, grassroots NGOs that work closely with tribal communities in the Nilgiri district of Tamil Nadu were recruited to address vaccine hesitancy in these communities ( 50 ). Another way the OH approach is implemented in communities is through community outreach and education. GADVASU has established the Center for OH that engages with the community to educate them about zoonotic diseases, AMR, food safety, and biosecurity measures in various training and education programs. Under ICAR's Farmers FIRST Programme, a multi-disciplinary team comprising of animal scientists, an agronomist, and a vegetable expert from GADVASU and Punjab Agricultural University worked closely with sarpanches of five villages to educate and help them in quality crop, vegetable, and fodder production, clean milk production, vector control (flies and ticks), mastitis detection, and in the construction of foot baths at farm gates for stronger biosecurity (J. Singh, personal communication, August 2022). In 2016, GADVASU and the State health department organized an awareness camp at the livestock farmers' fair to inform the visitors about symptoms and preventive measures for prevalent diseases such as brucellosis. They also provided free brucellosis testing for farmers, testing 125 samples ( 51 ). Community outreach has also been extended to school children, educating them about rabies, food safety, hygienic practices and AMR ( 52 – 55 ). Aside from in-person events and training, the university disseminates knowledge on OH topics through print (books, magazines), digital tools (apps, YouTube channels, Facebook, etc.), and mass media (newspaper, radio, and television) (J. Singh, personal communication, August 2022). Discussion The OH approach to addressing zoonotic diseases amid increasing enablers like globalization, urbanization, climate change, etc., is critical globally and especially for India. As evidenced by the number of initiatives and publications describing initiatives or suggested frameworks for OH implementation identified through our literature review, there appears to be excellent support for implementing OH in India. Addressing zoonotic diseases in India is complicated because of the complex and non-integrated infrastructure of its health (animal and human) and environmental agencies. Zoonotic diseases in humans, domestic animals and wild animals fall under the purview of several ministries, including the MoHFW, the Ministry of Agriculture, and the Wildlife Institute of India, respectively ( 56 ). Furthermore, issues relating to forests and climate change come under the purview of the Ministry of Environment, Forest and Climate Change. These different ministries often have focused and non-overlapping objectives and priorities, challenging coordinated operationalization of the OH approach for zoonotic diseases in India. Moreover, while overarching policies, regulations, and outbreak response are managed from the federal level, animal and human health initiatives are implemented by state governments, which set their own priorities based on the local context and budget. Central frameworks, dedicated funding, or data-sharing mechanisms to implement OH across sectors are lacking because of this governmental structure. Other factors such as infrastructure, zoonotic disease knowledge, training, response capacity, inter-sector politics, and disparate human and animal disease reporting systems make implementing OH in India challenging from local to central levels ( 14 ). Efforts to improve the implementation of OH through enhanced cross-sectoral coordination and collaboration at the national and state level are ongoing ( 14 ). The Roadmap to Combat Zoonoses in India was launched in 2008 to facilitate coordination among human, animal and wildlife health sectors, and to identify priority research areas for zoonoses ( 56 ). In 2017, India prepared a National Action Plan on AMR, taking the OH approach involving human, animal and environmental sectors ( 10 ). Additionally, a national-level OH Consortium led by the Department of Biotechnology -National Institute of Animal Biotechnology, Hyderabad has recently been established, consisting of 27 organizations, including several medical, veterinary, and wildlife agencies ( 57 ), and zoonotic research toward the establishment of a National Institute of OH in India has been jointly supported by the ICMR and the ICAR ( 11 ). The objective of our literature review was to identify examples of local OH activities and coordination. We focused primarily on published peer-reviewed literature and did not include gray literature in our search methods, which might have limited the activities we found reported. In general, we found that most of the literature focusing on OH in India could be characterized as (1) Studies describing public health and laboratory research, including reviews, (2) Policy or social sciences research or analysis of Indian OH frameworks, and (3) Case studies of OH relevant programs or activities. Within these publications, we found that many of local OH activities occur through public health or laboratory research, often involving animal and human health researchers focusing on zoonotic diseases. We also found that outbreaks instigated local intersectoral coordination and collaboration, and there are some examples of community involvement or outreach in local OH activities. Given the importance of OH action at the local level, it is encouraging to see research-related activities and implementation at this level. The documented research has largely focused on zoonotic diseases or AMR for public health surveillance. Intersectoral collaboration mostly happens between animal and human health researchers, though occasionally environmental sampling occurs, primarily for AMR studies. In general, we noted that the ecological sector was rarely engaged even when studies had a OH approach. Here, there is an opportunity for increased involvement of researchers outside of animal and public health fields, particularly within the environmental health, agricultural, and wildlife management sectors. As factors related to the transmission, emergence, and general etiology of zoonotic diseases are influenced by local environmental dynamics, this is a critical sector to engage both in research studies and in OH activity coordination. A combination of survey and laboratory-based methods is often used in documented studies, including some incorporating economic analysis. Not only is multidisciplinary engagement critical for OH research, but similarly are multidisciplinary methods and input, which should be considered in future studies. The OH Consortium is an excellent initiative to enable intersectoral collaboration at multiple levels on OH activities through research; research collaborations established through other OH studies could support and integrate into this initiative. There are several examples in which the veterinary community has supported diagnostics for diseases in human patients. The success of veterinary-supported COVID diagnostic testing and training demonstrates that veterinary health professionals can be important partners during human disease outbreaks, supporting the health system and workforce through laboratory diagnostics, clinical screening, and other logistics. Furthermore, they can also perform routine testing for zoonotic diseases in human samples outside of health emergencies. Cross-functional training in laboratory diagnostics could benefit both sectors in times of emergency. Intersectoral collaboration is most common during disease outbreaks and stems from the urgent demand for action and human resources. A network analysis by Yasobant et al. shows that intersectoral interactions between actors at different levels (administrative, service providers, and community) of the health system in Ahmedabad, India are greater during an outbreak situation than during non-outbreak periods and they are directed in a top-down approach ( 22 ). Outside of outbreaks, there is little incentive for intersectoral collaboration. Yasobant et al. also found that solution-based OH initiatives or activities in India during outbreaks or health emergencies are robust and implemented at the grassroots level, but more level-based and integrated into third-party collaborations (example inter-ministerial OH task force) outside of outbreaks are necessary for a robust and resilient Indian health system ( 24 ). Community-level engagement in or involvement in OH activities is lacking, as evidenced by our literature review. A qualitative and quantitative stakeholder analysis reveals that better hygiene and practices at the community level, communication at the grass root level, and development of OH Cell at the community level were among the top essential strategies for the operationalization of OH in the prevention and control of zoonotic diseases in Ahmedabad, India ( 32 ). Dasgupta et al. similarly emphasize the importance of strengthening community surveillance and community engagement through a bottom-to-top approach operationalized by community-based human and animal health workers, forest officers and rangers, farmers, and domestic animal owners across genders ( 21 ). Engaging community health workers may require incentives, as female (ASHA) and Male Multipurpose Health workers interviewed in Ahmedabad expressed low interest in serving as OH activists without financial compensation, recognition, and operational support or mandates ( 16 ). While not documented in the published literature, the community outreach that veterinary institutions provide are vital. OH outreach to communities and improving awareness about zoonotic diseases, particularly in rural areas and farming communities, could have a tremendous impact on public engagement to address zoonotic diseases and prevention at the local level. Most Indian farmers are small livestock holders, raising animals for their sustainability. Their animals are usually kept in or near the same premises, increasing contact between animals and family members. Children also assist their parents in raising and caring for animals, activities that may expose them to zoonotic pathogens. However, there are gaps in zoonotic disease awareness in both adults ( 58 – 61 ) and children. A study conducted on school children (aged 11–16 years) in a semi-urban area of Karnataka showed that 31% of them ( n = 320) were not aware of the initial steps to be taken after a dog bite, demonstrating the need to educate children about rabies, an endemic zoonotic disease in India ( 62 ). Therefore, zoonotic or OH outreach should be delivered to communities, including primary and secondary school education. Conclusion The heterogeneity of India necessitates the OH framework that is robust at the local level. Factors that drive the emergence and persistence of zoonotic diseases, including high animal-human interactions, climate factors, etc., must be addressed at the grassroots level, the level at which they occur. From our review of the literature, it is apparent that OH-driven intersectoral coordination is built through research activities or ramped up ad hoc at the local level, for example, during outbreaks. This coordination should be formalized and sustained for ongoing, day-to-day activities like surveillance and disease notification that are important for detecting and responding to zoonotic diseases. Finally, the sustainability and success of OH frameworks depend upon the operationalization at the local level, including being conducted by local intersectoral teams with the knowledge and experience in the socio-cultural and political context of the specific area or community. Data availability statement The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Author contributions JT, FP, BS, and ND conceptualized the manuscript topic. JT, FP, BS, ND, RS, JS, and AP contributed to the writing and editing of the manuscript. JT, RS, AP, JS, PK, BS, ND, and FP reviewed and extracted information from selected publications in literature review. RS, AP, JS, and BS contributed key inputs on local One Health activities occurring in India, including GADVASU and TANUVAS activities. JG and DG provided oversight and review of content and manuscript writing process. All authors contributed to the article and approved the submitted version. Conflict of interest JT and FP are employed by EpiPointe. ND is a consultant to EpiPointe. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6637489/

Establishing a theoretical foundation for measuring global health security: a scoping review

Background Since the 2014–2016 West Africa Ebola epidemic, the concept of measuring health security capacity has become increasingly important within the broader context of health systems-strengthening, enhancing responses to public health emergencies, and reducing global catastrophic biological risks. Efforts to regularly and sustainably track the evolution of health security capabilities and capacities over time – while also accounting for political, social, and environmental risks – could help countries progress toward eliminating sources of health insecurity. We sought to aggregate evidence-based principles that capture a country's baseline public health and healthcare capabilities, its health security system performance before and during infectious disease crises, and its broader social, political, security, and ecological risk environments. Methods We conducted a scoping review of English-language scholarly and gray literature to identify evidence- and practice-based indicators and proxies for measuring health security at the country level over time. We then used a qualitative coding framework to identify recurrent themes in the literature and synthesize foundational principles for measuring global health security. Documents reviewed included English-language literature published after 2001 until the end of the research period—September 2017—to ensure relevance to the current global health security landscape; literature examining acute infectious disease threats with potential for transnational spread; and literature addressing global health security efforts at the country level. Results We synthesized four foundational principles for measuring global health security: measurement requires assessment of existing capacities, as well as efforts to build core public health, healthcare, and biosecurity capabilities; assessments of national programs and efforts to mitigate a critical subset of priority threats could inform efforts to generate useful metrics for global health security; there are measurable enabling factors facilitating health security-strengthening efforts; and finally, measurement requires consideration of social, political, and ecological risk environments. Conclusion The themes identified in this review could inform efforts to systematically assess the impacts and effectiveness of activities undertaken to strengthen global health security. Electronic supplementary material The online version of this article (10.1186/s12889-019-7216-0) contains supplementary material, which is available to authorized users. Background Since the 2014–2016 West Africa Ebola epidemic, the concept of measuring health security capacity has become increasingly important within the broader context of health systems-strengthening, enhancing responses to public health emergencies, and reducing global catastrophic biological risks. Efforts to regularly and sustainably track the evolution of health security capabilities and capacities over time – while also accounting for political, social, and environmental risks – could help countries progress toward eliminating sources of health insecurity. We sought to aggregate evidence-based principles that capture a country's baseline public health and healthcare capabilities, its health security system performance before and during infectious disease crises, and its broader social, political, security, and ecological risk environments. Methods We conducted a scoping review of English-language scholarly and gray literature to identify evidence- and practice-based indicators and proxies for measuring health security at the country level over time. We then used a qualitative coding framework to identify recurrent themes in the literature and synthesize foundational principles for measuring global health security. Documents reviewed included English-language literature published after 2001 until the end of the research period—September 2017—to ensure relevance to the current global health security landscape; literature examining acute infectious disease threats with potential for transnational spread; and literature addressing global health security efforts at the country level. Results We synthesized four foundational principles for measuring global health security: measurement requires assessment of existing capacities, as well as efforts to build core public health, healthcare, and biosecurity capabilities; assessments of national programs and efforts to mitigate a critical subset of priority threats could inform efforts to generate useful metrics for global health security; there are measurable enabling factors facilitating health security-strengthening efforts; and finally, measurement requires consideration of social, political, and ecological risk environments. Conclusion The themes identified in this review could inform efforts to systematically assess the impacts and effectiveness of activities undertaken to strengthen global health security. Electronic supplementary material The online version of this article (10.1186/s12889-019-7216-0) contains supplementary material, which is available to authorized users. Introduction "Global health security" refers to prevention, detection, and response to naturally emerging, accidental, and deliberate biological threats [ 1 ]. Since the 2014 West Africa Ebola epidemic, the concept of health security has become increasingly important within the broader context of health systems-strengthening, enhancing responses to public health emergencies, and global catastrophic biological risks [ 2 ]. In this vein, the World Health Organization's (WHO) Ebola Interim Assessment Panel, the International Working Group on Financing Preparedness, Chatham House, Harvard University's Global Health Institute, the National Academy of Medicine, and the World Bank Group have issued calls to improve monitoring and measurement efforts around global health security [ 1 , 3 – 5 ]. The WHO's Joint External Evaluation (JEE) tool partially addresses this need by articulating country-level capacities required to mitigate infectious disease threats; the JEE also establishes a scoring system for quantifying progress made toward meeting benchmarks specified in the International Health Regulations (IHR). However, the JEE process is voluntary and relies on an in-country assessment and in-kind contributions of personnel who conduct the evaluation. While this process remains vital, additional universal approaches to measuring baseline, country-level health security are needed. Efforts to regularly and sustainably track and reproducibly compare the evolution of health security capabilities and capacities over time – while also accounting for political, social, security, and environmental risks – could help countries progress toward eliminating sources of health insecurity. To inform ongoing efforts to strengthen health systems and establish new mechanisms for monitoring health security -- and as a preliminary step in an ongoing project to develop a Global Health Security Index -- we performed a scoping literature review to articulate foundational principles for measuring global health security. Our objective was to identify evidence-based principles that not only capture health security capabilities before and during infectious disease crises, but also a country's baseline public health and healthcare capacities and its broader social, environmental, and political risk environments. Themes that emerged from the literature informed our selection of indicators and sub-indicators that could help conceptualize and quantify health security capacities at the country level. In this paper, we summarize the themes we identified in the literature and also offer suggestions for improving future efforts to measure country-level health security. Methods We conducted a scoping review of the biomedical and social science scholarly literature, as well as the gray literature. We searched PubMed, Web of Science, and OAIster using the search terms and search limits outlined in Fig. 1 . Because this review extracted data from secondary sources and did not involve human subjects research, ethical approval was not required. Fig. 1 Search Terms Search Terms Documents eligible for review included only English-language literature published after 2001—to ensure relevance to the current global health security landscape—until the end of the research period (September 2017). Sources were selected if they examined acute infectious disease threats with potential for transnational spread and addressed health security-strengthening efforts at the country level. Documents were excluded if they addressed health security at subnational levels (e.g. county-, district, and/or province-level); biological threats without national or international consequences; or plant, animal, or marine infectious disease threats without known implications for human health. After exclusion of duplicate titles (i.e., those titles that came up in more than one search), all titles were reviewed by one researcher for relevancy using the above inclusion and exclusion criteria. For those included articles, the abstracts were then reviewed to determine relevancy. All articles deemed relevant were then read in its entirety by a researcher to identify recurrent themes and proxies for measuring global health security. Using NVivo software and a qualitative coding framework developed from a priori themes derived from the JEE and previous global health security research, we coded the documents iteratively, adding new codes to the framework as we identified additional global health security themes and indicators. From the coding process, we synthesized foundational principles for measuring global health security. Availability of data and materials The qualitative coding framework containing all of the themes we identified, and a full list of documents reviewed are provided in Additional file 1 and Additional file 2 , respectively. Results Our search initially yielded 1092 articles from PubMed, 440 articles from OAIster, and 356 articles from Web of Science (1888 total). We eliminated 396 duplicate documents, and, using the aforementioned criteria, eliminated another 1255 documents that were deemed irrelevant upon review of their titles and abstracts, producing a final set of 237 documents which were subsequently coded using NVivo 11 Pro qualitative software (see Fig. 2 ). Fig.2 Study Selection Following are major thematic findings synthesized from our review of the literature, which could serve as foundational principles for measuring global health security. While these thematic findings were derived from our analysis of all 237 documents, we have cited only the documents that we deemed to be most illustrative of said themes within the body of this review. For reference, Additional fil 1 contains a full list of themes, matched with all of the corresponding documents from which they were extracted. Measuring global health security requires analysis of existing prevention, detection, and response capacities, as well as efforts to build core public health, healthcare, and biosecurity capacities Our review broadly affirmed the importance of assessing baseline country capacities for preventing, detecting, and responding to infectious disease threats, an approach widely adopted by existing assessment tools, including the JEE. We identified themes targeting three broad areas: performance of critical health security systems, biosafety and biosecurity, and public health preparedness. With respect to measuring system performance, the literature underscored the importance of assessing biosurveillance systems, emergency response systems, and public health laboratories [ 6 – 8 ]. Specific biosurveillance system capacities and capabilities include the presence of formal programs for monitoring influenza, foodborne pathogens, and wildlife; robust reporting mechanisms; and indicator-based, sentinel surveillance and early warning systems for outbreak detection [ 9 , 10 ]. We found few descriptions of the operational capabilities required for emergency response; those identified include the ability to coordinate communication between emergency response partners, healthcare surge capacity, the presence of business continuity plans, and sustaining essential services during a crisis [ 11 ]. Many of the laboratory system performance themes we identified – including, but not limited to, the presence of national reference laboratories, the quality of diagnostic capacities for priority diseases, protocols for shipping hazardous specimens, and the presence of accreditation and biosafety policies – originated from the IHR Core Capacity Monitoring Framework (2013) [ 16 ]. Our review also highlighted the interdependence between global health security, biosecurity, and biosafety. As such, we sought to determine how best to capture the performance of national biosecurity and biosafety mechanisms [ 12 – 14 ]. Given that the definition of "biosecurity" often varies between countries, some of the themes we identified were similarly divergent [ 15 ]. Despite some differences in definition, the literature broadly affirmed that robust biosecurity and biosafety mechanisms are critical components of global health security. National laws, regulations, policies, and protocols for enforcing biosecurity and biosafety standards were broadly cited as important components of biosecurity and biosafety at the country level [ 16 – 18 ]. Additionally, the literature consistently underscored the importance of oversight and governance, particularly in the context of reducing risks in the life sciences: national select agent programs, institutional biosafety committees and other deliberative oversight bodies, codes of conduct and ethics, and educational initiatives for scientists and policymakers emerged as important features of robust oversight and governance mechanisms [ 18 – 20 ]. The literature also underscored the essential role of public health preparedness capacities in enhancing health security. Capacities for medical countermeasure development, deployment, and stockpiling – particularly of vaccines – were widely cited as important indicators of health security [ 21 ]. Other sources highlighted the importance of access to nonmedical countermeasures (e.g. personal protective equipment, masks, and respirators) in ensuring robust health sector responses to emergent threats [ 22 ]. Notably, though our review elicited some indicators for assessing healthcare delivery during infectious disease crises, few documents examined the roles of healthcare in global health security-strengthening efforts. Those that did addressed infection control in clinical settings; the merits and downsides of isolation and quarantine during severe outbreaks; surge capacities during public health crises and mass-casualty events; and coalition-building as a strategy for enhancing regional healthcare capacities [ 11 , 22 – 24 ]. Our review also highlighted the importance of risk assessment, which was widely cited as an important tool for characterizing threats across the spectrum of biological risk [ 25 – 27 ]. The literature also emphasized the importance of measuring risk communication capabilities, given the social and economic costs associated with public anxiety, panic, and unrest that often accompanies health crises [ 28 ]. Finally, the themes of workforce availability and training cross-cut nearly every health security capacity identified in our review. Healthcare workforces and public health professionals in particular were singled out as critical frontline defenses against emergent threats [ 28 , 29 ]. Assessments of national programs and efforts to mitigate a critical subset of priority threats could serve as useful proxies for measuring global health security Zoonotic diseases trigger devastating economic losses in the agricultural industry and pose threats to human health, particularly for those working in poultry and swine operations. The literature indicated that infection control and occupational guidelines within these operations – including disinfection, vaccination, and use of personal protective equipment – could reduce the risk of disease transmission [ 30 – 32 ]. Programs and policies within the wild game industry (e.g. safe animal handling and regulation of trade between hunters and market owners) are also instrumental in reducing zoonotic transmission [ 33 ]. Disease surveillance among wild and domestic animal populations and collaborative approaches to threat mitigation between the human and animal health sectors were also cited as important safeguards against zoonotic disease outbreaks [ 33 ]. Widespread emergence of antimicrobial-resistant (AMR) pathogens has diminished the effectiveness of many first-line drugs. Thus, national standards, policies, and programs promoting antimicrobial stewardship (both in human and animal populations) could help mitigate AMR threats, thereby strengthening health security [ 31 , 34 ]. Additionally, sentinel testing for drug resistance – including among pervasive infectious diseases such as tuberculosis – and increasing access to the diagnostic services needed to detect these pathogens are critical measures for reducing burdens of AMR pathogens [ 34 ]. Mass gatherings present additional health security challenges by amplifying the risk of disease transmission both in host countries and countries of returning attendees. As such, the frequency of mass gatherings, as well as the size, location, duration, and season of gathering, could serve as another indicator for measuring countries' health security vulnerabilities; for example, outbreaks of bacterial meningitis linked to annual Hajj pilgrimages resulted in global spread in 1987 and 2000 [ 21 , 35 ]. Implementing mandatory vaccination policies, as done by the Saudi Ministry of Health, could help mitigate disease transmission during mass gatherings [ 21 ]. Additionally, adherence to global standards for planning for mass gatherings (such as those developed by the WHO) emerged as another important factor to consider when assessing country-level health security [ 35 ]. Lastly, the literature noted that mass gatherings could become targets of deliberate biological attacks. In advance of the 2009 Beijing Olympics, for example, China enacted enhanced emergency preparedness measures, including stockpiling pharmaceuticals. Global health security measurements should ideally capture country-level capacities for implementing such measures against deliberate attacks [ 36 ]. In addition to efforts addressing mass gatherings, the literature cited national programs targeting risks associated with biotechnology and the life sciences as similarly important components of global health security [ 37 – 40 ]. The life sciences are a critical tool for advancing health security, but could pose threats in the hands of a malicious actor intending to cause deliberate harm. Policies and programs addressing dual research of concern, DURC (i.e. research that could be misused in a way that endangers the public's health) could ensure proper oversight of entities working with potentially dangerous pathogens. Keys to successful oversight include cooperation between government, academia, the private sector, and law enforcement [ 41 , 42 ]. There are measurable enabling factors that facilitate global health security-strengthening efforts Historical experiences with infectious disease outbreaks can act as an impetus for biosecurity programs, policies, and funding needed to prevent future crises. For example, the September 2001 terrorist and anthrax attacks in the United States catalyzed increased spending and support for biodefense programs [ 8 ]. However, there are notable exceptions wherein prior experiences with catastrophic outbreaks in low- and middle-income countries do not always culminate in full preparedness and response capacities across a country or region. This phenomenon has been most recently illustrated by a major outbreak of Ebola in the Democratic Republic of the Congo and an outbreak of Lassa fever in Nigeria, despite the 2014–16 West Africa Ebola outbreak highlighting critical shortcomings in the region's health security capacities [ 43 , 44 ]. Still, examining how past crises subsequently trigger changes in spending and programmatic support could elucidate how country-level health security evolves over time. Additionally, early disease detection and prevention depends on collaboration and communication between health authorities at local, national, regional, and international levels. International norms and strategies play important roles in promoting international collaboration and communication; the IHR, for example, include a directive for signatories to help build health security capacities in resource-poor countries [ 35 , 45 ]. Other efforts, such as the One Health Initiative, have highlighted intersections between human, environmental, and animal health and the need for greater coordination between these sectors [ 10 ]. Evaluating adherence to established norms and incorporation of new approaches to preventing infectious disease crises could aid in determining a country's collaborative efforts with international partners, as well as the extent to which its animal, human, and environmental health sectors have aligned to tackle emergent threats. In addition to norms, laws, policies, and regulations also shape health security approaches and outcomes (and vice-versa), and their presence or absence could further modulate a country's ability to mitigate infectious disease threats. Anema notes, for instance, that many IHR signatories have met the specified core capacity requirements for establishing national legislation and policy; among these states, those which "centralized and harmonized their public health policies and practices" demonstrated greater capacities for overall IHR compliance [ 16 ]. Additionally, geopolitical and economic instability were also found to modulate state vulnerability to health security threats; Linacre, for example, notes that countries with low GDPs and primarily agrarian economies are uniquely vulnerable to the threat of agroterrorism, given its potential to slow economic growth. Poor economies, in turn, could subsequently give rise to social unrest and insurgent activity [ 15 ]. In addition to highlighting linkages between law, policy, and global health security, the literature broadly affirmed the value of participation in global multilateral institutions (e.g. the World Health Organization; the Global Fund to Fight AIDS, Tuberculosis and Malaria; GAVI, the Vaccine Alliance; and UNAIDS) and compliance with international agreements aimed at strengthening global health security (e.g. the IHR, the Biological and Toxin Weapons Convention, the Cartagena Protocol on Biosafety) [ 8 , 17 , 46 , 47 ]. However, at least one article noted that multilateral health initiatives run the risk of establishing parallel health service delivery systems and financing schemes that could disincentivize efforts to build and strengthen in-country mechanisms for mitigating infectious disease threats [ 17 ]. Therefore, metrics for evaluating the global risk environment should ideally assess a given country's health security capabilities against its reliance on supranational governance structures and non-governmental funding streams. With respect to local and regional collaboration around global health security, the literature highlighted the importance of engaging civil society and private-sector stakeholders, law enforcement, the intelligence community, academia, and political leaders [ 41 , 42 , 48 , 49 ]. The extent to which these non-public health entities could serve as another indicator of the robustness of a country's collaborative health security efforts. Formulation of national strategic plans that coordinate multisector efforts to prevent infectious disease crises could serve as an additional indicator for country-level health security. For example, the U.S. National Strategy for Countering Biological Threats offers guidance for averting catastrophic biological events that could threaten national security [ 50 ]. Additionally, funds and resources offered through national programs – such as those offered through the U.S.'s Hospital Preparedness Program – could further incentivize multisector collaboration [ 51 ]. Political leadership and commitment are instrumental in ensuring that health security remains a top priority. The U.S. federal government, for example, has launched global health programs, such as the President's Emergency Plan for AIDS Relief, the President's Malaria Initiative, the Global Disease Detection Program of the U.S. Centers for Disease Control and Prevention, and the Emerging Pandemic Threats Program through the United States Agency for International Development. Political support also helps ensure adequate funding for biosecurity programs, such as the U.S. Department of Defense Cooperative Threat Reduction Program Biological Threat Reduction Program and the U.S. Department of State Biosecurity Engagement Program. As such, the federal budget typically includes funds for "both biodefense and non-biodefense goals and applications," which address a range of public health, healthcare, national security, and international security issues in addition to biosecurity, and improve preparedness and response [ 52 , 51 ]. Globally, sustained financial investments also facilitate country progress toward meeting IHR benchmarks. Besides the U.S., other countries have made financial commitments to strengthening global health security, including Australia, which recently established an Indo-Pacific Centre for Health Security; Finland, which has assumed a leading role in advancing global health security efforts worldwide; Canada, through the Global Affairs Canada Weapons Threat Reduction Program and Public Health Agency Canada; and the Republic of Korea, made an early pledge of USD$100 million to support the Global Health Security Agenda [ 53 – 56 ]. Finally, legislative frameworks for biosecurity and biosafety may also be useful measures of country-level health security. The U.S. Federal Select Agent Program, for example, defines a set of microorganisms and toxins deemed to carry high risks of deliberate misuse and establishes strict security requirements for facilities working with designated high-priority agents; in 2017, the U.S. also implemented policies addressing oversight of pathogens with pandemic potential and enacted a brief moratorium on federally funded gain-of-function studies involving such pathogens [ 57 , 58 ]. Similarly, in 2008, Israel passed the Regulation of Research into Biological Disease Agents Law, which calls for an oversight body to monitor research involving select, high-risk pathogens [ 59 ]. In addition, any legislative frameworks underpinning research should respect the rights of the individual, including ethical considerations when conducting human subject research, and also when implementing public health policies (i.e. quarantine and monitoring policies). Integration of such ethical considerations into legal frameworks will likely become increasingly important in the context of risks associated with biotechnology and the life sciences [ 60 , 61 ]. Measuring global health security requires consideration of the risk environment from which infectious disease threats might emerge The interplay between built and natural environments, climate, and human and animal activity is a critical determinant of global health security [ 62 – 64 , 46 ]. Variable economic and political conditions, regulatory environments, modes of governance, laws, and policies concurrently shape state vulnerability to infectious disease threats. As such, the concept of the "risk environment" (i.e. the socioeconomic, political, regulatory, and ecological factors that could give rise to health insecurity) is a useful paradigm for measuring global health security in the context of a country's unique baseline conditions. Additionally, geopolitical factors, such as the presence of extremist groups, ongoing conflict, and modes of governance should be considered when identifying sources of health insecurity. Globalization was widely cited as a driving force behind transnational disease spread. Increasing travel, tourism, and trade were described as important promoters of economic growth, but ones that concomitantly elevate the risk of disease transmission among highly mobile populations connected through increasingly accessible modes of global transit [ 35 ]. Efforts to measure global health security might endeavor to capture this tension between protecting public health and ensuring unrestricted commercial activity, such as the presence of health checkpoints at ports of entry, trade embargoes, laws or policies restricting the trade of high-risk products, and safety standards for commercial goods. For example, following a 1986 outbreak of bovine spongiform encephalopathy in the United Kingdom, several countries banned imports of live ruminants and beef originating from the UK [ 47 ]. The literature also identified urbanization, changes in land use, and evolving agricultural practices as important components of the health security risk environment, given their roles in modulating disease spread between humans and animals, and creating conditions that could enable emergence of novel zoonoses [ 33 , 65 , 66 ]. The risks posed by increased proximity between human and animal populations manifests in other contexts as well; as noted previously, several documents described mass gatherings as potential catalysts of disease transmission, citing public health risks associated with the Hajj and the Olympic Games [ 21 , 35 ]. Additionally, Brioudes and Gummow note that certain trade and agricultural practices – including swill feeding, illegal wildlife trading, and introductions of undeclared goods – could further heighten the risk of exposing emerging pathogens to susceptible human or animal populations [ 65 ]. Measuring global health security requires analysis of existing prevention, detection, and response capacities, as well as efforts to build core public health, healthcare, and biosecurity capacities Our review broadly affirmed the importance of assessing baseline country capacities for preventing, detecting, and responding to infectious disease threats, an approach widely adopted by existing assessment tools, including the JEE. We identified themes targeting three broad areas: performance of critical health security systems, biosafety and biosecurity, and public health preparedness. With respect to measuring system performance, the literature underscored the importance of assessing biosurveillance systems, emergency response systems, and public health laboratories [ 6 – 8 ]. Specific biosurveillance system capacities and capabilities include the presence of formal programs for monitoring influenza, foodborne pathogens, and wildlife; robust reporting mechanisms; and indicator-based, sentinel surveillance and early warning systems for outbreak detection [ 9 , 10 ]. We found few descriptions of the operational capabilities required for emergency response; those identified include the ability to coordinate communication between emergency response partners, healthcare surge capacity, the presence of business continuity plans, and sustaining essential services during a crisis [ 11 ]. Many of the laboratory system performance themes we identified – including, but not limited to, the presence of national reference laboratories, the quality of diagnostic capacities for priority diseases, protocols for shipping hazardous specimens, and the presence of accreditation and biosafety policies – originated from the IHR Core Capacity Monitoring Framework (2013) [ 16 ]. Our review also highlighted the interdependence between global health security, biosecurity, and biosafety. As such, we sought to determine how best to capture the performance of national biosecurity and biosafety mechanisms [ 12 – 14 ]. Given that the definition of "biosecurity" often varies between countries, some of the themes we identified were similarly divergent [ 15 ]. Despite some differences in definition, the literature broadly affirmed that robust biosecurity and biosafety mechanisms are critical components of global health security. National laws, regulations, policies, and protocols for enforcing biosecurity and biosafety standards were broadly cited as important components of biosecurity and biosafety at the country level [ 16 – 18 ]. Additionally, the literature consistently underscored the importance of oversight and governance, particularly in the context of reducing risks in the life sciences: national select agent programs, institutional biosafety committees and other deliberative oversight bodies, codes of conduct and ethics, and educational initiatives for scientists and policymakers emerged as important features of robust oversight and governance mechanisms [ 18 – 20 ]. The literature also underscored the essential role of public health preparedness capacities in enhancing health security. Capacities for medical countermeasure development, deployment, and stockpiling – particularly of vaccines – were widely cited as important indicators of health security [ 21 ]. Other sources highlighted the importance of access to nonmedical countermeasures (e.g. personal protective equipment, masks, and respirators) in ensuring robust health sector responses to emergent threats [ 22 ]. Notably, though our review elicited some indicators for assessing healthcare delivery during infectious disease crises, few documents examined the roles of healthcare in global health security-strengthening efforts. Those that did addressed infection control in clinical settings; the merits and downsides of isolation and quarantine during severe outbreaks; surge capacities during public health crises and mass-casualty events; and coalition-building as a strategy for enhancing regional healthcare capacities [ 11 , 22 – 24 ]. Our review also highlighted the importance of risk assessment, which was widely cited as an important tool for characterizing threats across the spectrum of biological risk [ 25 – 27 ]. The literature also emphasized the importance of measuring risk communication capabilities, given the social and economic costs associated with public anxiety, panic, and unrest that often accompanies health crises [ 28 ]. Finally, the themes of workforce availability and training cross-cut nearly every health security capacity identified in our review. Healthcare workforces and public health professionals in particular were singled out as critical frontline defenses against emergent threats [ 28 , 29 ]. Assessments of national programs and efforts to mitigate a critical subset of priority threats could serve as useful proxies for measuring global health security Zoonotic diseases trigger devastating economic losses in the agricultural industry and pose threats to human health, particularly for those working in poultry and swine operations. The literature indicated that infection control and occupational guidelines within these operations – including disinfection, vaccination, and use of personal protective equipment – could reduce the risk of disease transmission [ 30 – 32 ]. Programs and policies within the wild game industry (e.g. safe animal handling and regulation of trade between hunters and market owners) are also instrumental in reducing zoonotic transmission [ 33 ]. Disease surveillance among wild and domestic animal populations and collaborative approaches to threat mitigation between the human and animal health sectors were also cited as important safeguards against zoonotic disease outbreaks [ 33 ]. Widespread emergence of antimicrobial-resistant (AMR) pathogens has diminished the effectiveness of many first-line drugs. Thus, national standards, policies, and programs promoting antimicrobial stewardship (both in human and animal populations) could help mitigate AMR threats, thereby strengthening health security [ 31 , 34 ]. Additionally, sentinel testing for drug resistance – including among pervasive infectious diseases such as tuberculosis – and increasing access to the diagnostic services needed to detect these pathogens are critical measures for reducing burdens of AMR pathogens [ 34 ]. Mass gatherings present additional health security challenges by amplifying the risk of disease transmission both in host countries and countries of returning attendees. As such, the frequency of mass gatherings, as well as the size, location, duration, and season of gathering, could serve as another indicator for measuring countries' health security vulnerabilities; for example, outbreaks of bacterial meningitis linked to annual Hajj pilgrimages resulted in global spread in 1987 and 2000 [ 21 , 35 ]. Implementing mandatory vaccination policies, as done by the Saudi Ministry of Health, could help mitigate disease transmission during mass gatherings [ 21 ]. Additionally, adherence to global standards for planning for mass gatherings (such as those developed by the WHO) emerged as another important factor to consider when assessing country-level health security [ 35 ]. Lastly, the literature noted that mass gatherings could become targets of deliberate biological attacks. In advance of the 2009 Beijing Olympics, for example, China enacted enhanced emergency preparedness measures, including stockpiling pharmaceuticals. Global health security measurements should ideally capture country-level capacities for implementing such measures against deliberate attacks [ 36 ]. In addition to efforts addressing mass gatherings, the literature cited national programs targeting risks associated with biotechnology and the life sciences as similarly important components of global health security [ 37 – 40 ]. The life sciences are a critical tool for advancing health security, but could pose threats in the hands of a malicious actor intending to cause deliberate harm. Policies and programs addressing dual research of concern, DURC (i.e. research that could be misused in a way that endangers the public's health) could ensure proper oversight of entities working with potentially dangerous pathogens. Keys to successful oversight include cooperation between government, academia, the private sector, and law enforcement [ 41 , 42 ]. There are measurable enabling factors that facilitate global health security-strengthening efforts Historical experiences with infectious disease outbreaks can act as an impetus for biosecurity programs, policies, and funding needed to prevent future crises. For example, the September 2001 terrorist and anthrax attacks in the United States catalyzed increased spending and support for biodefense programs [ 8 ]. However, there are notable exceptions wherein prior experiences with catastrophic outbreaks in low- and middle-income countries do not always culminate in full preparedness and response capacities across a country or region. This phenomenon has been most recently illustrated by a major outbreak of Ebola in the Democratic Republic of the Congo and an outbreak of Lassa fever in Nigeria, despite the 2014–16 West Africa Ebola outbreak highlighting critical shortcomings in the region's health security capacities [ 43 , 44 ]. Still, examining how past crises subsequently trigger changes in spending and programmatic support could elucidate how country-level health security evolves over time. Additionally, early disease detection and prevention depends on collaboration and communication between health authorities at local, national, regional, and international levels. International norms and strategies play important roles in promoting international collaboration and communication; the IHR, for example, include a directive for signatories to help build health security capacities in resource-poor countries [ 35 , 45 ]. Other efforts, such as the One Health Initiative, have highlighted intersections between human, environmental, and animal health and the need for greater coordination between these sectors [ 10 ]. Evaluating adherence to established norms and incorporation of new approaches to preventing infectious disease crises could aid in determining a country's collaborative efforts with international partners, as well as the extent to which its animal, human, and environmental health sectors have aligned to tackle emergent threats. In addition to norms, laws, policies, and regulations also shape health security approaches and outcomes (and vice-versa), and their presence or absence could further modulate a country's ability to mitigate infectious disease threats. Anema notes, for instance, that many IHR signatories have met the specified core capacity requirements for establishing national legislation and policy; among these states, those which "centralized and harmonized their public health policies and practices" demonstrated greater capacities for overall IHR compliance [ 16 ]. Additionally, geopolitical and economic instability were also found to modulate state vulnerability to health security threats; Linacre, for example, notes that countries with low GDPs and primarily agrarian economies are uniquely vulnerable to the threat of agroterrorism, given its potential to slow economic growth. Poor economies, in turn, could subsequently give rise to social unrest and insurgent activity [ 15 ]. In addition to highlighting linkages between law, policy, and global health security, the literature broadly affirmed the value of participation in global multilateral institutions (e.g. the World Health Organization; the Global Fund to Fight AIDS, Tuberculosis and Malaria; GAVI, the Vaccine Alliance; and UNAIDS) and compliance with international agreements aimed at strengthening global health security (e.g. the IHR, the Biological and Toxin Weapons Convention, the Cartagena Protocol on Biosafety) [ 8 , 17 , 46 , 47 ]. However, at least one article noted that multilateral health initiatives run the risk of establishing parallel health service delivery systems and financing schemes that could disincentivize efforts to build and strengthen in-country mechanisms for mitigating infectious disease threats [ 17 ]. Therefore, metrics for evaluating the global risk environment should ideally assess a given country's health security capabilities against its reliance on supranational governance structures and non-governmental funding streams. With respect to local and regional collaboration around global health security, the literature highlighted the importance of engaging civil society and private-sector stakeholders, law enforcement, the intelligence community, academia, and political leaders [ 41 , 42 , 48 , 49 ]. The extent to which these non-public health entities could serve as another indicator of the robustness of a country's collaborative health security efforts. Formulation of national strategic plans that coordinate multisector efforts to prevent infectious disease crises could serve as an additional indicator for country-level health security. For example, the U.S. National Strategy for Countering Biological Threats offers guidance for averting catastrophic biological events that could threaten national security [ 50 ]. Additionally, funds and resources offered through national programs – such as those offered through the U.S.'s Hospital Preparedness Program – could further incentivize multisector collaboration [ 51 ]. Political leadership and commitment are instrumental in ensuring that health security remains a top priority. The U.S. federal government, for example, has launched global health programs, such as the President's Emergency Plan for AIDS Relief, the President's Malaria Initiative, the Global Disease Detection Program of the U.S. Centers for Disease Control and Prevention, and the Emerging Pandemic Threats Program through the United States Agency for International Development. Political support also helps ensure adequate funding for biosecurity programs, such as the U.S. Department of Defense Cooperative Threat Reduction Program Biological Threat Reduction Program and the U.S. Department of State Biosecurity Engagement Program. As such, the federal budget typically includes funds for "both biodefense and non-biodefense goals and applications," which address a range of public health, healthcare, national security, and international security issues in addition to biosecurity, and improve preparedness and response [ 52 , 51 ]. Globally, sustained financial investments also facilitate country progress toward meeting IHR benchmarks. Besides the U.S., other countries have made financial commitments to strengthening global health security, including Australia, which recently established an Indo-Pacific Centre for Health Security; Finland, which has assumed a leading role in advancing global health security efforts worldwide; Canada, through the Global Affairs Canada Weapons Threat Reduction Program and Public Health Agency Canada; and the Republic of Korea, made an early pledge of USD$100 million to support the Global Health Security Agenda [ 53 – 56 ]. Finally, legislative frameworks for biosecurity and biosafety may also be useful measures of country-level health security. The U.S. Federal Select Agent Program, for example, defines a set of microorganisms and toxins deemed to carry high risks of deliberate misuse and establishes strict security requirements for facilities working with designated high-priority agents; in 2017, the U.S. also implemented policies addressing oversight of pathogens with pandemic potential and enacted a brief moratorium on federally funded gain-of-function studies involving such pathogens [ 57 , 58 ]. Similarly, in 2008, Israel passed the Regulation of Research into Biological Disease Agents Law, which calls for an oversight body to monitor research involving select, high-risk pathogens [ 59 ]. In addition, any legislative frameworks underpinning research should respect the rights of the individual, including ethical considerations when conducting human subject research, and also when implementing public health policies (i.e. quarantine and monitoring policies). Integration of such ethical considerations into legal frameworks will likely become increasingly important in the context of risks associated with biotechnology and the life sciences [ 60 , 61 ]. Measuring global health security requires consideration of the risk environment from which infectious disease threats might emerge The interplay between built and natural environments, climate, and human and animal activity is a critical determinant of global health security [ 62 – 64 , 46 ]. Variable economic and political conditions, regulatory environments, modes of governance, laws, and policies concurrently shape state vulnerability to infectious disease threats. As such, the concept of the "risk environment" (i.e. the socioeconomic, political, regulatory, and ecological factors that could give rise to health insecurity) is a useful paradigm for measuring global health security in the context of a country's unique baseline conditions. Additionally, geopolitical factors, such as the presence of extremist groups, ongoing conflict, and modes of governance should be considered when identifying sources of health insecurity. Globalization was widely cited as a driving force behind transnational disease spread. Increasing travel, tourism, and trade were described as important promoters of economic growth, but ones that concomitantly elevate the risk of disease transmission among highly mobile populations connected through increasingly accessible modes of global transit [ 35 ]. Efforts to measure global health security might endeavor to capture this tension between protecting public health and ensuring unrestricted commercial activity, such as the presence of health checkpoints at ports of entry, trade embargoes, laws or policies restricting the trade of high-risk products, and safety standards for commercial goods. For example, following a 1986 outbreak of bovine spongiform encephalopathy in the United Kingdom, several countries banned imports of live ruminants and beef originating from the UK [ 47 ]. The literature also identified urbanization, changes in land use, and evolving agricultural practices as important components of the health security risk environment, given their roles in modulating disease spread between humans and animals, and creating conditions that could enable emergence of novel zoonoses [ 33 , 65 , 66 ]. The risks posed by increased proximity between human and animal populations manifests in other contexts as well; as noted previously, several documents described mass gatherings as potential catalysts of disease transmission, citing public health risks associated with the Hajj and the Olympic Games [ 21 , 35 ]. Additionally, Brioudes and Gummow note that certain trade and agricultural practices – including swill feeding, illegal wildlife trading, and introductions of undeclared goods – could further heighten the risk of exposing emerging pathogens to susceptible human or animal populations [ 65 ]. Discussion Our review highlighted country-level considerations that could inform efforts to monitor global health security, including the presence of existing prevention, detection, and response capacities and capabilities; the presence of national programs, policies, and laws to mitigate various kinds of threats; coordination and communication between health authorities at all levels of government, as well as with other sectors (including public and private); political leadership that supports health security issues; as well as engagement in global multilateral institutions and an understanding of the risk environment. There was considerable overlap between the themes identified in the literature and priorities outlined in the Global Health Security Agenda (GHSA), IHR, and JEE tool. Generally, the literature affirmed that many of the goals and indicators specified in both are valuable benchmarks for assessing baseline, country-level health security capacities. The prevent-detect-respond paradigm articulated in the GHSA serves as a useful organizing principle for conceptualizing global health security, and the JEE tool puts forth a valuable methodology for measuring both health security capacities and country progress toward IHR targets. The JEE also rightly prioritizes inter-sectoral discussion between agencies of national governments and is intended as a tool to begin the development of a national action plan for health security containing specific milestones to be filled and financed over a specific set of timeframes. Given that the JEE consists of subjectively assigned scores based largely on qualitative observation, the comparisons between countries and over years of the JEE process over time remains in question. Therefore, in addition to the JEE, efforts to measure global health security over time might benefit from focusing on outcome-based metrics that capture a country's demonstrated prevention, detection, and response capabilities during an infectious disease outbreak. With respect to syndromic surveillance, for example, the JEE considers the size and coverage of a given surveillance system, its electronic reporting capacities, and its methods of data validation; by contrast, Glassman suggests that a country measuring the effectiveness of its syndromic surveillance systems should consider metrics such as rates of disease underreporting or numbers of misidentified cases [ 67 ]. Other such metrics for measuring functionality and effectiveness of surveillance systems could include whether a country has an established mechanism for regular sharing of surveillance data between human, animal, and environmental health authorities. While both approaches have merits and limitations, focusing on metrics that assess demonstrated capabilities against desired outcomes could also help countries chart actionable paths toward increased IHR compliance and greater health security. We found little overlap between the global health security literature – which focuses primarily on acute infectious disease threats – and scholarly work on health system resilience, which encompasses a broader universe of core public health and healthcare assets and functions, such as health workforces, financing, universal health coverage, public risk communication, and integration between the sectors involved in mitigating infectious disease threats. Though our review yielded few tangible metrics for measuring healthcare sector performance, it did underscore the critical roles that healthcare capabilities play in strengthening country-level health security. As demonstrated during the 2014 Ebola outbreak, a healthcare sector's ability to scale up operations in response to an accelerating threat, diagnose and isolate sick patients, and prevent the spread of infection in clinical settings – all while minimizing disruptions to routine healthcare delivery – are crucial determinants of an outbreak's broader community impacts. However, current frameworks for assessing global health security – including the JEE – feature few indicators addressing core healthcare capacities and capabilities and their integration into emergency response activities. Given that a country's infectious disease response capabilities depend on the broader functionality of its health systems, these core public health and healthcare capacities should figure more prominently in efforts to conceptualize and measure global health security. For example, researchers have already sought to identify global and public health core competencies for nursing education; a similar explication of core public health and healthcare functions essential to strengthening health security could enhance efforts to improve infectious disease threat mitigation [ 68 ]. Notably, our review also highlighted the importance of the risk environment. Historically, the risk environment framework has been used in the context of drug-related harm reduction activities, HIV prevention, and assessments of global nuclear security [ 69 , 70 ]. However, many measurement efforts do not account for the full spectrum of social, political, and environmental risks that could give rise to, exacerbate, or mitigate infectious disease crises. Extrapolating the risk environment framework to encompass health security challenges at the population- and system-level underscores the interdependence between built, social, and natural environments, and better captures the potential for disease emergence at the human-animal-ecosystem interface. Furthermore, integrating economic, regulatory, and political considerations into measurements of global health security aligns with observed increases in global travel, migration, and commerce, as well as ongoing changes in land use, climate, and geopolitical stability. In addition to the conceptual challenges associated with measuring global health security, several practical barriers continue posing technical challenges. Even with a strong theoretical foundation, measurement efforts might still be hindered by limited data availability. Many of the metrics employed in the JEE and other tools – both qualitative and quantitative – are not regularly or systematically collected in a standardized manner. Though alternative metrics could support more conceptually sound methods of measurement, poor data availability would still preclude their adoption and meaningful use. International organizations that routinely collect needed metrics should accelerate efforts to identify high-priority data needs, collect said data, and make them readily accessible to the public; the private sector might play an important role in making greater data availability a reality. Finally, although measuring global health security is a critical step in progressing toward greater IHR compliance, measurement alone cannot improve country-level health security capacities. As such, measurement efforts should be linked to incentives (i.e. funding and programmatic support) that promote capacity-building across all sectors involved in infectious disease mitigation. Our investigation had a few limitations; namely, that the themes identified are a function of available published scholarship. The majority of the articles reviewed were produced by researchers representing predominantly high-income countries. As such, the literature skewed heavily toward themes (e.g. dual-use research of concern) that may not be top health security priorities for policymakers in lower-income countries with differing health security threat landscapes (e.g. greater risks from naturally emerging zoonoses) as compared to industrialized nations. Additionally, though we attempted to systematically review the gray literature, our search engines did not produce many of the foundational documents published by the WHO, ministries of health, and non-governmental groups in the health security space. Future efforts to measure global health security might benefit from additional, complementary modes of data collection, such as consultations with subject matter experts. This might also be useful in further refining those capacities and capabilities that are necessary in countries with poor healthcare and public health infrastructure. Finally, we acknowledge that there is ongoing debate over the definition of global health security itself, and that our inclusion and exclusion criteria – while reflective of the definition used by the WHO – might not have allowed us to identify literature proposing alternative definitions. Conclusions In light of ongoing international initiatives aiming to enhance global capacities and capabilities for infectious disease preparedness and response, the evidence-based themes identified in this review could inform efforts to systematically assess the impacts and effectiveness of activities undertaken to strengthen global health security. Additional files Additional file 1: Coding Framework Coding Framework. The formal coding framework used to identify metrics in the literature for measuring global health security. The definition of each code is provided, along with the number of sources in which the code was used, and the number of times the code was used across all sources. (DOCX 29 kb) Additional file 2: GHS Literature Review. Comprehensive List of Literature Reviewed (alphabetical) The full list of gray & scholarly publications identified in our review. (DOCX 35 kb) Additional file 1: Coding Framework Coding Framework. The formal coding framework used to identify metrics in the literature for measuring global health security. The definition of each code is provided, along with the number of sources in which the code was used, and the number of times the code was used across all sources. (DOCX 29 kb) Additional file 2: GHS Literature Review. Comprehensive List of Literature Reviewed (alphabetical) The full list of gray & scholarly publications identified in our review. (DOCX 35 kb)

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6307773/

Different functional states of fusion protein gB revealed on human cytomegalovirus by cryo electron tomography with Volta phase plate

Human cytomegalovirus (HCMV) enters host by glycoprotein B (gB)-mediated membrane fusion upon receptor-binding to gH/gL-related complexes, causing devastating diseases such as birth defects. Although an X-ray crystal structure of the recombinant gB ectodomain at postfusion conformation is available, the structures of prefusion gB and its complex with gH/gL on the viral envelope remain elusive. Here, we demonstrate the utility of cryo electron tomography (cryoET) with energy filtering and the cutting-edge technologies of Volta phase plate (VPP) and direct electron-counting detection to capture metastable prefusion viral fusion proteins and report the structures of glycoproteins in the native environment of HCMV virions. We established the validity of our approach by obtaining cryoET in situ structures of the vesicular stomatitis virus (VSV) glycoprotein G trimer (171 kD) in prefusion and postfusion conformations, which agree with the known crystal structures of purified G trimers in both conformations. The excellent contrast afforded by these technologies has enabled us to identify gB trimers (303kD) in two distinct conformations in HCMV tomograms and obtain their in situ structures at up to 21 à resolution through subtomographic averaging. The predominant conformation (79%), which we designate as gB prefusion conformation, fashions a globular endodomain and a Christmas tree-shaped ectodomain, while the minority conformation (21%) has a columnar tree-shaped ectodomain that matches the crystal structure of the "postfusion" gB ectodomain. We also observed prefusion gB in complex with an "L"-shaped density attributed to the gH/gL complex. Integration of these structures of HCMV glycoproteins in multiple functional states and oligomeric forms with existing biochemical data and domain organization of other class III viral fusion proteins suggests that gH/gL receptor-binding triggers conformational changes of gB endodomain, which in turn triggers two essential steps to actuate virus-cell membrane fusion: exposure of gB fusion loops and unfurling of gB ectodomain. Introduction Human cytomegalovirus (HCMV), a member of the Betaherpesvirinae subfamily of the Herpesviridae family, is a leading viral cause of birth defects [ 1 , 2 ] and a major contributor to life-threatening complications in immunocompromised individuals. As one of the largest membrane-containing viruses, HCMV shares a common multilayered organization with all other herpesviruses, composed of an icosahedrally ordered nucleocapsid enclosing a double-stranded DNA genome, a poorly defined tegument protein layer, and a pleomorphic, glycoprotein-embedded envelope [ 3 ]. During infection, herpesviruses fuse their envelopes with cell membranes, resulting in the delivery of nucleocapsid into the cytoplasm of the host cells. This complex process requires a number of viral glycoproteins and host receptors functioning in a coordinated manner. Glycoproteins gB and gH/gL are conserved across all herpesviruses and are essential for virus entry into cells [ 4 ]. Receptor-binding to gH/gL-containing complexes—the composition of which differs among clinical and laboratory-adapted HCMV strains and across different herpesviruses [ 5 ]—triggers conformational changes of fusion protein gB, leading to fusion of the viral envelope with cell membrane [ 6 ]. This use of both a fusion protein and a receptor-binding complex for herpesvirus entry differs from many other enveloped viruses, which use a single protein for both receptor binding and membrane fusion. Averaging up to tens of thousands of particle images by single-particle cryoEM method has resolved in situ structures of capsid proteins [ 7 – 9 ] and the capsid-associated tegument protein pp150 [ 10 ], up to atomic resolution [ 11 ]. However, such method is not applicable to the studies of herpesvirus gB and other glycoproteins due to their disorganized distribution on the pleomorphic viral envelope. Instead, the structures of gB ectodomains and various forms of gH/gL from herpes simplex virus (HSV) [ 12 , 13 ], Epstein-Barr virus (EBV) [ 14 , 15 ] and HCMV [ 16 , 17 ] have been solved by x-ray crystallography. The gB ectodomain structures from these studies share structural similarities to other class III viral fusion proteins in their postfusion conformation [ 18 – 20 ]. Among these proteins, vesicular stomatitis virus (VSV) G is the only one whose ectodomain structure has been solved for both prefusion [ 21 ] and postfusion [ 20 ] conformations, thanks to its pH-reversibility between the two conformations and amenability to crystallization at both high and low pH conditions. At pH 6.3–6.9 conditions, VSV G has also been observed to exist in monomeric forms both in solution and on virion envelope, possibly representing fusion intermediates [ 22 , 23 ]. By contrast, the prefusion conformation of herpesvirus gB is metastable and its structure has been elusive (even the recent crystal structure of the full-length HSV-1 gB is also in the postfusion conformation [ 24 ]). While cryo electron tomography (cryoET) of HSV-1 virions has revealed glycoproteins on the native viral envelope, poor contrast of cryoET reconstructions makes it difficult to distinguish different glycoprotein structures and conformations [ 25 ]. Recent efforts resorted to the use of purified HSV-1 gB-decorated vesicles to visualize the prefusion gB, but its domain assignments have been controversial due to difficulties in interpreting cryoET structures with poor contrast and signal/noise ratio (SNR) [ 26 , 27 ]. As a result, the mechanism underlying the complex process of receptor-triggered membrane fusion remains poorly understood for not only HCMV, but also for other herpesviruses. Recently, the technologies of electron-counting [ 28 , 29 ], energy filter and Volta phase plate (VPP) [ 30 ] have significantly improved contrast and SNR of cryoEM images and their combined use in cryoET has led to resolution of two functional states of 26S proteasome in neurons [ 31 ]. In this study, we first demonstrated the ability to distinguish prefusion and postfusion conformations of the VSV G trimer (171 kD) in situ by employing a combination of VPP, direct electron-counting, energy filtering and subtomographic averaging. Application of the same approach to HCMV virions has allowed us to identify different conformational states of HCMV gB (303 kD) in their native virion environments and to determine the in situ structure of prefusion gB at a resolution of ~21 à . Moreover, we also observed prefusion gB forming a complex with gH/gL in situ for the first time. Integration of these structures and knowledge of class III viral fusion proteins has led to a working model of how conformational changes drive membrane fusion during HCMV entry into host cells. Results Establishing VPP cryoET for obtaining in situ prefusion and postfusion structures We first established the validity of our cryoET method of combining VPP, direct electron detection, energy filtering, and subtomographic averaging by obtaining in situ structures of class III viral fusion proteins with known structures. Towards this end, we took advantage of the relative simplicity of VSV in having a single 57kD glycoprotein, G, on the viral envelope, with its trimeric structures known for both prefusion and postfusion conformations; and used VSV as a gold standard to validate our method. For VSV at pH = 7.5, tomograms reconstructed from tilt series obtained by 300kV Titan Krios equipped with VPP, energy filter and direct electron detection show excellent contrast, enabling the visualizations of G projecting from viral envelope, the helical nucleocapsid, as well as the internal densities corresponding to polymerases L ( Fig 1A and 1B ). Two conformations of G are readily differentiable based on the height and shape of the ectodomain: the majority is long (12.5nm) and slim, while the minority is short (8.7nm) and fat ( Fig 1B ). Subtomographic averages of 330 long-form particles and 65 short-form particles from five tomograms both contain a prominent ectodomain, with the long one (~28 à resolution) fit perfectly with the crystal structure of G ectodomain trimer in the postfusion conformation ( Fig 1C–1E ) and the short one with that in the prefusion conformation ( Fig 1G–1I ). Similar structures were observed in a previous electron tomography study performed on negatively stained sample [ 32 ]. 10.1371/journal.ppat.1007452.g001 Fig 1 In situ structures of two distinct conformations of VSV G. (A) A 10à -thick slab from the tomogram showing bullet-shaped VSV virions. (B) The VSV particle indicated by red arrow in (A), with a yellow arrow pointing to long-form G and a yellow arrow head pointing to short-form G. (C~F) Subtomographic average of the long-form G densities, whose ectodomain matches the crystal structure of G in postfusion conformation. The subtomographic average of the long-form G densities is shown either as shaded surfaces viewed from side (C) and top (D), or as semi-transparent gray (E) fitted with crystal structure of the G trimer (ribbon) in the postfusion conformation (PDB: 5I2M) [ 20 ]. For clarity, one of the subunits is shown alone in (F) with five domains (DI~DV) indicated. (G~J) Subtomographic average of the short-form G densities, whose ectodomain matches the crystal structure of G in prefusion conformation. The subtomographic average of the short-form G densities is shown either as shaded surfaces viewed from side (G) and top (H), or as semi-transparent gray (I) fitted with crystal structure of the G trimer (ribbon) in the prefusion conformation (PDB: 5I2S) [ 21 ]. For clarity, one of the subunits is shown alone in (J) with five domains (DI~DV) indicated. Both crystal structures of G contain five domains, DI through DV, despite drastic domain arrangements ( Fig 1E, 1F, 1I and 1J ). The dramatically different appearances between the two conformations are primarily due to the refolding of the short loop (residue 273 to 275) in DIII, resulting in the elongation of the central helix and a taller postfusion trimer. DIII form the trimeric core in both conformations, buried in the center of the cryoET density map ( Fig 1E and 1I ). The other domains (DI, DII and DIV) undergo a rigid-body type rearrangement—only changing the relative orientations and locations while retaining their domain structures [ 21 ] ( Fig 1F and 1J ). This analysis demonstrates that our cryoET approach incorporating the three cutting-edge technologies can distinguish the two forms of in situ structures of glycoprotein G and allows fitting existing domain structures of individual fusion protein into the density maps for functional interpretation. Unprecedented structural details revealed by VPP cryoET of HCMV particles Next, we applied the same strategy established above to obtain in situ structures of gB and its interaction with gH/gL complex. We imaged virions of the highly passaged laboratory HCMV strain AD169, taking advantage of its simplicity, as it has lost some glycoprotein genes and does not contain gH/gL/UL128/UL130/UL131A pentamers on its envelope [ 33 ]. We recorded cryoET tilt series of HCMV virions with and without VPP in a Titan Krios instrument equipped with an energy filter and a direct electron detector in electron-counting mode. Both the raw images in the tilt series and the reconstructed tomograms show significantly better contrast when VPP was used ( S1 Fig , S1 – S4 Movies). Typical in virions obtained by high-speed centrifugation, the viral envelopes are pleomorphic and often exhibit membrane blebs likely due to mechanical stress during purification ( Fig 2A ). [As discussed below, such mechanical stress might also be responsible for triggering some of the "spring-loaded"/higher-energy (prefusion) gB to its lower energy ("postfusion") form, which were used as an internal control to validate our cryoET subtomographic averaging method.] In the tomograms reconstructed from the tilt series obtained with VPP (referred to as VPP tomograms) ( Fig 2 ), three types of enveloped viral particles are readily recognized: virions with C-capsid containing densely-packed dsDNA genome ( Fig 2B ), non-infectious enveloped particles (NIEPs) with B-capsid containing a protein scaffold (red arrows in Fig 2A ) or with empty A-capsid (cyan arrow in Fig 2A ). Inside C-capsids, the dsDNA molecule occupies evenly throughout the entire interior of the capsid with the 20 à -diameter dsDNA duplex resolved ( Fig 2B )—the first time such detailed features ever observed directly by cryoET. In B-capsids, the scaffolding protein (pUL80, up to 1000 copies/capsid [ 34 , 35 ]) is organized into a density sphere with an outer and inner diameter of ~700 and ~400à , respectively. In capsids devoid of genome DNA, a portal complex for DNA translocation is visible at one of the 12 vertices of the capsid ( Fig 2C ). The viral envelope is pleomorphic ( Fig 2A ) and its membrane resolved into two leaflets 40à apart ( Fig 2D ), sporting sparsely and randomly located, and clearly identifiable glycoprotein spikes on the outer leaflet ( Fig 2A and Fig 3A ). 10.1371/journal.ppat.1007452.g002 Fig 2 CryoET of HCMV. (A) A 10à -thick slab from the tomogram of S2 Movie showing a digital slice of various HCMV particles, each comprising an outer envelope layer, an intermediate tegument compartment, and an icosahedral capsid inside. Though all wrapped a pleomorphic glycoprotein-decorated envelope, these particles differ inside: each virion (yellow arrows) containing a C capsid (with DNA genome) and non-infectious enveloped particle (NIEP) either an A capsid (empty, blue arrow), or a B capsid (containing scaffolding protein but no DNA, red arrows). (B) A zoom-in slice of a C capsid, with dsDNA duplexes resolved among the fingerprint-like pattern of the genome. (C) A slice from the particle indicated by the cyan arrow in (A) showing the unique portal complex (arrow) at one of the 12 vertices. (D) A zoom-in envelope region of the particle indicated by a yellow arrow in (A) showing the two resolved leaflets of the lipid bilayer envelope (inset: the enlarged boxed region). The side of the boxes in B-D is 120 nm. 10.1371/journal.ppat.1007452.g003 Fig 3 In situ structures of gB in the "postfusion" and prefusion conformations. (A) A representative virion showing various glycoprotein densities on its envelope. (B) Identifications of the columnar tree-shaped (red box) and Christmas tree-shaped (yellow box) glycoprotein densities that both match the expected volume of gB (see main text). Insets are the enlargements of the two forms with their corresponding shape schematic. (C~H) Sub-tomographic average of the columnar tree-shaped glycoprotein densities, whose ectodomain matches the crystal structure of gB ectodomain in the postfusion conformation (PDB: 5CXF) [ 16 ]. The subtomographic average of the columnar-shaped densities (yellow) and segmented membrane bilayer (blue, from I) are shown either as shaded surfaces viewed from side (C), top (D) and slanted bottom (E), or as semi-transparent gray fitted with the gB ectodomain trimer crystal structure (ribbon) at the postfusion conformation (PDB: 5CXF) [ 16 ] (F). Two subunits of the gB trimer crystal structure are shown as pink and gray surfaces, while the third subunit as ribbons with its domains colored as in [ 16 ] and its transmembrane helix as brown cylinder and the C-terminal flexible endodomain as a swinging dotted lines (G). For clarity, the third subunit is shown alone in (H) with five domains (DI~DV) indicated. (I~K) Sub-tomographic average of the Christmas tree-shaped densities (yellow) and associated membrane bilayer (blue) viewed from side (I), top (J) and slanted bottom (K). Identifications of gB trimers on the viral envelope We used the following three pieces of evidence to establish the identifications of gB trimers on the viral envelope. First, among HCMV glycoproteins, gB is known to only exist as homotrimer with a combined mass of ~300 kD [ 36 ] and is the most abundant complex over 100 kD [ 37 ]. This mass is expected to occupy an estimated extracellular volume of ~300 nm 3 . Among the density spikes decorating the outer leaflet of the viral membrane, only two differently shaped spikes with such volume were identified, suggesting that they might be gB trimer at different conformational states ( Fig 3B ). Second, the two distinctive side-view shapes—one triangular, Christmas-tree like (71%) and the other rectangular, columnar-tree like (29%) ( Fig 3B )—are similar to the side-views of the cryoET reconstructions of HSV-1 gB trimers on purified vesicles in their putative prefusion and postfusion conformations, respectively [ 27 ]. Third, we performed subtomographic averaging to these two types of spikes, respectively, in order to examine them with a higher SNR. Both of the averaged models exhibit apparent three-fold symmetry with the symmetric axis perpendicular to the plane of viral membrane, despite slight distortion arising from the inherent "missing wedge problem" of electron tomography ( S2 Fig ). These three pieces of evidence all point to our tentative assignment of the Christmas tree-shaped and the columnar tree-shaped densities on the HCMV envelope as gB trimers in the prefusion and "postfusion" (quotation marks are used here since the conformation is not really caused by fusion but likely triggered by mechanical stress during virion purification with high-speed centrifugation) conformations, respectively. Indeed, as shown below, the available crystal structure of gB in the postfusion conformation matches perfectly with our final subtomographic average of the columnar tree-shaped density, further validating our assignments. Subtomographic averages of the putative gB in prefusion and "postfusion" conformations As mentioned above, we performed subtomographic averaging to characterize the two putative gB conformations at a higher resolution. The significantly enhanced contrast afforded by imaging with VPP at a near-focus condition allowed the clear visualizations of different structures in the reconstructed tomograms. For direct comparison, we also obtained tilt series without using a VPP (referred to as non-VPP tomograms). For the latter data, we had to use a significantly larger defocus value (-4μm) to improve image contrast and record much more tilt series (28 total) in order to obtain a similar number of gB particles for subtomographic averaging due to greater difficulties in distinguishing different glycoprotein morphologies in the tomograms ( S1 Fig ). In addition, the use of large defocus has necessitated correction for contrast transfer function (CTF): the structure obtained without CTF correction contains phase-inverted, incorrect structure information beyond 25 à ( S4E Fig ), as reflected by the broken connections between the ectodomain and the viral membrane in the absence of CTF correction ( S3B , S4C and S4D Figs). In total, 350 particles of the columnar tree-shaped and 1509 particles of the Christmas tree-shaped densities were included for subtomographic averaging. For the columnar tree-shaped structure, all particles were extracted from the VPP tomograms due to ambiguities in distinguishing its slender shape from background noise in the non-VPP tomograms. For the Christmas tree-shaped structure, 874 particles, which came from VPP tomograms, were first used and 635 particles from non-VPP tomograms eventually were also included to further improve resolution. Three-fold symmetry was imposed subsequently to improve SNR and the resolution of the averaged structures. Fourier shell correlation (FSC) analyses indicate that the resolutions for the symmetrized 3D subtomographic average of the columnar tree-shaped and Christmas tree-shaped spikes are 26 à and 21 à , respectively, based on the gold-standard criterion ( S3A Fig ). The subtomographic average of the columnar tree-shaped spike resolves the two leaflets of the bilayer viral envelope and a prominent (161à in height) ectodomain ( Fig 3C–3E , S5 Movie ). The ectodomain density matches well with the crystal structure of the HCMV gB ectodomain trimer [ 16 ] ( Fig 3F–3H ), validating our initial assignment of the columnar tree-shaped density as gB structure in its "postfusion" conformation and re-establishing the validity of our approach. The subtomographic average of our putative prefusion gB densities reveals the two leaflets of the bilayer viral envelope with prominent gB densities attached to both: a prominent ectodomain attached to the outer leaflet (130à in height) and a globular (about 35à in height and 26à in width) endodomain to the inner leaflet ( Fig 3I–3K ). The ectodomain in the putative prefusion gB is shorter than that in the gB "postfusion" conformation and anchors to the membrane with three well-separated densities, forming a tripod ( Fig 3I–3K , S6 Movie ). Although no crystal structure of prefusion gB is available to fit into our subtomographic average to directly confirm or refute this prefusion gB assignment, it is believed that herpesvirus gB bears structural and mechanistic similarities to other class III viral fusion proteins, which can be used to aid our assignment. Indeed, the postfusion conformation of HCMV gB ectodomain is similar to the postfusion conformations of all other class III viral fusion proteins [ 18 ], including the postfusion VSV G ( Fig 1C–1F ). The lower portion of the prefusion conformation of the VSV G trimer ( Fig 1G–1I ) has a tripod shape similar to the lower portion of the Christmas tree-shaped density ( Fig 3I ). The prefusion VSV G trimer is shorter than—and undergoes drastic domain rearrangements towards—its postfusion conformation [ 20 , 21 ] ( Fig 1 ); likewise, the Christmas tree-shaped density is shorter than the columnar tree-shaped density. Taken together, these characteristic similarities to the prefusion structure of VSV G corroborate our initial assignment of the Christmas tree-shaped density as the in situ prefusion structure of HCMV gB trimer. Domain assignments of the in situ gB structures in both conformations Structure-guided sequence analysis ( Fig 4A ) indicates that each full-length gB protomer contains an N-terminal ectodomain (residues 87–705), a membrane proximal region (MPR, residues 706–750), a single transmembrane helix (residues 751–771) and a C-terminal endodomain (residues 772–906). For the "postfusion" gB trimer, the ectodomain in the subtomographic average can be divided into a base in contact with the membrane, and two lobes—middle and crown—connected by a neck ( Fig 3F ). The crystal structure of the ectodomain trimer shows that each protomer consists of five domains: DI, DII, DIII, DIV and DV ( Fig 3G and 3H ) [ 16 ]. Except for DV, these domains can be located in our subtomographic average of the "postfusion" gB ( Fig 3F ). DI, each containing two fusion loops, is located at the base of the trimer; DII and DIV reside, respectively, in the middle and crown lobes, which are connected by DIII in the neck. DV contains a long loop connected by two short helices and is buried, thus is not resolved in our subtomographic average gB trimer due to the limited resolution. 10.1371/journal.ppat.1007452.g004 Fig 4 Schematic illustration of the full-length HCMV gB. (A) Mapping of domains to the full-length HCMV gB primary sequence (SS = signal sequence, MPR = membrane proximal region, TM = transmembrane domain, Endo = endodomain). Upper inset: in the prefusion (Pre) conformation, the sequence of the central helix in DIII resolved in the postfusion gB crystal structure (Post) is predicted to fold into two helices joined by a short loop around residues 498–500. Lower inset: predicted secondary structures of the sequence encompassing the MPR, TM and endodomain of gB in the prefusion conformation. (B) Helical wheel diagram of the first 15 amino acids of MPR (sequence in red dashed box in (A)), showing one side with a cluster of hydrophobic amino acids. As detailed in the Method, we employed a combination of manual rigid-body fitting of known domain structures from the existing HCMV gB postfusion structure [ 16 ], comparative modeling of DIII based on the homologous DIII from VSV G prefusion conformation [ 21 ], followed by optimization by the molecular dynamics flexible fitting (MDFF) method [ 38 ], to put forward a provisional domain arrangement model of the prefusion gB ( Fig 5 ). DV was not considered in our domain modeling of HCMV gB prefusion conformation due to the lack of a template structure, since DV was truncated in the crystal structure of postfusion VSV G. MDFF not only optimized the chemical interactions among the fitted domains, but also improved overall model to map correlation coefficient from 0.83 to 0.94 ( Fig 5H and 5I , S7 Movie ). 10.1371/journal.ppat.1007452.g005 Fig 5 Domain fitting for prefusion gB. (A~C) Domain fitting for DI. (A) Scores of 28 models for DI. Models are ranked in ascending order of scores. Red dot indicates the model of highest score. Dots with green and orange circles indicate the models of medium and low score, respectively. The DI models, superimposed in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray) and viewed from side and top, are shown in green and orange dashed boxes next to the corresponding dots, respectively. (B, C) DI structure, indicated as red dot in (A), is superimposed in the subtomographic average of the Christmas tree-shaped density (semitransparent gray), viewed from side (B) and top (C). (D) VSV G domain rearrangement of crystal monomer structure from prefusion (left, PDB: 2J6J) [ 21 ] to postfusion (right, PDB: 2CMZ) [ 21 ]. The red dotted lines represent the unresolved domain DV between transmembrane helix and the ectodomain. (E) the crystal structure of one protomer of the postfusion HCMV gB ectodomain (left, PDB: 5CXF) [ 16 ] is shown as ribbon next to the predicted prefusion gB structure (right) with domains arranged according to those in the prefusion VSV G. α4 and α5 represent the long central helix and the following short helix in DIII in postfusion gB structure. α4a and α4b represent the two helix breaking from α4. Helices are labeled as in [ 16 ]. (F, G) The predicted prefusion gB structure shown in (E, right) is superposed with two other symmetric copies (gray ribbon) in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray), viewed from side (F) and top (G). (H, I) The MDFF-simulated prefusion gB structure is superimposed with two other symmetric copies (gray ribbon) in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray), viewed from side (H) and top (I). The epitopes of HSV-1 antibodies SS55/SS56 and R240 are DI and fusion loop 2 of HSV-1 gB, respectively ([ 26 ]); the corresponding locations of these two epitopes in our domain model of HCMV gB are indicated. The model from MDFF does not include the MPR (residues 706–750), which is proposed to lie between the ectodomain and the transmembrane helix ( Fig 4A ) and "mask" the fusion loops to prevent their premature (non-productive) association with lipid [ 39 ]. Helical wheel projection of the first 15 amino acids of the MPR shows an amphipathic helix ( Fig 4B ) whose hydrophobic side could interact with the fusion loops. This notion is consistent with our interpretation of DI in the subtomographic averages of both prefusion and "postfusion" conformations, with the fusion loops pointing to and in close proximity to the membrane. Visualization of gB interacting with putative receptor-binding gH/gL complex Among herpesviruses, gB and gH/gL are highly conserved and known to form a fusion machinery for virus entry [ 40 ]. Previous biochemical studies have indicated that gH/gL regulated fusion activity of gB [ 41 ] and might form a complex with gB in virions on the basis of co-immunoprecipitation experiments [ 42 ]. Besides gB trimer densities mentioned above, "L"-shaped spikes were also observed protruding outwards from the viral envelope, which we interpret as gH/gL complexes on the basis of size and shape similarities to the gH/gL crystal structure [ 13 , 17 ]. Moreover, among such "L"-shaped spikes, ~7% were observed to be in contact with the Christmas tree-shaped, prefusion gB trimer, forming a gB-gH/gL complex ( Fig 6B and 6C ), while others were unbound. No "postfusion" gB trimer have been observed involving in gB-gH/gL complex. 10.1371/journal.ppat.1007452.g006 Fig 6 In situ structure of gH/gL complex adjacent to prefusion gB. (A) A series of slices in a tomogram showing a HCMV particle at different sections. (B) A 3D surface view from the HCMV particle in (A) with subtomographic average of prefusion and "postfusion" gB and gH/gL placed back on the viral envelope segmented from the tomogram. Black arrow indicates the gB/gH/gL complex. Unidentified glycoprotein densities are indicated as rods. Tegument proteins are shown as cyan densities. The recently published icosahedral reconstruction of capsid [ 11 ] was low-passed to 10à , radially colored and placed back in its location. (C~F) The subtomographic average (C) showing a putative gH/gL complex adjacent to prefusion gB. The subtomographic average is also shown fitted with crystal structure of gH/gL (ribbon) (PDB: 5VOB) [ 17 ] in semitransparent surface viewed from side (D), top (E) and bottom (F). A subtomographic average was obtained by aligning and averaging 49 such gB-gH/gL complexes to investigate the contact sites between prefusion gB and gH/gL ( Fig 6C–6F ), with a resolution around 30à reported by calcFSC in PEET . The HCMV gH/gL crystal structure [ 17 ] fits well in the "L"-shaped density in the subtomographic average (0.75 of the cross-correlation coefficient between the cryoET map and the model filtered to 30à , Fig 6D and 6E ). This fitting, together with the predicted domain arrangement in the prefusion gB structure ( Fig 5H and 5I ), reveals that DI of gB may contact the gH subunit of gH/gL ( Fig 6D ). The contact sites on gB and gH are consistent with the gH-binding site on HSV-1 gB suggested by blocking gH binding to gB with SS55 and SS56 antibodies (epitopes mapped to residues 153–363 of gB) ( Fig 5H ) [ 43 ] and the gB-binding sites on gH/gL suggested by anti-gH/gL antibody LP11 for HSV [ 13 ], respectively. Mutagenesis of gH cytotail has led to its proposed role of acting as a "wedge" to split the gB endodomain "clamp" to trigger gB ectodomain refolding [ 44 ]. Though the details of their interactions in the endodomain are yet to be resolved, this first observation of gH/gL complex making contact with prefusion gB in situ ( Fig 6 ) supports the notion that receptor binding to gH/gL triggers transformation of gB from prefusion to postfusion conformation. Establishing VPP cryoET for obtaining in situ prefusion and postfusion structures We first established the validity of our cryoET method of combining VPP, direct electron detection, energy filtering, and subtomographic averaging by obtaining in situ structures of class III viral fusion proteins with known structures. Towards this end, we took advantage of the relative simplicity of VSV in having a single 57kD glycoprotein, G, on the viral envelope, with its trimeric structures known for both prefusion and postfusion conformations; and used VSV as a gold standard to validate our method. For VSV at pH = 7.5, tomograms reconstructed from tilt series obtained by 300kV Titan Krios equipped with VPP, energy filter and direct electron detection show excellent contrast, enabling the visualizations of G projecting from viral envelope, the helical nucleocapsid, as well as the internal densities corresponding to polymerases L ( Fig 1A and 1B ). Two conformations of G are readily differentiable based on the height and shape of the ectodomain: the majority is long (12.5nm) and slim, while the minority is short (8.7nm) and fat ( Fig 1B ). Subtomographic averages of 330 long-form particles and 65 short-form particles from five tomograms both contain a prominent ectodomain, with the long one (~28 à resolution) fit perfectly with the crystal structure of G ectodomain trimer in the postfusion conformation ( Fig 1C–1E ) and the short one with that in the prefusion conformation ( Fig 1G–1I ). Similar structures were observed in a previous electron tomography study performed on negatively stained sample [ 32 ]. 10.1371/journal.ppat.1007452.g001 Fig 1 In situ structures of two distinct conformations of VSV G. (A) A 10à -thick slab from the tomogram showing bullet-shaped VSV virions. (B) The VSV particle indicated by red arrow in (A), with a yellow arrow pointing to long-form G and a yellow arrow head pointing to short-form G. (C~F) Subtomographic average of the long-form G densities, whose ectodomain matches the crystal structure of G in postfusion conformation. The subtomographic average of the long-form G densities is shown either as shaded surfaces viewed from side (C) and top (D), or as semi-transparent gray (E) fitted with crystal structure of the G trimer (ribbon) in the postfusion conformation (PDB: 5I2M) [ 20 ]. For clarity, one of the subunits is shown alone in (F) with five domains (DI~DV) indicated. (G~J) Subtomographic average of the short-form G densities, whose ectodomain matches the crystal structure of G in prefusion conformation. The subtomographic average of the short-form G densities is shown either as shaded surfaces viewed from side (G) and top (H), or as semi-transparent gray (I) fitted with crystal structure of the G trimer (ribbon) in the prefusion conformation (PDB: 5I2S) [ 21 ]. For clarity, one of the subunits is shown alone in (J) with five domains (DI~DV) indicated. Both crystal structures of G contain five domains, DI through DV, despite drastic domain arrangements ( Fig 1E, 1F, 1I and 1J ). The dramatically different appearances between the two conformations are primarily due to the refolding of the short loop (residue 273 to 275) in DIII, resulting in the elongation of the central helix and a taller postfusion trimer. DIII form the trimeric core in both conformations, buried in the center of the cryoET density map ( Fig 1E and 1I ). The other domains (DI, DII and DIV) undergo a rigid-body type rearrangement—only changing the relative orientations and locations while retaining their domain structures [ 21 ] ( Fig 1F and 1J ). This analysis demonstrates that our cryoET approach incorporating the three cutting-edge technologies can distinguish the two forms of in situ structures of glycoprotein G and allows fitting existing domain structures of individual fusion protein into the density maps for functional interpretation. Unprecedented structural details revealed by VPP cryoET of HCMV particles Next, we applied the same strategy established above to obtain in situ structures of gB and its interaction with gH/gL complex. We imaged virions of the highly passaged laboratory HCMV strain AD169, taking advantage of its simplicity, as it has lost some glycoprotein genes and does not contain gH/gL/UL128/UL130/UL131A pentamers on its envelope [ 33 ]. We recorded cryoET tilt series of HCMV virions with and without VPP in a Titan Krios instrument equipped with an energy filter and a direct electron detector in electron-counting mode. Both the raw images in the tilt series and the reconstructed tomograms show significantly better contrast when VPP was used ( S1 Fig , S1 – S4 Movies). Typical in virions obtained by high-speed centrifugation, the viral envelopes are pleomorphic and often exhibit membrane blebs likely due to mechanical stress during purification ( Fig 2A ). [As discussed below, such mechanical stress might also be responsible for triggering some of the "spring-loaded"/higher-energy (prefusion) gB to its lower energy ("postfusion") form, which were used as an internal control to validate our cryoET subtomographic averaging method.] In the tomograms reconstructed from the tilt series obtained with VPP (referred to as VPP tomograms) ( Fig 2 ), three types of enveloped viral particles are readily recognized: virions with C-capsid containing densely-packed dsDNA genome ( Fig 2B ), non-infectious enveloped particles (NIEPs) with B-capsid containing a protein scaffold (red arrows in Fig 2A ) or with empty A-capsid (cyan arrow in Fig 2A ). Inside C-capsids, the dsDNA molecule occupies evenly throughout the entire interior of the capsid with the 20 à -diameter dsDNA duplex resolved ( Fig 2B )—the first time such detailed features ever observed directly by cryoET. In B-capsids, the scaffolding protein (pUL80, up to 1000 copies/capsid [ 34 , 35 ]) is organized into a density sphere with an outer and inner diameter of ~700 and ~400à , respectively. In capsids devoid of genome DNA, a portal complex for DNA translocation is visible at one of the 12 vertices of the capsid ( Fig 2C ). The viral envelope is pleomorphic ( Fig 2A ) and its membrane resolved into two leaflets 40à apart ( Fig 2D ), sporting sparsely and randomly located, and clearly identifiable glycoprotein spikes on the outer leaflet ( Fig 2A and Fig 3A ). 10.1371/journal.ppat.1007452.g002 Fig 2 CryoET of HCMV. (A) A 10à -thick slab from the tomogram of S2 Movie showing a digital slice of various HCMV particles, each comprising an outer envelope layer, an intermediate tegument compartment, and an icosahedral capsid inside. Though all wrapped a pleomorphic glycoprotein-decorated envelope, these particles differ inside: each virion (yellow arrows) containing a C capsid (with DNA genome) and non-infectious enveloped particle (NIEP) either an A capsid (empty, blue arrow), or a B capsid (containing scaffolding protein but no DNA, red arrows). (B) A zoom-in slice of a C capsid, with dsDNA duplexes resolved among the fingerprint-like pattern of the genome. (C) A slice from the particle indicated by the cyan arrow in (A) showing the unique portal complex (arrow) at one of the 12 vertices. (D) A zoom-in envelope region of the particle indicated by a yellow arrow in (A) showing the two resolved leaflets of the lipid bilayer envelope (inset: the enlarged boxed region). The side of the boxes in B-D is 120 nm. 10.1371/journal.ppat.1007452.g003 Fig 3 In situ structures of gB in the "postfusion" and prefusion conformations. (A) A representative virion showing various glycoprotein densities on its envelope. (B) Identifications of the columnar tree-shaped (red box) and Christmas tree-shaped (yellow box) glycoprotein densities that both match the expected volume of gB (see main text). Insets are the enlargements of the two forms with their corresponding shape schematic. (C~H) Sub-tomographic average of the columnar tree-shaped glycoprotein densities, whose ectodomain matches the crystal structure of gB ectodomain in the postfusion conformation (PDB: 5CXF) [ 16 ]. The subtomographic average of the columnar-shaped densities (yellow) and segmented membrane bilayer (blue, from I) are shown either as shaded surfaces viewed from side (C), top (D) and slanted bottom (E), or as semi-transparent gray fitted with the gB ectodomain trimer crystal structure (ribbon) at the postfusion conformation (PDB: 5CXF) [ 16 ] (F). Two subunits of the gB trimer crystal structure are shown as pink and gray surfaces, while the third subunit as ribbons with its domains colored as in [ 16 ] and its transmembrane helix as brown cylinder and the C-terminal flexible endodomain as a swinging dotted lines (G). For clarity, the third subunit is shown alone in (H) with five domains (DI~DV) indicated. (I~K) Sub-tomographic average of the Christmas tree-shaped densities (yellow) and associated membrane bilayer (blue) viewed from side (I), top (J) and slanted bottom (K). Identifications of gB trimers on the viral envelope We used the following three pieces of evidence to establish the identifications of gB trimers on the viral envelope. First, among HCMV glycoproteins, gB is known to only exist as homotrimer with a combined mass of ~300 kD [ 36 ] and is the most abundant complex over 100 kD [ 37 ]. This mass is expected to occupy an estimated extracellular volume of ~300 nm 3 . Among the density spikes decorating the outer leaflet of the viral membrane, only two differently shaped spikes with such volume were identified, suggesting that they might be gB trimer at different conformational states ( Fig 3B ). Second, the two distinctive side-view shapes—one triangular, Christmas-tree like (71%) and the other rectangular, columnar-tree like (29%) ( Fig 3B )—are similar to the side-views of the cryoET reconstructions of HSV-1 gB trimers on purified vesicles in their putative prefusion and postfusion conformations, respectively [ 27 ]. Third, we performed subtomographic averaging to these two types of spikes, respectively, in order to examine them with a higher SNR. Both of the averaged models exhibit apparent three-fold symmetry with the symmetric axis perpendicular to the plane of viral membrane, despite slight distortion arising from the inherent "missing wedge problem" of electron tomography ( S2 Fig ). These three pieces of evidence all point to our tentative assignment of the Christmas tree-shaped and the columnar tree-shaped densities on the HCMV envelope as gB trimers in the prefusion and "postfusion" (quotation marks are used here since the conformation is not really caused by fusion but likely triggered by mechanical stress during virion purification with high-speed centrifugation) conformations, respectively. Indeed, as shown below, the available crystal structure of gB in the postfusion conformation matches perfectly with our final subtomographic average of the columnar tree-shaped density, further validating our assignments. Subtomographic averages of the putative gB in prefusion and "postfusion" conformations As mentioned above, we performed subtomographic averaging to characterize the two putative gB conformations at a higher resolution. The significantly enhanced contrast afforded by imaging with VPP at a near-focus condition allowed the clear visualizations of different structures in the reconstructed tomograms. For direct comparison, we also obtained tilt series without using a VPP (referred to as non-VPP tomograms). For the latter data, we had to use a significantly larger defocus value (-4μm) to improve image contrast and record much more tilt series (28 total) in order to obtain a similar number of gB particles for subtomographic averaging due to greater difficulties in distinguishing different glycoprotein morphologies in the tomograms ( S1 Fig ). In addition, the use of large defocus has necessitated correction for contrast transfer function (CTF): the structure obtained without CTF correction contains phase-inverted, incorrect structure information beyond 25 à ( S4E Fig ), as reflected by the broken connections between the ectodomain and the viral membrane in the absence of CTF correction ( S3B , S4C and S4D Figs). In total, 350 particles of the columnar tree-shaped and 1509 particles of the Christmas tree-shaped densities were included for subtomographic averaging. For the columnar tree-shaped structure, all particles were extracted from the VPP tomograms due to ambiguities in distinguishing its slender shape from background noise in the non-VPP tomograms. For the Christmas tree-shaped structure, 874 particles, which came from VPP tomograms, were first used and 635 particles from non-VPP tomograms eventually were also included to further improve resolution. Three-fold symmetry was imposed subsequently to improve SNR and the resolution of the averaged structures. Fourier shell correlation (FSC) analyses indicate that the resolutions for the symmetrized 3D subtomographic average of the columnar tree-shaped and Christmas tree-shaped spikes are 26 à and 21 à , respectively, based on the gold-standard criterion ( S3A Fig ). The subtomographic average of the columnar tree-shaped spike resolves the two leaflets of the bilayer viral envelope and a prominent (161à in height) ectodomain ( Fig 3C–3E , S5 Movie ). The ectodomain density matches well with the crystal structure of the HCMV gB ectodomain trimer [ 16 ] ( Fig 3F–3H ), validating our initial assignment of the columnar tree-shaped density as gB structure in its "postfusion" conformation and re-establishing the validity of our approach. The subtomographic average of our putative prefusion gB densities reveals the two leaflets of the bilayer viral envelope with prominent gB densities attached to both: a prominent ectodomain attached to the outer leaflet (130à in height) and a globular (about 35à in height and 26à in width) endodomain to the inner leaflet ( Fig 3I–3K ). The ectodomain in the putative prefusion gB is shorter than that in the gB "postfusion" conformation and anchors to the membrane with three well-separated densities, forming a tripod ( Fig 3I–3K , S6 Movie ). Although no crystal structure of prefusion gB is available to fit into our subtomographic average to directly confirm or refute this prefusion gB assignment, it is believed that herpesvirus gB bears structural and mechanistic similarities to other class III viral fusion proteins, which can be used to aid our assignment. Indeed, the postfusion conformation of HCMV gB ectodomain is similar to the postfusion conformations of all other class III viral fusion proteins [ 18 ], including the postfusion VSV G ( Fig 1C–1F ). The lower portion of the prefusion conformation of the VSV G trimer ( Fig 1G–1I ) has a tripod shape similar to the lower portion of the Christmas tree-shaped density ( Fig 3I ). The prefusion VSV G trimer is shorter than—and undergoes drastic domain rearrangements towards—its postfusion conformation [ 20 , 21 ] ( Fig 1 ); likewise, the Christmas tree-shaped density is shorter than the columnar tree-shaped density. Taken together, these characteristic similarities to the prefusion structure of VSV G corroborate our initial assignment of the Christmas tree-shaped density as the in situ prefusion structure of HCMV gB trimer. Domain assignments of the in situ gB structures in both conformations Structure-guided sequence analysis ( Fig 4A ) indicates that each full-length gB protomer contains an N-terminal ectodomain (residues 87–705), a membrane proximal region (MPR, residues 706–750), a single transmembrane helix (residues 751–771) and a C-terminal endodomain (residues 772–906). For the "postfusion" gB trimer, the ectodomain in the subtomographic average can be divided into a base in contact with the membrane, and two lobes—middle and crown—connected by a neck ( Fig 3F ). The crystal structure of the ectodomain trimer shows that each protomer consists of five domains: DI, DII, DIII, DIV and DV ( Fig 3G and 3H ) [ 16 ]. Except for DV, these domains can be located in our subtomographic average of the "postfusion" gB ( Fig 3F ). DI, each containing two fusion loops, is located at the base of the trimer; DII and DIV reside, respectively, in the middle and crown lobes, which are connected by DIII in the neck. DV contains a long loop connected by two short helices and is buried, thus is not resolved in our subtomographic average gB trimer due to the limited resolution. 10.1371/journal.ppat.1007452.g004 Fig 4 Schematic illustration of the full-length HCMV gB. (A) Mapping of domains to the full-length HCMV gB primary sequence (SS = signal sequence, MPR = membrane proximal region, TM = transmembrane domain, Endo = endodomain). Upper inset: in the prefusion (Pre) conformation, the sequence of the central helix in DIII resolved in the postfusion gB crystal structure (Post) is predicted to fold into two helices joined by a short loop around residues 498–500. Lower inset: predicted secondary structures of the sequence encompassing the MPR, TM and endodomain of gB in the prefusion conformation. (B) Helical wheel diagram of the first 15 amino acids of MPR (sequence in red dashed box in (A)), showing one side with a cluster of hydrophobic amino acids. As detailed in the Method, we employed a combination of manual rigid-body fitting of known domain structures from the existing HCMV gB postfusion structure [ 16 ], comparative modeling of DIII based on the homologous DIII from VSV G prefusion conformation [ 21 ], followed by optimization by the molecular dynamics flexible fitting (MDFF) method [ 38 ], to put forward a provisional domain arrangement model of the prefusion gB ( Fig 5 ). DV was not considered in our domain modeling of HCMV gB prefusion conformation due to the lack of a template structure, since DV was truncated in the crystal structure of postfusion VSV G. MDFF not only optimized the chemical interactions among the fitted domains, but also improved overall model to map correlation coefficient from 0.83 to 0.94 ( Fig 5H and 5I , S7 Movie ). 10.1371/journal.ppat.1007452.g005 Fig 5 Domain fitting for prefusion gB. (A~C) Domain fitting for DI. (A) Scores of 28 models for DI. Models are ranked in ascending order of scores. Red dot indicates the model of highest score. Dots with green and orange circles indicate the models of medium and low score, respectively. The DI models, superimposed in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray) and viewed from side and top, are shown in green and orange dashed boxes next to the corresponding dots, respectively. (B, C) DI structure, indicated as red dot in (A), is superimposed in the subtomographic average of the Christmas tree-shaped density (semitransparent gray), viewed from side (B) and top (C). (D) VSV G domain rearrangement of crystal monomer structure from prefusion (left, PDB: 2J6J) [ 21 ] to postfusion (right, PDB: 2CMZ) [ 21 ]. The red dotted lines represent the unresolved domain DV between transmembrane helix and the ectodomain. (E) the crystal structure of one protomer of the postfusion HCMV gB ectodomain (left, PDB: 5CXF) [ 16 ] is shown as ribbon next to the predicted prefusion gB structure (right) with domains arranged according to those in the prefusion VSV G. α4 and α5 represent the long central helix and the following short helix in DIII in postfusion gB structure. α4a and α4b represent the two helix breaking from α4. Helices are labeled as in [ 16 ]. (F, G) The predicted prefusion gB structure shown in (E, right) is superposed with two other symmetric copies (gray ribbon) in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray), viewed from side (F) and top (G). (H, I) The MDFF-simulated prefusion gB structure is superimposed with two other symmetric copies (gray ribbon) in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray), viewed from side (H) and top (I). The epitopes of HSV-1 antibodies SS55/SS56 and R240 are DI and fusion loop 2 of HSV-1 gB, respectively ([ 26 ]); the corresponding locations of these two epitopes in our domain model of HCMV gB are indicated. The model from MDFF does not include the MPR (residues 706–750), which is proposed to lie between the ectodomain and the transmembrane helix ( Fig 4A ) and "mask" the fusion loops to prevent their premature (non-productive) association with lipid [ 39 ]. Helical wheel projection of the first 15 amino acids of the MPR shows an amphipathic helix ( Fig 4B ) whose hydrophobic side could interact with the fusion loops. This notion is consistent with our interpretation of DI in the subtomographic averages of both prefusion and "postfusion" conformations, with the fusion loops pointing to and in close proximity to the membrane. Visualization of gB interacting with putative receptor-binding gH/gL complex Among herpesviruses, gB and gH/gL are highly conserved and known to form a fusion machinery for virus entry [ 40 ]. Previous biochemical studies have indicated that gH/gL regulated fusion activity of gB [ 41 ] and might form a complex with gB in virions on the basis of co-immunoprecipitation experiments [ 42 ]. Besides gB trimer densities mentioned above, "L"-shaped spikes were also observed protruding outwards from the viral envelope, which we interpret as gH/gL complexes on the basis of size and shape similarities to the gH/gL crystal structure [ 13 , 17 ]. Moreover, among such "L"-shaped spikes, ~7% were observed to be in contact with the Christmas tree-shaped, prefusion gB trimer, forming a gB-gH/gL complex ( Fig 6B and 6C ), while others were unbound. No "postfusion" gB trimer have been observed involving in gB-gH/gL complex. 10.1371/journal.ppat.1007452.g006 Fig 6 In situ structure of gH/gL complex adjacent to prefusion gB. (A) A series of slices in a tomogram showing a HCMV particle at different sections. (B) A 3D surface view from the HCMV particle in (A) with subtomographic average of prefusion and "postfusion" gB and gH/gL placed back on the viral envelope segmented from the tomogram. Black arrow indicates the gB/gH/gL complex. Unidentified glycoprotein densities are indicated as rods. Tegument proteins are shown as cyan densities. The recently published icosahedral reconstruction of capsid [ 11 ] was low-passed to 10à , radially colored and placed back in its location. (C~F) The subtomographic average (C) showing a putative gH/gL complex adjacent to prefusion gB. The subtomographic average is also shown fitted with crystal structure of gH/gL (ribbon) (PDB: 5VOB) [ 17 ] in semitransparent surface viewed from side (D), top (E) and bottom (F). A subtomographic average was obtained by aligning and averaging 49 such gB-gH/gL complexes to investigate the contact sites between prefusion gB and gH/gL ( Fig 6C–6F ), with a resolution around 30à reported by calcFSC in PEET . The HCMV gH/gL crystal structure [ 17 ] fits well in the "L"-shaped density in the subtomographic average (0.75 of the cross-correlation coefficient between the cryoET map and the model filtered to 30à , Fig 6D and 6E ). This fitting, together with the predicted domain arrangement in the prefusion gB structure ( Fig 5H and 5I ), reveals that DI of gB may contact the gH subunit of gH/gL ( Fig 6D ). The contact sites on gB and gH are consistent with the gH-binding site on HSV-1 gB suggested by blocking gH binding to gB with SS55 and SS56 antibodies (epitopes mapped to residues 153–363 of gB) ( Fig 5H ) [ 43 ] and the gB-binding sites on gH/gL suggested by anti-gH/gL antibody LP11 for HSV [ 13 ], respectively. Mutagenesis of gH cytotail has led to its proposed role of acting as a "wedge" to split the gB endodomain "clamp" to trigger gB ectodomain refolding [ 44 ]. Though the details of their interactions in the endodomain are yet to be resolved, this first observation of gH/gL complex making contact with prefusion gB in situ ( Fig 6 ) supports the notion that receptor binding to gH/gL triggers transformation of gB from prefusion to postfusion conformation. Discussion Since the postfusion conformation of gB is energetically favorable and structurally more stable, it is not surprising that purified recombinant gB so far have all adopted the "postfusion" conformation [ 12 , 16 , 45 ]. Therefore, imaging gB in its native, virion environment by cryoET seems to be the necessary approach to obtain the in situ structure in its metastable, prefusion conformation. However, a major challenge in interpreting in situ cryoET structures is the intrinsic poor contrast of tomographic reconstructions due to the use of low electron dose in order to avoid radiation damage to specimen. Poor contrast makes it difficult to identify different molecules or structures for subtomographic averaging. Normally for cellular tomography without phase plate, one could image with a large defocus value to achieve better contrast, aiding in distinguishing densities with different characteristics for subtomographic averaging. However, such approach only offers limited improvements in contrast ( S1 Fig ), and difficulties still exist in identifying the slender gB in postfusion conformation in our tomograms. This experience is consistent with two previous cryoET studies on HSV-1 gB structures, in which large defocus values were used to increase contrast to facilitate subsequent subtomographic averaging, yet the resulting structure either is at much lower resolution [ 26 ] than reported here or has led to controversial interpretations [ 27 ]. The greatly improved contrast afforded by VPP technology allowed the differentiation of various glycoprotein structures based on their characteristic appearances on the virion membrane ( Fig 7A and 7B ; S1 Fig ). Therefore, cryoET with VPP offers a clear advantage in resolving structures of proteins in the native environments, enabling their identifications and subtomographic averaging to obtain structures of multi-functional states, as also demonstrated by the existence of two states of 26S proteasome inside neurons [ 31 ]. 10.1371/journal.ppat.1007452.g007 Fig 7 Schematic illustration of conformation changes of gB during membrane fusion. (A, B) Subtomographic averages of prefusion gB (A) with domains illustrated as in Fig 3 and "postfusion" gB (B) with domains colored as in [ 16 ]. (C) A working model of gB conformational change during membrane fusion. In step 1, destabilization of the endodomain of prefusion gB either by cytotail conformational changes following gH/gL receptor-binding or by other means ( e . g ., mechanical stress such as high-speed centrifugation during viral purification) triggers DI and DII to reorient, exposing the fusion loops on DI. Subsequently, the exposed fusion loops could make contact either with cell membrane in close proximity (in the case of receptor binding) (step 2) or with viral membrane. Finally (step 3), DV refolds into an extended form, transforming gB into its "postfusion" conformation: in the presence of cell membrane, the C-terminal part and the fusion loops come together and the membranes fuse; in the absence of cell membrane, the exposed fusion loops insert into the viral membrane. A vital step of herpesvirus infections is the fusion of viral and cell membranes, a complicated process involving at least three conserved proteins—gB, gH and gL. The in situ structures of gB at both prefusion and "postfusion" conformations reported here can shed lights on conformational changes of gB during membrane fusion and inform how herpesvirus entry into cell ( Fig 7 ). Prior to fusion, gB needs to be maintained at its inactive, metastable prefusion conformation ( Fig 7A ). The maintenance of this metastable conformation possibly involves a properly-folded endodomain of gB, since removal of the endodomain caused gB ectodomain to adopt the postfusion conformation [ 46 ]. In addition, the direct observation in our cryoET structure of gB-gH/gL complex ( Fig 6 ) and its isolation by co-immunoprecipitation [ 42 ] both suggest that the metastable ectodomain of gB might also be stabilized through the interaction with the ectodomain of gH subunit ( Fig 7C ). Host receptor-binding to gH/gL complex would trigger a conformational change in gH/gL cytotail and its dissociation from, and the destabilization of, the endodomain of gB, which in turn triggers the massive conformational changes of gB ectodomain to expose its fusion loops (step 1). Subsequently, DIII central helix extends, unfurling other domains and swinging the fusion loops to engage with the host membrane (step 2). Facilitated by the intrinsic fluidity in the plane of the membrane, the refolding of gB domains to the lower-energy, postfusion conformation, in which its ectodomain C-terminal end and the fusion loops must come together, leads to fusion of the two membranes and the release of viral DNA-containing capsid into cytoplasm (step 3). In the absence of receptor binding as in the situation of this study, mechanical stress to the membrane caused by such means as high-speed centrifugation could also destabilize the membrane-associated endodomain, triggering metastable prefusion gB to undergo the cascade of transformation events, possibly accompanied by the exposure of the fusion loops (step 1). Lacking host cell membrane, these events, with exposed fusion loops eventually encountering and inserting its hydrophobic moieties into the viral membrane, will be followed by refolding of other domains into the stable, "postfusion" conformation (step 3). Notably, the topology of the conformational change during step 2 to step 3 would preclude transiting from prefusion to postfusion conformation without breaking the three-fold symmetry. Indeed, monomeric intermediates of VSV G have been observed both in solution and on the surface of virions at intermediate pH conditions [ 22 , 23 ]. In our model, the fusion loops of prefusion gB point to and are in close proximity to the viral membrane, possibly buried within a hydrophobic "mask" of MPR, which is attached to the C-terminal end of the gB ectodomain crystal structure. This membrane-proximal location of the gB fusion loops is the same as that based on the cryoET structure of the HSV-1 gB/anti-fusion loop 2-antibody at 5nm resolution [ 26 ] and is consistent with the fusion loop locations in all known atomic structures of classes I and III viral fusion proteins, including influenza HA [ 47 ], HIV env trimer [ 48 ], VSV G [ 21 ] and others [ 6 ]. Notably, our model is in stark contrast to the exposed fusion loops assigned to the membrane-distal tips of the "short-form" HSV-1 gB structures [ 27 ], which were obtained by cryoET of purified gB-containing vesicles. The ectodomain of the "short-form" vesicular HSV-1 gB structure is 15% shorter in height and 23% wider in diameter than that of our in situ HCMV gB structure, despite both sharing the Christmas tree shape ( S5 Fig ). Superposition of the domain assignment obtained by the hierarchical fitting approach [ 27 ] into the "short-form" HSV-1 gB structure shows that the densities projecting from the lower whorl of the Christmas tree-shaped trimer were unaccounted for ( S5B Fig ). Moreover, placing the same domain assignment into our in situ HCMV gB prefusion structure reveals that the fusion loops in this assignment are projecting out of the cryoET map, yet the leader density of the map is not accounted for ( S5C Fig ). When filtering the pseudoatomic model to 25à , the cross-correlation coefficient is 0.74, as compared to 0.93 of our prefusion structure. We believe that an exposed fusion loop orientation of prefusion gB is unlikely for both chemical and biological reasons—exposed hydrophobic moieties are chemically unfavorable in solution and can lead to unproductive membrane insertion during infection. Indeed, the "short-form" HSV-1 gB structure was cautiously interpreted as an ambiguous "prefusion and/or intermediate" conformation [ 27 ], probably to reconcile these contradictory considerations. Secondary structure prediction indicates that the endodomain is helix-rich (~50%) ( Fig 4A ). Our results suggest that gB endodomain undergoes significant conformational changes, from prominently visible/stable in the prefusion structure ( Fig 3I and 3K ), to invisible/flexible in the "postfusion" structure ( Fig 3C and 3E ). Proteolysis and circular dichroism analyses of the endodomain of the highly homologous HSV-1 gB posit that gB endodomain clamps the viral membrane and stabilizes gB in its prefusion conformation [ 44 , 49 ]. This proposed model is supported by studies on truncation and substitution mutations in endodomain [ 44 , 46 ]. The structured endodomain resolved in the recent crystal structure of full-length gB was thought to be similar to that in prefusion gB [ 24 ]. Detergent solubilization of the membrane may be responsible for the postfusion conformation of its ectodomain. Our observation of the endodomain structure of HCMV gB changing from a stable, prefusion conformation ( Fig 3I ) to a flexible, postfusion conformation ( Fig 3E ) is consistent with its proposed role in stabilization of gB prefusion conformation on native viral membrane [ 24 ]. Materials and methods HCMV virion preparation Human fibroblast MRC-5 cells (ATCC) were cultured in Eagle's Minimum Essential Medium (EMEM, ATCC) with 10% fetal bovine serum (FBS, Omega scientific: FB-11). Cells were grown in T-175 cm2 flasks to 90% confluence and infected with HCMV strain AD169 (ATCC, Rockville, MD) at a multiplicity of infection (MOI) of 0.1–0.5, and incubated for about 7 days. Once the cells showed 100% cytopathic effect, the media were collected and centrifuged at 10,000 g for 15 min to remove cells and large cell debris. The clarified supernatant was collected and centrifuged at 60, 000 g for 1 hour to pellet HCMV virions. Pellets were resuspended in 20mM phosphate buffered saline (PBS, pH 7.4), loaded on a 15%–50% (w/w) sucrose density gradient, and centrifuged at 60,000 g for 1 hr. After the density gradient centrifugation, three light-scattering bands were observed in the density gradient: top, middle and bottom. The middle band contained both HCMV virions and NIEPs (particles with intact viral envelopes as judged by negative-staining EM) and was collected, diluted in PBS and then centrifuged at 60,000 g for 1 hour. The final pellet was resuspended in PBS for further cryoET sample preparation. VSV virion preparation VSV virion (Indiana serotype, San Juan strain) samples were produced as previously described [ 50 ]. Particularly, the inoculum was passaged multiple times in Hela cells with a very low multiplicity of infection (MOI), 0.001, to suppress the truncated defective-interference particles. The full VSV particles were isolated in a sucrose gradient and the final inoculum was also plaque-purified in Hela cells. We then pelleted the VSV virions at 30,000g for 2 hours and resuspended them in PBS. The stock was subjected to another low speed centrifugation at 12,000g for 5min in a desktop centrifuge to remove large aggregates. After resuspension, the pellets were banded on a 10ml density gradient containing 0–50% potassium tartrate and 30–0% glycerol. The virions-containing band was collected, diluted in PBS, pelleted at 30,000g for 2 hours, resuspended in PBS and kept in 4°C refrigerator for further cryoET sample preparation. CryoET sample preparation and data collection An aliquot of 2.5 μl of the sample mixed with 5-nm diameter gold beads were applied onto freshly glow-discharged Quantifoil Holey Carbon Grids. Grids were blotted and plunge-frozen in liquid ethane cooled by liquid nitrogen using an FEI Mark IV Vitrobot cryo-sample plunger and were stored in liquid nitrogen before subsequent usage. CryoEM imaging and cryoET tilt series acquisition were performed with SerialEM [ 51 ] on an FEI Titan Krios 300kV transmission electron microscope equipped with a Gatan imaging filter (GIF), a Gatan K2 Summit direct electron detector, and with or without a Volta phase plate (VPP). Tilt series were recorded by tilting the specimen covering the angular range of -66° to +60° (starting tilt from -48° to +60°, then from -50° to -66°) with 2° or 3° interval, with a nominal magnification of x53,000 (corresponding to a calibrated pixel size of 2.6 à ) and a cumulative electron dose of 100~110 e - /à 2 . Exposure time was multiplied by a factor of the square root of 1/cosα (in which α = tilt angle), and the exposure time at 0° was set at 1.2s for the tilt step-size of 2° or 1.6s for the tilt step-size of 3°. Movies were recorded with the frame rate of 0.2 frame/s on a Gatan K2 Summit direct electron detector operated in counting mode with the dose rate of 8–10 e - /pixel/s. An energy filter slit of 20 eV was chosen for the GIF. For imaging with VPP, defocus value was targeted at -0.6μm. Note, one of the benefits of using a phase plate is that the CTF is insensitive to the sign of the defocus value being negative (underfocus) or positive (overfocus) [ 52 ]. VPP was advanced to a new position every tilt series, followed by a 2 min waiting for stabilization, and pre-conditioned by electron illumination with a total dose of 12 nC for 60s to achieve a phase shift of ~54° as previously described [ 53 ]. For tilt series obtained without VPP, the defocus value was maintained at around -4μm while other imaging parameters were kept the same as those for the tilt series with VPP. 3D reconstruction Frames in each movie of the raw tilt series were aligned, drift-corrected and averaged with Motioncorr [ 54 ] to produce a single image for each tilt angle. Both sets of tilt series, collected with and without VPP, were reconstructed with IMOD 4.8 software package [ 55 ] in the following six steps. All images in a tilt series were coarsely aligned by cross-correlation (step 1) and then finely aligned by tracking selected gold fiducial beads (step 2). The positions of each bead in all images of the tilt series were fitted into a specimen-movements mathematical model, resulting in a series of predicted positions. The mean residual error (mean distance between the actual and predicted positions) was recorded to facilitate bead tracking and poorly-modeled-bead fixing (step 3). With the boundary box reset and the tilt axis readjusted (step 4), images were realigned (step 5). Finally, two tomograms were generated by weighted back projection and simultaneous iterative reconstruction technique (SIRT) method, respectively (step 6). For data collected without VPP, contrast transfer function (CTF) was corrected with the ctfphaseflip program [ 56 ] of IMOD in step5. The defocus value for each image in one tilt series was determined by CTFTILT [ 57 ], and the estimated defocus value of each image was used as input for ctfphaseflip . Subtomographic averaging Subtomographic averaging was performed using PEET 1.11 [ 58 , 59 ]. High contrast SIRT tomograms were 4× binned by the binvol program of IMOD to facilitate particle picking. Particles were picked manually in IMOD as follows. For distinct conformations of VSV G and HCMV gB on viral envelope, two points ( head and tail ) in one contour were used to define one particle (glycoprotein)— head is the membrane-proximal end of the protrusion density while tail is the membrane-distal end. An initial motive list file, a RotAxes file and three model files containing the coordinates of head , centroid and tail for each particle were generated by stalkInit in PEET . In total, we manually picked 337 long-form particles from 5 VPP tomograms of VSV, and 350 columnar tree-shaped particles and 886 Christmas tree-shaped particles from 11 VPP tomograms of HCMV. Besides, 637 Christmas tree-shaped particles were picked from 28 non-VPP tomograms, averaged either alone or together with those from the VPP tomograms for prefusion gB. For the reconstruction of the long-form VSV G, subtomographic averaging was performed first with 4× binned SIRT tomograms using the sum of all particles as the initial reference. Through stalkInit , each particle's tilt orientation ( i . e ., the axes normal to the membrane) was already coarsely aligned to Y axis, but its twist orientation ( i . e ., the angle around the axis) was randomized. Therefore, in the first refinement cycle, we set the angular search range 180° max (-180° to 180°) with 9° step in Phi (Y axis), and 5° (-5° to 5°) max with 1° step in both Theta (Z axis) and Psi (X axis), and search distance 3 pixels along all three axes. Due to the known symmetry of postfusion VSV G, the resulting averaged structure was then trimerized and used as the reference of the next refinement cycle. The trimerized structure was the sum of each refined particle and its two symmetrical copies—the two symmetrical copies have the same position and tilt orientation as the refined particle, but twist orientation differed by either 120° or 240°. For subsequent refinement cycles, the newly trimerized structure from the last refinement cycle was used as reference, with both angular and distance search ranges narrowing down gradually. After four refinement cycles, the averaged structure converged based on no further improvement in resolution. The following refinement cycles were performed with 2× binned tomograms reconstructed by weighted back projection, after up-sampling (generations of 2× binned model files and updates of corresponding motive list files from the latest refinement cycle), with small search distance range (4 pixels) and narrow angular search range (-20° to 20°). The reference was updated from the averaged structure of the last refinement cycle (trimerized). For particles with distance of <1 pixel and twist angle difference of <1°, the one representative with lower cross-correlation coefficient was treated as duplicate particle and removed during the refinement. The averaged structure, contributed by 330 particles, converged after eight refinement cycles and was filtered to the final resolution, calculated by calcFSC in PEET based on the 0.143 FSC criterion. Reconstructions of columnar tree-shaped and Christmas tree-shaped particles on HCMV envelope followed the same refinement procedure as the reconstruction of long-form VSV G, except that trimerization was only applied after three-fold symmetry became apparent in the averaged structures. With the removal of duplicate particles, the final averaged structures of the postfusion (columnar tree-shaped) and prefusion (Christmas tree-shaped) conformations were obtained from 350 particles and 1509 particles, respectively. Furthermore, gold-standard FSC calculations for the structures were performed afterwards by splitting the original dataset of each conformation into two independent groups. The same refinement procedure used above was applied to the two newly-generated groups independently. Upon the convergence of the averaged structures, FSC were calculated by calcUnbiasedFSC in PEET ( S3A Fig .). For the reconstruction of the short-form VSV G, 65 particles were manually picked from five tomograms with single point to define the centroid position. Each particle was manually rotated around X, Y, Z axes to a similar orientation (both the tilt orientation and twist angle) in IMOD slicer window. By slicer2MOTL in PEET , the initial motive list files for subtomographic averaging were generated from the corresponding X, Y, Z rotation degrees. For the Angular Search Range, small search range was set during all seven refinement cycles. The final subtomographic average was Gaussian filtered with width 7 using the "volume filter" tool in UCSF Chimera [ 60 ]. Due to the limited number of particles (49 particles), HCMV gB-gH/gL complex was reconstructed with the same strategy above. 3D visualization We used IMOD [ 61 ] to visualize reconstructed tomograms and UCSF Chimera to visualize the subtomographic averages in three dimensions. The crystal structures of prefusion VSV G (PDB: 5I2S) [ 21 ], postfusion VSV G (PDB: 5I2M) [ 20 ], HCMV postfusion gB (PDB: 5CXF) [ 16 ] and gH/gL part from HCMV pentamer (PDB: 5VOB) [ 17 ] were fitted into subtomographic averages of prefusion G, postfusion G, postfusion gB and gB-gH/gL complex, respectively, with the tool fit in map in Chimera . Segmentation and surface rendering for the membrane and tegument proteins were done by the tools volume tracer and color zone in Chimera . All membrane glycoproteins were placed back on the viral membrane according to their locations in the original tomogram. A published structure of HCMV capsid with inner tegument protein [ 11 ] was filtered to 10 à and placed back at the same position of the capsid in tomogram. Domain modeling and structure prediction As outlined below, we employed a combination of initial manual fitting of known domain structures, followed by simulation with MDFF program [ 38 ] to generate a gB prefusion model based on our cryoET prefusion gB trimer density map and the existing gB ectodomain postfusion crystal structure (PDB: 5CXF) [ 16 ]. First, the ectodomain in the subtomographic averaged density map of prefusion gB trimer was segmented out and its symmetric axis obtained with Chimera 's "volume eraser" tool and "measure symmetry" command, respectively. Second, Chimera 's "fitmap" command with "global search" and 15à -resolution options was used to refine 1000 initial random DI placements, resulting in 28 refined fitted positions, each with a correlation coefficient (between the fitted model and the density map) and a "clash volume fraction" value (between symmetry-related copies). We chose the fitted position with the largest fitting score, defined as the correlation coefficient subtracted by the "clash volume fraction" penalty value ( Fig 5A ). Third, we obtained our initial DIII by computationally mutating the DIII model from the existing hypothetic model of EBV prefusion gB [ 14 ], as it is known to differ substantially from its postfusion conformation for both herpesvirus gB [ 14 , 62 ] and homologous VSV G [ 21 ]. Compared to that in the postfusion gB, the central helix α4 in DIII in the prefusion gB is bent in order to fit into the top of the Christmas tree-shaped density. This bent varies from only ~30° in our proposed HCMV gB prefusion structure to ~90° in VSV G ( Fig 5D ) [ 21 ] and ~180° in influenza HA [ 47 ] and HIV env [ 48 ]. This DIII model, and the models of DII and DIV from the gB postfusion crystal structure were manually fitted as rigid bodies into our prefusion gB trimer cryoET density to produce a composite model with the above obtained DI trimer model by referencing the prefusion VSV G crystal structure. Connecting loops were then added to this composite model through the Modloop server [ 63 ]. Fourth, the resulting trimer model was used as the initial model for MDFF simulations [ 38 ] with grid force scale of 0.3. Secondary structure, cis peptide and chirality restraints were imposed during MDFF simulations. Simulations were performed with NAMD 2.12 [ 64 ], using the CHARMM36 force field with CMAP corrections [ 65 ]. Secondary structures for residues 707–906 of gB were predicted with Phyre2 [ 66 ]. HCMV virion preparation Human fibroblast MRC-5 cells (ATCC) were cultured in Eagle's Minimum Essential Medium (EMEM, ATCC) with 10% fetal bovine serum (FBS, Omega scientific: FB-11). Cells were grown in T-175 cm2 flasks to 90% confluence and infected with HCMV strain AD169 (ATCC, Rockville, MD) at a multiplicity of infection (MOI) of 0.1–0.5, and incubated for about 7 days. Once the cells showed 100% cytopathic effect, the media were collected and centrifuged at 10,000 g for 15 min to remove cells and large cell debris. The clarified supernatant was collected and centrifuged at 60, 000 g for 1 hour to pellet HCMV virions. Pellets were resuspended in 20mM phosphate buffered saline (PBS, pH 7.4), loaded on a 15%–50% (w/w) sucrose density gradient, and centrifuged at 60,000 g for 1 hr. After the density gradient centrifugation, three light-scattering bands were observed in the density gradient: top, middle and bottom. The middle band contained both HCMV virions and NIEPs (particles with intact viral envelopes as judged by negative-staining EM) and was collected, diluted in PBS and then centrifuged at 60,000 g for 1 hour. The final pellet was resuspended in PBS for further cryoET sample preparation. VSV virion preparation VSV virion (Indiana serotype, San Juan strain) samples were produced as previously described [ 50 ]. Particularly, the inoculum was passaged multiple times in Hela cells with a very low multiplicity of infection (MOI), 0.001, to suppress the truncated defective-interference particles. The full VSV particles were isolated in a sucrose gradient and the final inoculum was also plaque-purified in Hela cells. We then pelleted the VSV virions at 30,000g for 2 hours and resuspended them in PBS. The stock was subjected to another low speed centrifugation at 12,000g for 5min in a desktop centrifuge to remove large aggregates. After resuspension, the pellets were banded on a 10ml density gradient containing 0–50% potassium tartrate and 30–0% glycerol. The virions-containing band was collected, diluted in PBS, pelleted at 30,000g for 2 hours, resuspended in PBS and kept in 4°C refrigerator for further cryoET sample preparation. CryoET sample preparation and data collection An aliquot of 2.5 μl of the sample mixed with 5-nm diameter gold beads were applied onto freshly glow-discharged Quantifoil Holey Carbon Grids. Grids were blotted and plunge-frozen in liquid ethane cooled by liquid nitrogen using an FEI Mark IV Vitrobot cryo-sample plunger and were stored in liquid nitrogen before subsequent usage. CryoEM imaging and cryoET tilt series acquisition were performed with SerialEM [ 51 ] on an FEI Titan Krios 300kV transmission electron microscope equipped with a Gatan imaging filter (GIF), a Gatan K2 Summit direct electron detector, and with or without a Volta phase plate (VPP). Tilt series were recorded by tilting the specimen covering the angular range of -66° to +60° (starting tilt from -48° to +60°, then from -50° to -66°) with 2° or 3° interval, with a nominal magnification of x53,000 (corresponding to a calibrated pixel size of 2.6 à ) and a cumulative electron dose of 100~110 e - /à 2 . Exposure time was multiplied by a factor of the square root of 1/cosα (in which α = tilt angle), and the exposure time at 0° was set at 1.2s for the tilt step-size of 2° or 1.6s for the tilt step-size of 3°. Movies were recorded with the frame rate of 0.2 frame/s on a Gatan K2 Summit direct electron detector operated in counting mode with the dose rate of 8–10 e - /pixel/s. An energy filter slit of 20 eV was chosen for the GIF. For imaging with VPP, defocus value was targeted at -0.6μm. Note, one of the benefits of using a phase plate is that the CTF is insensitive to the sign of the defocus value being negative (underfocus) or positive (overfocus) [ 52 ]. VPP was advanced to a new position every tilt series, followed by a 2 min waiting for stabilization, and pre-conditioned by electron illumination with a total dose of 12 nC for 60s to achieve a phase shift of ~54° as previously described [ 53 ]. For tilt series obtained without VPP, the defocus value was maintained at around -4μm while other imaging parameters were kept the same as those for the tilt series with VPP. 3D reconstruction Frames in each movie of the raw tilt series were aligned, drift-corrected and averaged with Motioncorr [ 54 ] to produce a single image for each tilt angle. Both sets of tilt series, collected with and without VPP, were reconstructed with IMOD 4.8 software package [ 55 ] in the following six steps. All images in a tilt series were coarsely aligned by cross-correlation (step 1) and then finely aligned by tracking selected gold fiducial beads (step 2). The positions of each bead in all images of the tilt series were fitted into a specimen-movements mathematical model, resulting in a series of predicted positions. The mean residual error (mean distance between the actual and predicted positions) was recorded to facilitate bead tracking and poorly-modeled-bead fixing (step 3). With the boundary box reset and the tilt axis readjusted (step 4), images were realigned (step 5). Finally, two tomograms were generated by weighted back projection and simultaneous iterative reconstruction technique (SIRT) method, respectively (step 6). For data collected without VPP, contrast transfer function (CTF) was corrected with the ctfphaseflip program [ 56 ] of IMOD in step5. The defocus value for each image in one tilt series was determined by CTFTILT [ 57 ], and the estimated defocus value of each image was used as input for ctfphaseflip . Subtomographic averaging Subtomographic averaging was performed using PEET 1.11 [ 58 , 59 ]. High contrast SIRT tomograms were 4× binned by the binvol program of IMOD to facilitate particle picking. Particles were picked manually in IMOD as follows. For distinct conformations of VSV G and HCMV gB on viral envelope, two points ( head and tail ) in one contour were used to define one particle (glycoprotein)— head is the membrane-proximal end of the protrusion density while tail is the membrane-distal end. An initial motive list file, a RotAxes file and three model files containing the coordinates of head , centroid and tail for each particle were generated by stalkInit in PEET . In total, we manually picked 337 long-form particles from 5 VPP tomograms of VSV, and 350 columnar tree-shaped particles and 886 Christmas tree-shaped particles from 11 VPP tomograms of HCMV. Besides, 637 Christmas tree-shaped particles were picked from 28 non-VPP tomograms, averaged either alone or together with those from the VPP tomograms for prefusion gB. For the reconstruction of the long-form VSV G, subtomographic averaging was performed first with 4× binned SIRT tomograms using the sum of all particles as the initial reference. Through stalkInit , each particle's tilt orientation ( i . e ., the axes normal to the membrane) was already coarsely aligned to Y axis, but its twist orientation ( i . e ., the angle around the axis) was randomized. Therefore, in the first refinement cycle, we set the angular search range 180° max (-180° to 180°) with 9° step in Phi (Y axis), and 5° (-5° to 5°) max with 1° step in both Theta (Z axis) and Psi (X axis), and search distance 3 pixels along all three axes. Due to the known symmetry of postfusion VSV G, the resulting averaged structure was then trimerized and used as the reference of the next refinement cycle. The trimerized structure was the sum of each refined particle and its two symmetrical copies—the two symmetrical copies have the same position and tilt orientation as the refined particle, but twist orientation differed by either 120° or 240°. For subsequent refinement cycles, the newly trimerized structure from the last refinement cycle was used as reference, with both angular and distance search ranges narrowing down gradually. After four refinement cycles, the averaged structure converged based on no further improvement in resolution. The following refinement cycles were performed with 2× binned tomograms reconstructed by weighted back projection, after up-sampling (generations of 2× binned model files and updates of corresponding motive list files from the latest refinement cycle), with small search distance range (4 pixels) and narrow angular search range (-20° to 20°). The reference was updated from the averaged structure of the last refinement cycle (trimerized). For particles with distance of <1 pixel and twist angle difference of <1°, the one representative with lower cross-correlation coefficient was treated as duplicate particle and removed during the refinement. The averaged structure, contributed by 330 particles, converged after eight refinement cycles and was filtered to the final resolution, calculated by calcFSC in PEET based on the 0.143 FSC criterion. Reconstructions of columnar tree-shaped and Christmas tree-shaped particles on HCMV envelope followed the same refinement procedure as the reconstruction of long-form VSV G, except that trimerization was only applied after three-fold symmetry became apparent in the averaged structures. With the removal of duplicate particles, the final averaged structures of the postfusion (columnar tree-shaped) and prefusion (Christmas tree-shaped) conformations were obtained from 350 particles and 1509 particles, respectively. Furthermore, gold-standard FSC calculations for the structures were performed afterwards by splitting the original dataset of each conformation into two independent groups. The same refinement procedure used above was applied to the two newly-generated groups independently. Upon the convergence of the averaged structures, FSC were calculated by calcUnbiasedFSC in PEET ( S3A Fig .). For the reconstruction of the short-form VSV G, 65 particles were manually picked from five tomograms with single point to define the centroid position. Each particle was manually rotated around X, Y, Z axes to a similar orientation (both the tilt orientation and twist angle) in IMOD slicer window. By slicer2MOTL in PEET , the initial motive list files for subtomographic averaging were generated from the corresponding X, Y, Z rotation degrees. For the Angular Search Range, small search range was set during all seven refinement cycles. The final subtomographic average was Gaussian filtered with width 7 using the "volume filter" tool in UCSF Chimera [ 60 ]. Due to the limited number of particles (49 particles), HCMV gB-gH/gL complex was reconstructed with the same strategy above. 3D visualization We used IMOD [ 61 ] to visualize reconstructed tomograms and UCSF Chimera to visualize the subtomographic averages in three dimensions. The crystal structures of prefusion VSV G (PDB: 5I2S) [ 21 ], postfusion VSV G (PDB: 5I2M) [ 20 ], HCMV postfusion gB (PDB: 5CXF) [ 16 ] and gH/gL part from HCMV pentamer (PDB: 5VOB) [ 17 ] were fitted into subtomographic averages of prefusion G, postfusion G, postfusion gB and gB-gH/gL complex, respectively, with the tool fit in map in Chimera . Segmentation and surface rendering for the membrane and tegument proteins were done by the tools volume tracer and color zone in Chimera . All membrane glycoproteins were placed back on the viral membrane according to their locations in the original tomogram. A published structure of HCMV capsid with inner tegument protein [ 11 ] was filtered to 10 à and placed back at the same position of the capsid in tomogram. Domain modeling and structure prediction As outlined below, we employed a combination of initial manual fitting of known domain structures, followed by simulation with MDFF program [ 38 ] to generate a gB prefusion model based on our cryoET prefusion gB trimer density map and the existing gB ectodomain postfusion crystal structure (PDB: 5CXF) [ 16 ]. First, the ectodomain in the subtomographic averaged density map of prefusion gB trimer was segmented out and its symmetric axis obtained with Chimera 's "volume eraser" tool and "measure symmetry" command, respectively. Second, Chimera 's "fitmap" command with "global search" and 15à -resolution options was used to refine 1000 initial random DI placements, resulting in 28 refined fitted positions, each with a correlation coefficient (between the fitted model and the density map) and a "clash volume fraction" value (between symmetry-related copies). We chose the fitted position with the largest fitting score, defined as the correlation coefficient subtracted by the "clash volume fraction" penalty value ( Fig 5A ). Third, we obtained our initial DIII by computationally mutating the DIII model from the existing hypothetic model of EBV prefusion gB [ 14 ], as it is known to differ substantially from its postfusion conformation for both herpesvirus gB [ 14 , 62 ] and homologous VSV G [ 21 ]. Compared to that in the postfusion gB, the central helix α4 in DIII in the prefusion gB is bent in order to fit into the top of the Christmas tree-shaped density. This bent varies from only ~30° in our proposed HCMV gB prefusion structure to ~90° in VSV G ( Fig 5D ) [ 21 ] and ~180° in influenza HA [ 47 ] and HIV env [ 48 ]. This DIII model, and the models of DII and DIV from the gB postfusion crystal structure were manually fitted as rigid bodies into our prefusion gB trimer cryoET density to produce a composite model with the above obtained DI trimer model by referencing the prefusion VSV G crystal structure. Connecting loops were then added to this composite model through the Modloop server [ 63 ]. Fourth, the resulting trimer model was used as the initial model for MDFF simulations [ 38 ] with grid force scale of 0.3. Secondary structure, cis peptide and chirality restraints were imposed during MDFF simulations. Simulations were performed with NAMD 2.12 [ 64 ], using the CHARMM36 force field with CMAP corrections [ 65 ]. Secondary structures for residues 707–906 of gB were predicted with Phyre2 [ 66 ]. Supporting information S1 Fig Comparison of tomograms obtained with and without VPP. (A~C) A slice (A) and zoom-in envelope regions (B, C) of a tomogram reconstructed from tilt series obtained with VPP, showing greatly improved contrast that is sufficient to distinguish columnar tree-shaped ("postfusion") gB (yellow arrows in B) from the Christmas tree-shaped (prefusion) gB (red arrows in C). (D~F) A slice (D) and zoom-in envelope regions (E, F) of a tomogram reconstructed from tilt series obtained without VPP, showing the relatively poor contrast and great ambiguity to distinguish columnar tree-shaped ("postfusion") gB (yellow arrow in E) from the Christmas tree-shaped (prefusion) gB (red arrow in F). Consequently, significantly more tilt series without VPP than with VPP had to been recorded in order to obtain similar number of particles for subtomographic averaging. (TIF) Click here for additional data file. S2 Fig Subtomographic averages of gB without imposing symmetry. (A, B) Subtomographic averages of gB in its postfusion conformation without imposing symmetry viewed from side (A) and top (B). (C, D) Subtomographic averages of gB in its prefusion conformation without imposing symmetry viewed from side (C) and top (D). (TIF) Click here for additional data file. S3 Fig Fourier shell correlation (FSC) analyses and resolution comparisons. (A) FSC coefficients as a function of spatial frequency for the gold-standard resolution determined for final subtomographic averages of prefusion (black) and "postfusion" (red) gB trimers. (B) FSC coefficients as a function of spatial frequency between subtomographic averages of prefusion gB trimers obtained with VPP and without VPP. For the average obtained without VPP, CTF correction is necessary as indicated by the negative correlation coefficients in the range from 1/26 à -1 to 1/20 à -1 spatial frequencies. (TIF) Click here for additional data file. S4 Fig Tomograms and subtomographic averages from tilt series obtained without VPP. (A, B) Comparison of corresponding slices from CTF-uncorrected (A) and CTF-corrected (B) tomograms. The viral envelope region of the particle indicated by the dashed boxes in (A, red) and (B, yellow) are enlarged, showing that the membrane bilayer is better resolved after CTF correction in the yellow zoom-in inset. (C, D) Subtomographic average of the Christmas tree-shaped densities (yellow) and associated membrane bilayer (blue) viewed from side, top and slanted bottom. (C) is obtained without CTF-correction and (D) is with CTF-correction. (E) FSC coefficients as a function of spatial frequency between subtomographic averages of prefusion gB trimers obtained with CTF correction and without CTF correction. For the subtomographic average obtained without VPP, CTF correction is necessary as indicated by the negative correlation coefficients (grey zone) for some spatial frequencies. (TIF) Click here for additional data file. S5 Fig Direct comparisons of the averaged map and fitted pseudoatomic model between prefusion HCMV gB and previous "short-form" HSV-1 gB. (A) Christmas tree-shaped prefusion gB on HCMV virion (this study), reviewed from its side and top, as in Figs 3 and 5 . (B) "Short-form" gB of HSV-1 [ 27 ] showing as in (A), colored surface (upper panel) and superposition (lower panel) of semitransparent density (gray) and fitted domains (ribbons). Red dotted ellipses indicate the densities that were unaccounted for by the fitted psedoatomic model. (C) Superposition of the proposed HSV-1 gB domain (DI and DII) arrangement from [ 27 ] into the density of the prefusion gB trimer on HCMV virion (this study) showing mismatch, which is indicated by red dotted ellipses. (TIF) Click here for additional data file. S1 Movie An example of aligned tilt series obtained with VPP. (Scale bar: 100nm.) (AVI) Click here for additional data file. S2 Movie Slices through a tomogram reconstructed by simultaneous iterative reconstruction technique (SIRT) from tilt series in S1 Movie . (Scale bar: 100nm.) (AVI) Click here for additional data file. S3 Movie An example of aligned tilt series obtained without VPP. (Scale bar: 100nm.) (AVI) Click here for additional data file. S4 Movie Slices through a tomogram reconstructed by SIRT from tilt series in S3 Movie . (Scale bar: 100nm.) (AVI) Click here for additional data file. S5 Movie Surface rendering of the subtomographic average of "postfusion" gB (yellow) and associated membrane bilayer (blue). (AVI) Click here for additional data file. S6 Movie Surface rendering of the subtomographic average of prefusion gB (yellow) and associated membrane bilayer (blue). (AVI) Click here for additional data file. S7 Movie MDFF-simulated prefusion gB structure, colored as [ 16 ], is superimposed with two other symmetric copies (gray ribbon) in the subtomographic average of the Christmas tree-shaped density (semi-transparent gray). (AVI) Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9477280/

Single domain antibodies against enteric pathogen virulence factors are active as curli fiber fusions on probiotic E . coli Nissle 1917

Enteric microbial pathogens, including Escherichia coli , Shigella and Cryptosporidium species, take a particularly heavy toll in low-income countries and are highly associated with infant mortality. We describe here a means to display anti-infective agents on the surface of a probiotic bacterium. Because of their stability and versatility, VHHs, the variable domains of camelid heavy-chain-only antibodies, have potential as components of novel agents to treat or prevent enteric infectious disease. We isolated and characterized VHHs targeting several enteropathogenic E . coli (EPEC) virulence factors: flagellin (Fla), which is required for bacterial motility and promotes colonization; both intimin and the translocated intimin receptor (Tir), which together play key roles in attachment to enterocytes; and E . coli secreted protein A (EspA), an essential component of the type III secretion system (T3SS) that is required for virulence. Several VHHs that recognize Fla, intimin, or Tir blocked function in vitro . The probiotic strain E . coli Nissle 1917 (EcN) produces on the bacterial surface curli fibers, which are the major proteinaceous component of E . coli biofilms. A subset of Fla-, intimin-, or Tir-binding VHHs, as well as VHHs that recognize either a T3SS of another important bacterial pathogen ( Shigella flexneri ), a soluble bacterial toxin (Shiga toxin or Clostridioides difficile toxin TcdA), or a major surface antigen of an important eukaryotic pathogen ( Cryptosporidium parvum ) were fused to CsgA, the major curli fiber subunit. Scanning electron micrographs indicated CsgA-VHH fusions were assembled into curli fibers on the EcN surface, and Congo Red binding indicated that these recombinant curli fibers were produced at high levels. Ectopic production of these VHHs conferred on EcN the cognate binding activity and, in the case of anti-Shiga toxin, was neutralizing. Taken together, these results demonstrate the potential of the curli-based pathogen sequestration strategy described herein and contribute to the development of novel VHH-based gut therapeutics. Introduction Enteric pathogens, which include viruses, bacteria, and eukaryotic microbes, are a major cause of global morbidity and mortality. These pathogens take a particularly heavy toll in low-income countries where diarrheal disease remains a major cause of infant mortality [ 1 , 2 ]. Traditional interventions such as antibiotics and vaccines suffer from limited efficacy, distribution and implementation challenges, and the rise of antimicrobial resistance [ 3 ]. Virulence factors have been identified for many important enteric microbes, but conventional measures to prevent or treat diarrheal disease based on these factors have proved difficult to develop. Therefore, new therapeutic strategies are needed. One of the leading causes of infant diarrheal disease and associated mortality in low- and middle-income countries is enteropathogenic Escherichia coli (EPEC) [ 1 , 4 – 6 ]. Colonization by EPEC is facilitated by flagella- (Fla-) driven motility that promotes penetration of the mucus layer and association with the intestinal epithelium, where bacteria induce the formation of 'attaching and effacing' (AE) lesions [ 7 , 8 ]. These lesions, which enable epithelial colonization, are characterized by the effacement of microvilli and the induction of filamentous actin 'pedestals' beneath bacteria closely associated with intestinal epithelial cells [ 9 , 10 ]. To generate AE lesions, EPEC utilizes a type III secretion system (T3SS) to translocate the bacterial effector Tir (translocated intimin receptor) into host cells, where it localizes to the plasma membrane and binds to the EPEC surface adhesin intimin. Intimin-mediated clustering of Tir triggers the assembly of filamentous actin beneath bound bacteria. The related pathogen, Shiga toxin-producing enterohemorrhagic E . coli (EHEC), a food-borne pathogen which causes systemic illness in high-income regions such as the U.S. and Europe, generates AE lesions by a similar mechanism [ 9 , 11 – 13 ], as do some veterinary pathogens such as rabbit enteropathogenic E . coli (REPEC) and the mouse pathogen Citrobacter rodentium [ 14 – 16 ]. The direct administration of antibodies or antibody fragments has been proposed as a potential treatment for enteric diseases of diverse etiology [ 17 – 23 ]. VHHs, the variable domain of camelid heavy-chain-only antibodies (also known as 'nanobodies'), appear particularly well suited for this application [ 18 , 24 – 28 ]. Unlike conventional antibodies, VHH antibodies can be efficiently and functionally expressed in E . coli thanks to their small size and single-domain structure. Furthermore, VHHs are effectively expressed as fusion proteins with other VHHs, thus potentially enhancing avidity, increasing specificity, and enabling binding to multiple targets [ 29 ]. Fusion with other functional domains also adds further versatility to their use as therapeutic agents. Together, these properties confer the potential to reduce production costs, improve scalability, and enable novel therapeutic applications. Although VHHs have opened many novel therapeutic avenues, several challenges remain for their implementation as intestinal therapeutics. Despite the inherent stability and robustness of many VHHs [ 30 ], the harsh chemical and enzymatic conditions and continuous flow found in the GI tract will likely promote the degradation and clearance of VHHs before they reach their target. The delivery of sufficient, stable, and functional VHHs to the gut environment therefore constitutes a substantial hurdle. Additionally, producing, purifying, and formulating large amounts of VHHs is likely to be resource- and labor-intensive, effectively limiting the practicality of such approaches, a particularly relevant limitation for implementation in low-income nation where the enteric disease burden is highest. Engineered living therapeutics are an alternative strategy for localized production and delivery of molecules to the gut. By genetically modifying a suitable nonpathogenic bacterial strain, heterologous proteins of interest can be produced in situ , circumventing the challenges associated with traditional drug delivery strategies [ 31 – 33 ]. The ability to utilize bacteria as a therapeutic agent, bypassing the need for protein purification, can potentially render engineered living therapeutics inexpensive and scalable. E . coli Nissle 1917 (EcN) has emerged in recent years as a leading candidate for such approaches [ 34 , 35 ]. EcN has an excellent track record of safety through decades of use as a probiotic, and has also been shown to reduce the severity of ulcerative colitis symptoms [ 36 ], as well as interfere with the pathogenicity of several enteric pathogens [ 37 ], in part due to its ability to colonize the human gastrointestinal tract [ 38 , 39 ]. Transient colonization of humans has also been shown using engineered EcN [ 40 ]. Notably, EcN and other laboratory E . coli strains can produce VHHs, as demonstrated by numerous studies [ 41 , 42 ]. Curli fibers are the main proteinaceous components of E . coli biofilms. In previous work, we used engineered EcN to display modified curli fibers in vivo [ 43 ]. By fusing heterologous protein domains to CsgA–the major curli subunit–we were able to construct a cell-anchored mesh of robust protein fibers endowed with novel functionalities, ranging from the display of anti-inflammatory peptides [ 43 ] to the nucleation of gold nanoparticles [ 44 ]. By fusing pathogen surface-binding VHHs to CsgA, we sought to adapt this strategy to enable EcN to bind enteric pathogens in situ , thereby interfering with their pathogen-host interactions and possibly resulting in pathogen elimination. We call this approach "curli-based pathogen sequestration", drawing an analogy to the polymer sequestrants used to remove excess ions from the gut in chronic kidney disease and a handful of other disorders [ 45 , 46 ]. Here, we describe the generation and characterization of novel VHHs targeting the Fla, Tir, intimin and EspA antigens of several EPEC, REPEC, EHEC, and Citrobacter strains. We then fused a subset of these VHHs, along with several previously described VHHs that bind virulence factors from other enteric pathogens, to CsgA. By expressing these modified curli fibers in EcN and testing their function, we demonstrated the efficacy of the curli-based sequestration approach in vitro against several pathogenic E . coli virulence factors. Finally, we showed that EcN producing CsgA-VHH fusions are capable of recognizing surface antigens on two other major enteric pathogens, Shigella flexneri and the eukaryotic pathogen Cryptosporidium parvum . Results Generation and initial characterization of VHHs that recognize Fla, Tir, intimin, or EspA With the goal of obtaining VHHs that bind to selected virulence factors of AE members of the pathogenic E . coli family ( Fig 1 ), we immunized alpacas with these virulence factor antigens, either purified from selected enteric pathogens or prepared as recombinant proteins. For anti-flagella VHHs, we immunized with both purified REPEC or EPEC flagella or purified recombinant FliC proteins from multiple pathogenic E . coli species. Recombinant REPEC EspA, EHEC and C . rodentium intimins, and EHEC Tir protein were also employed as additional virulence factor immunogens. Unlike flagella, the EspA, Tir, and portions of the intimin proteins are relatively well conserved [ 47 ], increasing the likelihood of identifying VHHs that recognize diverse AE pathogens. Following immunization, phage-displayed VHHs prepared from the alpaca B cells were panned and then screened for binding to the immunizing antigens and to orthologous antigens from related AE pathogens as described in Materials and Methods. VHH DNA sequences were then determined, and one or two representative VHHs from each different VHH family (i.e., a family constitutes those VHHs apparently derived from a common B cell progenitor) were selected for soluble protein expression. The selected VHHs and their binding properties are summarized in Table 1 . Enzyme-linked immunosorbent assays (ELISAs) were used to estimate the apparent affinity (EC 50 value) of the VHHs for their original antigen, as well as their apparent affinity for homologous targets from other AE pathogens. VHHs varied widely with regard to their binding capacities and cross-specificities ( Table 1 and S1 Fig ). Note that some of the selected VHHs were assigned a simplified name related to their target antigen, e.g., "αInt-12" (i.e., "anti-intimin 12"). 10.1371/journal.ppat.1010713.g001 Fig 1 Schematic overview of bacterial virulence factors used as VHH targets in this study. 10.1371/journal.ppat.1010713.t001 Table 1 Selected VHHs. (a) Anti-Fla VHHs VHH name Vector name Simplified name Immunogen Panned on EC 50 a REPEC motility inhibition REPEC flagella EPEC flagella EPEC FliC JUV-B11 JVE-2 αFla-1 REPEC flagella; EPEC rFliC REPEC flagella Trace b NB Trace - JUV-C4 JVE-4 αFla-2 0.5 1 0.3 ND JUV-E8 JVE-5 αFla-3 0.5 Trace NB + JUV-G8 JVE-7 αFla-4 5 NB 10 + JUV-H1 JVE-10 αFla-5 NB NB NB - JUV-H5 JVE-11 αFla-6 5 NB 25 + JWU-F3 JXA-1 MC1061/EPEC intimin 10 0.2 ND ND JWU-H4 JXA-5 3 10 ND ND JXE-B1 JXK-1 EPEC flagella EPEC flagella NB 0.1 ND ND (b) Anti-EspA VHHs VHH name Vector name Simplified name Immunogen Panned on EC 50 a Pedestal blocking activity REPEC EspA C . rodentium EspA JXF-D7 JXM-6 αEspA-1 REPEC EspA REPEC EspA 0.7 0.7 - JXF-D8 JXM-8 αEspA-2 0.7 0.7 - JXF-H9 JXM-12 αEspA-3 0.7 0.7 - JXF-C4 JXM-15 αEspA-4 Trace 0.7 - JYB-B1 JYE-1 αEspA-5 αEspA-1-captured REPEC EspA c 0.3 0.3 ND JYB-B8 JYE-2 αEspA-6 0.7 0.7 ND JYB-D1 JYE-3 αEspA-7 3 3 ND JYB-H4 JYE-4 αEspA-8 0.5 0.5 ND JYB-H6 JYE-5 αEspA-9 0.3 0.3 ND (c) Anti-Tir VHHs VHH name Vector name Simplified name Immunogen Panned on EC 50 a Tir-intimin blocking activity d EHEC Tir EPEC Tir REPEC Tir JVB-C6 JVG-1 αTir-1 EHEC Tir REPEC Tir trace 50 trace - JVB-G4 JVG-2 αTir-2 0.1 0.1 0.1 ++ e JVB-G8 JVG-3 αTir-3 0.2 0.2 0.2 - JVC-C6 JVI-1 αTir-4 EHEC Tir 0.1 0.2 0.1 + JVC-D10 JVI-2 αTir-5 0.1 0.2 0.2 - JVC-E5 JVI-3 αTir-6 0.1 0.2 0.2 + JVA-A1 JVF-1 αTir-7 0.7 10 3 - JVA-C8 JVF-2 αTir-8 0.1 0.2 0.2 +++ JVA-C9 JVF-3 αTir-9 0.1 5 0.2 - JVA-D4 JVF-4 αTir-10 0.5 0.5 0.5 - JVA-F6 JVF-7 αTir-11 0.1 25 0.2 - JVA-D11 JVF-8 αTir-12 0.2 0.2 0.2 + JVA-E10 JVF-12 αTir-13 0.5 0.2 0.1 - JVA-G1 JVF-14 αTir-14 0.1 0.2 0.2 +++ (d) Anti-intimin VHHs VHH name Vector name Simplified name Immunogen Panned on EC 50 a Pedestal blocking activity f EHEC intimin EPEC intimin MC1061 /EPEC intimin JWS-H4 JWZ-5 αInt-12 EHEC intimin E . coli 1061/pInt (EHEC) Trace Trace 10 - JWT-C1 JWZ-7 αInt-13 0.5 Trace NB + JWU-D8 JWZ-9 αInt-14 E . coli 1061/pInt (EPEC) Trace Trace 0.5 + JWU-G8 JWZ-15 αInt-17 NB NB 0.5 + JXN-E2 g JXS-2 C . rodentium intimin DH5α/pInt ( C . rodentium ) NB NB NB ND a EC 50 estimates based on dilution ELISAs such as shown in S1 Fig b Trace–EC 50 >125 nM, i.e., poor but detectable c Panning employed JXF-D7-captured REPEC EspA target d Tir-intimin binding inhibition from Fig 3 : + p125 nM, i.e., poor but detectable c Panning employed JXF-D7-captured REPEC EspA target d Tir-intimin binding inhibition from Fig 3 : + p95% containing VHH inserts. Identification and purification of VHHs Phage library panning methods have been previously described [ 90 ]. Typically, the virulence factor proteins were coated onto plastic at 10 μg/mL of target in the first panning round, followed by a second round of panning at high stringency, with virulence factor proteins coated at 1 μg/mL, and using a 10-fold lower titer of input phage, shorter binding times, and longer washes. In some cases, the virulence factor targets were captured onto plastic by previously coated VHHs which bind the target or its fusion partner. VHH capture panning was also used in some cases to block isolation of VHHs to immunodominant epitopes on monomeric targets. Random clones from the selected populations were then screened by ELISA for expression of VHHs that bound to the virulence factor targets. Clones producing the strongest signals or showing broader target specificity were characterized by DNA fingerprinting. The coding sequences of VHHs selected as having unique fingerprints and the strongest ELISA signals were obtained. Based on sequence homology, one VHH representing each homology group (having no evidence of a common B cell clonal origin) was selected for expression and characterization. These VHHs were expressed individually in pET32 vectors and purified as recombinant E . coli thioredoxin fusions with a carboxy-terminal E-tag, as previously described [ 90 ]. Dilution ELISA ELISAs were performed using Nunc Maxisorp 96 well plates (Thermo Fisher Scientific). Virulence factor targets were typically coated overnight at 4°C, 1 μg/mL in PBS, then blocked for at least an hour at 37°C with 4% milk in PBS, 0.1% Tween. For capture ELISAs, plates were first coated with 5 μg/mL of VHHs that recognized the virulence factor or its fusion partner. The captured VHHs lacked both the thioredoxin partner and E-tag. After blocking, the virulence factor was then incubated at 1 μg/mL for one hour at 37°C with 4% milk in PBS, 0.1% Tween and washed. Dilution ELISAs were then initiated by diluting the VHH (expressed in a pET-32 vector with an amino terminal thioredoxin and a carboxyl terminal E-tag) to 125 nM and performing serial dilutions of 1:5. After incubation for one hour at 37°C, plates were washed and then incubated with 1:10,000 rabbit HRP/anti-E-tag (Bethyl Laboratories) for one hour, washed, developed with TMB (Sigma Aldrich) as recommended by the manufacturer and measured for absorbance at 450 nm. ELISA measuring the effect of anti-Tir VHHs on the intimin-Tir interaction The ability of anti-Tir VHHs derived from EHEC to interfere with intimin-Tir binding was measured by ELISA. High-binding assay plates (Corning) were coated with 5 μg/mL of recombinant his-tagged Tir diluted in 1x coating buffer (50 mM Na2CO3, 50 mM NaHCO3, pH 9.6) in a volume of 100 μL per well and incubated overnight at 4°C. Plates were washed three times with 300 μL of wash buffer (0.05% Tween in PBS) and then blocked with BSA (3% in PBS) for 2 hours at room temperature (RT). Plates were washed and 100 μL of 500 nM anti-Tir VHH were added to each well. 0.1% BSA was used as a negative control. Plates were incubated at RT for 2 hours or at 4°C overnight. Wells were then probed with 150 nM GST-tagged intimin or with GST alone, and incubated at RT for 2 hours or at 4°C overnight. Plates were washed again and then fixed with 3.7% paraformaldehyde at RT for 20 mins at 4°C. Following another wash, plates were blocked with 5% milk in PBS for 30 min at RT. After washing, plates were incubated with goat anti-GST (GE Healthcare) for an hour and GST binding was detected kinetically using an alkaline-phosphatase-linked rabbit anti-goat IgG secondary antibody (diluted 1:2000 in 0.1% BSA/PBS). Binding of the secondary antibody was detected colorimetrically (AP substrate N1891, Sigma Aldrich) at 405 nm, and the average reaction rate (V mean ) was calculated. REPEC motility assay Motility assays were performed as described previously [ 73 ], with slight modification, to measure the ability of anti-Fla VHHs to inhibit Rabbit Enteropathogenic Escherichia coli (REPEC) motility. Briefly, REPEC cultures were streaked on LB agar plates and incubated for 16 hours at 37°C. The next day, a single colony was transferred into 5 mL LB broth and incubated overnight at 37°C with continuous shaking. On the following day, the culture was diluted 1:50 into LB broth and grown at 37°C with continuous shaking to an OD 600 of 0.5. A 1:1 dilution of bacteria and VHHs was then prepared (6 μL of bacterial culture was mixed with 6 μL of VHH concentrations ranging from 0 to 6.4 μM), mixed gently with a pipette, and incubated at 4°C for 2 h. The 12 μL mixture was then transferred to the center surface of a 0.3% semi solid agar plate and incubated for 24 h at room temperature. The diameter of bacterial growth was measured by first placing the plate on a dark background to enhance the contrast between bacterial growth and the agar medium. A metric scale ruler was then used to measure the growth diameter. Images were captured using a Syngene imager. EPEC pedestal assay The ability of anti-intimin and anti-Tir VHHs to inhibit EPEC pedestal formation was assessed after infection of HeLa cells, as described previously [ 91 ], with slight modification. Briefly, 30,000 HeLa cells were inoculated into the wells of 24-well plates (Invitro Scientific,) and incubated overnight at 37°C in an incubator with 5% CO2. On the same day, a single EPEC colony was inoculated into 5 mL DMEM in 100mM HEPES medium (pH 7.4) and incubated overnight in 5% CO2 at 37°C without shaking. The next day, the EPEC culture was diluted 1:16 into new infection medium (0.6 μL EPEC added to 9.4 μL media containing DMEM, 20 mM HEPES, and 3.5% FBS; pH 7.4), and 3.33 μL of the EPEC suspension were incubated either alone or with 100 nM anti-Tir or anti-intimin VHH in 0.5 mL DMEM at 4°C for 2 hours, in 1.5 mL Eppendorf tubes on a rocker. HeLa cell monolayers were then washed with 0.5 mL PBS, and EPEC suspensions were added to the monolayers. Plates were then centrifuged at 500 RPM for 5 min and incubated for 3 h at 37°C in a 5% CO2 incubator. Next, cells were washed twice with PBS, fixed with 0.5 mL 2.5% paraformaldehyde in PBS for 10 mins at RT on a shaker, washed twice with PBS for 5 min on a shaker at RT, permeabilized with 0.5 mL of 0.1% TritonX-100 for 5 min, and washed twice again before staining with DAPI (Thermo Fisher Scientific) and Alexa Fluor-488 Phalloidin (Thermo Fisher Scientific) at RT for 1.5 hours. Monolayers were then washed and 7 μL prolong gold anti-fade reagent (Thermo Fisher Scientific) was used to mount coverslips on wells before imaging with a fluorescent microscope. EPEC pedestal formation was blindly scored, as follows; 1: Very few pedestals are present on the edges of the wells, 2: More pedestals present, only at the edges of the well, 3: Most cells have no pedestals, but a few pedestals present in the center and edges of wells, 4: Most cells have pedestals, but a few empty cells are present, 5: The majority of the cells have pedestals. Using the above numbering criteria, pedestals were scored blindly by a second researcher. CsgA-VHH plasmid construction and cloning The cloning of the synthetic curli operon csgBACEFG onto the pL6FO vector was described in detail elsewhere [ 55 ]. DNA sequences of desired VHHs and corresponding primers were synthesized by and purchased from Integrated DNA Technologies. Plasmid construction was carried out using Gibson Assembly [ 92 ]. Quantitative Congo Red binding assay Curli fiber formation was quantified using a Congo Red binding assay as previously described [ 43 ]. Briefly, 1 mL of induced EcN CsgA-VHH culture was pelleted at 4000 × g for 10 minutes at room temperature and resuspended in a 25 μM Congo Red PBS solution. After a 10-minute incubation, the cell suspension was pelleted again, and the unbound Congo Red dye was quantified by measuring the supernatant absorbance at 490 nm. The signal was subtracted from a Congo Red blank, divided by the culture's OD 600 measurement to reflect curli production per cell, and normalized with respect to a EcN CsgA (no VHH) positive control. Electron microscopy Field emission scanning electron microscope (FESEM) samples were prepared by fixing with 2% (w/v) glutaraldehyde and 2% (w/v) paraformaldehyde at room temperature, overnight. The samples were gently washed with water, and the solvent was gradually exchanged with ethanol with an increasing ethanol 15-minute incubation step gradient (25, 50, 75 and 100% (v/v) ethanol). The samples were dried in a critical point dryer, placed onto SEM sample holders using silver adhesive (Electron Microscopy Sciences) and sputtered until they were coated in a 10–20 nm layer of Pt/Pd. Images were acquired using a Zeiss Ultra55 FESEM equipped with a field emission gun operating at 5–10 kV. GFP pull-down assay For each condition, 1 mL of induced overnight culture was centrifuged at 4000 × g for 10 minutes. The supernatant was aspirated, and the pellets were resuspended in 150 nM GFP in fasted-state simulated colonic fluid, which was prepared as described by Vertzoni et al . [ 93 ]. The cells were incubated on a shaking platform (225 RPM) at 37°C for 15 minutes and pelleted again. The GFP remaining in solution was assayed by measuring the fluorescent signal (485nm/528nm) using a plate reader (Spectramax M5, Molecular Devices). GFP concentration was estimated based on a calibration curve using known GFP concentrations. Shiga toxin pull-down assay Induced EcN cultures were pelleted at 4000 × g for 10 minutes and resuspended in 10 ng/mL of Stx2 in PBS. The bacterial suspensions were then serially diluted tenfold (from 1:10 to 1:10 4 ) in 10 ng/mL Stx2 PBS solution, maintaining a constant Stx2 concentration. The suspensions were incubated at 37°C on a 225 RPM rotating platform for 1 hour and pelleted again. For each condition, 10 μL of supernatant were added to 90 μL of Vero cell medium in its corresponding well. After a 48-hour incubation at 37°C with 5% CO 2 , 10 μL of PrestoBlue Cell Viability Reagent (Thermo Fisher Scientific) was added into each well, followed by a 10-minute incubation and measurement of fluorescent signal at 560nm/590nm. REPEC and REHEC aggregation assays REPEC, REHEC and PBP8 were cultured as previously described. Cultures were pelleted and resuspended in PBS to obtain 3x, 1x, 0.3x or 0.1x suspensions as compared to the original culture density. Cell suspensions were subsequently mixed and added into 96-well conical-bottom microwell plates (Thermo Fisher Scientific) and allowed to settle overnight at room temperature prior to imaging. Generation of anti-IpaD VHH trimer VHH heterotrimer was designed and generated as previously described [ 94 , 95 ]. Briefly, DNA encoding the 20ipaD, JMJ-F5, and JPS-G3 VHHs [ 20 ] separated by 15-amino acid flexible glycine-serine linkers ((GGGGS) 3 ) was synthesized (GenScript Biotech, Piscataway, NJ) and ligated into pET32b(+) vector between the N-terminal thioredoxin (trx) fusion partner and a C-terminal E-tag epitope. VHH trimer was expressed in E . coli Rosetta-gami 2(DE3)pLac1 (Novagen) by overnight incubation in 1 mM IPTG, followed by lysis and purification on nickel agarose resin (Invitrogen). Bound VHHs were eluted from resin using increasing concentrations of imidazole ranging from 10mM to 250mM. Shigella contact-mediated hemolysis assay To determine the ability of CsgA-VHH to inhibit Shigella virulence activity, a contact-mediated hemolysis assay was carried out as previously described [ 65 ], with slight modification. Prior to exposure of S . flexneri to red blood cells, induced PBP8 cultures were pelleted and resuspended in PBS to obtain a 10x concentrated cell suspension, which was subsequently serially diluted to yield 5x, 2.5x, 1.25x and 0.625x suspensions. The pathogen was then incubated for 30 min at room temperature with either the PBP8 suspensions or the abovementioned anti-IpaD VHH heterotrimer (trx/20ipaD/JMJ-F5/JPS-G3/E) as a positive control, in concentrations between 3.2–2000 nM. Preparation of Cryptosporidium lysate Pre-bleached C . parvum oocysts were excysted in 0.75% taurocholic acid suspension in PBS for 1h at 37°C. Following centrifugation (18,000 × g, 2 min), supernatant was collected and the pelleted sample consisting of sporozoites, unexcysted oocysts and oocyst shells was then sonicated (Qsonica CL5, Qsonica Sonicators, USA) with thirty cycles, 20 seconds each. Sonicated pellet was resuspended in supernatant and saved as ' C . parvum whole lysate'. The concentration of the antigen fractions was determined by measurement of optical density using a Nanodrop instrument (ND-1000, NanoDrop Technologies). Pull-down of C . parvum antigens using E . coli Nissle A variety of modified pull-down studies utilizing the principles of ELISA, Western blot and immunofluorescence were applied to test the ability of anti-gp900 VHHs fused to PBP8 curli to bind their targets. All experiments used a nonspecific control PBP8 construct which expressed curli in fusion with a VHH targeting the green fluorescent protein (CsgA-αGFP). For the ELISA, the goal was to pull-down C . parvum antigens by PBP8 immobilized to a plastic surface. Briefly, 100 μl of induced overnight PBP8 cultures expressing CsgA-αgp900-1 and -2 were coated on 96-well MaxiSorb plates at 2x concentration and incubated overnight at 4°C. The following day, plates were washed with TBS-0.1% Tween and blocked with 4% milk-TBS-0.1% Tween solution for 1 h at 37°C. Plates were washed and C . parvum whole lysate was applied in 2-fold dilutions starting with 50 μg/mL concentration, and then incubated for 1 h at 37°C. After washing, specifically bound C . parvum antigen was incubated with a second E-tagged detection VHH that binds to the same C . parvum antigen recognized by the PBP8 displayed VHH, but to a non-competing epitope, at 1 μg/mL for 1 h at 37°C. Plates were then washed and incubated with an anti-E-tag HRP antibody (Bethyl Laboratories) at 1:10,000 for 1 h at 37°C. Plates were washed a final time and OPD was added to each well for 20 minutes. The reaction was stopped with 1 M H2SO4 and absorbance was measured at 490 nm using a microplate reader. For the Western blot, the goal was to quantify depletion of the target in the soluble whole lysate of C . parvum after incubation with PBP8 displaying an anti-gp900 VHH and removal from the solution by centrifugation. For target pull down, 50 μL of the induced (for VHH display) and blocked PBP8 was suspended in PBS at 2x concentration and incubated with 30 μg of C . parvum whole lysate for 1 h with rotation at room temperature. Samples were centrifuged (5000 ×g, 1 min) and supernatant was collected for analysis. Fifteen μL supernatant aliquots of supernatant were diluted with 4xLDS buffer (Novagen) to achieve 1x concentration and denatured at 70°C for 10 minutes. Samples were loaded into the wells of 4–12% Bis-Tris gel (Novex) and electrophoresed in 1x MOPS buffer at 100 V for 10 minutes and then at 200 V for 40 minutes. The gel was transferred on the nitrocellulose membrane using a wet transfer system (395 mA, 4h). Membranes were blocked with 4% milk-TBS 0.1% Tween for 1h and washed with TBS-T before blotting with a second detection VHH (recognizing a non-competing epitope on the target) at 1 μg/ml for 1 h with rotation. Membranes were then washed and incubated with a secondary anti-E-tag HRP antibody at 1:5,000 dilutions for 1 h with rotation. Western blots were developed using chemiluminescent substrate (GE Healthcare) and imaged using a ChemiDoc system (Bio-Rad). A densitometry analysis was performed using Image Lab software to report the percent of target band depletion as normalized to the loading control. Immunofluorescent imaging was used to quantify PBP8 bacteria attached to C . parvum parasite and its trails immobilized on the surface. Pre-bleached C . parvum oocysts were suspended in 0.75% taurocholic acid and excysted in a 37°C water bath for 30 minutes to release sporozoites. Aliquots of excysted 10,000 oocysts were transferred onto poly-L-lysine slides (Chromaview) and incubated for another 30 minutes at 37°C under humidified conditions to allow for further excystation and gliding of sporozoites. Slides were then dried, fixed with 4% paraformaldehyde at room temperature (20 min) and washed with PBS. Such prepared parasites were then probed with 200 μL of 2x PBP8 suspensions and incubated for 1h at room temperature, after which they were washed with PBS to remove unbound PBP8. To detect PBP8 bound to the sporozoites and trails, slides were probed with anti-LPS Mab (ThermoFisher Scientific) at 1:200 dilution, followed by an anti-mouse IgG Alexa Fluor 568 antibody (Invitrogen) at 1:500 dilution, both incubated for 1h at room temperature. Sporozoites were counterstained with an E-tagged VHH targeting gp900 at the apical complex and trails (CsgA-αgp900-2) at 1 μg/mL concentration, followed by an anti-E-tag-FITC antibody (Bethyl Laboratories) at 1:100 dilution, both incubated for 1h at room temperature. Lastly, slides were washed, dried, and mounted with antifade medium. Fluorescing sporozoites were imaged under epifluorescence (Nikon Eclipse Ti-E microscope, Nikon Instruments Inc.). The number of fluorescent foci was quantified using ImageJ 1.48v particle analyzer (U.S. National Institutes of Health, Bethesda, Maryland, USA). Statistical analysis All statistical analyses were performed using Prism 9.1.0 (GraphPad Software). Data are presented as mean ± standard error of mean (SEM), unless otherwise specified. Statistical significance was assessed using one-way or two-way analysis of variance (ANOVA), followed by Welsh's t-test, as described in figure legends. Ethics statements All animal experiments were approved by the Tufts University Institutional Animal Care Use Committee in accordance with the Guide for the Care and Use of Laboratory Animals of the National Research Council. Cryptosporidium parasites used in this study were generated in animals in compliance with protocols No. G2017-107 and No. G2017-120. The VHH-display phage library was generated and derived from an alpaca in accordance with the protocol No. G2019-142. Cell strains and plasmids All strains and plasmids used in this study are summarized in S1 Table . Bacterial culture All E . coli and C . rodentium strains were cultured in LB broth at 37°C at 225 RPM, unless otherwise specified. EcN (PBP8) strains were streaked from frozen stock onto selective lysogeny broth (LB) agar plates and grown overnight at 37°C. Cultures were subsequently started from single colonies into 5 mL LB supplemented with 50 μg/mL kanamycin and grown overnight at 37°C with shaking at 225 RPM. The following day, overnight cultures were diluted 1:100 into 10 mL fresh media and grown at 37°C and 225 RPM, and protein expression was induced by adding 100 μM isopropyl β-D-1-thiogalactopyranoside (IPTG). Induced cultures were grown overnight. Shigella flexneri was streaked from frozen stock onto tryptic soy agar (TSA) plates supplemented with 0.02% Congo Red (CR) and grown overnight at 37°C. The following day, 3 colonies were used to inoculate 50 mL of tryptic soy broth (TSB) in a baffled flask. Only colonies stained red by CR were used. The culture was grown at 37°C and 225 RPM to OD at 600 nm of 0.8–1.0, placed on ice upon reaching the desired OD, pelleted at 3500 RPM for 10 minutes at 4°C, and resuspended in 5 mL to obtain a 10x suspension. Mammalian cell culture HeLa cells (ATCC CCL-2) were maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS) and 1% Penicillin-Streptomycin antibiotics in a 5% CO2 incubator at 37°C. For infection, 30,000 cells were seeded on 24-well plates in a volume of 0.5 mL/well. The next day, cells were gently washed with PBS before inoculating with bacteria. Vero cells (ATCC CCL-81) were grown in Eagle's Minimum Essential Medium (EMEM) supplemented with 10% FBS in a 5% CO2 incubator at 37°C. For Stx2 toxicity assays, Vero cells were seeded on 96-well plates one day prior to incubation with toxin. Parasite propagation C . parvum oocysts, MD isolate originating from deer and passaged repeatedly in sheep and mice [ 86 ], were generated at Tufts University by propagation in CD-1 mice as described elsewhere [ 87 ], in compliance with study protocol No. G2017-107 approved by the Tufts University Institutional Animal Care Use Committee. Prior to excystation, oocysts were bleached on ice for 7 minutes using 5% dilution of commercial bleach (Clorox Original, The Clorox Company, CA). To remove bleach, oocysts were washed three times by suspension in PBS and centrifugation (18,000 × g, 2 min). Purification of flagella Flagella from REPEC (E22), EPEC (E2348/69), and E10 (O119:H6) were isolated as described previously [ 88 ], with slight modification. Briefly, a single colony was transferred into 5 mL LB broth and incubated overnight at 37°C with continuous shaking. The next day, the culture was diluted 1:100 into LB broth and grown at 37°C to OD 600 of 0.5. 100 μL of the culture was plated onto the surface of eighty 100mm diameter LB agar plates and incubated for 24h at 37°C. 500 μL of PBS was then added to each plate and a glass slide was used to gently scrape bacteria from the agar plate. Bacteria were collected in a centrifuge bottle. To shear flagella from the bacteria, the centrifuge bottle was manually shaken for 2 min and then shaken for 5 mins at 4°C at 220 RPM. The bottle was then centrifuged at 7025 × g for 20 min at 4°C to remove cell debris. Bacteria-free supernatant was transferred to a new centrifuge bottle, which was further centrifuged at 25,402 × g for 1 hour at 4°C to precipitate flagella. To recover flagella, the supernatant was removed, and the pellet was resuspended in 500 μL of ice-cold PBS. To confirm that the purified flagella encompassed flagellin monomers of 60 kDa, flagella were visualized by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and by Western blotting using Rabbit anti-H6 flagella antibody. Note: EPEC E2348/69 produces fewer flagella filaments when grown in LB media [ 52 ]. Therefore, to maximize shearing of flagella from E2348/69, bacteria were either passed through a syringe and a 22-gauge needle or heat treated at 65°C for 30 mins. Alpaca immunizations Immunizations were performed essentially as described by Vrentas et al . [ 89 ]. Two different pairs of alpacas were each immunized in two separate rounds of immunization with various combinations of purified REPEC or EPEC flagella, and/or recombinant proteins MBP/EHEC intimin, or 6xHis/EHEC Tir. For each round of immunization, five successive multi-site subcutaneous injections were employed at about 3-week intervals. Blood was obtained for lymphocyte preparation 3–5 days after the fifth immunization and RNA was prepared from lymphocytes using the RNeasy kit (Qiagen, Valencia, CA). A VHH-display phage library was prepared essentially as described previously [ 73 ] following each of the rounds of alpaca immunization, yielding libraries with complexities of about 1–2 x 10 7 independent clones, and >95% containing VHH inserts. Identification and purification of VHHs Phage library panning methods have been previously described [ 90 ]. Typically, the virulence factor proteins were coated onto plastic at 10 μg/mL of target in the first panning round, followed by a second round of panning at high stringency, with virulence factor proteins coated at 1 μg/mL, and using a 10-fold lower titer of input phage, shorter binding times, and longer washes. In some cases, the virulence factor targets were captured onto plastic by previously coated VHHs which bind the target or its fusion partner. VHH capture panning was also used in some cases to block isolation of VHHs to immunodominant epitopes on monomeric targets. Random clones from the selected populations were then screened by ELISA for expression of VHHs that bound to the virulence factor targets. Clones producing the strongest signals or showing broader target specificity were characterized by DNA fingerprinting. The coding sequences of VHHs selected as having unique fingerprints and the strongest ELISA signals were obtained. Based on sequence homology, one VHH representing each homology group (having no evidence of a common B cell clonal origin) was selected for expression and characterization. These VHHs were expressed individually in pET32 vectors and purified as recombinant E . coli thioredoxin fusions with a carboxy-terminal E-tag, as previously described [ 90 ]. Dilution ELISA ELISAs were performed using Nunc Maxisorp 96 well plates (Thermo Fisher Scientific). Virulence factor targets were typically coated overnight at 4°C, 1 μg/mL in PBS, then blocked for at least an hour at 37°C with 4% milk in PBS, 0.1% Tween. For capture ELISAs, plates were first coated with 5 μg/mL of VHHs that recognized the virulence factor or its fusion partner. The captured VHHs lacked both the thioredoxin partner and E-tag. After blocking, the virulence factor was then incubated at 1 μg/mL for one hour at 37°C with 4% milk in PBS, 0.1% Tween and washed. Dilution ELISAs were then initiated by diluting the VHH (expressed in a pET-32 vector with an amino terminal thioredoxin and a carboxyl terminal E-tag) to 125 nM and performing serial dilutions of 1:5. After incubation for one hour at 37°C, plates were washed and then incubated with 1:10,000 rabbit HRP/anti-E-tag (Bethyl Laboratories) for one hour, washed, developed with TMB (Sigma Aldrich) as recommended by the manufacturer and measured for absorbance at 450 nm. ELISA measuring the effect of anti-Tir VHHs on the intimin-Tir interaction The ability of anti-Tir VHHs derived from EHEC to interfere with intimin-Tir binding was measured by ELISA. High-binding assay plates (Corning) were coated with 5 μg/mL of recombinant his-tagged Tir diluted in 1x coating buffer (50 mM Na2CO3, 50 mM NaHCO3, pH 9.6) in a volume of 100 μL per well and incubated overnight at 4°C. Plates were washed three times with 300 μL of wash buffer (0.05% Tween in PBS) and then blocked with BSA (3% in PBS) for 2 hours at room temperature (RT). Plates were washed and 100 μL of 500 nM anti-Tir VHH were added to each well. 0.1% BSA was used as a negative control. Plates were incubated at RT for 2 hours or at 4°C overnight. Wells were then probed with 150 nM GST-tagged intimin or with GST alone, and incubated at RT for 2 hours or at 4°C overnight. Plates were washed again and then fixed with 3.7% paraformaldehyde at RT for 20 mins at 4°C. Following another wash, plates were blocked with 5% milk in PBS for 30 min at RT. After washing, plates were incubated with goat anti-GST (GE Healthcare) for an hour and GST binding was detected kinetically using an alkaline-phosphatase-linked rabbit anti-goat IgG secondary antibody (diluted 1:2000 in 0.1% BSA/PBS). Binding of the secondary antibody was detected colorimetrically (AP substrate N1891, Sigma Aldrich) at 405 nm, and the average reaction rate (V mean ) was calculated. REPEC motility assay Motility assays were performed as described previously [ 73 ], with slight modification, to measure the ability of anti-Fla VHHs to inhibit Rabbit Enteropathogenic Escherichia coli (REPEC) motility. Briefly, REPEC cultures were streaked on LB agar plates and incubated for 16 hours at 37°C. The next day, a single colony was transferred into 5 mL LB broth and incubated overnight at 37°C with continuous shaking. On the following day, the culture was diluted 1:50 into LB broth and grown at 37°C with continuous shaking to an OD 600 of 0.5. A 1:1 dilution of bacteria and VHHs was then prepared (6 μL of bacterial culture was mixed with 6 μL of VHH concentrations ranging from 0 to 6.4 μM), mixed gently with a pipette, and incubated at 4°C for 2 h. The 12 μL mixture was then transferred to the center surface of a 0.3% semi solid agar plate and incubated for 24 h at room temperature. The diameter of bacterial growth was measured by first placing the plate on a dark background to enhance the contrast between bacterial growth and the agar medium. A metric scale ruler was then used to measure the growth diameter. Images were captured using a Syngene imager. EPEC pedestal assay The ability of anti-intimin and anti-Tir VHHs to inhibit EPEC pedestal formation was assessed after infection of HeLa cells, as described previously [ 91 ], with slight modification. Briefly, 30,000 HeLa cells were inoculated into the wells of 24-well plates (Invitro Scientific,) and incubated overnight at 37°C in an incubator with 5% CO2. On the same day, a single EPEC colony was inoculated into 5 mL DMEM in 100mM HEPES medium (pH 7.4) and incubated overnight in 5% CO2 at 37°C without shaking. The next day, the EPEC culture was diluted 1:16 into new infection medium (0.6 μL EPEC added to 9.4 μL media containing DMEM, 20 mM HEPES, and 3.5% FBS; pH 7.4), and 3.33 μL of the EPEC suspension were incubated either alone or with 100 nM anti-Tir or anti-intimin VHH in 0.5 mL DMEM at 4°C for 2 hours, in 1.5 mL Eppendorf tubes on a rocker. HeLa cell monolayers were then washed with 0.5 mL PBS, and EPEC suspensions were added to the monolayers. Plates were then centrifuged at 500 RPM for 5 min and incubated for 3 h at 37°C in a 5% CO2 incubator. Next, cells were washed twice with PBS, fixed with 0.5 mL 2.5% paraformaldehyde in PBS for 10 mins at RT on a shaker, washed twice with PBS for 5 min on a shaker at RT, permeabilized with 0.5 mL of 0.1% TritonX-100 for 5 min, and washed twice again before staining with DAPI (Thermo Fisher Scientific) and Alexa Fluor-488 Phalloidin (Thermo Fisher Scientific) at RT for 1.5 hours. Monolayers were then washed and 7 μL prolong gold anti-fade reagent (Thermo Fisher Scientific) was used to mount coverslips on wells before imaging with a fluorescent microscope. EPEC pedestal formation was blindly scored, as follows; 1: Very few pedestals are present on the edges of the wells, 2: More pedestals present, only at the edges of the well, 3: Most cells have no pedestals, but a few pedestals present in the center and edges of wells, 4: Most cells have pedestals, but a few empty cells are present, 5: The majority of the cells have pedestals. Using the above numbering criteria, pedestals were scored blindly by a second researcher. CsgA-VHH plasmid construction and cloning The cloning of the synthetic curli operon csgBACEFG onto the pL6FO vector was described in detail elsewhere [ 55 ]. DNA sequences of desired VHHs and corresponding primers were synthesized by and purchased from Integrated DNA Technologies. Plasmid construction was carried out using Gibson Assembly [ 92 ]. Quantitative Congo Red binding assay Curli fiber formation was quantified using a Congo Red binding assay as previously described [ 43 ]. Briefly, 1 mL of induced EcN CsgA-VHH culture was pelleted at 4000 × g for 10 minutes at room temperature and resuspended in a 25 μM Congo Red PBS solution. After a 10-minute incubation, the cell suspension was pelleted again, and the unbound Congo Red dye was quantified by measuring the supernatant absorbance at 490 nm. The signal was subtracted from a Congo Red blank, divided by the culture's OD 600 measurement to reflect curli production per cell, and normalized with respect to a EcN CsgA (no VHH) positive control. Electron microscopy Field emission scanning electron microscope (FESEM) samples were prepared by fixing with 2% (w/v) glutaraldehyde and 2% (w/v) paraformaldehyde at room temperature, overnight. The samples were gently washed with water, and the solvent was gradually exchanged with ethanol with an increasing ethanol 15-minute incubation step gradient (25, 50, 75 and 100% (v/v) ethanol). The samples were dried in a critical point dryer, placed onto SEM sample holders using silver adhesive (Electron Microscopy Sciences) and sputtered until they were coated in a 10–20 nm layer of Pt/Pd. Images were acquired using a Zeiss Ultra55 FESEM equipped with a field emission gun operating at 5–10 kV. GFP pull-down assay For each condition, 1 mL of induced overnight culture was centrifuged at 4000 × g for 10 minutes. The supernatant was aspirated, and the pellets were resuspended in 150 nM GFP in fasted-state simulated colonic fluid, which was prepared as described by Vertzoni et al . [ 93 ]. The cells were incubated on a shaking platform (225 RPM) at 37°C for 15 minutes and pelleted again. The GFP remaining in solution was assayed by measuring the fluorescent signal (485nm/528nm) using a plate reader (Spectramax M5, Molecular Devices). GFP concentration was estimated based on a calibration curve using known GFP concentrations. Shiga toxin pull-down assay Induced EcN cultures were pelleted at 4000 × g for 10 minutes and resuspended in 10 ng/mL of Stx2 in PBS. The bacterial suspensions were then serially diluted tenfold (from 1:10 to 1:10 4 ) in 10 ng/mL Stx2 PBS solution, maintaining a constant Stx2 concentration. The suspensions were incubated at 37°C on a 225 RPM rotating platform for 1 hour and pelleted again. For each condition, 10 μL of supernatant were added to 90 μL of Vero cell medium in its corresponding well. After a 48-hour incubation at 37°C with 5% CO 2 , 10 μL of PrestoBlue Cell Viability Reagent (Thermo Fisher Scientific) was added into each well, followed by a 10-minute incubation and measurement of fluorescent signal at 560nm/590nm. REPEC and REHEC aggregation assays REPEC, REHEC and PBP8 were cultured as previously described. Cultures were pelleted and resuspended in PBS to obtain 3x, 1x, 0.3x or 0.1x suspensions as compared to the original culture density. Cell suspensions were subsequently mixed and added into 96-well conical-bottom microwell plates (Thermo Fisher Scientific) and allowed to settle overnight at room temperature prior to imaging. Generation of anti-IpaD VHH trimer VHH heterotrimer was designed and generated as previously described [ 94 , 95 ]. Briefly, DNA encoding the 20ipaD, JMJ-F5, and JPS-G3 VHHs [ 20 ] separated by 15-amino acid flexible glycine-serine linkers ((GGGGS) 3 ) was synthesized (GenScript Biotech, Piscataway, NJ) and ligated into pET32b(+) vector between the N-terminal thioredoxin (trx) fusion partner and a C-terminal E-tag epitope. VHH trimer was expressed in E . coli Rosetta-gami 2(DE3)pLac1 (Novagen) by overnight incubation in 1 mM IPTG, followed by lysis and purification on nickel agarose resin (Invitrogen). Bound VHHs were eluted from resin using increasing concentrations of imidazole ranging from 10mM to 250mM. Shigella contact-mediated hemolysis assay To determine the ability of CsgA-VHH to inhibit Shigella virulence activity, a contact-mediated hemolysis assay was carried out as previously described [ 65 ], with slight modification. Prior to exposure of S . flexneri to red blood cells, induced PBP8 cultures were pelleted and resuspended in PBS to obtain a 10x concentrated cell suspension, which was subsequently serially diluted to yield 5x, 2.5x, 1.25x and 0.625x suspensions. The pathogen was then incubated for 30 min at room temperature with either the PBP8 suspensions or the abovementioned anti-IpaD VHH heterotrimer (trx/20ipaD/JMJ-F5/JPS-G3/E) as a positive control, in concentrations between 3.2–2000 nM. Preparation of Cryptosporidium lysate Pre-bleached C . parvum oocysts were excysted in 0.75% taurocholic acid suspension in PBS for 1h at 37°C. Following centrifugation (18,000 × g, 2 min), supernatant was collected and the pelleted sample consisting of sporozoites, unexcysted oocysts and oocyst shells was then sonicated (Qsonica CL5, Qsonica Sonicators, USA) with thirty cycles, 20 seconds each. Sonicated pellet was resuspended in supernatant and saved as ' C . parvum whole lysate'. The concentration of the antigen fractions was determined by measurement of optical density using a Nanodrop instrument (ND-1000, NanoDrop Technologies). Pull-down of C . parvum antigens using E . coli Nissle A variety of modified pull-down studies utilizing the principles of ELISA, Western blot and immunofluorescence were applied to test the ability of anti-gp900 VHHs fused to PBP8 curli to bind their targets. All experiments used a nonspecific control PBP8 construct which expressed curli in fusion with a VHH targeting the green fluorescent protein (CsgA-αGFP). For the ELISA, the goal was to pull-down C . parvum antigens by PBP8 immobilized to a plastic surface. Briefly, 100 μl of induced overnight PBP8 cultures expressing CsgA-αgp900-1 and -2 were coated on 96-well MaxiSorb plates at 2x concentration and incubated overnight at 4°C. The following day, plates were washed with TBS-0.1% Tween and blocked with 4% milk-TBS-0.1% Tween solution for 1 h at 37°C. Plates were washed and C . parvum whole lysate was applied in 2-fold dilutions starting with 50 μg/mL concentration, and then incubated for 1 h at 37°C. After washing, specifically bound C . parvum antigen was incubated with a second E-tagged detection VHH that binds to the same C . parvum antigen recognized by the PBP8 displayed VHH, but to a non-competing epitope, at 1 μg/mL for 1 h at 37°C. Plates were then washed and incubated with an anti-E-tag HRP antibody (Bethyl Laboratories) at 1:10,000 for 1 h at 37°C. Plates were washed a final time and OPD was added to each well for 20 minutes. The reaction was stopped with 1 M H2SO4 and absorbance was measured at 490 nm using a microplate reader. For the Western blot, the goal was to quantify depletion of the target in the soluble whole lysate of C . parvum after incubation with PBP8 displaying an anti-gp900 VHH and removal from the solution by centrifugation. For target pull down, 50 μL of the induced (for VHH display) and blocked PBP8 was suspended in PBS at 2x concentration and incubated with 30 μg of C . parvum whole lysate for 1 h with rotation at room temperature. Samples were centrifuged (5000 ×g, 1 min) and supernatant was collected for analysis. Fifteen μL supernatant aliquots of supernatant were diluted with 4xLDS buffer (Novagen) to achieve 1x concentration and denatured at 70°C for 10 minutes. Samples were loaded into the wells of 4–12% Bis-Tris gel (Novex) and electrophoresed in 1x MOPS buffer at 100 V for 10 minutes and then at 200 V for 40 minutes. The gel was transferred on the nitrocellulose membrane using a wet transfer system (395 mA, 4h). Membranes were blocked with 4% milk-TBS 0.1% Tween for 1h and washed with TBS-T before blotting with a second detection VHH (recognizing a non-competing epitope on the target) at 1 μg/ml for 1 h with rotation. Membranes were then washed and incubated with a secondary anti-E-tag HRP antibody at 1:5,000 dilutions for 1 h with rotation. Western blots were developed using chemiluminescent substrate (GE Healthcare) and imaged using a ChemiDoc system (Bio-Rad). A densitometry analysis was performed using Image Lab software to report the percent of target band depletion as normalized to the loading control. Immunofluorescent imaging was used to quantify PBP8 bacteria attached to C . parvum parasite and its trails immobilized on the surface. Pre-bleached C . parvum oocysts were suspended in 0.75% taurocholic acid and excysted in a 37°C water bath for 30 minutes to release sporozoites. Aliquots of excysted 10,000 oocysts were transferred onto poly-L-lysine slides (Chromaview) and incubated for another 30 minutes at 37°C under humidified conditions to allow for further excystation and gliding of sporozoites. Slides were then dried, fixed with 4% paraformaldehyde at room temperature (20 min) and washed with PBS. Such prepared parasites were then probed with 200 μL of 2x PBP8 suspensions and incubated for 1h at room temperature, after which they were washed with PBS to remove unbound PBP8. To detect PBP8 bound to the sporozoites and trails, slides were probed with anti-LPS Mab (ThermoFisher Scientific) at 1:200 dilution, followed by an anti-mouse IgG Alexa Fluor 568 antibody (Invitrogen) at 1:500 dilution, both incubated for 1h at room temperature. Sporozoites were counterstained with an E-tagged VHH targeting gp900 at the apical complex and trails (CsgA-αgp900-2) at 1 μg/mL concentration, followed by an anti-E-tag-FITC antibody (Bethyl Laboratories) at 1:100 dilution, both incubated for 1h at room temperature. Lastly, slides were washed, dried, and mounted with antifade medium. Fluorescing sporozoites were imaged under epifluorescence (Nikon Eclipse Ti-E microscope, Nikon Instruments Inc.). The number of fluorescent foci was quantified using ImageJ 1.48v particle analyzer (U.S. National Institutes of Health, Bethesda, Maryland, USA). Statistical analysis All statistical analyses were performed using Prism 9.1.0 (GraphPad Software). Data are presented as mean ± standard error of mean (SEM), unless otherwise specified. Statistical significance was assessed using one-way or two-way analysis of variance (ANOVA), followed by Welsh's t-test, as described in figure legends. Supporting information S1 Table Strains and plasmids. (DOCX) Click here for additional data file. S2 Table Aligned sequences of VHHs binding pathogenic E . coli virulence factors. Complementarity determining regions (CDR1, 2 and 3) are highlighted and appear in order from left to right. (PDF) Click here for additional data file. S1 Fig Assessing the binding properties of selected VHHs via ELISA. Antigens used were homologues of Fla (a-b), Tir (c-g), Int (h) or EspA (i-l) corresponding to either EPEC (b, c, h), EHEC (e, g), REPEC (a, d, i, j) or C . rodentium (f, k, l). In each assay, antigen was either directly added to the plate in purified form (a-b, e-g, i-l), bound to the plate by an adsorbed noncompeting VHH (c-d), or displayed on the surface of MC1061 (h). (TIFF) Click here for additional data file. S2 Fig Anti-EspA VHHs did not inhibit pedestal formation. HeLa cells were exposed to EPEC incubated with VHH, fixed and stained with DAPI (blue) and Alexa Fluor-488 Phalloidin (green). Similar to the "no VHH" negative control (a, c, e), all anti-EspA VHHs tested (b, d, f) resulted in the formation of pedestals (though not all anti-EspA VHHs were tested) (scale bar = 100 μm). (TIFF) Click here for additional data file. S3 Fig Representative FESEM images of PBP8 expressing CsgA and CsgA-VHH. (a) PBP8 with no plasmid, expressing no curli fibers. (b-h) PBP8 expressing CsgA-VHH, exhibiting a range of fiber morphologies. (b) CsgA-αStx2, (c) CsgA-αInt-12, (d) CsgA-αInt-17, (e) CsgA-αFla-3, (f) CsgA-αFla-4, (g) CsgA-αIpaD-1, (h) CsgA-αgp900-2 (scale bar = 2 μm). (TIFF) Click here for additional data file. S4 Fig PBP8 expressing CsgA-ɑFla can induce REPEC aggregation. Suspensions of PBP8 expressing CsgA-VHH were mixed with either REPEC (a) or REHEC (b) and allowed to settle overnight in conical 96-well plates. Aggregation was only observed when REPEC was mixed with CsgA-αFla-3 and -4. (TIFF) Click here for additional data file. S5 Fig PBP8 expressing CsgA-ɑIpaD can bind soluble IpaD. ELISA demonstrated the ability of CsgA-αIpaD to bind soluble IpaD. PBP8 was adsorbed onto a well plate, followed by incubation with varying IpaD concentrations. Binding of IpaD to the adsorbed PBP8 was then detected by a specific non-competing VHH (JMK-H2, Barta et , al ., 2017 [ 20 ]), followed by an anti-Etag IgG-HRP conjugate. PBP8 expressing either CsgA-αIpaD-1 or CsgA-αIpaD-2 significantly outperformed the off-target negative control (CsgA-αGFP). Data presented as mean ± SD. Two-way ANOVA (P < 0.0001) was performed to test the presence of difference between conditions, P-values calculated by Welch's t-test. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001. (TIFF) Click here for additional data file. S6 Fig PBP8 expressing CsgA-ɑgp900 exhibit increased attachment to and colocalization with C . parvum sporozoites. (a) Fluorescent micrographs demonstrating increased attachment of PBP8 (red) to C . parvum sporozoites, counterstained in green in the bottom panels (scale bar = 50 μm). (b) While PBP8 (red) expressing CsgA-ɑgp900-1 (and to a lesser extent CsgA-ɑgp900-2) consistently colocalize with sporozoites (green), the CsgA-ɑGFP negative control was often observed away from the green fluorescent foci, consistent with nonspecific binding (scale bar = 5 μm). (TIFF) Click here for additional data file. S7 Fig Curli fiber production in growing bacterial cultures as measured by Congo Red fluorescence. A shift in emission maximum for Congo Red upon binding to curli fibers in growing PBP8 cultures was used to help confirm fiber formation for two CsgA-VHH fusions. "No curli" refers to a negative control wherein PBP8 was transformed with a plasmid bearing the same antibiotic selection markers but no curli genes. "LB" refers to a negative control containing no cells, only LB medium and Congo Red dye. (PDF) Click here for additional data file. S8 Fig WAXS of dried films composed of CsgA and CsgA-VHH fusions. (a) CsgA-6xHis; (b) CsgA-αFla-3; (c) CsgA-αgp900-2. (PDF) Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5221479/

In vitro exposure system for study of aerosolized influenza virus

Infection of adherent cell monolayers using a liquid inoculum represents an established method to reliably and quantitatively study virus infection, but poorly recapitulates the exposure and infection of cells in the respiratory tract that occurs during infection with aerosolized pathogens. To better simulate natural infection in vitro, we adapted a system that generates viral aerosols similar to those exhaled by infected humans to the inoculation of epithelial cell monolayers. Procedures for cellular infection and calculation of exposure dose were developed and tested using viruses characterized by distinct transmission and pathogenicity phenotypes: an HPAI H5N1, an LPAI H7N9, and a seasonal H3N2 virus. While all three aerosolized viruses were highly infectious in a human bronchial epithelial cell line (Calu-3) cultured submerged in media, differences between the viruses were observed in primary human alveolar epithelial cells and in Calu-3 cells cultured at air-liquid interface. This system provides a novel enhancement to traditional in vitro experiments, particularly those focused on the early stages of infection. 1 Introduction Infection of adherent cell monolayers using a liquid inoculum represents an established method to reliably and quantitatively study virus infection. Relatively straightforward and inexpensive, this method allows for the frequent collection of viral samples and the testing of a variety of experimental conditions and discrete cell types including those of human origin. Unfortunately, traditional in vitro replication studies poorly recapitulate the exposure and infection of cells in the respiratory tract that occurs during natural exposure to aerosolized pathogens. Not only does infection occur while cells' apical surface is immersed in liquid, but at typical cell densities, the often-used "low" multiplicity of infection (MOI) of 0.01 corresponds to a dose of over a thousand PFU per square centimeter. Available evidence suggests that in the case of aerosol transmission, natural human influenza infection is likely initiated by substantially fewer particles. Studies of infected patients found low viral concentrations in aerosols generated by breathing, coughing, and/or sneezing ( Fabian et al., 2008 , Milton et al., 2013 , Yang et al., 2011 ), and fewer than five TCID 50 are capable of initiating symptomatic infection in experimentally exposed volunteers ( Alford et al., 1966 ). Similar results have been observed in the ferret model; these animals can be infected with fewer than ten PFU and subsequently exhale under five PFU per minute ( Gustin et al., 2015 , Gustin et al., 2011 , Gustin et al., 2013 , Roberts et al., 2011 ). Using a library of barcoded viruses, Varble et al. found that respiratory droplet transmission between ferrets involved only single-digit numbers of virions ( Varble et al., 2014 ). Reports of A(H7N9) cases developing subsequent to patient visits to live bird markets despite lack of poultry contact, and the detection of virus in air sampled from such markets, indicate that zoonotic infection may also occur after human exposure to low quantities of aerosolized virus ( Li et al., 2015 , Liu et al., 2014 , Zhou et al., 2016 ). In order to better study the effects of potentially damaging aerosols on human cells, the toxicology field has begun to expose cultured respiratory epithelial cells to aerosolized, rather than liquid-suspended, chemical and particulate matter. Cells have been shown to be more sensitive to the effects of the former ( Bitterle et al., 2006 , Raemy et al., 2012 ). In these studies, aerosol concentration can be measured by the use of optical or gravimetric methods. These types of methods are not effective for the measure of virus-containing aerosols, however, because they detect liquid droplet nuclei rather than the virus within them, and cannot differentiate between infectious and non-infectious virions. Microbiologists have developed aerosolization systems to overcome these challenges, and have used them for experimental infections of animals and to study the effect of environmental conditions on viability of numerous pathogens including Mycobacterium tuberculosis , Bacillus anthracis , measles virus, and influenza virus ( Clark et al., 2011 , Gustin et al., 2011 , Lemon et al., 2011 , Savransky et al., 2013 ). This work has provided important insights into the intra- and inter-host spread of these pathogens by facilitating the observation and manipulation of near-natural infection within a controlled laboratory environment. However, despite the frequent employment of in vitro studies to complement animal experimentation, use of an aerosol system for in vitro infection with any pathogen has not, to our knowledge, been previously described. We combined aspects of the toxicological and microbiological approaches to establish a novel method to expose adherent mammalian cell monolayers in air-liquid interface to defined quantities of aerosolized influenza virus and compared this with traditional liquid inoculation. In order to most effectively mimic the conditions of natural infection, we explored the use of very low MOI infection and culture techniques designed to promote cell differentiation in conjunction with virus aerosolization. Using highly pathogenic avian influenza (HPAI), low pathogenic avian influenza (LPAI), and seasonal influenza viruses, we demonstrate that infection of respiratory epithelial cells with physiologically low concentrations of aerosolized virus can be successfully recreated inside the laboratory. In conjunction with research using animal models, these techniques facilitate a closer study of the infectivity of aerosolized influenza virus in the context of human infection. The approach described here is not restricted to influenza virus and would also be applicable to the study of other respiratory viruses of public health concern. 2 Materials and methods 2.1 Viruses Influenza A viruses were propagated in the allantoic cavity of 10-day-old embryonated hens' eggs and titered via standard plaque assay using Madin-Darby canine kidney (MDCK) cells as previously described ( Maines et al., 2005 , Zeng et al., 2007 ). All experiments were conducted under biosafety level 3 containment, including enhancements as required by the U.S. Department of Agriculture and the Federal Select Agent Program ( Chosewood et al., 2009 ). 2.2 Cell culture and liquid inoculations The bronchial epithelial cell line Calu-3 (ATCC) was cultured as previously described ( Zeng et al., 2007 ). Primary human alveolar epithelial cells (Cell Biologics) were cryopreserved at passage 3, then grown and expanded per manufacturer's instructions. All cells were seeded on 24 mm diameter (6-well format) or 12 mm diameter (12-well format) semipermeable membrane inserts with a 0.4 µm pore size (Corning) and grown to confluence under submerged conditions. After reaching a transepithelial resistance of >1000 Ω 2 ( Zeng et al., 2007 ), apical media was removed from selected Calu-3 cells to create an air-liquid interface (ALI), which was maintained for three weeks to facilitate cell differentiation and the establishment of a mucin layer. Prior to inoculation, apical media (if present) was removed from the cell monolayer and cells cultured under submerged conditions were washed to remove serum present in culture medium. Liquid inoculation was performed using 300 μL of virus, diluted as specified in the results, and incubated on the cell surface for one hour before washing. After infection, cells were cultured in cell type-specific serum-free medium to which 1 μg/mL N-p-tosyl-ʟ-phenylalanine chloromethyl ketone (TPCK)-treated trypsin (Sigma-Aldrich) was added for alveolar cell cultures. Aliquots of apical culture supernatant or wash media (incubated atop cells cultured at ALI for 20 min) were collected at the indicated times post-infection (p.i.) and immediately frozen at −80 °C until titration. Growth curves were generated and analyzed using Prism 6.0.7 (GraphPad Software Inc). 2.3 Aerosol inoculations The automated bioaerosol system used for all experiments has been previously described in detail ( Gustin et al., 2011 , Hartings and Roy, 2004 ) and all conditions were maintained here unless specified otherwise. Briefly, virus suspended in a solution of PBS-0.03% (w/v) BSA was aerosolized using a three-jet Collision nebulizer (BGI, Inc.) and passed through an exposure chamber at a rate of 20 L/min. Cells placed inside the exposure chamber on a wire mesh shelf were inoculated under air-liquid interface conditions ( Fig. 1 ). Using the AeroMP (Biaera Technologies) aerosol management platform, aerosol exposures were conducted at 21 °C and 50% relative humidity for 15 min followed by a 5 min purge to allow evacuation of the aerosolized virus from the chamber ( Gustin et al., 2011 ). Prior to in vitro exposure, spray factor (SF) values were determined for stocks of all viruses to estimate the starting virus concentration in the nebulizer needed to obtain a desired concentration of virus in the aerosol. A Biosampler (SKC Inc) was used to quantify the virus actually aerosolized within the chamber during each exposure. Following aerosol exposure, membrane inserts were transferred to clean 6-well or 12-well plates and serum-free medium was added to apical and basolateral compartments as described above; cells maintained under ALI conditions had medium added to the basolateral compartment only. Confirmation of cell viability during aerosol exposure was performed using the WST-1 cell proliferation reagent (Roche Applied Science), according to manufacturer's instructions, with four independent samples tested for each condition. Fig. 1 Graphic representation of aerosol system for in vitro use. Depiction of human cells cultured on transwell inserts and exposed to aerosolized influenza virus using a previously characterized system ( Gustin et al., 2011 ). Cell culture dishes rest in the exposure chamber on a wire shelf under air-liquid interface conditions for the duration of the exposure. Inset, individual transwell inserts are transferred to sterile plates once removed from the exposure chamber. Fig. 1. 2.4 Quantitation of Exposure and Infectious Doses The total number of infectious virions passed through the chamber during the exposure session ( N cham ) was calculated as N cham = ( C samp ) ( V samp ) ( Q cham ) Q samp where C samp is the concentration of virus in the sampler, V samp is the volume of media in the sampler, and Q cham and Q samp represent the flow rates of chamber and sampler, respectively (see Supplemental methods for derivation). N cham was multiplied by the ratio of the surface area of each transwell ( SA ) to the cross-sectional area of the chamber ( XA ) to yield exposure dose ( ED ). ID 50 s were calculated according to the method of Reed and Muench, with the proportional distance multiplied by log 10 (dose above 50%)/log 10 (dose below 50%) to account for the deviation of exposure doses from exact 10-fold dilutions ( Reed and Muench, 1938 ). Variability around the mean exposure dose was quantified using a binomial model with n viable virions passing through the chamber and a probability of success (virion lands on well) equal to SA/XA . Upper and lower critical values at the 95% confidence level were calculated using R 3.2.3 (R Foundation for Statistical Computing); the true exposure dose for any particular well has a 95% chance of falling between these two values. The cumulative probability distribution indicated that the minimum dose had to be three or greater for 95% of all wells to be inoculated with at least one virion. In order minimize the chance of including a well not exposed to any virus, we therefore aimed not to use mean exposure doses under 5 PFU. When this did occur due to variations in aerosolization efficiency, the minimum 50% infectious dose is reported as ≤3 PFU. 2.5 Real-time RT-PCR Total RNA was extracted from mock-infected or virus-infected cell monolayers after removal of supernatant using the RNeasy mini kit (Qiagen). RT-PCR was performed with a QuantiTect SYBR green RT-PCR kit (Qiagen) in duplicate reactions from duplicate samples using an influenza A virus M1 gene primer set ( Zeng et al., 2007 ). Influenza virus M gene RNA copy numbers were extrapolated using a standard curve based on samples of known M gene copy number. 2.1 Viruses Influenza A viruses were propagated in the allantoic cavity of 10-day-old embryonated hens' eggs and titered via standard plaque assay using Madin-Darby canine kidney (MDCK) cells as previously described ( Maines et al., 2005 , Zeng et al., 2007 ). All experiments were conducted under biosafety level 3 containment, including enhancements as required by the U.S. Department of Agriculture and the Federal Select Agent Program ( Chosewood et al., 2009 ). 2.2 Cell culture and liquid inoculations The bronchial epithelial cell line Calu-3 (ATCC) was cultured as previously described ( Zeng et al., 2007 ). Primary human alveolar epithelial cells (Cell Biologics) were cryopreserved at passage 3, then grown and expanded per manufacturer's instructions. All cells were seeded on 24 mm diameter (6-well format) or 12 mm diameter (12-well format) semipermeable membrane inserts with a 0.4 µm pore size (Corning) and grown to confluence under submerged conditions. After reaching a transepithelial resistance of >1000 Ω 2 ( Zeng et al., 2007 ), apical media was removed from selected Calu-3 cells to create an air-liquid interface (ALI), which was maintained for three weeks to facilitate cell differentiation and the establishment of a mucin layer. Prior to inoculation, apical media (if present) was removed from the cell monolayer and cells cultured under submerged conditions were washed to remove serum present in culture medium. Liquid inoculation was performed using 300 μL of virus, diluted as specified in the results, and incubated on the cell surface for one hour before washing. After infection, cells were cultured in cell type-specific serum-free medium to which 1 μg/mL N-p-tosyl-ʟ-phenylalanine chloromethyl ketone (TPCK)-treated trypsin (Sigma-Aldrich) was added for alveolar cell cultures. Aliquots of apical culture supernatant or wash media (incubated atop cells cultured at ALI for 20 min) were collected at the indicated times post-infection (p.i.) and immediately frozen at −80 °C until titration. Growth curves were generated and analyzed using Prism 6.0.7 (GraphPad Software Inc). 2.3 Aerosol inoculations The automated bioaerosol system used for all experiments has been previously described in detail ( Gustin et al., 2011 , Hartings and Roy, 2004 ) and all conditions were maintained here unless specified otherwise. Briefly, virus suspended in a solution of PBS-0.03% (w/v) BSA was aerosolized using a three-jet Collision nebulizer (BGI, Inc.) and passed through an exposure chamber at a rate of 20 L/min. Cells placed inside the exposure chamber on a wire mesh shelf were inoculated under air-liquid interface conditions ( Fig. 1 ). Using the AeroMP (Biaera Technologies) aerosol management platform, aerosol exposures were conducted at 21 °C and 50% relative humidity for 15 min followed by a 5 min purge to allow evacuation of the aerosolized virus from the chamber ( Gustin et al., 2011 ). Prior to in vitro exposure, spray factor (SF) values were determined for stocks of all viruses to estimate the starting virus concentration in the nebulizer needed to obtain a desired concentration of virus in the aerosol. A Biosampler (SKC Inc) was used to quantify the virus actually aerosolized within the chamber during each exposure. Following aerosol exposure, membrane inserts were transferred to clean 6-well or 12-well plates and serum-free medium was added to apical and basolateral compartments as described above; cells maintained under ALI conditions had medium added to the basolateral compartment only. Confirmation of cell viability during aerosol exposure was performed using the WST-1 cell proliferation reagent (Roche Applied Science), according to manufacturer's instructions, with four independent samples tested for each condition. Fig. 1 Graphic representation of aerosol system for in vitro use. Depiction of human cells cultured on transwell inserts and exposed to aerosolized influenza virus using a previously characterized system ( Gustin et al., 2011 ). Cell culture dishes rest in the exposure chamber on a wire shelf under air-liquid interface conditions for the duration of the exposure. Inset, individual transwell inserts are transferred to sterile plates once removed from the exposure chamber. Fig. 1. 2.4 Quantitation of Exposure and Infectious Doses The total number of infectious virions passed through the chamber during the exposure session ( N cham ) was calculated as N cham = ( C samp ) ( V samp ) ( Q cham ) Q samp where C samp is the concentration of virus in the sampler, V samp is the volume of media in the sampler, and Q cham and Q samp represent the flow rates of chamber and sampler, respectively (see Supplemental methods for derivation). N cham was multiplied by the ratio of the surface area of each transwell ( SA ) to the cross-sectional area of the chamber ( XA ) to yield exposure dose ( ED ). ID 50 s were calculated according to the method of Reed and Muench, with the proportional distance multiplied by log 10 (dose above 50%)/log 10 (dose below 50%) to account for the deviation of exposure doses from exact 10-fold dilutions ( Reed and Muench, 1938 ). Variability around the mean exposure dose was quantified using a binomial model with n viable virions passing through the chamber and a probability of success (virion lands on well) equal to SA/XA . Upper and lower critical values at the 95% confidence level were calculated using R 3.2.3 (R Foundation for Statistical Computing); the true exposure dose for any particular well has a 95% chance of falling between these two values. The cumulative probability distribution indicated that the minimum dose had to be three or greater for 95% of all wells to be inoculated with at least one virion. In order minimize the chance of including a well not exposed to any virus, we therefore aimed not to use mean exposure doses under 5 PFU. When this did occur due to variations in aerosolization efficiency, the minimum 50% infectious dose is reported as ≤3 PFU. 2.5 Real-time RT-PCR Total RNA was extracted from mock-infected or virus-infected cell monolayers after removal of supernatant using the RNeasy mini kit (Qiagen). RT-PCR was performed with a QuantiTect SYBR green RT-PCR kit (Qiagen) in duplicate reactions from duplicate samples using an influenza A virus M1 gene primer set ( Zeng et al., 2007 ). Influenza virus M gene RNA copy numbers were extrapolated using a standard curve based on samples of known M gene copy number. 3 Results 3.1 System establishment and exposure dose determination Rather than use the type of exposure system described in studies of cellular responses to aerosolized particulate matter (eg. CULTEX Radial Flow System), we employed one intended for use in animal infection. Previously optimized for use with influenza virus ( Gustin et al., 2011 ), this system and its components are designed to maintain the viability of biological aerosols. This system features a large exposure chamber, which allows for simultaneous exposure of several plates of cells, facilitating comparisons of different cell types or growth conditions between uniformly exposed wells ( Fig. 1 ). In vitro inoculations of cells grown on semi-permeable membrane inserts (transwells) were conducted as follows: apical media (if present) was removed immediately prior to exposure, then plates holding the inserts and basolateral media were placed in the chamber and exposed to ten-fold serial dilutions of aerosolized virus for a duration of 15 min. Following exposure, inserts were transferred to a sterile tissue culture plate containing fresh basolateral media and apical media was replaced. Mock infection led to no significant decrease in cell viability (data not shown), indicating that the presence of basolateral media was sufficient to keep the cells from drying out during exposure. The ability to quantitate the dose to which an animal or cell monolayer is exposed is a critical component of experiments utilizing aerosol-based exposure systems ( Hartings and Roy, 2004 ), and is necessary for comparison of aerosol infections with those conducted by the more traditional liquid route. We found that titration of virus collected in wells of media placed in the chamber alongside the cells was insufficiently sensitive to reliably quantify the low doses to which each monolayer was exposed (data not shown), possibly because deposition efficiency differed between the cell surface and the liquid media. We therefore modeled our approach on that used in aerosol inoculation of animals. Presented dose for an animal can be expressed as the total number of infectious virions passing through the chamber during an exposure session multiplied by the ratio of the volume of aerosol inhaled by the animal to the total volume of aerosol passed through the chamber (see Supplemental methods for derivation). For in vitro quantification, we substituted the ratio of the surface area of each transwell insert to the cross-sectional area of the chamber for the ratio of inhaled to total aerosol volume. Because the efficiency of particle deposition on the cell surface was estimated to be approximately 100%, presented dose and exposure dose were considered equivalent. No correlation between plate position within the chamber or well position within the plate and virological outcome was observed. 3.2 Validation of experimental approach Initial characterization studies were conducted using the Calu-3 human bronchial epithelial cell line. This cell type is relevant to respiratory infection and has previously been shown to support replication of a variety of influenza A viruses, though published studies have typically used MOIs of 0.01 (equivalent to 10,000 TCID 50 or PFU per 24 mm well) or higher ( Zeng et al., 2007 , Zhou et al., 2011 ). In order to generate comparison data for aerosol infections, we first conducted an analogous experiment using traditional liquid inoculation at a wide range of doses. Two influenza viruses known to replicate with high efficiency (A/Thailand/16/2004 [Thai/16, HPAI A(H5N1)] and A/Anhui/1/2013 [Anhui/1, LPAI A(H7N9)], both isolated from fatal human cases) were serially diluted and used to inoculate quintuplicate wells. Because of the potential for both random and systemic error in making repeated serial dilutions, we titrated all inocula, enabling us to more precisely estimate the number of infectious virions to which each well was exposed. RNA was collected from the cell monolayers of two wells 24 h post-inoculation and assayed for the presence of viral nucleic acid via RT-PCR with primers specific to the M1 gene. Growth kinetics in the remaining three wells were monitored by titration of cell supernatants collected between 2 and 96 h post-infection ( Fig. 2 ). At inoculum doses above our limit of detection (10 PFU/mL or 3 PFU/well), both viruses consistently infected all replicate wells and replicated to high titer, though growth was somewhat delayed at lower inoculation doses relative to higher ones ( Figs. 2 , 3 B). We also observed robust replication in cultures inoculated with approximately 1 PFU (dose estimated from serial dilution) of Anhui/1 virus. At doses of less than one PFU, infection was infrequent, characterized by low titers and undetectable levels of viral nucleic acid in the cell monolayer 24 h after inoculation. Fig. 2 Replication of influenza A viruses in Calu-3 cells. A) Calu-3 cells were infected by the traditional liquid route (dashed line) or the aerosol route (solid line) at the target MOI with the viruses shown, and cultured at 37 °C or 33 °C. Culture supernatants were collected at the indicated times p.i., and titers were determined by standard plaque assay to quantify infectious virus. The limit of detection was 10 PFU. Error bars indicate standard deviation. Lines represent positive wells (infectious virus detected at two sequential timepoints or at 96 h alone, 3/3 unless otherwise noted) only. Cultures with 2/3 positive wells: Thai/16 33 °C aerosol 1×10 −6 , liquid 1×10 −7 ; Anhui/1 33 °C aerosol 1×10 −6 , liquid 1×10 −7 ; Panama/99 37 °C aerosol 1×10 −6 . Cultures with 1/3 positive wells: Thai/16 37 °C liquid 1×10 −8 ; Panama/99 33 °C aerosol 1×10 −5 and 1×10 −6. B) Exact inoculum dose (PFU) and MOI for each infection shown in panel.. Fig. 2. Fig. 3 Infection and replication in Calu-3 cells after aerosol exposure. A) Peak viral titers detected in each well inoculated via the aerosol route. Titers (log 10 PFU/mL) are provided for each well that showed evidence of productive replication, defined as infectious virus detected at two sequential timepoints or at 96 h alone (full replication curves are shown in Fig. 2 ). Exposure dose (PFU) varied slightly between viruses and is therefore listed as a range. Exact exposure doses for each virus are listed in Table S1 . Cells cultured at 33 and 37 °C were exposed concurrently. CID 50 indicates the 50% cellular infectious dose, or MOI required to achieve 50% infectivity, calculated by dividing ID 50 by the cell number. Limit of detection was 10 PFU. B) Comparison of viral supernatant titer (left Y axis) and M copy number (right Y axis) present in the cell monolayer between aerosol and liquid inoculation at 24 h p.i. Cells were cultured at 37 °C. Supernatants were collected immediately prior to lysis of the cell monolayer for RNA collection. Each parameter is expressed as mean ± standard deviation of two independent wells. Limit of detection for M segment RNA was 10 copies.. Fig. 3. Aerosol experiments were conducted with three viruses chosen to represent a diversity of mammalian in vivo pathogenicity and transmissibility phenotypes: Thai/16, Anhui/1, and A/Panama/2007/99 (Panama/1999, seasonal A(H3N2)). All three viruses infected Calu-3 cells with high efficiency following aerosol exposure ( Fig. 3 ). We observed rates of productive infection comparable to those seen after inoculation using a liquid suspension, with 50% infectious doses (ID 50 s) for all three viruses of under five PFU. High peak viral titers (10 8 PFU/mL) were detected in the supernatant regardless of exposure dose, though, as with liquid inoculum, replication was delayed at lower inoculation doses ( Fig. 3 ). Parallel cultures were incubated post-exposure at 33 °C, a temperature thought to represent that of the mammalian upper respiratory tract, after infection to see whether infectivity was temperature-dependent. We found that while 24-h titers of the two avian viruses were slightly lower at this temperature than at 37 °C, infectivity of these cultures was reduced only for Panama/99 virus, and only slightly (ID 50 of 12 vs ≤3 PFU). The concordance in infectious dose between aerosol and liquid inoculations suggested that our calculated exposure dose for each well accurately represented the average number of virions to which a well was truly exposed. To confirm this, we conducted a series of exposures at doses near 1 PFU (range 0.02-7 PFU) per well. In light of our liquid exposure data with Thai/16 and Anhui/1, we reasoned that if the calculated exposure doses were accurate, the majority of Calu-3 wells should be infected with either of these viruses at doses ≥1 PFU, whereas few wells would be infected at doses 0.05). This suggested that our estimation of 100% deposition efficiency did not compromise our exposure dose estimates. Table 1 Concordance between observed and expected infection rates of wells exposed by the aerosol route to doses near 1 PFU. Table 1. Exposure dose a Virus Observed # infected b Expected # infected c p -value d Power e >1 PFU Thai/16 9/12 ≥10.5/12 0.17 0.99 Anhui/1 15/18 ≥16.5/18 0.18 1 0.05). This suggested that our estimation of 100% deposition efficiency did not compromise our exposure dose estimates. Table 1 Concordance between observed and expected infection rates of wells exposed by the aerosol route to doses near 1 PFU. Table 1. Exposure dose a Virus Observed # infected b Expected # infected c p -value d Power e >1 PFU Thai/16 9/12 ≥10.5/12 0.17 0.99 Anhui/1 15/18 ≥16.5/18 0.18 1 5 µm) droplets or virus present on surfaces, reaches the lower respiratory tract when inhaled ( BeruBe et al., 2009 ). We were therefore particularly interested in the ability of aerosolized virus to infect primary human alveolar cells. We found that both the seasonal virus Panama/99 and an outbreak-associated H7N9 virus, Anhui/1, replicated productively in primary human pneumocytes after inoculation via the liquid, but not aerosol, route whereas the H5N1 virus Thai/16 was highly infectious and replicated to high titer regardless of inoculation method. For Panama/99 virus, these in vitro results are consistent with both ferret studies, which do not detect virus replication in the lungs, and with human seasonal virus infections, which are typically limited to the upper respiratory tract. In contrast, severe human A(H7N9) cases have been characterized by symptoms of lung infection ( Chen et al., 2013 , Gao et al., 2013 , Hu et al., 2013 , Yang et al., 2014 , Yu et al., 2013b ), and virus has been detected in the lungs of experimentally infected animals ( Belser et al., 2013 , de Wit et al., 2014 , Gabbard et al., 2014 , Watanabe et al., 2013 , Xu et al., 2014 , Zhang et al., 2013 , Zhu et al., 2013 ). Our findings raise the possibility that the development of viral pneumonia associated with H7N9 virus develops not upon initial exposure, but subsequent to viral spread from adjacent tissues, and warrant further investigation regarding the dynamics of H7N9 virus infection throughout the respiratory tract. The need for such spread may provide a window of opportunity for the immune system to restrict the virus before it causes severe disease, which would explain the apparent prevalence of clinically inapparent and mild infection with this virus ( Chen et al., 2014 , Ip et al., 2013 , Yu et al., 2013a ). Using the Calu-3 cell line, we demonstrated that growth under ALI conditions reduced the efficiency of both initial infection and subsequent viral replication. The high viscosity of mucus and abundance of virus-binding sialic acids may limit the diffusion of virus between cells, thereby reducing viral titers. Notably, abrogation in infectivity resulting from culture at ALI was more pronounced with the two avian viruses tested than the seasonal virus Panama/99. This finding is consistent with the hypothesis that respiratory mucus serves as an important barrier to the ability of avian influenza viruses to transmit between humans, possibly because their specificity for α2,3-linked sialic acids makes them more susceptible to binding and entrapment by mucus, which some studies have suggested contains glycans primarily in the α2,3 conformation ( Baum and Paulson, 1990 , Couceiro et al., 1993 ). Use of reverse genetics techniques to compare viruses differing only in the sialic acid binding preferences of the hemagglutinin and/or neuraminidase proteins will allow for further investigation of this phenomenon. Aerosol inoculation, particularly when used in conjunction with increasingly sophisticated techniques for in vitro cell culture, offers a unique opportunity to study virus-cell interactions in an environment resembling that of the human respiratory tract. Our studies suggest that aerosol inoculation may enhance studies of viral tropism and improve our understanding of the effects of environmental conditions on the ability of influenza virus to initiate infection. Continued investigation regarding the role of inoculation route in viral binding and entry processes will further our understanding of the infectivity of influenza viruses with distinct phenotypes. In addition to influenza, the methods outlined here could be used for the study of other respiratory viruses such as severe acute respiratory syndrome (SARS) virus, varicella zoster virus (VZV), and measles virus, as well as the risk assessment of novel pathogens. The ability to combine aerosol inoculation with the benefits of in vitro study, notably the ability to study specific cell types in isolation, will facilitate a greater understanding of the infectivity and tropism of respiratory pathogens of public health concern. Appendix A Supplementary material Supplementary material .

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9227656/

An Electronic Data Capture Tool for Data Collection During Public Health Emergencies: Development and Usability Study

Background The Discovery Critical Care Research Network Program for Resilience and Emergency Preparedness (Discovery PREP) partnered with a third-party technology vendor to design and implement an electronic data capture tool that addressed multisite data collection challenges during public health emergencies (PHE) in the United States. The basis of the work was to design an electronic data capture tool and to prospectively gather data on usability from bedside clinicians during national health system stress queries and influenza observational studies. Objective The aim of this paper is to describe the lessons learned in the design and implementation of a novel electronic data capture tool with the goal of significantly increasing the nation's capability to manage real-time data collection and analysis during PHE. Methods A multiyear and multiphase design approach was taken to create an electronic data capture tool, which was used to pilot rapid data capture during a simulated PHE. Following the pilot, the study team retrospectively assessed the feasibility of automating the data captured by the electronic data capture tool directly from the electronic health record. In addition to user feedback during semistructured interviews, the System Usability Scale (SUS) questionnaire was used as a basis to evaluate the usability and performance of the electronic data capture tool. Results Participants included Discovery PREP physicians, their local administrators, and data collectors from tertiary-level academic medical centers at 5 different institutions. User feedback indicated that the designed system had an intuitive user interface and could be used to automate study communication tasks making for more efficient management of multisite studies. SUS questionnaire results classified the system as highly usable (SUS score 82.5/100). Automation of 17 (61%) of the 28 variables in the influenza observational study was deemed feasible during the exploration of automated versus manual data abstraction. The creation and use of the Project Meridian electronic data capture tool identified 6 key design requirements for multisite data collection, including the need for the following: (1) scalability irrespective of the type of participant; (2) a common data set across sites; (3) automated back end administrative capability (eg, reminders and a self-service status board); (4) multimedia communication pathways (eg, email and SMS text messaging); (5) interoperability and integration with local site information technology infrastructure; and (6) natural language processing to extract nondiscrete data elements. Conclusions The use of the electronic data capture tool in multiple multisite Discovery PREP clinical studies proved the feasibility of using the novel, cloud-based platform in practice. The lessons learned from this effort can be used to inform the improvement of ongoing global multisite data collection efforts during the COVID-19 pandemic and transform current manual data abstraction approaches into reliable, real time, and automated information exchange. Future research is needed to expand the ability to perform automated multisite data extraction during a PHE and beyond. Background The Discovery Critical Care Research Network Program for Resilience and Emergency Preparedness (Discovery PREP) partnered with a third-party technology vendor to design and implement an electronic data capture tool that addressed multisite data collection challenges during public health emergencies (PHE) in the United States. The basis of the work was to design an electronic data capture tool and to prospectively gather data on usability from bedside clinicians during national health system stress queries and influenza observational studies. Objective The aim of this paper is to describe the lessons learned in the design and implementation of a novel electronic data capture tool with the goal of significantly increasing the nation's capability to manage real-time data collection and analysis during PHE. Methods A multiyear and multiphase design approach was taken to create an electronic data capture tool, which was used to pilot rapid data capture during a simulated PHE. Following the pilot, the study team retrospectively assessed the feasibility of automating the data captured by the electronic data capture tool directly from the electronic health record. In addition to user feedback during semistructured interviews, the System Usability Scale (SUS) questionnaire was used as a basis to evaluate the usability and performance of the electronic data capture tool. Results Participants included Discovery PREP physicians, their local administrators, and data collectors from tertiary-level academic medical centers at 5 different institutions. User feedback indicated that the designed system had an intuitive user interface and could be used to automate study communication tasks making for more efficient management of multisite studies. SUS questionnaire results classified the system as highly usable (SUS score 82.5/100). Automation of 17 (61%) of the 28 variables in the influenza observational study was deemed feasible during the exploration of automated versus manual data abstraction. The creation and use of the Project Meridian electronic data capture tool identified 6 key design requirements for multisite data collection, including the need for the following: (1) scalability irrespective of the type of participant; (2) a common data set across sites; (3) automated back end administrative capability (eg, reminders and a self-service status board); (4) multimedia communication pathways (eg, email and SMS text messaging); (5) interoperability and integration with local site information technology infrastructure; and (6) natural language processing to extract nondiscrete data elements. Conclusions The use of the electronic data capture tool in multiple multisite Discovery PREP clinical studies proved the feasibility of using the novel, cloud-based platform in practice. The lessons learned from this effort can be used to inform the improvement of ongoing global multisite data collection efforts during the COVID-19 pandemic and transform current manual data abstraction approaches into reliable, real time, and automated information exchange. Future research is needed to expand the ability to perform automated multisite data extraction during a PHE and beyond. Introduction Knowledge sharing during public health emergencies (PHE) is critical to managing swift and appropriate responses by key decision makers. Moreover, clinical responsibilities are typically increased, and dedicated research personnel may be lacking during a PHE. Despite the call to action from the medical community placed on data sharing for effective response, there remains a lack of standard best practice on information exchange during PHE, with no widely available platform mechanism to facilitate data sharing [ 1 - 4 ]. The absence of standards and technology challenges the ability of clinicians to develop a unified treatment plan to confront patients exposed to the PHE at hand. This has been evident since 2001, when the US Public Health System was challenged with the threat of an Anthrax outbreak [ 5 ]. Disparate information sources and unclear jurisdiction across local, state, and federal agencies prevent accurate knowledge sharing and aligned recommendations from decision makers [ 5 ]. The lack of information during PHE is a global challenge, as demonstrated in the data collection efforts during the Zika virus epidemic, Ebola outbreak [ 6 ], and most recently the COVID-19 pandemic [ 7 - 9 ]. This has been exacerbated during the COVID-19 pandemic, where data are needed to guide treatment protocols, but data sharing across a global spectrum is nonexistent or delayed [ 9 - 12 ]. Global standards and a system that allows for real-time learning during public health crisis are critical to our health care community's ability to respond to PHE [ 7 - 9 , 13 - 18 ]. Optimal responses to PHE require data-driven approaches that allow for prospective and real-time clinical data collection and dissemination that overcome the various challenges in data quality [ 18 ]. The current systems suffer from inadequate infrastructure for multisite clinical data capture [ 8 , 16 , 19 , 20 ], delays in dissemination of data due to lack of technical capacity [ 21 ], a lack of tools to manage the quality of data [ 20 ], and the absence of simple and straightforward interfaces that do not add to clinical burden of data collection during PHE [ 18 ]. To mitigate the known barriers to data collection during PHE, the Discovery Critical Care Research Network Program for Resilience and Emergency Preparedness (Discovery PREP [ 17 ]) partnered with Akido Labs, a third-party technology vendor, to develop a platform known as Project Meridian, a tool designed for data capture and dissemination during PHE. Discovery PREP's experience with current research data capture platforms during national health system stress tests, and other PHE tabletop exercises, indicated excessive person-hour effort required to coordinate data collection from multiple sites in a simulated PHE [ 22 - 25 ]. Thus, Discovery PREP began investigating novel methods toward multisite clinical data extraction with the goal of significantly increasing the nation's capability to manage real-time clinical data collection and analysis during PHE. Exploration proceeded with the design and development of a technology-agnostic electronic data capture tool that could facilitate multisite automated data extraction and storage. Following the development of the electronic data capture tool, the feasibility of advancing data capture using automated data extraction compared to manual data entry was assessed in 2 observational studies [ 26 - 29 ]. This paper describes the technical process and lessons learned from this effort, concluding with recommendations for improvement of data sharing platforms during PHE. Methods Overview A multiyear and multiphase approach was taken to develop the electronic data capture tool as visualized in the design timeline ( Figure 1 ). The tool was first developed and piloted for rapid data capture and then expanded to assess the feasibility of automated clinical data extraction. The design and evaluation of the electronic data capture tool spanned from January 2017 to April 2018. Reducing the burden of data collection was a key design principle for the electronic data capture tool, as clinical responsibilities typically increase during PHE, and the availability of research personnel is insufficient to capture the volume of data need for robust clinical trials and their analysis, especially for the critically ill or injured [ 2 , 3 , 18 ]. The electronic data capture tool was designed with an intuitive data entry interface to reduce time and effort for data entry with the added capability to enter data on a smartphone. Ease of use was combined with considerations for scalability across multiple institutions to eliminate manual administration processes and bridge the gap created by disparate platforms. Figure 1 Timeline. Platform Design and Development The participants included Discovery PREP physicians, their local administrators, and data collectors from tertiary-level academic medical centers at 5 geographically distributed institutions (University of Southern California, Washington University in St. Louis, Baylor University, Mayo Clinic, and Duke University). Design, development, rapid prototyping, and user feedback took place between January 2017 and July 2017. Information was gathered prior to the development of the electronic data capture tool to identify the unmet needs and solidify design specifications. Individual semistructured interviews were conducted over a 45-minute duration. A total of 11 participants were interviewed, including 3 (27%) physician researchers, 3 (27%) data collectors, 4 (36%) administrators, and 1 (9%) biostatistician to understand workflows and data collection challenges during a PHE. The initial predevelopment interviews were qualitative in nature and elicited information on data collection processes and limitations of current electronic data capture tools ( Table 1 ). Table 1 Qualitative interview questions prior to electronic data capture tool development. Question # Question detail 1 When do you complete CRFs a with respect to enrollment time? 2 How are new study subjects identified? 3 How are new subjects communicated to data collector? 4 What is your process for collecting data for the CRF? 5 What are the pain points you experience with REDCap? b 6 Pain points with the last study you participated in? a CRF: case report form. b REDCap: Research Electronic Data Capture. Project Meridian was designed to be powered by the Akido Labs Development Environment. The latter was designed to enable modern development in a health care environment by abstracting four core unique complexities specific to this industry, including security of patient health information, compliance, interoperability, and governance ( Multimedia Appendix 1 ). User-centered design practices with an eye toward a simple user interface were the basis of the design of the user interface and prototypes ( Figure 2 ; Multimedia Appendices 2 and 3 ). Postdevelopment, user prototyping interviews focused on feedback regarding electronic data capture tool prototypes. An agile approach of rapid iteration following user feedback was taken to enhance the electronic data capture tool following each interview in preparation for the next, following a hypothesize-design-test learning loop. Interviews were performed both in person and remotely via screen sharing, as needed. The interviews were semistructured, and users conducted standardized tasks while observed by the investigator team including the following: Discovery PREP administrative team, Akido Labs engineers, and a notetaker to capture user feedback on functionality, user experience, and messaging to guide usage. All clicks, mouse movements, and time required to accomplish specific tasks were recorded for analysis and used to refine the platform design. Figure 2 Project Meridian mobile capability screenshots. ECMO: extracorporeal membrane oxygenation; eCRF: Electronic Case Report Form; FDA: Food and Drug Administration; ICU: intensive care unit; IV: intravenous. Usability and Pilot Study Testing Additional feedback was gathered during user acceptance testing (UAT). UAT was performed using the two following scenarios: (1) a Discovery PREP health system stress query over 401 participants in August 2017 ( Figure 3 ), and (2) 34 Society of Critical Care Medicine participants affected by Hurricane Harvey in the state of Texas in September 2017 ( Figure 4 ). The chart on the left for both Figures 3 and 4 show the breakdown of responder practice setting. The map in Figure 3 illustrates the map of responders superimposed on population density. The map in Figure 4 displays the number of responses to the Hurricane Harvey query. Additional feature enhancements were assessed based on the feedback gathered during UAT. Following UAT, the electronic data capture tool was used to facilitate data collection for 2 clinical studies encompassing 403 users across the United States. Both clinical studies involved gathering information on the impact of seasonal influenza on health system stress. The studies were conducted with 12 sites for 17 weeks. The first study involved a predefined set of users with a large data collection form including 151 patient-level clinical data elements. The second encompassed a brief data collection form with 20 questions with census and health system stress level data. Data were collected weekly from health care systems using the Project Meridian platform. Following these studies, a subset of users (n=20) completed the System Usability Scale (SUS) questionnaire [ 30 , 31 ], 19 (95%) participated in debriefing sessions, and 13 (65%) completed a poststudy survey. Figure 3 User acceptance testing (UAT) map of responders, National Health System Stress Query (n=401). ED: emergency department; ICU: intensive care unit. Figure 4 User acceptance testing (UAT) map of responders, Hurricane Harvey (n=20). ED: emergency department; ICU: intensive care unit. Ethics Approval Both studies were approved under the University of Southern California Investigational Review Board (HS-16-00948). Automated Versus Manual Data Extraction Seasonal influenza was used as a proxy for a PHE during the comparison of automated and manual data collection [ 29 ]. Patients in an intensive care unit (ICU) with laboratory-confirmed influenza were enrolled into an investigational review board–approved observational study (HS-17-00837). At a single institution, patient selection and data collection were completed using two methods in parallel: research personnel effort (manual) and querying of institutional clinical data warehouse (automated). Data were collected over a 2-week period using a consensus, previously reported tiered case report form (CRF). Tier 1 of the CRF sought demographics, diagnoses, and lab results as well as supportive care details from the first 24 hours of ICU visit. Tier 2 sought more detailed clinical data from disease onset to patient discharge. Tier 1 was used for comparison in the feasibility test. The automated approach required the identification of relevant patients and gathering of key data elements by executing daily automated queries to an institutional clinical data warehouse. Data were stored and compared for accuracy following the 2-week period. Overview A multiyear and multiphase approach was taken to develop the electronic data capture tool as visualized in the design timeline ( Figure 1 ). The tool was first developed and piloted for rapid data capture and then expanded to assess the feasibility of automated clinical data extraction. The design and evaluation of the electronic data capture tool spanned from January 2017 to April 2018. Reducing the burden of data collection was a key design principle for the electronic data capture tool, as clinical responsibilities typically increase during PHE, and the availability of research personnel is insufficient to capture the volume of data need for robust clinical trials and their analysis, especially for the critically ill or injured [ 2 , 3 , 18 ]. The electronic data capture tool was designed with an intuitive data entry interface to reduce time and effort for data entry with the added capability to enter data on a smartphone. Ease of use was combined with considerations for scalability across multiple institutions to eliminate manual administration processes and bridge the gap created by disparate platforms. Figure 1 Timeline. Platform Design and Development The participants included Discovery PREP physicians, their local administrators, and data collectors from tertiary-level academic medical centers at 5 geographically distributed institutions (University of Southern California, Washington University in St. Louis, Baylor University, Mayo Clinic, and Duke University). Design, development, rapid prototyping, and user feedback took place between January 2017 and July 2017. Information was gathered prior to the development of the electronic data capture tool to identify the unmet needs and solidify design specifications. Individual semistructured interviews were conducted over a 45-minute duration. A total of 11 participants were interviewed, including 3 (27%) physician researchers, 3 (27%) data collectors, 4 (36%) administrators, and 1 (9%) biostatistician to understand workflows and data collection challenges during a PHE. The initial predevelopment interviews were qualitative in nature and elicited information on data collection processes and limitations of current electronic data capture tools ( Table 1 ). Table 1 Qualitative interview questions prior to electronic data capture tool development. Question # Question detail 1 When do you complete CRFs a with respect to enrollment time? 2 How are new study subjects identified? 3 How are new subjects communicated to data collector? 4 What is your process for collecting data for the CRF? 5 What are the pain points you experience with REDCap? b 6 Pain points with the last study you participated in? a CRF: case report form. b REDCap: Research Electronic Data Capture. Project Meridian was designed to be powered by the Akido Labs Development Environment. The latter was designed to enable modern development in a health care environment by abstracting four core unique complexities specific to this industry, including security of patient health information, compliance, interoperability, and governance ( Multimedia Appendix 1 ). User-centered design practices with an eye toward a simple user interface were the basis of the design of the user interface and prototypes ( Figure 2 ; Multimedia Appendices 2 and 3 ). Postdevelopment, user prototyping interviews focused on feedback regarding electronic data capture tool prototypes. An agile approach of rapid iteration following user feedback was taken to enhance the electronic data capture tool following each interview in preparation for the next, following a hypothesize-design-test learning loop. Interviews were performed both in person and remotely via screen sharing, as needed. The interviews were semistructured, and users conducted standardized tasks while observed by the investigator team including the following: Discovery PREP administrative team, Akido Labs engineers, and a notetaker to capture user feedback on functionality, user experience, and messaging to guide usage. All clicks, mouse movements, and time required to accomplish specific tasks were recorded for analysis and used to refine the platform design. Figure 2 Project Meridian mobile capability screenshots. ECMO: extracorporeal membrane oxygenation; eCRF: Electronic Case Report Form; FDA: Food and Drug Administration; ICU: intensive care unit; IV: intravenous. Usability and Pilot Study Testing Additional feedback was gathered during user acceptance testing (UAT). UAT was performed using the two following scenarios: (1) a Discovery PREP health system stress query over 401 participants in August 2017 ( Figure 3 ), and (2) 34 Society of Critical Care Medicine participants affected by Hurricane Harvey in the state of Texas in September 2017 ( Figure 4 ). The chart on the left for both Figures 3 and 4 show the breakdown of responder practice setting. The map in Figure 3 illustrates the map of responders superimposed on population density. The map in Figure 4 displays the number of responses to the Hurricane Harvey query. Additional feature enhancements were assessed based on the feedback gathered during UAT. Following UAT, the electronic data capture tool was used to facilitate data collection for 2 clinical studies encompassing 403 users across the United States. Both clinical studies involved gathering information on the impact of seasonal influenza on health system stress. The studies were conducted with 12 sites for 17 weeks. The first study involved a predefined set of users with a large data collection form including 151 patient-level clinical data elements. The second encompassed a brief data collection form with 20 questions with census and health system stress level data. Data were collected weekly from health care systems using the Project Meridian platform. Following these studies, a subset of users (n=20) completed the System Usability Scale (SUS) questionnaire [ 30 , 31 ], 19 (95%) participated in debriefing sessions, and 13 (65%) completed a poststudy survey. Figure 3 User acceptance testing (UAT) map of responders, National Health System Stress Query (n=401). ED: emergency department; ICU: intensive care unit. Figure 4 User acceptance testing (UAT) map of responders, Hurricane Harvey (n=20). ED: emergency department; ICU: intensive care unit. Ethics Approval Both studies were approved under the University of Southern California Investigational Review Board (HS-16-00948). Automated Versus Manual Data Extraction Seasonal influenza was used as a proxy for a PHE during the comparison of automated and manual data collection [ 29 ]. Patients in an intensive care unit (ICU) with laboratory-confirmed influenza were enrolled into an investigational review board–approved observational study (HS-17-00837). At a single institution, patient selection and data collection were completed using two methods in parallel: research personnel effort (manual) and querying of institutional clinical data warehouse (automated). Data were collected over a 2-week period using a consensus, previously reported tiered case report form (CRF). Tier 1 of the CRF sought demographics, diagnoses, and lab results as well as supportive care details from the first 24 hours of ICU visit. Tier 2 sought more detailed clinical data from disease onset to patient discharge. Tier 1 was used for comparison in the feasibility test. The automated approach required the identification of relevant patients and gathering of key data elements by executing daily automated queries to an institutional clinical data warehouse. Data were stored and compared for accuracy following the 2-week period. Results Platform Design and Development During the design phase, the results of the initial qualitative interviews highlighted the following themes: (1) the need to automate data entry; (2) the need to automate frequent study communications and coordination tasks; (3) the importance of ease-of-access and usability; and (4) the need to enable real-time data reporting to stakeholders during a PHE. Identification of these needs led to the inclusion of multiple feature enhancements within the Project Meridian platform prior to product launch ( Textbox 1 ). Project Meridian feature enhancements. Feature description Gamification—leaderboard for number of responses Text message–based survey initiation Refer a colleague (if primary responder not on clinical service, or new responder) Redesign of automated survey email (improving call to action) Improving visibility of case report form completion rate Common view for members of one study team (one institution). All case report forms visible to all study data collectors at a given institution Advanced query functions to prompt individuals or their sites Usability and Pilot Study Testing Design, development, and UAT of the platform occurred over a 9-month period. During usability testing, using the observational studies, data entry personnel reported increased awareness of data entry completeness with the use of site level summary dashboards. Additionally, Discovery PREP study administrators reported that the automation of scheduled personalized emails to the study participants reduced study administration time by an estimated 80% compared to previous studies. The results of the SUS questionnaire [ 30 , 31 ] classified the system among the 90th percentile of a broad class of systems evaluated [ 30 ] and was therefore highly usable (SUS score 82.5/100). Automated Versus Manual Data Extraction The automated and manual data extraction pilot for patient selection independently identified the correct patients (N=4) during the 2-week study period. Completion of Tier 1 of the CRF per patient was 100% (28/28) via manual approach and 61% (17/28) via automation. Compared with manually collected data, automated data were 50% (70/141) identical and 13% (18/141) different. Variables such as demographics, ventilator status, and availability of lab values were identical. The individual lab values pulled in the first 24 hours of ICU admission were not always identical as there were multiple values available for some patients within that first 24-hour period. Values for pregnancy status, preadmit events, coinfections, and means of identification were missing. Data obtained through automated means had an inherent delay of up to 24 hours due to the use of the data warehouse infrastructure. Manually collected data had an average delay of 2-days between fulfillment of inclusion criteria and enrollment into the study. Platform Design and Development During the design phase, the results of the initial qualitative interviews highlighted the following themes: (1) the need to automate data entry; (2) the need to automate frequent study communications and coordination tasks; (3) the importance of ease-of-access and usability; and (4) the need to enable real-time data reporting to stakeholders during a PHE. Identification of these needs led to the inclusion of multiple feature enhancements within the Project Meridian platform prior to product launch ( Textbox 1 ). Project Meridian feature enhancements. Feature description Gamification—leaderboard for number of responses Text message–based survey initiation Refer a colleague (if primary responder not on clinical service, or new responder) Redesign of automated survey email (improving call to action) Improving visibility of case report form completion rate Common view for members of one study team (one institution). All case report forms visible to all study data collectors at a given institution Advanced query functions to prompt individuals or their sites Usability and Pilot Study Testing Design, development, and UAT of the platform occurred over a 9-month period. During usability testing, using the observational studies, data entry personnel reported increased awareness of data entry completeness with the use of site level summary dashboards. Additionally, Discovery PREP study administrators reported that the automation of scheduled personalized emails to the study participants reduced study administration time by an estimated 80% compared to previous studies. The results of the SUS questionnaire [ 30 , 31 ] classified the system among the 90th percentile of a broad class of systems evaluated [ 30 ] and was therefore highly usable (SUS score 82.5/100). Automated Versus Manual Data Extraction The automated and manual data extraction pilot for patient selection independently identified the correct patients (N=4) during the 2-week study period. Completion of Tier 1 of the CRF per patient was 100% (28/28) via manual approach and 61% (17/28) via automation. Compared with manually collected data, automated data were 50% (70/141) identical and 13% (18/141) different. Variables such as demographics, ventilator status, and availability of lab values were identical. The individual lab values pulled in the first 24 hours of ICU admission were not always identical as there were multiple values available for some patients within that first 24-hour period. Values for pregnancy status, preadmit events, coinfections, and means of identification were missing. Data obtained through automated means had an inherent delay of up to 24 hours due to the use of the data warehouse infrastructure. Manually collected data had an average delay of 2-days between fulfillment of inclusion criteria and enrollment into the study. Discussion The electronic data capture tool designed and tested proved highly usable and capable of collecting critical information during PHE test scenarios. One of the lessons learned globally during the recent COVID-19 pandemic is the importance of standardized real-time data collection, analysis, and reporting [ 7 , 32 ]. Prior to the pandemic, Discovery PREP investigators and federal partners developed a novel data capture system to manage multisite data collection to address the all-hazards core data set used to characterize serious illness, injuries, and resource requirements during PHE [ 18 ]. The design and implementation of the Project Meridian electronic data capture tool was Discovery PREP's successful solution to enhance coordinated data collection capabilities during PHEs by addressing the pain points experienced by the clinical community during multisite data collection. Discovery PREP continued to leverage and report on the use of the Project Meridian platform in subsequent national studies [ 33 , 34 ]. Throughout this design and use process, Discovery PREP learned that specific design tenets need to be addressed to successfully gather essential information during a PHE. These tenets include the following: Gathering data to assess a nationwide health system stress during influenza seasons involved collecting data from a bedside clinician (N=403) or an individual institution (N=12) [ 26 - 28 ]. Thus, the data collection system needs to be scalable and adaptable to the number and type of participants. The data gathered during an event may include multiple types of case report forms with a combination of similar and differing variables that often require repeat measurements. For example, one of the observational studies was a weekly query to assess health system stress, while the other study was a single-report event with the same set of variables but with additional clinical content. Furthermore, a common data dictionary was created across Discovery PREP participating institutions to ensure the alignment of data collection across the sites. Thus, a data collection system must be able to accommodate a common consensus data set, with repeated measures across studies, and aggregate data for analysis and reporting to regional and federal government agencies. Automating study administrative and communication tasks (eg, reminder emails) reduced the amount of manual administration for the study. Additionally, the status board (eg, leaderboard) served as a self-service visual to assess individual responses compared to others and to drive an increase in participant response. Thus, a data collection system should automate communication tasks and incorporate a status board for self-service and to encourage participation, especially during PHEs such as the COVID-19 pandemic. The participants noted that during a busy clinical shift, text messaging was a more effective way to obtain a rapid response. Thus, a data collection system needs to adapt to the preferred communication method of the participant, which may vary across time and institutions. Automation of data and reduction of data acquisition time requires a highly interoperable system that integrates with the variety of platforms used at various institutions. Thus, a data collection system should provide the flexibility and functionality to integrate with local information technology infrastructure for automated and near–real time data capture. In a single institution, the identification of eligible patients was reliably accomplished using automation. Additionally, 50% of the data collected manually for one of the observational studies was identically gathered through automation. However, when comparing the manual versus automated data extraction process, only discrete, categorical data fields were available. Text blocks within progress, operative, and discharge notes or the history and physical notes could not be automated for our purposes. Thus, an optimal data collection system should include natural language processing capability with access to these types of domains to fully automate local data extraction. Extensive work is needed to meet the needs of rapid data collection during a PHE. This has been evident during the COVID-19 pandemic, where surveillance efforts have underlined the benefits of creating a Clinical Informatics Digital Hub for monitoring and for clinical trial data management [ 32 ]. To expand the findings in this report, more investigation is needed to assess the following: feasibility of real-time automation; the use of synchronization protocols as needed in areas challenged by unreliable or slow internet access [ 35 ]; the use of natural language processing to capture unstructured data [ 36 ]; and application of artificial intelligence to expand our ability to respond to a rapidly evolving disease [ 37 ]. With a lack of common regional or federal PHE reporting standards in the United States, third-party integration platforms such as Project Meridian can provide essential flexible infrastructure. Rapid data collection is critical to an optimized national and international response [ 6 , 32 , 36 ]. Discovery PREP addressed this need by building and piloting an electronic data capture tool that was successful in collecting coordinated and real-time multisite data to assess health system stress and evaluated treatment protocols for seasonal influenza across the United States. The lessons learned from this report should be leveraged to improve data collection efforts and provide the foundation for further investigations focused on the evolution of manual data abstraction into reliable, real-time, and automated information exchange. Disclaimer The contents of this publication are the sole responsibility of the authors and do not necessarily reflect the views, opinions, or policies of any government entity. Mention of trade names, commercial products, or organizations does not imply endorsement by the US government. Some of the information described herein was presented in abstract form at the 48th Critical Care Congress of the Society of Critical Care Medicine, February 17-20, 2019, San Diego, California, United States, and the 49th Annual Critical Care Congress, February 16-19, 2020, Orlando, Florida, United States.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5097160/

Growth inhibition of cytosolic Salmonella by caspase-1 and caspase-11 precedes host cell death

Sensing bacterial products in the cytosol of mammalian cells by NOD-like receptors leads to the activation of caspase-1 inflammasomes, and the production of the pro-inflammatory cytokines interleukin (IL)-18 and IL-1β. In addition, mouse caspase-11 (represented in humans by its orthologs, caspase-4 and caspase-5) detects cytosolic bacterial LPS directly. Activation of caspase-1 and caspase-11 initiates pyroptotic host cell death that releases potentially harmful bacteria from the nutrient-rich host cell cytosol into the extracellular environment. Here we use single cell analysis and time-lapse microscopy to identify a subpopulation of host cells, in which growth of cytosolic Salmonella Typhimurium is inhibited independently or prior to the onset of cell death. The enzymatic activities of caspase-1 and caspase-11 are required for growth inhibition in different cell types. Our results reveal that these proteases have important functions beyond the direct induction of pyroptosis and proinflammatory cytokine secretion in the control of growth and elimination of cytosolic bacteria. Results Caspase-mediated growth inhibition of cytosolic Salmonella 3T3 fibroblasts are relatively non-permissive for the growth of sifA mutant Salmonella ( ΔsifA) 27 ( Fig. 1a ), but the mechanism for this is unclear. In DMSO treated samples, the similar levels of lactate dehydrogenase (LDH) released from fibroblasts invaded by WT or ΔsifA Salmonella ( Fig. 1b ) suggest that host cell death is not sufficient to explain the reduced growth of ΔsifA bacteria ( Fig. 1a ). To explore the contribution of caspases in the growth inhibition of cytosolic Salmonella we exposed 3T3 fibroblasts to the irreversible pan-caspase inhibitor zVAD-FMK. Addition of zVAD-FMK resulted in greater LDH release from cells infected with either WT or ΔsifA Salmonella from 6 h post-invasion (p.i.) ( Fig. 1b ). Nevertheless, intracellular bacterial growth of ΔsifA Salmonella, measured by CFU (10 h p.i.), increased significantly in cells treated with zVAD-FMK ( Fig. 1a ). Microscopic analysis revealed that inhibition of caspases led to an increase in the % of infected cells harbouring >30 bacteria, for both WT and ΔsifA Salmonella ( Fig. 1c ). Following SPI-1 T3SS-mediated invasion of cells, between 10 and 20% of vacuoles were ruptured ( Supplementary Fig. 1B ), yielding a naturally occurring population of cytosolic WT bacteria 7 . ΔsifA bacteria yield a greater population of cytosolic bacteria, as a result of SPI-1 T3SS-mediated instability of the early vacuole (1–2 h) and deregulated activities of SPI-2 T3SS effectors (4–6 h onwards) 27 . To determine if caspase-mediated inhibition of WT Salmonella affected cytosolic and/or vacuolar bacteria, bacterial growth was analysed after exposure of infected cells to chloroquine (CQ), which accumulates to bactericidal concentrations selectively within acidic endosomes and vacuoles, including the SCV (refs 33 , 34 ). A prgH mutant is defective for SPI-1 function but can enter non-phagocytic cells if it carries a plasmid expressing Yersinia Invasin 35 . As non-SPI-1-mediated entry results in decreased vacuole rupture in epithelial cells 7 , the Invasin-producing prgH mutant was used as a control to establish a concentration of CQ sufficient to kill >99% of bacteria. At this concentration, ∼5% (±2%) of WT Salmonella survived, and only underwent replication when zVAD-FMK was added prior to infection ( Fig. 1d , cytosolic population). In contrast, when only the vacuolar population (total CFU counts—cytosolic population) was analysed ( Fig. 1d , vacuolar population), there was no statistical difference in CFU following the addition of zVAD-FMK. Furthermore, zVAD-FMK had no effect on the recovery of the Invasin-producing prgH mutant, confirming that caspase inhibition does not influence vacuolar Salmonella ( Supplementary Fig. 1A ). Therefore, when only cytosolic bacteria were analysed, a strong caspase-dependent inhibition on bacterial numbers was detected. To visualize cytosolic replication of WT bacteria after inhibition of caspase activity, infected 3T3 cells expressing GFP-tagged galectin-8 (a marker of vacuole integrity 36 ) were imaged over time in the presence of the membrane impermeant dye propidium iodide, so that viable cells could be distinguished from dying cells. zVAD-FMK did not affect recruitment of galectin-8 or LC3B (an autophagy protein that is frequently recruited to bacteria following vacuole rupture 7 ) to ruptured SCVs ( Supplementary Fig. 1B,C ). Pan-caspase inhibition resulted in a dramatic increase in Salmonella replication in cells containing ruptured SCVs ( Fig. 1e ). Importantly, when bacterial replication was observed, it preceded the uptake of PI ( Fig. 1e ). Time-lapse imaging in GFP-LC3B-expressing fibroblasts following exposure to zVAD-FMK confirmed the striking degree of bacterial replication, even in cells where LC3B was recruited to bacteria ( Fig. 1f ). As zVAD-FMK does not influence SCV stability, our results strongly suggest that caspases inhibit growth of WT and ΔsifA Salmonella in the cytosol of 3T3 fibroblasts. To investigate the contribution of individual caspases, we tested more specific peptide-based inhibitors of caspases and siRNA-mediated depletion of caspase-11 or caspase-7. None of these peptide inhibitors had an effect on bacterial growth ( Supplementary Fig. 1D ). Only knockdown of caspase-11 resulted in increased WT and ΔsifA bacterial numbers ( Fig. 1g ). Caspase-11 production in MEFs reduces bacterial replication The Salmonella ΔsifA mutant replicates in the cytosol of HeLa cells 27 37 and certain mouse embryonic fibroblast lines (MEFs) also appear permissive for the growth of this strain ( Fig. 2a ; 38 ). Interestingly, in MEFs this was correlated with the absence of detectable caspase-11 ( Supplementary Fig. 2A ) and almost 100-fold less basal caspase-11 messengerRNA (mRNA) (but not caspase-7), compared with RAW macrophages, immortalized bone-marrow-derived macrophages (iBMDMs) or 3T3 fibroblasts ( Supplementary Fig. 2A,B ). Transcriptional upregulation of caspase-11 occurs following activation of TLR4 by extracellular LPS 14 15 . MEFs infected with Salmonella failed to induce detectable levels of caspase-11 ( Fig. 2b ), which might explain the lack of ΔsifA Salmonella infection-induced cell death in MEFs compared with iBMDMs ( Fig. 2c ). To test if the permissiveness of MEFs was due to the low levels of caspase-11, these cells were transduced to express either WT caspase-11 (C11) or catalytic mutant caspase-11 where the critical cysteine 39 was replaced for glycine at position 254 (C11CM) ( Fig. 2d ). Neither form of caspase-11 affected LDH release following invasion by WT or ΔsifA Salmonella (compare Fig. 2c right panel with Fig. 2e ). However, catalytically active caspase-11 partially reduced the replication of both WT and ΔsifA Salmonella ( Fig. 2f ). Therefore, when introduced into MEFs, caspase-11 restricts intracellular bacterial replication in the absence of detectable host cell death. Caspase-11 restrains ΔsifA mutant growth in macrophages To investigate if the apparent intra-macrophage growth defect observed for ΔsifA bacteria 27 is dependent on caspase-11, immortalized bone-marrow derived macrophages (iBMDM) from C57BL/6 and caspase-11 ( Casp11 −/− ) knock-out mice were analysed. There was a greater than 10-fold increase in WT Salmonella over a 12 h time period in C57BL/6 and Casp11 −/− iBMDMs. As expected, far fewer ΔsifA bacteria were recovered from C57BL/6 cells; this growth inhibition was partially alleviated in Casp11 −/− ( Fig. 3a ), but not Casp7 −/− macrophages ( Supplementary Fig. 3A ). Therefore, as in 3T3 fibroblasts, the low recovery of ΔsifA bacteria is at least partially dependent on caspase-11. To determine if caspase-11 affected macrophage vacuole stability, cytosolic bacteria were quantified by microscopy after selective permeabilisation of the plasma membrane. Less than 5% of WT Salmonella were cytosolic in both C57BL/6 and Casp11 −/− iBMDMs at 2, 6 and 10 h post-uptake ( Fig. 3b ). At 6 h, ∼25% of ΔsifA bacteria were cytosolic in C57BL/6 and Casp11 −/− iBMDMs ( Fig. 3b ). Therefore, caspase-11 does not affect vacuole escape of ΔsifA Salmonella. Whereas the percentage of cytosolic ΔsifA bacteria did not increase in C57BL/6 iBMDMs between 6 and 10 h, a significant increase occurred in Casp11 −/− iBMDMs ( Fig. 3b ). Furthermore, chloroquine-resistant ΔsifA bacteria showed a 2-fold growth increase in Casp11 −/− but not C57BL/6 iBMDMs between 7 and 10 h ( Supplementary Fig. 3B ). Therefore, an increase in the proportion of cytosolic bacteria probably accounts for the increased CFU counts obtained from Casp11 −/− iBMDMs. In macrophages, caspase-11 initiates cytosolic LPS-dependent cell death 16 19 23 , characterized by plasma membrane pore formation, cell swelling and lysis, which could result in a reduced number of host cells available for analysis by CFU. In addition, antibiotic present in the culture medium could kill intracellular bacteria after entering host cells through plasma membrane pores 40 . To analyse the extent of caspase-11-dependent cell death following infection with ΔsifA Salmonella, Casp11 −/− iBMDMs were stably transduced with vectors expressing either WT caspase-11 or a catalytically dead mutant 41 (C 254 G; Supplementary Fig. 3C ). As expected, Casp11 −/− cells expressing WT caspase-11 exhibited greater IL-1β release ( Supplementary Fig. 3D ) than cells expressing the catalytic mutant. Similarly, caspase-11 catalytic activity was required for cell lysis (indicated by LDH release) up to 6 h post-uptake, when a similar percentage of ΔsifA bacteria were cytosolic in the presence or absence of caspase-11 ( Fig. 3b,c ). However, by 12 h, cell lysis and membrane damage (indicated by uptake of PI) were independent of caspase-11 ( Fig. 3c ; Supplementary Fig. 3E ). Therefore, the increased CFU counts in macrophages lacking caspase-11 ( Fig. 3a ) could be attributed (at least in part) to decreased early cell death, but might also involve alleviation of a cytosolic growth inhibitory mechanism. To investigate this further, we measured bacterial loads within individual cells with intact plasma membranes. iBMDMs were infected with GFP-expressing ΔsifA bacteria in the presence of PI and were analysed by flow cytometry. By 10 h post-uptake, 8% (±1%) of PI-negative Casp11 −/− iBMDMs contained a high bacterial load (defined as greater than 750 arbitrary units, corresponding to greater than ∼30 bacteria per cell) compared with 0.4% (±0.1%) in PI-negative C57BL/6 iBMDMs ( Fig. 3d , left and centre panels). Analysis of the geometric mean of GFP fluorescence per PI-negative iBMDM revealed a significant increase in Casp11 −/− iBMDMs at 10 h compared with C57BL/6 iBMDMs ( Fig. 3d , right hand panel). Expression in Casp11 −/− iBMDMs of functional but not of catalytically inactive caspase-11 reduced the intracellular load of ΔsifA bacteria at 10 h post-uptake to the level of that observed in C57BL/6 iBMDMs ( Fig. 3e ). These data suggest that ΔsifA bacteria failed to grow in PI-negative WT iBMDMs, at a time when ∼25% of bacteria were cytosolic ( Fig. 3b ). We then used time-lapse microscopy to provide a more detailed analysis of bacterial growth in intact cells and the onset of PI uptake over time. The bacterial load per cell (represented as the % of infected macrophages containing low, medium or high bacterial loads) was recorded when macrophages switched from PI-negative to PI-positive ( Fig. 3f , left hand graph). So that the bacterial load in all cells was recorded, any infected macrophage that remained PI-negative by 16 h post-uptake was also recorded ( Fig. 3f , right hand graph). WT Salmonella replicated within C57BL/6 iBMDMs over time ( Fig. 3f left panel, 3G top panel and Supplementary Movie 1 ) and occasionally cells became PI-positive ( Fig. 3g , black arrow). In agreement with results obtained by flow cytometry ( Fig. 3d ), ΔsifA Salmonella displayed little replication in PI-negative C57BL/6 iBMDMs. Up to 50% of infected cells became PI-positive by 15.5 h, but again the bacterial burden was low ( Fig. 3f and Fig. 3g middle panel and Supplementary Movie 2 ). However, in ∼20% of infected Casp11 −/− iBMDMs, ΔsifA Salmonella had undergone medium (10–30 bacteria/cell) to high (30+ bacteria/cell) levels of replication prior to PI uptake ( Fig. 3f right panel, 3G bottom panel and Supplementary Movie 3 ). This replication might provide an explanation for the increase in LDH release from 6 to 12 h in Casp11 −/− iBMDMs ( Fig. 3c ). Furthermore, these time-lapse microscopy experiments reveal (i) a non-synchronous loss of plasma membrane integrity and (ii) caspase-11 mediated growth restriction of ΔsifA bacteria in a sub-population of cells that are not PI-positive. Caspase-1 and caspase-11 inhibit Δ sifA mutant growth As alleviation of growth inhibition in Casp11 −/− iBMDMs was relatively mild, ( Fig. 4a ) we investigated if other caspase family members might inhibit intracellular bacterial growth. Caspase-1 is constitutively expressed in macrophages but very weakly expressed in the permissive MEF cell type ( Fig. 2b , Supplementary Fig. 2A,B ), suggesting that caspase-1 might also contribute to growth inhibition of cytosolic bacteria. The involvement of caspase-1 was tested in several ways. First, exposure of iBMDMs to YVAD-FMK (a caspase-1 inhibitor) resulted in increased recovery of ΔsifA Salmonella ( Fig. 4a ) but not WT Salmonella ( Supplementary Fig. 4A ). Second, significantly more ΔsifA bacterial CFUs were recovered from infected Casp1/11 −/− iBMDMs at 17 h ( Fig. 4a ) and 12 h post-uptake when compared with Casp11 −/− iBMDMs ( Figs 3a and 4b , for which Casp11 −/− data were acquired at the same time as Casp1/11 −/− data). Third, analysis of intracellular ΔsifA bacteria by flow cytometry in PI-negative cells revealed that by 10 h post-uptake, 15% of Casp1/11 −/− iBMDMs contained a high bacterial load ( Fig. 4c ), compared with 8% of Casp11 −/− iBMDMs and 0.4% of C57BL/6 iBMDMs ( Fig. 3d ). As a further test of the contribution of both caspase-1 and caspase-11, WT proteins, catalytically inactive mutants (CM) C1-C 284 A and C11-C 254 G 41 or a non-cleavable but catalytically active form of caspase-1 (mutated at 6 aspartate residues that become cleaved during autoproteolysis (6D-N) 42 ) were expressed individually in Casp1/11 −/− iBMDMs ( Supplementary Fig. 4B ). WT but not catalytically inactive caspase-1 and caspase-11 reduced ΔsifA bacterial loads significantly ( Fig. 4d ), implicating both proteases in growth inhibition of cytosolic bacteria. By 6 h, a greater number of Casp1/11 −/− iBMDMs expressing active caspase-1 underwent ΔsifA Salmonella-induced cell death, compared with cells expressing caspase-11. However, by 8 h, caspase-11-dependent cell death was also observed in response to infection by ΔsifA ( Fig. 4e ) but not WT bacteria ( Supplementary Fig. 4C ). By 12 h, cell death appeared to be independent of caspase-1 and caspase-11 activity ( Fig. 4e ). Time-lapse microscopy experiments showed that ΔsifA bacteria underwent medium to high levels of replication within Casp1/11 −/− iBMDMs ( Fig. 4f,g and Supplementary Movie 4 ). This replication, which occurred in ∼80% of infected cells, preceded the uptake of PI and represented a far greater intracellular population than occurred in WT C57BL/6 (30 bacteria/cell), despite a similar number of Casp1/11 −/− cells harbouring cytosolic bacteria as C57BL/6 iBMDMs ( Fig. 6e ). These results show that inhibition of cytosolic bacterial growth can occur prior to cell death and that this requires the activities of caspase-1 and caspase-11. Caspase-1 and -11 repress growth of cytosolic WT Salmonella In the course of these experiments we detected a small population (30 bacteria, for both WT and ΔsifA Salmonella ( Fig. 1c ). Following SPI-1 T3SS-mediated invasion of cells, between 10 and 20% of vacuoles were ruptured ( Supplementary Fig. 1B ), yielding a naturally occurring population of cytosolic WT bacteria 7 . ΔsifA bacteria yield a greater population of cytosolic bacteria, as a result of SPI-1 T3SS-mediated instability of the early vacuole (1–2 h) and deregulated activities of SPI-2 T3SS effectors (4–6 h onwards) 27 . To determine if caspase-mediated inhibition of WT Salmonella affected cytosolic and/or vacuolar bacteria, bacterial growth was analysed after exposure of infected cells to chloroquine (CQ), which accumulates to bactericidal concentrations selectively within acidic endosomes and vacuoles, including the SCV (refs 33 , 34 ). A prgH mutant is defective for SPI-1 function but can enter non-phagocytic cells if it carries a plasmid expressing Yersinia Invasin 35 . As non-SPI-1-mediated entry results in decreased vacuole rupture in epithelial cells 7 , the Invasin-producing prgH mutant was used as a control to establish a concentration of CQ sufficient to kill >99% of bacteria. At this concentration, ∼5% (±2%) of WT Salmonella survived, and only underwent replication when zVAD-FMK was added prior to infection ( Fig. 1d , cytosolic population). In contrast, when only the vacuolar population (total CFU counts—cytosolic population) was analysed ( Fig. 1d , vacuolar population), there was no statistical difference in CFU following the addition of zVAD-FMK. Furthermore, zVAD-FMK had no effect on the recovery of the Invasin-producing prgH mutant, confirming that caspase inhibition does not influence vacuolar Salmonella ( Supplementary Fig. 1A ). Therefore, when only cytosolic bacteria were analysed, a strong caspase-dependent inhibition on bacterial numbers was detected. To visualize cytosolic replication of WT bacteria after inhibition of caspase activity, infected 3T3 cells expressing GFP-tagged galectin-8 (a marker of vacuole integrity 36 ) were imaged over time in the presence of the membrane impermeant dye propidium iodide, so that viable cells could be distinguished from dying cells. zVAD-FMK did not affect recruitment of galectin-8 or LC3B (an autophagy protein that is frequently recruited to bacteria following vacuole rupture 7 ) to ruptured SCVs ( Supplementary Fig. 1B,C ). Pan-caspase inhibition resulted in a dramatic increase in Salmonella replication in cells containing ruptured SCVs ( Fig. 1e ). Importantly, when bacterial replication was observed, it preceded the uptake of PI ( Fig. 1e ). Time-lapse imaging in GFP-LC3B-expressing fibroblasts following exposure to zVAD-FMK confirmed the striking degree of bacterial replication, even in cells where LC3B was recruited to bacteria ( Fig. 1f ). As zVAD-FMK does not influence SCV stability, our results strongly suggest that caspases inhibit growth of WT and ΔsifA Salmonella in the cytosol of 3T3 fibroblasts. To investigate the contribution of individual caspases, we tested more specific peptide-based inhibitors of caspases and siRNA-mediated depletion of caspase-11 or caspase-7. None of these peptide inhibitors had an effect on bacterial growth ( Supplementary Fig. 1D ). Only knockdown of caspase-11 resulted in increased WT and ΔsifA bacterial numbers ( Fig. 1g ). Caspase-11 production in MEFs reduces bacterial replication The Salmonella ΔsifA mutant replicates in the cytosol of HeLa cells 27 37 and certain mouse embryonic fibroblast lines (MEFs) also appear permissive for the growth of this strain ( Fig. 2a ; 38 ). Interestingly, in MEFs this was correlated with the absence of detectable caspase-11 ( Supplementary Fig. 2A ) and almost 100-fold less basal caspase-11 messengerRNA (mRNA) (but not caspase-7), compared with RAW macrophages, immortalized bone-marrow-derived macrophages (iBMDMs) or 3T3 fibroblasts ( Supplementary Fig. 2A,B ). Transcriptional upregulation of caspase-11 occurs following activation of TLR4 by extracellular LPS 14 15 . MEFs infected with Salmonella failed to induce detectable levels of caspase-11 ( Fig. 2b ), which might explain the lack of ΔsifA Salmonella infection-induced cell death in MEFs compared with iBMDMs ( Fig. 2c ). To test if the permissiveness of MEFs was due to the low levels of caspase-11, these cells were transduced to express either WT caspase-11 (C11) or catalytic mutant caspase-11 where the critical cysteine 39 was replaced for glycine at position 254 (C11CM) ( Fig. 2d ). Neither form of caspase-11 affected LDH release following invasion by WT or ΔsifA Salmonella (compare Fig. 2c right panel with Fig. 2e ). However, catalytically active caspase-11 partially reduced the replication of both WT and ΔsifA Salmonella ( Fig. 2f ). Therefore, when introduced into MEFs, caspase-11 restricts intracellular bacterial replication in the absence of detectable host cell death. Caspase-11 restrains ΔsifA mutant growth in macrophages To investigate if the apparent intra-macrophage growth defect observed for ΔsifA bacteria 27 is dependent on caspase-11, immortalized bone-marrow derived macrophages (iBMDM) from C57BL/6 and caspase-11 ( Casp11 −/− ) knock-out mice were analysed. There was a greater than 10-fold increase in WT Salmonella over a 12 h time period in C57BL/6 and Casp11 −/− iBMDMs. As expected, far fewer ΔsifA bacteria were recovered from C57BL/6 cells; this growth inhibition was partially alleviated in Casp11 −/− ( Fig. 3a ), but not Casp7 −/− macrophages ( Supplementary Fig. 3A ). Therefore, as in 3T3 fibroblasts, the low recovery of ΔsifA bacteria is at least partially dependent on caspase-11. To determine if caspase-11 affected macrophage vacuole stability, cytosolic bacteria were quantified by microscopy after selective permeabilisation of the plasma membrane. Less than 5% of WT Salmonella were cytosolic in both C57BL/6 and Casp11 −/− iBMDMs at 2, 6 and 10 h post-uptake ( Fig. 3b ). At 6 h, ∼25% of ΔsifA bacteria were cytosolic in C57BL/6 and Casp11 −/− iBMDMs ( Fig. 3b ). Therefore, caspase-11 does not affect vacuole escape of ΔsifA Salmonella. Whereas the percentage of cytosolic ΔsifA bacteria did not increase in C57BL/6 iBMDMs between 6 and 10 h, a significant increase occurred in Casp11 −/− iBMDMs ( Fig. 3b ). Furthermore, chloroquine-resistant ΔsifA bacteria showed a 2-fold growth increase in Casp11 −/− but not C57BL/6 iBMDMs between 7 and 10 h ( Supplementary Fig. 3B ). Therefore, an increase in the proportion of cytosolic bacteria probably accounts for the increased CFU counts obtained from Casp11 −/− iBMDMs. In macrophages, caspase-11 initiates cytosolic LPS-dependent cell death 16 19 23 , characterized by plasma membrane pore formation, cell swelling and lysis, which could result in a reduced number of host cells available for analysis by CFU. In addition, antibiotic present in the culture medium could kill intracellular bacteria after entering host cells through plasma membrane pores 40 . To analyse the extent of caspase-11-dependent cell death following infection with ΔsifA Salmonella, Casp11 −/− iBMDMs were stably transduced with vectors expressing either WT caspase-11 or a catalytically dead mutant 41 (C 254 G; Supplementary Fig. 3C ). As expected, Casp11 −/− cells expressing WT caspase-11 exhibited greater IL-1β release ( Supplementary Fig. 3D ) than cells expressing the catalytic mutant. Similarly, caspase-11 catalytic activity was required for cell lysis (indicated by LDH release) up to 6 h post-uptake, when a similar percentage of ΔsifA bacteria were cytosolic in the presence or absence of caspase-11 ( Fig. 3b,c ). However, by 12 h, cell lysis and membrane damage (indicated by uptake of PI) were independent of caspase-11 ( Fig. 3c ; Supplementary Fig. 3E ). Therefore, the increased CFU counts in macrophages lacking caspase-11 ( Fig. 3a ) could be attributed (at least in part) to decreased early cell death, but might also involve alleviation of a cytosolic growth inhibitory mechanism. To investigate this further, we measured bacterial loads within individual cells with intact plasma membranes. iBMDMs were infected with GFP-expressing ΔsifA bacteria in the presence of PI and were analysed by flow cytometry. By 10 h post-uptake, 8% (±1%) of PI-negative Casp11 −/− iBMDMs contained a high bacterial load (defined as greater than 750 arbitrary units, corresponding to greater than ∼30 bacteria per cell) compared with 0.4% (±0.1%) in PI-negative C57BL/6 iBMDMs ( Fig. 3d , left and centre panels). Analysis of the geometric mean of GFP fluorescence per PI-negative iBMDM revealed a significant increase in Casp11 −/− iBMDMs at 10 h compared with C57BL/6 iBMDMs ( Fig. 3d , right hand panel). Expression in Casp11 −/− iBMDMs of functional but not of catalytically inactive caspase-11 reduced the intracellular load of ΔsifA bacteria at 10 h post-uptake to the level of that observed in C57BL/6 iBMDMs ( Fig. 3e ). These data suggest that ΔsifA bacteria failed to grow in PI-negative WT iBMDMs, at a time when ∼25% of bacteria were cytosolic ( Fig. 3b ). We then used time-lapse microscopy to provide a more detailed analysis of bacterial growth in intact cells and the onset of PI uptake over time. The bacterial load per cell (represented as the % of infected macrophages containing low, medium or high bacterial loads) was recorded when macrophages switched from PI-negative to PI-positive ( Fig. 3f , left hand graph). So that the bacterial load in all cells was recorded, any infected macrophage that remained PI-negative by 16 h post-uptake was also recorded ( Fig. 3f , right hand graph). WT Salmonella replicated within C57BL/6 iBMDMs over time ( Fig. 3f left panel, 3G top panel and Supplementary Movie 1 ) and occasionally cells became PI-positive ( Fig. 3g , black arrow). In agreement with results obtained by flow cytometry ( Fig. 3d ), ΔsifA Salmonella displayed little replication in PI-negative C57BL/6 iBMDMs. Up to 50% of infected cells became PI-positive by 15.5 h, but again the bacterial burden was low ( Fig. 3f and Fig. 3g middle panel and Supplementary Movie 2 ). However, in ∼20% of infected Casp11 −/− iBMDMs, ΔsifA Salmonella had undergone medium (10–30 bacteria/cell) to high (30+ bacteria/cell) levels of replication prior to PI uptake ( Fig. 3f right panel, 3G bottom panel and Supplementary Movie 3 ). This replication might provide an explanation for the increase in LDH release from 6 to 12 h in Casp11 −/− iBMDMs ( Fig. 3c ). Furthermore, these time-lapse microscopy experiments reveal (i) a non-synchronous loss of plasma membrane integrity and (ii) caspase-11 mediated growth restriction of ΔsifA bacteria in a sub-population of cells that are not PI-positive. Caspase-1 and caspase-11 inhibit Δ sifA mutant growth As alleviation of growth inhibition in Casp11 −/− iBMDMs was relatively mild, ( Fig. 4a ) we investigated if other caspase family members might inhibit intracellular bacterial growth. Caspase-1 is constitutively expressed in macrophages but very weakly expressed in the permissive MEF cell type ( Fig. 2b , Supplementary Fig. 2A,B ), suggesting that caspase-1 might also contribute to growth inhibition of cytosolic bacteria. The involvement of caspase-1 was tested in several ways. First, exposure of iBMDMs to YVAD-FMK (a caspase-1 inhibitor) resulted in increased recovery of ΔsifA Salmonella ( Fig. 4a ) but not WT Salmonella ( Supplementary Fig. 4A ). Second, significantly more ΔsifA bacterial CFUs were recovered from infected Casp1/11 −/− iBMDMs at 17 h ( Fig. 4a ) and 12 h post-uptake when compared with Casp11 −/− iBMDMs ( Figs 3a and 4b , for which Casp11 −/− data were acquired at the same time as Casp1/11 −/− data). Third, analysis of intracellular ΔsifA bacteria by flow cytometry in PI-negative cells revealed that by 10 h post-uptake, 15% of Casp1/11 −/− iBMDMs contained a high bacterial load ( Fig. 4c ), compared with 8% of Casp11 −/− iBMDMs and 0.4% of C57BL/6 iBMDMs ( Fig. 3d ). As a further test of the contribution of both caspase-1 and caspase-11, WT proteins, catalytically inactive mutants (CM) C1-C 284 A and C11-C 254 G 41 or a non-cleavable but catalytically active form of caspase-1 (mutated at 6 aspartate residues that become cleaved during autoproteolysis (6D-N) 42 ) were expressed individually in Casp1/11 −/− iBMDMs ( Supplementary Fig. 4B ). WT but not catalytically inactive caspase-1 and caspase-11 reduced ΔsifA bacterial loads significantly ( Fig. 4d ), implicating both proteases in growth inhibition of cytosolic bacteria. By 6 h, a greater number of Casp1/11 −/− iBMDMs expressing active caspase-1 underwent ΔsifA Salmonella-induced cell death, compared with cells expressing caspase-11. However, by 8 h, caspase-11-dependent cell death was also observed in response to infection by ΔsifA ( Fig. 4e ) but not WT bacteria ( Supplementary Fig. 4C ). By 12 h, cell death appeared to be independent of caspase-1 and caspase-11 activity ( Fig. 4e ). Time-lapse microscopy experiments showed that ΔsifA bacteria underwent medium to high levels of replication within Casp1/11 −/− iBMDMs ( Fig. 4f,g and Supplementary Movie 4 ). This replication, which occurred in ∼80% of infected cells, preceded the uptake of PI and represented a far greater intracellular population than occurred in WT C57BL/6 (30 bacteria/cell), despite a similar number of Casp1/11 −/− cells harbouring cytosolic bacteria as C57BL/6 iBMDMs ( Fig. 6e ). These results show that inhibition of cytosolic bacterial growth can occur prior to cell death and that this requires the activities of caspase-1 and caspase-11. Caspase-1 and -11 repress growth of cytosolic WT Salmonella In the course of these experiments we detected a small population (<5%) of cytosolic WT Salmonella ( Figs 3b and 6c ). A chloroquine protection assay was used to kill vacuolar bacteria, enabling analysis of this subpopulation in the presence or absence of caspase-1 and caspase-11. In Casp1/11 −/− iBMDMs exposed to chloroquine, twice as much growth of cytosolic WT bacteria occurred between 4 and 8 h compared with C57BL/6 cells ( Fig. 6f ). Altogether with our findings in 3T3 fibroblasts ( Fig. 1 ), these results indicate that both caspase-1 and caspase-11 contribute to cytosolic growth inhibition of WT Salmonella. Effects of caspases on ΔsifA bacteria in primary macrophages To determine if primary bone-marrow-derived macrophages (BMDM) inhibit growth of cytosolic Salmonella, bacterial loads were measured by flow cytometry in PI-negative cells. Similar to our observations in iBMDMs, primary BMDMs inhibited growth of ΔsifA Salmonella in a caspase-1 and caspase-11-dependent manner ( Fig. 7a ). Furthermore, by 12 h, the burden of WT Salmonella had increased in the Casp1/11 −/− BMDMs. At 8 h and 10 h post-uptake, LDH release following infection by ΔsifA Salmonella was dependent on both caspase-1 and caspase-11. However, from 12 h onwards, cell death was independent of caspase-1 and 11 ( Fig. 7b ), similar to our findings in immortalized BMDMs ( Figs 3c and 4h,i ). Therefore, our results with immortalized cells are reflected in primary cells and unlikely to be an artefact of the immortalization process. Finally, we examined bacterial loads in splenocytes obtained from C57BL/6 and Casp1/11 −/− mice at 48 h following intraperitoneal inoculation of GFP-expressing Salmonella strains. As expected, in C57BL/6 mice ΔsifA Salmonella were severely attenuated for overall growth compared with WT bacteria and this defect was rescued in Casp1/11 −/− mice ( Fig. 7c ). Analysis of bacterial loads by flow cytometry revealed far fewer ΔsifA Salmonella in CD11b(+) macrophages compared with WT Salmonella ( Fig. 7d ). However, macrophages from Casp1/11 −/− mice harboured numbers of ΔsifA Salmonella that were similar to those of WT bacteria in CD11b(+) cells from C57BL/6 mice ( Fig. 7d ), indicating caspase-1 and 11-dependent growth inhibition of cytosolic Salmonella in vivo . Discussion In the present work we analysed the fate of host cells and cytosolic bacterial growth at both whole population and single cell levels. Our two major findings are that (1) cells undergo a heterogeneous response upon bacterial infection: over a time course of several hours, not all cells containing cytosolic bacteria undergo cell lysis, and even in cells that lyse, the timing of loss of plasma membrane integrity varies widely. (2) Intracellular cytosolic bacterial growth can be inhibited either before or independently of the onset of host cell death; this process requires activity of both caspase-1 and 11. Control of cytosolic bacterial growth also involves the receptors NLRC4 and possibly NLRP3. In contrast, the absence of Gsdmd was not sufficient to alleviate growth attenuation up to 10 h. Furthermore, caspase-mediated processing of cytokines did not appear to be required as the absence of the adaptor protein ASC, the cytokine IL-18 or the receptor for IL-1β (IL-1r) did not yield increased growth of cytosolic bacteria. The Salmonella sifA mutant provides a convenient if artificial means to expose bacterial surface ligands to cytosolic receptors and to analyse the fate of cytosolic bacteria in macrophages. However, following phagocytosis, a small proportion of WT bacteria also lose their vacuolar membranes ( Fig. 3b ) 9 , and we found that caspase-1 and caspase-11 inhibited their cytosolic growth in both immortalized and primary macrophages. Therefore, the experiments involving ΔsifA Salmonella are applicable to WT bacteria, which could be potentially very detrimental to the host if they were to replicate in the nutrient-rich macrophage cytosol. The importance of caspase-mediated growth attenuation of cytosolic Salmonella was also revealed in non-phagocytic cells, where SPI-1 T3SS-dependent invasion results in a greater proportion of cytosolic bacteria. In addition, lack of growth inhibition in MEFs and inhibition of growth in 3T3 fibroblasts were directly correlated with the absence and presence of caspase-11, respectively. Increased cytosolic replication of Salmonella in human colonic epithelial cells following knock-down of caspase-4 was reported by Knodler et al . 24 34 . This was attributed to delayed shedding of host cells; however loss of a cytosolic growth inhibition might also have contributed to this phenotype. Previous studies have shown that caspases prevent cytosolic growth of Salmonella 23 24 and Legionella 49 through pyroptosis. In the latter case, degradation of cytosolic bacteria was also observed, and this was reduced in cells exposed to zVAD-FMK. However, it is not clear if cells containing degraded bacteria were intact or undergoing cell death. In agreement with several previous experiments 13 16 18 23 we found that not all cells that were exposed to cytosolic Salmonella undergo cell death. Several experiments showed that by 10 h post-uptake of ΔsifA Salmonella in C57BL/6 macrophages, cell death ranged from 20 to 30%, even though ∼70% of cells contained cytosolic bacteria ( Figs 3c,f and 6e ). If the macrophage cytosol is normally permissive for bacterial growth, then growth would be expected to occur in the cytosol of non-pyroptotic cells. However, by analysing bacterial load in single cells by flow cytometry and time-lapse microscopy in the presence of PI, it was clear that bacterial growth could be inhibited prior to the loss of plasma membrane integrity. In contrast, in the absence of caspase-1 and caspase-11 the macrophage cytosol was very permissive for bacterial growth, indicating additional functions for these inflammatory caspases. In support of this, Gsdmd, which is required for pyroptosis 46 47 , did not contribute significantly to inhibition of bacterial growth up to 10 h. This provides compelling evidence that the macrophage cytosol can restrict bacterial growth and that caspase-1 and caspase-11 have functions beyond the onset of cell death to mediate this activity. Other lines of evidence support a dual role for caspases in the control of cytosolic Salmonella. First, although WT and ΔsifA bacteria induced similar levels of cell death in 3T3 fibroblasts, growth of ΔsifA Salmonella only occurred after caspase inhibition ( Fig. 1a,b ). Second, expression of caspase-11 in MEFs was insufficient to elicit cell death (presumably due to the absence of other infection-induced proteins required for pyroptosis 9 26 50 ) but nevertheless reduced bacterial numbers significantly ( Fig. 2 ). Interestingly, caspase activation without concomitant cell death has also been reported to occur in Salmonella-infected neutrophils 51 . Non-pyroptotic caspase-mediated bacterial growth inhibition has previously been reported for vacuolar bacteria: caspase-1 regulates macrophage phagosome acidification (thereby contributing to killing of Staphylococcus aureus 31 ) and it promotes Legionella phagosome fusion with lysosomes 30 52 . Caspase-11 also appears to have additional functions, enhancing lysosomal fusion of Legionella vacuoles through modulation of cofilin, a regulator of actin polymerization 29 53 . Finally, Casp1/11 −/− iBMDMs produce reduced mROS and hydrogen peroxide, required for effective control of vacuolar Salmonella 32 . However, these mechanisms are unlikely to explain our findings on cytosolic Salmonella as caspase inhibition would be expected to influence vacuolar WT bacteria to a similar or greater extent. Autophagy can inhibit bacterial growth following the rupture of pathogen-containing vacuoles in epithelial cells 7 36 . In addition, members of the guanylate-binding protein family control intracellular bacterial growth by pyroptotic and non-pyroptotic mechanisms, including antibacterial autophagy and the induction of bacterial cell lysis by an unknown mechanism 9 54 55 . Autophagy is unlikely to account for the activities we described here for the following reasons: ΔsifA Salmonella are not targeted to the autophagic machinery in HeLa cells 7 and we detected a similar level of association of the autophagy marker LC3B to Salmonella after the addition of zVAD-FMK ( Supplementary Fig. 1C ). Therefore, our results suggest that additional substrate(s) of both caspase-1 and caspase-11 generate the production of antimicrobial activity within the cytosol before or without the onset of pyroptosis, adding to the increasing roles of these caspases beyond pyroptosis and cytokine processing. Several studies have identified putative caspase-1 substrates 56 57 including transcription factors, cytoskeletal components and glycolytic enzymes. Whether cytosolic antimicrobial activity might be due to an antimicrobial peptide, such as ubiquicidin 58 , limited cellular glycolysis when caspase-1 is active or through modulation of the cytoskeleton as described for vacuolar bacteria 29 32 52 awaits further investigation. The growth of sifA mutant Salmonella remained attenuated in Gsdmd −/− iBMDMs up to 10 h, but it is possible that direct GSDMD-mediated killing of bacteria 48 could occur at later time points. Mechanistically, it is noteworthy that self-cleavage of caspase-1, which is not required for cell death 42 is also not required for restriction of bacterial growth, highlighting the different functions of cleaved and uncleaved caspase-1. This suggests that substrates of processed caspase-1, such as pro-IL-1β and IL-18 are insufficient to explain our results. In line with this, the absence of ASC did not result in enhanced growth of the sifA mutant and IL-1r −/− or IL-18 −/− iBMDMs were still able to control replication of ΔsifA Salmonella. Many factors could account for the heterogeneous response to cytosolic bacteria. These include concentrations and/or availability of appropriate ligand, sensor, caspase enzyme and its substrate(s) all of which could vary from cell to cell. In addition, variation in caspase-11 levels could result from differential transcriptional upregulation after priming by agonists including LPS. Whereas caspase-11 can be activated through direct binding to cytosolic LPS (ref. 18 ), infection with Salmonella also activates caspase-1 via the NLRP3 and NLRC4 inflammasomes 44 . In this respect, absence of NLRC4 partially alleviated growth inhibition of the sifA mutant, which was further alleviated by exposure to KCl. This could represent a requirement for NLRP1 or NLRP3 but as C57BL/6 macrophages have been shown to have dysfunctional NLRP1b (ref. 59 ), it seems more likely that NLRP3 is involved. In Casp11 −/− macrophages, the addition of KCl did not significantly alter intracellular bacterial growth, suggesting non-canonical NLRP3 activation. Although caspase-11 contributes to the release of IL-1α, its cell death-inducing function appears to be independent to that of caspase-1 (refs 13 , 20 , 60 ). Our evidence indicates that caspase-1 and caspase-11 also function independently in their cell autonomous bacterial growth-suppressive activities. Growth of cytosolic Salmonella was significantly greater in Casp1/11 −/− compared with Casp11 −/− macrophages. In addition, microscopic analysis of infected cells also suggested independent activities of caspase-1 and caspase-11: caspase-1 was found predominantly in single inflammasome 'specks' following infection whereas caspase-11 was diffusely cytosolic and infrequently (<5%) associated with bacteria. Altogether, this suggests that caspase-1 and caspase-11 could employ distinct mechanisms to restrict bacterial replication within the cytosol. The relative contributions of caspase-1 and caspase-11 in the control of cytosolic Salmonella have been analysed following mixed infections of WT and sifA mutant bacteria in WT, Casp11 −/− and Casp1/11 −/− knock-out mice 23 . The equivalent competitive index values of the sifA mutant in Casp11 −/− and Casp1/11 −/− backgrounds suggested a major role for caspase-11 but not caspase-1 in the growth inhibition of sifA mutant bacteria. However, loss of caspase-1-mediated growth attenuation of WT Salmonella 16 could have masked an effect of caspase-1 on sifA mutant bacteria within the mixed infection. If so, this would be consistent with our results that show a clear function for both caspase-1 and caspase-11 in the growth control of cytosolic ΔsifA Salmonella. In conclusion, our experiments have revealed a surprising degree of heterogeneity in the response of host cells to cytosolic bacteria. We found that the catalytic activities of both caspase-1 and caspase-11 function to control growth of cytosolic Salmonella by both pyroptotic and non-pyroptotic mechanisms. Since vacuoles containing pathogenic or commensal bacteria can be ruptured through pathogen or host-dependent mechanisms 9 , caspase-dependent cytosolic growth inhibitory activity could prevent a wide variety of bacteria from cytosolic replication. Methods Antibodies and reagents Antibodies for immunoblotting were from Sigma (actin, AC-74 used at 1:5,000 dilution and caspase-11, 17D9 used at 1:1,000 dilution), Cell Signaling (caspase-7, 9492 used at 1:1,000 dilution), Adipogen (caspase-1 p20, AG-20B-0042 used at 1:1,000 dilution), DSHB (tubulin, E7 used at 1:5,000 dilution) and Santa Cruz Biotechnology (ASC sc-22514-R used at 1:1,000 dilution). Propidium iodide was from Life Technologies and Lipofectamine2000 from Invitrogen. siRNAs for caspase-7 and caspase-11 were purchased from Santa Cruz and used at 40 pmol. zVAD-FMK and YVAD-FMK were from R&D systems. Bacterial infections Salmonella enterica serovar Typhimurium (strain 12023) was grown overnight in LB. GFP-expressing Salmonella carry plasmid pFPV25.1, mCherry-expressing Salmonella carry plasmid pDiGc (ref. 61 ). prgH mutant Salmonella carry the plasmid pRI203, expressing Yersinia InvA (ref. 62 ). Bacteria (20 μl) were opsonized with 20 μl mouse serum (Sigma) in 170 μl DMEM for 20 min before addition of 600 μl DMEM. Macrophages (in 500 μl media in 24 well plates) were infected with 40 μl of opsonized bacteria (MOI 5–10), centrifuged at 110 g for 5 min and incubated for 25 min at 37 °C. Following two washes with PBS, cells were incubated with 100 μg ml −1 gentamicin for 2 h and then 20 μg ml −1 , or directly incubated with 20 μg ml −1 gentamicin. For SPI-1 T3SS-mediated invasion of 3T3 fibroblasts or MEFs, stationary phase bacterial cultures were sub-cultured (1:33) in fresh LB and grown for 3.5 h at 37 °C before inoculation. Cells in 24 well plates (500 μl media/well) were infected with 7 μl of sub-cultured bacteria for 7 min. After two PBS washes cells were incubated with 100 μg ml −1 gentamicin for 2 h and 20 μg ml −1 gentamicin thereafter. Cell culture C57BL/6 WT, Asc −/− , Casp11 −/− , Casp1/11 −/− , Casp7 −/− , IL-18 −/− , IL-1r −/− and Nlrc4 −/− BMDM were infected with the v-myc/v-raf expressing J2 retrovirus 63 , and differentiated in 20% L929-MCSF supernatant. Cells were then maintained in Dulbecco's modified Eagle medium (DMEM, Sigma), 10% fetal calf serum (FCS), 20% L929-MCSF and 1 mM sodium pyruvate at 37 °C, 5% CO 2 . 3T3 fibroblasts, MEFs, 293ETs and RAW 264.7 macrophages (ATCC) were cultured in DMEM containing 10% FCS. C57BL/6 control and Gsdmd −/− iBMDMs were maintained as described above. Cell lines, tested for mycoplasma, were chosen for ease of Salmonella infection, enabling analysis after both invasive and non-invasive uptake. Primary BMDMs were differentiated in 20% L929-MCSF supernatant for 1 week after isolation. For assays investigating the effect of caspase inhibitors, cells were incubated in DMEM with 10% FCS supplemented with 20 μM zVAD-FMK, 20 μM YVAD-FMK or 20 μM of other peptide inhibitors ( Supplementary Fig. 1D ) or DMSO (1:1000) as vehicle control, for 1 h before infection. When indicated KCl was added at 50 mM, 1 h before infection. Constructs and retroviral transductions Plasmids encoding GFP-tagged galectin-8 and LC3B were kind gifts from Dr Felix Randow and have previously been described 36 . Genes encoding murine caspase-1 or caspase-11 were ligated into a replication-defective retroviral plasmid (m6p) (ref. 64 ). Site directed mutagenesis was used to introduce mutations, which were verified by sequencing. Caspase-1 6D-N comprises 6 Asp to Asn mutations, preventing self-cleavage, while maintaining catalytic activity 42 . For transduction, retroviral particles were packaged into vesicular stomatitis virus pseudotyped virus after co-transfection of 293ET cells. After 48 h, cells were selected in puromycin (2.5 μg ml −1 ) or blasticidin (5 μg ml −1 ) so that all cells within a population expressed the transgene. Where GFP fusions were used, cells were sorted by Fluorescence-Activated Cell Sorting to obtain a 100% GFP-positive population. Colony forming unit assay To enumerate intracellular bacteria, cells from duplicate or triplicate wells of a 24 well plate, infected as above, were lysed in 1 ml of ice cold PBS containing 0.1% Triton X100 for 5 min. Serial dilutions were plated on duplicate LB agar and plates were incubated overnight at 37 °C. Colonies were counted using an Acolyte colony counter. Where CQ treatment was used (Sigma, 250 μM) it was added between 1.5 and 3 h (3T3 fibroblasts) or 2 and 4 h (WT infected iBMDMs) or 6 and 7 h ( ΔsifA Salmonella). For 3T3 fibroblasts the colony counts are represented as the fold growth in vacuolar bacteria (total—CQ resistant) and cytosolic bacteria (CQ resistant). For iBMDMs the fold growth in CQ-resistant bacteria (cytosolic) are shown. ELISA Concentrations of IL-1β in macrophage culture supernatants were measured using mouse IL-1β kits according to manufacturer's recommendations (Affymetrix ebioscience) after uptake of Salmonella Flow cytometry To measure the replication of GFP-expressing Salmonella in intact cells, cells were infected as above and harvested following trypsin treatment, washed and re-suspended in Optimem (Invitrogen) containing 1 μg ml −1 Propidium Iodide (PI). Data, consisting of at least 10,000 events, were acquired on a FACs Calibur and analysed using FlowJo 8.8.6. Data are represented as the fold-change (from 1 or 2 h p.u.) in geometric mean of cells harbouring GFP-expressing bacteria. Immunoblotting Proteins in post nuclear supernatants from 1 × 10 6 cells were separated on either 10% or 12% Tris polyacrylamide gels. Proteins were transferred to Nitrocellulose membranes, which were then blocked in 5% milk in TBST (100 mM Tris Cl pH 7.4, 150 mM NaCl, 0.1% Tween20). Membranes were incubated overnight at 4 °C with primary antibodies, washed three times with TBST and then incubated for 2 h with secondary antibodies at room temperature. Visualization was done using ECL+detection regents (GE Healthcare). Uncropped blots are shown in Supplementary Fig. 6 . LDH cytotoxicity assay Host cell death was measured as a percentage of total LDH release, according to the recommended protocol (Promega). Medium was used as a blank control to obtain background measurements and supernatants from non-infected samples were subtracted from infected conditions. Total LDH release was measured after cell lysis at −80 °C. Mice For primary BMDMs, Caspase-1/11 double knockout mice were from the Swiss Immunological Mouse Repository (SwImMR) and caspase-7 −/− mice were purchased from Jackson Laboratories. Casp11 −/− primary BMDMs had been isolated from previously described mice 65 . C57BL/6 control mice were from Charles Rivers. All animals were bred in accordance with accredited animal facility regulations at Imperial College London. Imperial College Animal Welfare and Ethical Review Body (AWERB) granted approval for all mouse work. iBMDMs, that have been previously described were prepared from Nlrc4 −/− , Asc −/− (ref. 45 ), Casp1/11 −/− (ref. 66 ) and Casp11 −/− (ref. 65 ) mice. For mouse infections, mice (8–10 week old, C57BL/6 or Casp1/11 −/− ) were inoculated intraperitoneal with 1 × 10 5 CFUs with GFP expressing WT or ΔsifA Salmonella. After 48 h, spleens were collected, homogenized and splenic CD11b(+) cells enriched using magnetic beads according to the manufacturer instructions (Miltenyi Biotec). Purified cells were analysed by flow cytometry in Optimem containing PI. Mice showing very poor infection of the spleen were excluded. Randomization and blinding were not used. Microscopy and digitonin assays Cells were seeded on glass cover slips one-day prior to infection and fixed in 4% paraformaldehyde for 20 min. Confocal images were taken on a Zeiss 710 microscope with a × 100 objective. For digitonin-mediated permeabilisation of the plasma membrane, live cells were treated with 40 μg ml −1 digitonin for 5 min on ice prior to immunolabelling with anti-CSA1 (1:400, Kirkegaard and Perry Laboratories), anti-GM130 (1:500, BD Transduction laboratories) and anti-PDI (Protein disulfide-isomerase, 1:100, Enzo) for 30 min on ice. Cells were then washed twice in PBS and fixed in 4% paraformaldehyde. After permeabilisation in PBS, 0.1% Triton X100 and 10% horse serum, cover slips were incubated with appropriate AlexaFluor secondary antibodies (Invitrogen) and DAPI (4′,6-Diamidino-2-Phenylindole, Dihydrochloride) for 30 min before mounting onto glass slides. PI uptake PI uptake was used to determine plasma membrane integrity. Macrophages (3 × 10 5 cells per ml) were seeded in white clear-bottomed 96-well plates (Greiner) and infected with opsonised late stationary phase Salmonella (MOI 10:1) for 30 min at 37 °C. Following infection, cells were washed twice with PBS and 200 μl Optimem medium containing 10% FCS, 20 μg ml −1 gentamicin and 1 μg ml −1 PI was added. Triton X-100 (0.1%) was included in Optimem medium in wells used for positive controls. Optimem medium without PI was added to negative control wells. Plates were incubated at 37 °C in 5% CO 2 within a Tecan Infinite M200PRO fluorescent plate reader throughout infection, with PI fluorescence measured every 15 min. Non-infected controls were subtracted from infected samples and then divided by the fluorescence of wells treated with Triton-X100 to give the relative PI uptake. Quantitative reverse transcriptase (RT)-PCR Total RNA was isolated from 1 × 10 6 cells (Qiagen RNAeasy mini kit) and 400 ng was used to synthesize complementary DNA (cDNA) according to manufactures recommendations (QuantiTect RT kit, Qiagen). cDNA (0.5 μl) was used in quantitative RT-PCRs (SybrGreen PCR master mix, Applied biosystems) containing 0.2 μM gene-specific primers. After determining the cycle threshold (Ct) required to reach a significant emission of Sybr Green reporter dye (Rotor-Gene 3000, Corbett Research), relative mRNA was calculated from a titration curve of cDNA. Data represent the relative amounts of mRNA, normalized to rps9 house keeping gene. The following primers were used: Caspase-1 For 5′- ACTGGGACCCTCAAGTTTTG, Rev 5′- CATCTCCAGAGCTGTGAG Caspase-7 For 5′-TGGAAAAGGTGGATTCTTCC, Rev 5′-CTTTGTCGAAGTTCTTGTTG Caspase-11 For 5′-AAACACCCTGACAAACCACT, Rev 5′-TTCCTCCATTTCCAGATTAG Rps9 For 5′- CTGGACGAGGGCAAGATGAAGC-3′ Rev 5′- TGACGTTGGCGGATGAGCACA-3′ Time-lapse microscopy Cells seeded in dishes (Matek) with an embedded glass cover slip were infected as above. Prior to imaging, medium was replaced with Optimem (Invitrogen) containing 10% FCS, 40 μM Hepes (Sigma), 20 μg ml −1 gentamicin and 1 μg ml −1 PI. Cells were maintained at 37 °C in a heated chamber and images were acquired at 20-min intervals using the × 40 objective on a Zeiss 710 confocal microscope. At least 100 infected cells were scored per experiment. Statistics Either an one-way ANOVA with Dunnett's multiple comparisons test or one and two-tailed unpaired equal variance Students t -test were used for statistical comparison from three or more independent experiments as indicated. * P <0.05, ** P <0.01, *** P <0.001. Experiments were performed with at least three independent replicates, except for the analysis of time-lapse microscopy. When possible, pilot data with a type I error rate of 5% was used to determine an appropriate sample size. Data availability The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information Files . Antibodies and reagents Antibodies for immunoblotting were from Sigma (actin, AC-74 used at 1:5,000 dilution and caspase-11, 17D9 used at 1:1,000 dilution), Cell Signaling (caspase-7, 9492 used at 1:1,000 dilution), Adipogen (caspase-1 p20, AG-20B-0042 used at 1:1,000 dilution), DSHB (tubulin, E7 used at 1:5,000 dilution) and Santa Cruz Biotechnology (ASC sc-22514-R used at 1:1,000 dilution). Propidium iodide was from Life Technologies and Lipofectamine2000 from Invitrogen. siRNAs for caspase-7 and caspase-11 were purchased from Santa Cruz and used at 40 pmol. zVAD-FMK and YVAD-FMK were from R&D systems. Bacterial infections Salmonella enterica serovar Typhimurium (strain 12023) was grown overnight in LB. GFP-expressing Salmonella carry plasmid pFPV25.1, mCherry-expressing Salmonella carry plasmid pDiGc (ref. 61 ). prgH mutant Salmonella carry the plasmid pRI203, expressing Yersinia InvA (ref. 62 ). Bacteria (20 μl) were opsonized with 20 μl mouse serum (Sigma) in 170 μl DMEM for 20 min before addition of 600 μl DMEM. Macrophages (in 500 μl media in 24 well plates) were infected with 40 μl of opsonized bacteria (MOI 5–10), centrifuged at 110 g for 5 min and incubated for 25 min at 37 °C. Following two washes with PBS, cells were incubated with 100 μg ml −1 gentamicin for 2 h and then 20 μg ml −1 , or directly incubated with 20 μg ml −1 gentamicin. For SPI-1 T3SS-mediated invasion of 3T3 fibroblasts or MEFs, stationary phase bacterial cultures were sub-cultured (1:33) in fresh LB and grown for 3.5 h at 37 °C before inoculation. Cells in 24 well plates (500 μl media/well) were infected with 7 μl of sub-cultured bacteria for 7 min. After two PBS washes cells were incubated with 100 μg ml −1 gentamicin for 2 h and 20 μg ml −1 gentamicin thereafter. Cell culture C57BL/6 WT, Asc −/− , Casp11 −/− , Casp1/11 −/− , Casp7 −/− , IL-18 −/− , IL-1r −/− and Nlrc4 −/− BMDM were infected with the v-myc/v-raf expressing J2 retrovirus 63 , and differentiated in 20% L929-MCSF supernatant. Cells were then maintained in Dulbecco's modified Eagle medium (DMEM, Sigma), 10% fetal calf serum (FCS), 20% L929-MCSF and 1 mM sodium pyruvate at 37 °C, 5% CO 2 . 3T3 fibroblasts, MEFs, 293ETs and RAW 264.7 macrophages (ATCC) were cultured in DMEM containing 10% FCS. C57BL/6 control and Gsdmd −/− iBMDMs were maintained as described above. Cell lines, tested for mycoplasma, were chosen for ease of Salmonella infection, enabling analysis after both invasive and non-invasive uptake. Primary BMDMs were differentiated in 20% L929-MCSF supernatant for 1 week after isolation. For assays investigating the effect of caspase inhibitors, cells were incubated in DMEM with 10% FCS supplemented with 20 μM zVAD-FMK, 20 μM YVAD-FMK or 20 μM of other peptide inhibitors ( Supplementary Fig. 1D ) or DMSO (1:1000) as vehicle control, for 1 h before infection. When indicated KCl was added at 50 mM, 1 h before infection. Constructs and retroviral transductions Plasmids encoding GFP-tagged galectin-8 and LC3B were kind gifts from Dr Felix Randow and have previously been described 36 . Genes encoding murine caspase-1 or caspase-11 were ligated into a replication-defective retroviral plasmid (m6p) (ref. 64 ). Site directed mutagenesis was used to introduce mutations, which were verified by sequencing. Caspase-1 6D-N comprises 6 Asp to Asn mutations, preventing self-cleavage, while maintaining catalytic activity 42 . For transduction, retroviral particles were packaged into vesicular stomatitis virus pseudotyped virus after co-transfection of 293ET cells. After 48 h, cells were selected in puromycin (2.5 μg ml −1 ) or blasticidin (5 μg ml −1 ) so that all cells within a population expressed the transgene. Where GFP fusions were used, cells were sorted by Fluorescence-Activated Cell Sorting to obtain a 100% GFP-positive population. Colony forming unit assay To enumerate intracellular bacteria, cells from duplicate or triplicate wells of a 24 well plate, infected as above, were lysed in 1 ml of ice cold PBS containing 0.1% Triton X100 for 5 min. Serial dilutions were plated on duplicate LB agar and plates were incubated overnight at 37 °C. Colonies were counted using an Acolyte colony counter. Where CQ treatment was used (Sigma, 250 μM) it was added between 1.5 and 3 h (3T3 fibroblasts) or 2 and 4 h (WT infected iBMDMs) or 6 and 7 h ( ΔsifA Salmonella). For 3T3 fibroblasts the colony counts are represented as the fold growth in vacuolar bacteria (total—CQ resistant) and cytosolic bacteria (CQ resistant). For iBMDMs the fold growth in CQ-resistant bacteria (cytosolic) are shown. ELISA Concentrations of IL-1β in macrophage culture supernatants were measured using mouse IL-1β kits according to manufacturer's recommendations (Affymetrix ebioscience) after uptake of Salmonella Flow cytometry To measure the replication of GFP-expressing Salmonella in intact cells, cells were infected as above and harvested following trypsin treatment, washed and re-suspended in Optimem (Invitrogen) containing 1 μg ml −1 Propidium Iodide (PI). Data, consisting of at least 10,000 events, were acquired on a FACs Calibur and analysed using FlowJo 8.8.6. Data are represented as the fold-change (from 1 or 2 h p.u.) in geometric mean of cells harbouring GFP-expressing bacteria. Immunoblotting Proteins in post nuclear supernatants from 1 × 10 6 cells were separated on either 10% or 12% Tris polyacrylamide gels. Proteins were transferred to Nitrocellulose membranes, which were then blocked in 5% milk in TBST (100 mM Tris Cl pH 7.4, 150 mM NaCl, 0.1% Tween20). Membranes were incubated overnight at 4 °C with primary antibodies, washed three times with TBST and then incubated for 2 h with secondary antibodies at room temperature. Visualization was done using ECL+detection regents (GE Healthcare). Uncropped blots are shown in Supplementary Fig. 6 . LDH cytotoxicity assay Host cell death was measured as a percentage of total LDH release, according to the recommended protocol (Promega). Medium was used as a blank control to obtain background measurements and supernatants from non-infected samples were subtracted from infected conditions. Total LDH release was measured after cell lysis at −80 °C. Mice For primary BMDMs, Caspase-1/11 double knockout mice were from the Swiss Immunological Mouse Repository (SwImMR) and caspase-7 −/− mice were purchased from Jackson Laboratories. Casp11 −/− primary BMDMs had been isolated from previously described mice 65 . C57BL/6 control mice were from Charles Rivers. All animals were bred in accordance with accredited animal facility regulations at Imperial College London. Imperial College Animal Welfare and Ethical Review Body (AWERB) granted approval for all mouse work. iBMDMs, that have been previously described were prepared from Nlrc4 −/− , Asc −/− (ref. 45 ), Casp1/11 −/− (ref. 66 ) and Casp11 −/− (ref. 65 ) mice. For mouse infections, mice (8–10 week old, C57BL/6 or Casp1/11 −/− ) were inoculated intraperitoneal with 1 × 10 5 CFUs with GFP expressing WT or ΔsifA Salmonella. After 48 h, spleens were collected, homogenized and splenic CD11b(+) cells enriched using magnetic beads according to the manufacturer instructions (Miltenyi Biotec). Purified cells were analysed by flow cytometry in Optimem containing PI. Mice showing very poor infection of the spleen were excluded. Randomization and blinding were not used. Microscopy and digitonin assays Cells were seeded on glass cover slips one-day prior to infection and fixed in 4% paraformaldehyde for 20 min. Confocal images were taken on a Zeiss 710 microscope with a × 100 objective. For digitonin-mediated permeabilisation of the plasma membrane, live cells were treated with 40 μg ml −1 digitonin for 5 min on ice prior to immunolabelling with anti-CSA1 (1:400, Kirkegaard and Perry Laboratories), anti-GM130 (1:500, BD Transduction laboratories) and anti-PDI (Protein disulfide-isomerase, 1:100, Enzo) for 30 min on ice. Cells were then washed twice in PBS and fixed in 4% paraformaldehyde. After permeabilisation in PBS, 0.1% Triton X100 and 10% horse serum, cover slips were incubated with appropriate AlexaFluor secondary antibodies (Invitrogen) and DAPI (4′,6-Diamidino-2-Phenylindole, Dihydrochloride) for 30 min before mounting onto glass slides. PI uptake PI uptake was used to determine plasma membrane integrity. Macrophages (3 × 10 5 cells per ml) were seeded in white clear-bottomed 96-well plates (Greiner) and infected with opsonised late stationary phase Salmonella (MOI 10:1) for 30 min at 37 °C. Following infection, cells were washed twice with PBS and 200 μl Optimem medium containing 10% FCS, 20 μg ml −1 gentamicin and 1 μg ml −1 PI was added. Triton X-100 (0.1%) was included in Optimem medium in wells used for positive controls. Optimem medium without PI was added to negative control wells. Plates were incubated at 37 °C in 5% CO 2 within a Tecan Infinite M200PRO fluorescent plate reader throughout infection, with PI fluorescence measured every 15 min. Non-infected controls were subtracted from infected samples and then divided by the fluorescence of wells treated with Triton-X100 to give the relative PI uptake. Quantitative reverse transcriptase (RT)-PCR Total RNA was isolated from 1 × 10 6 cells (Qiagen RNAeasy mini kit) and 400 ng was used to synthesize complementary DNA (cDNA) according to manufactures recommendations (QuantiTect RT kit, Qiagen). cDNA (0.5 μl) was used in quantitative RT-PCRs (SybrGreen PCR master mix, Applied biosystems) containing 0.2 μM gene-specific primers. After determining the cycle threshold (Ct) required to reach a significant emission of Sybr Green reporter dye (Rotor-Gene 3000, Corbett Research), relative mRNA was calculated from a titration curve of cDNA. Data represent the relative amounts of mRNA, normalized to rps9 house keeping gene. The following primers were used: Caspase-1 For 5′- ACTGGGACCCTCAAGTTTTG, Rev 5′- CATCTCCAGAGCTGTGAG Caspase-7 For 5′-TGGAAAAGGTGGATTCTTCC, Rev 5′-CTTTGTCGAAGTTCTTGTTG Caspase-11 For 5′-AAACACCCTGACAAACCACT, Rev 5′-TTCCTCCATTTCCAGATTAG Rps9 For 5′- CTGGACGAGGGCAAGATGAAGC-3′ Rev 5′- TGACGTTGGCGGATGAGCACA-3′ Time-lapse microscopy Cells seeded in dishes (Matek) with an embedded glass cover slip were infected as above. Prior to imaging, medium was replaced with Optimem (Invitrogen) containing 10% FCS, 40 μM Hepes (Sigma), 20 μg ml −1 gentamicin and 1 μg ml −1 PI. Cells were maintained at 37 °C in a heated chamber and images were acquired at 20-min intervals using the × 40 objective on a Zeiss 710 confocal microscope. At least 100 infected cells were scored per experiment. Statistics Either an one-way ANOVA with Dunnett's multiple comparisons test or one and two-tailed unpaired equal variance Students t -test were used for statistical comparison from three or more independent experiments as indicated. * P <0.05, ** P <0.01, *** P <0.001. Experiments were performed with at least three independent replicates, except for the analysis of time-lapse microscopy. When possible, pilot data with a type I error rate of 5% was used to determine an appropriate sample size. Data availability The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information Files . Additional information How to cite this article: Thurston, T. L. M. et al . Growth inhibition of cytosolic Salmonella by caspase-1 and caspase-11 precedes host cell death. Nat. Commun. 7, 13292 doi: 10.1038/ncomms13292 (2016). Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary Material Supplementary Information Supplementary Figures 1-6 Supplementary Movie 1 Time lapse imaging of C57BL/6 iBMDMs from 2 hours post-uptake with GFP-WT Salmonella in the presence of Propidium iodide (red). Images were acquired every 15 minutes on a Zeiss 710 confocal microscope with the 40x objective. Supplementary Movie 2 Time lapse imaging of C57BL/6 iBMDMs from 2 hours post-uptake with GFP-ΔsifA Salmonella in the presence of Propidium iodide (red). Images were acquired every 15 minutes on a Zeiss 710 confocal microscope with the 40x objective. Supplementary Movie 3 Time lapse imaging of Casp11-/- iBMDMs from 2 hours post-uptake with GFP-ΔsifA Salmonella in the presence of Propidium iodide (red). Images were acquired every 15 minutes on a Zeiss 710 confocal microscope with the 40x objective. Supplementary Movie 4 Time lapse imaging of Casp1/11-/- iBMDMs from 2 hours post-uptake with GFP-ΔsifA Salmonella in the presence of Propidium iodide (red). Images were acquired every 15 minutes on a Zeiss 710 confocal microscope with the 40x objective.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4012540/

IN VIVO EFFICACY OF A PHOSPHODIESTER TLR-9 APTAMER AND ITS BENEFICIAL EFFECT IN A PULMANARY ANTHRAX INFECTION MODEL

Immunostimulatory oligonucleotide (ISS-ODN) used as adjuvants are commonly modified with phosphorothioate (PS). The PS backbone prevents nuclease degradation, but confers undesired side effects, including systemic cytokine release. Previously, R10-60, a phosphodiester (PO) ISS-ODN, was structurally optimized as an intracellular Toll-like receptor-9 agonist. Here intravenous, intradermal and intranasal administration of PO R10-60 elicit local or adaptive immune responses with minimal systemic effects compared to a prototypic PS ISS-ODN in mice. Furthermore, prophylactic intranasal administration of PO R10-60 significantly delayed death in mice exposed to respiratory anthrax comparable to the PS ISS-ODN. The pattern of cytokine release suggested that early IL-1 β production might contribute to this protective effect, which was replicated with recombinant IL-1 β injections during infection. Hence, the transient effects from a PO TLR-9 agonist may be beneficial for protection in a bacterial bioterrorism attack, by delaying the onset of systemic infection without the induction of a cytokine syndrome.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8535822/

NLRC4 inflammasome–dependent cell death occurs by a complementary series of three death pathways and determines lethality in mice

The death of cells elicited by NLRC4 inflammasome overactivation plays a decisive role in lethal outcome of the mice. INTRODUCTION Inflammasomes are multiprotein complexes assembled by pattern recognition receptor (PRR) proteins in response to pathogen-associated molecular patterns and sterile stress signals ( 1 ). Canonical inflammasomes serve as key cellular activation platforms for caspase-1, whereas noncanonical inflammasomes elicit the activation of murine caspase-11 or human caspase-4 and caspase-5 ( 2 – 4 ). Activation of both types of inflammasomes leads to maturation of pro–interleukin-1β (IL-1β), pro–IL-18, and pyroptosis ( 1 , 2 , 5 , 6 ). Cleavage of gasdermin D (GSDMD) into N-terminal (N-GSDMD) and C-terminal (C-GSDMD) by caspase-1 is required for pyroptosis ( 7 – 9 ). The cleaved N-GSDMD fragment forms large oligomeric membrane pores that lead to pyroptosis, which facilitates IL-1β secretion ( 10 , 11 ). IL-1β and IL-18 secretions are part of the innate immune defense functions of inflammasomes, and pyroptosis could further enhance the release of these cytokines ( 7 ). Although "cytokine storm" is widely proposed to be responsible for the acute lethality of animal induced by systematic inflammation ( 5 , 12 ), inflammatory lipid mediators may also participate in the lethal inflammatory responses in vivo ( 13 ). It has been demonstrated that inflammasome-mediated cell death is important for controlling infected pathogens, and limiting pathogen(s) in vivo is critical for preventing host from mortality ( 14 – 18 ). However, hyperactivation of inflammasomes alone can also lead to organ damage and host death ( 13 , 17 ). The contribution of inflammasome-mediated cell death to animal death has remained unclear. A canonical inflammasome complex is typically composed of a PRR protein, such as NLRP3 [NACHT, leucine-rich repeat, and pyrin domain (PYD)–containing protein 3], NLRP1, NLRC4 [NLR family caspase recruitment domain (CARD)–containing protein 4] or absent in melanoma-2 (AIM2) and pyrin, the adaptor protein ASC (apoptosis-associated speck-like protein), and caspase-1 ( 1 , 19 ). In the cases of NLRP3 and AIM2 inflammasomes, their interaction with caspase-1 is mediated by ASC, which interacts with NLRP3 or AIM2 through its PYD and with caspase-1 through its CARD. These proteins form structures known as "specks," where the activation and cleavage of caspase-1 occur ( 1 , 19 ). As for NLRC4 and murine NLRP1b that lack a PYD domain, specks can be formed by recruiting ASC directly through their CARD. Different from NLRP3 and AIM2, NLRC4 and NLRP1b can directly recruit caspase-1 via their CARD and activate caspase-1–mediated pyroptosis in an ASC-independent manner. However, specks formed with ASC still have promoting effects on the activation of NLRC4 and NLRP1b inflammasomes ( 1 , 19 ). It was reported that genetic deletion of Gsdmd or Caspase - 1 ( Casp1 ) blocked NLRP3 inflammasome–mediated pyroptosis, but cells still died via an apoptotic pathway ( 7 , 20 , 21 ). In this case, caspase-8 and ASC have been shown to be responsible for the apoptosis ( 7 , 21 – 23 ). Caspase-8 was also reported to compensate the loss of caspase-1 in NLRC4 inflammasome, and the Rip3 −/− Casp1 −/− Casp8 −/− or Asc −/− Casp1 −/− Casp11 −/− mice were resistant to NLRC4-elicited hypothermia ( 17 ), which is known to often precede mouse death. A very recent report also showed flexible usage and switches of diverse cell death pathways in Salmonella -treated macrophages ( 15 ). To fully understand inflammasome-mediated cell death, we systematically studied switches from pyroptosis to apoptosis in canonical inflammasomes, including those inflammasomes where ASC is not obligatory. We showed that there are three cell death pathways that are adopted in a specific and preferential order downstream of inflammasome activation, and only blocking all of them can prevent cell death. We used flagellin-mediated NLRC4 inflammasome hyperactivation to evaluate whether and how cell death influences the outcome of mice. Strikingly, we found that as long as there is a remaining inflammasome-dependent cell death pathway, regardless of pyroptosis or apoptosis, the tissue/organ damages occur, and animals succumb. By this, we mean that it is cell death that plays a decisive role in the lethality of animals caused by inflammasome overactivation. In line with this result, impairment of the death pathways downstream of NLRC4 attenuated mice death caused by infection of nonpropagative Salmonella . These data suggest that blocking death pathways might be beneficial to the lethal infections when the propagation of pathogen(s) has been effectively inhibited by other therapeutic approaches. RESULTS The occurrence of cell death and subsequent cytolysis correlates with tissue damage and mouse death triggered by NLRC4 inflammasome Flagellin is a subunit protein of bacterial flagellum. FlaTox, a combination of recombinant flagellin fused to the amino-terminal domain of Bacillus anthracis lethal factor (LFn-Fla) and anthrax protective antigen (PA), was developed to selectively activate NLRC4 ( 13 ). LFn-Fla can enter cells by PA-mediated endocytosis, leading to NLRC4-dependent pyroptosis, small intestinal damage, and mouse death ( Fig. 1, A and B , and fig. S1, A to F) ( 13 , 16 , 17 , 24 ). Pyroptosis is a type of necrotic cell death with membrane disruption, which can be measured by lactate dehydrogenase (LDH) release. NLRC4-dependent pyroptosis and IL-1β release were detected in peritoneal lavage and blood of FlaTox-treated mice (fig. S1, G, H, K, and L). It was reported that extracellular flagellin induces tumor necrosis factor (TNF) and IL-6 production via an NLRC4-independent but Toll-like receptor 5 (TLR5)–dependent mechanism ( 25 ). As expected, Nlrc4 deletion only slightly reduced the TNF and IL-6 levels in peritoneal lavage from FlaTox-treated mice (fig. S1, I and J). The levels of TNF and IL-6 in the blood of FlaTox-treated Nlrc4 −/− mice were notable lower than those in wild-type (WT) mice (fig. S1, M and N). The LFn-Fla that we used can induce TNF and IL-6 production in bone marrow–derived macrophages independent of NLRC4 in the absence of PA (fig. S1, O to R). The local (peritoneal cavity) TNF and IL-6 should be induced by intraperitoneally injected FlaTox, and the circulating (blood) TNF and IL-6 ought to be controlled by other factor(s) that was directly or indirectly influenced by NLRC4 inflammasome activation. Fig. 1. Release of LDH, a cytolysis indicator, but not inflammatory cytokines or lipid mediators into peritoneal lavages is correlated with NLRC4 inflammasome–triggered mouse death. ( A ) Mice of indicated genotypes were intraperitoneally injected with FlaTox [LFn-Fla (4 μg/g body weight) and PA (2 μg/g body weight)] and monitored for survival rate ( n = 5). ( B ) Hematoxylin and eosin (H&E) staining of small intestinal tissue of mice treated as indicated for 2 hours. Scale bars, 100 μm. ( C to F ) Mice of indicated genotypes were intraperitoneally injected with FlaTox. Peritoneal lavages were isolated by using 1 ml of PBS at 1 and 2 hours after FlaTox challenge and then subjected to measurements of LDH (C), IL-1β (D), TNF (E), and IL-6 (F). n = 1 for PBS group (Ctrl) and n = 3 for FlaTox-treated group. ( G and H ) Mice of indicated genotypes were intraperitoneally injected with FlaTox for indicated times, and peritoneal lavages were collected for liquid chromatography tandem mass spectrometry (LC/MS/MS)–based lipidomics ( n = 3). ( I to K ) Mice of indicated genotypes were pretreated intraperitoneally for 30 min with of COX-1 inhibitor (SC-560) (5 μg/g body weight). Mice were intraperitoneally injected with FlaTox for indicated times, and peritoneal lavages were collected for LC/MS/MS-based lipidomics ( n = 3). Survival rate was monitored (K) ( n = 5). Since caspase-1 and GSDMD are downstream of NLRC4 to mediate pyroptosis and the release of IL-1β and IL-18, we sought to examine the effect of Gsdmd and Casp1 deletion in vivo. As we have Casp1 and Casp11 double knockout (KO) ( Casp1/11 −/− ) mice, and knowing that caspase-11 does not regulate FlaTox-induced NLRC4 inflammasome activation ( 17 ), we used Casp1/11 −/− mice in our experiments. Slightly different from the published data on FlaTox-induced pathology and body temperature drop ( 17 ), Gsdmd −/− and Casp1/11 −/− mice behaved more like WT mice in FlaTox-induced small intestinal damage and death ( Fig. 1, A and B ). It is known that deletion of Gsdmd or Casp1 blocks pyroptosis as well as IL-1β and IL-18 release ( 1 , 7 – 9 ), but we found that their deletion is quite different from Nlrc4 deletion, which totally blocks FlaTox-induced mouse death ( Fig. 1, A and B ). Thus, a more complicated mechanism is underneath the in vivo responses to FlaTox. To address the mechanism of in vivo responses to FlaTox, we analyzed FlaTox-induced pyroptosis ex vivo. We collected peritoneal lavages at 1 and 2 hours after FlaTox injection and measured the release of LDH, IL-1β, TNF, and IL-6. Deletion of Gsdmd or Casp1 effectively blocked LDH release at 1 hour. However, the level of LDH release into peritoneal cavity became comparable among Gsdmd −/− , Casp1 −/− , and WT mice at later times ( Fig. 1C ), indicating that cytolysis still occurred. It is known that in cultured cells, pyroptotic stimuli can trigger apoptosis when Casp1 or Gsdmd was deleted ( 7 , 20 , 21 ). Therefore, LDH release in Gsdmd −/− , Casp1 −/− mice at later times could be attributed to secondary necrotic lysis of the apoptotic cells or caspase-3–mediated pyroptosis via GSDME cleavage ( 26 , 27 ). However, the latter hypothesis is unlikely as no difference was observed between Gsdmd −/− mice and Gsdmd −/− Gsdme −/− mice (fig. S2, A to H). IL-1β secretion into peritoneal cavity was inhibited to different extents by deletion of Gsdmd −/− or Casp1 −/− at 1 hour, and continued inhibition was observed in Casp1 −/− but not Gsdmd −/− mice ( Fig. 1D ). The blockade of IL-1β release in Casp1 −/− mice should be resulted from failure to process IL-1β to its mature form ( 28 ), and the inhibited IL-1β release did not reduce animal death, supporting the conclusion by a previous report that flagellin-induced mouse death is independent of IL-1β and IL-18 production ( 13 ). The release of TNF and IL-6 into peritoneal cavity was not affected by gene deletion of Gsdmd or Casp1 ( Fig. 1, E and F ) because FlaTox-induced TNF and IL-6 production is NLRC4 pathway independent ( Fig. 1, E and F ). We then analyzed blood from FlaTox-treated mice and found that deletion of Gsdmd did not affect the increase of circulating LDH, IL-1β, TNF, and IL-6, and the Casp1 deletion also had no effect except for affecting IL-1β release (fig. S1, S to V). By measuring the correlations, we observed that the increase of blood TNF and IL-6 in FlaTox-treated mice is associated with LDH release but not IL-1β release. Here, it should be noted that TNF is not required for the death of FlaTox-treated mice as FlaTox was equally toxic to WT and Tnfr1 −/− mice (fig. S1W). Since inflammatory signaling lipids such as prostaglandins and leukotrienes are known to play a role in inflammasome-mediated pathological effects, we measured representatives of these lipids in peritoneal lavages at 0, 20, and 60 min after FlaTox injection. The induction of the eicosanoid lipids generated by lipoxygenase was almost completely blocked in both Gsdmd −/− and Casp1/11 −/− mice ( Fig. 1G ). Some of the cyclooxygenase (COX)–produced eicosanoids, prostaglandin D 2 (PGD 2 ), and thromboxane B2 (TXB 2 ) were significantly inhibited in Gsdmd −/− and Casp1/11 −/− mice, whereas the inhibition was slight on the production of prostaglandin E 2 (PGE 2 ) ( Fig. 1H ). We used COX inhibitor SC-560, which effectively inhibited the increase of PGE 2 and TXB 2 and, to a less extent, of PGD 2 in peritoneal lavages of FlaTox-treated mice ( Fig. 1I ), and had little or no effect on the other lipids ( Fig. 1J ). FlaTox-induced death of mice was not affected by SC-560 ( Fig. 1K ). Thus, none of these eicosanoid lipids is required for the lethality of FlaTox-treated mice ( Fig. 1, A and G to K ). Collectively, as for FlaTox-induced mouse death per se, TNF, IL-1β/18 induction, and "eicosanoid storm" seem to be dispensable, and only LDH release was found to have correlation with FlaTox-induced mouse death. Thus, cytolysis in vivo is likely the major risk of animal death. Cell death driven by NLRC4 and other canonical inflammasomes could be mediated by apoptosis when pyroptosis was blocked To understand why LDH release still occurred when pyroptotic pathway was already blocked in FlaTox-treated mice, we explored possible switches in death pathways. It was reported that ASC in NLRP3 or AIM2 complex can recruit caspase-8 to trigger apoptosis when Casp1 was deleted ( 20 , 21 ). We showed previously that Gsdmd deletion also switched NLRP3 inflammasome–mediated pyroptosis to apoptosis ( 7 ) and here that Gsdmd deletion has the same effect on cell death induced by AIM2 inflammasome (fig. S3, A to C). Caspase-1–GSDMD axis controls pyroptosis, and deletion of either Casp1 or Gsdmd switches NLRP3- and AIM2-mediated pyroptosis to apoptosis. To determine whether the switch to apoptosis also occurs in the activation of inflammasomes such as NLRC4 where ASC is not required, we used peritoneal macrophages from Gsdmd −/− mice. FlaTox induced pyroptosis in vitro in peritoneal macrophages as measured by LDH release, which was blocked by Gsdmd gene deletion ( Fig. 2A ). However, cytolysis still occurred at later times and reached a comparable level to that of WT macrophages at 8 hours ( Fig. 2A ). To understand what caused the later cytolysis, we used intracellular adenosine triphosphate (ATP) loss as a general indicator of cell death. Deletion of Gsdmd only inhibited about 20% of cell death at 4 hours ( Fig. 2B ), suggesting that cell death in a form other than pyroptosis occurred in Gsdmd −/− macrophages, and this type of cell death might be responsible for the later cytolysis. To better define the cell death pathway, we measured activities of different caspases. Gsdmd deletion did not affect FlaTox-induced autocleavage of caspase-1 as revealed by Western blotting ( Fig. 2C ). In contrast, activation of caspase-8 and caspase-3 was detected in FlaTox-treated Gsdmd −/− but not WT cells ( Fig. 2, D and E ), indicating that activation of pathways leads to apoptosis in Gsdmd −/− cells. Western blotting analysis also showed the cleavage of caspase -8 and caspase-3 in Gsdmd −/− cells ( Fig. 2C ), supporting our speculation that there is activation of an apoptotic pathway. Fig. 2. The effects of genetic deletion of Gsdmd , Caspase-1/11 , or Asc on NLRC4 inflammasome–mediated IL-1β secretion, cell death and type of cell death, and mice death. ( A to E ) Peritoneal macrophages of indicated genotypes were treated with FlaTox [LFn-Fla (2 μg/ml) and PA (2 μg/ml)] for indicated times. LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). Caspase-8 activity (D) and caspase-3/7 activity (E) in cell lysates were measured. Graphs show means ± SE from three independent experiments. ( F ) Model of the death paths. Blue color is used to indicate that this pathway is activated if the other death pathway was blocked. Note that LDH release is an indicator of pyroptosis, ATP loss is an indicator of cell death of any types, and caspase-8 and caspase-3 activations are indicators of apoptosis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. As caspase-1 is essential for NLRC4-induced pyroptosis, we used Casp1/11 −/− peritoneal macrophages to analyze the role of caspase-1 in FlaTox-induced NLRC4 inflammasome activation. Similar to the loss of Gsdmd , deletion of Casp1 effectively blocked FlaTox-induced pyroptosis ( Fig. 2A ), but we still could detect cell death ( Fig. 2B ). As expected, we cannot detect caspase-1 in caspase-1–deficient cells ( Fig. 2C ). Activation of caspase -8 and caspase-3 was observed in caspase-1–deficient cells after FlaTox treatment ( Fig. 2, C to E ), supporting our contention that when pyroptosis was blocked, cells switched to apoptosis. Collectively, we proposed that since the secondary necrotic lysis of apoptotic cells occurs in vitro, the release of LDH at later time in Fig. 2A should be caused by apoptosis. ASC is dispensable for NLRC4 inflammasome activation, but its presence can enhance NLRC4 inflammasome–mediated pyroptosis and cytokine production ( 14 , 29 , 30 ). We used peritoneal macrophages from Asc KO mice to address the role of ASC in FlaTox-induced pyroptosis. Deletion of Asc had almost no effect on FlaTox-induced pyroptosis (LDH release) ( Fig. 2A ) and apoptosis ( Fig. 2, B to E ). Autocleavage of caspase-1 was not detected in FlaTox-treated Asc KO cells, but the cleavage of GSDMD was detected ( Fig. 2C ). These data are consistent with published findings that pro–caspase-1 in inflammasome cleaves GSDMD ( 7 ). Unlike the deletion of Gsdmd or Casp1 , caspase-8 and caspase-3 activation cannot be detected in Asc -deficient cells ( Fig. 2, C to E ). These data clearly show that deletion of Asc did not alter the mechanisms of cell death in FlaTox-treated macrophages. We then explored whether the deletion of Gsdmd or Casp1 switches NLRP1b inflammasome–induced pyroptosis to apoptosis. RAW264.7 cell is a murine macrophage cell line in which NLRP1b inflammasome activation can be triggered by lethal toxin (LT; lethal factor plus PA) ( 31 – 35 ). Note that RAW264.7 cells lack ASC; RAW-asc is a RAW264.7-derived cell line containing ectopically expressed ASC ( 7 ). We confirmed that LT-induced pyroptosis in RAW264.7 and RAW-asc cells can be completely abolished by Nlrp1b deletion (fig. S4, A to C). Similar to what we observed in NLRC4 inflammasome activation, Casp1 or Gsdmd deletion blocked pyroptosis, but cells died via apoptosis instead (fig. S5, A to E). Similar data were obtained when J774, another murine cell line, was used (fig. S5, F to J). The data aforementioned support the idea that LDH release can be resulted directly from pyroptosis and indirectly from apoptosis. To explain the underlying mechanisms on the basis of the above results, we summarized pathway switches from pyroptosis to apoptosis in canonical inflammasome activation and their proposed function in vivo ( Fig. 2F ). Blockage of caspase-1–GSDMD path [named (first) path] inhibits pyroptosis and allows apoptosis instead. Inhibition of the first path blocks IL-1β/18 release and eicosanoids production. ASC is known to recruit caspase-8 [named (second) death path here] when the first path was impaired during NLRP3 or AIM2 inflammasome activation. On the basis of the data described above, we do not know at this moment which molecule(s) is responsible for caspase-8 activation in FlaTox- or LT-treated Casp1 - and/or Gsdmd -deficient cells. Nonetheless, we can conclude that apoptosis backs up pyroptosis in canonical inflammasomes, regardless of whether ASC is an indispensable or dispensable component of the given inflammasome, and not only pyroptosis but also apoptosis can be downstream of inflammasome activation to elicit cytolysis, which is likely the cause of animal death. Completely preventing cell death by simultaneously impairing ASC–caspase-8 apoptotic and caspase-1–GSDMD pyroptotic pathways inhibits NLRC4 overactivation–mediated animal death Although deletion of Asc did not affect FlaTox- or LT-induced pyroptosis, we next asked whether ASC is needed for caspase-8 activation and subsequent apoptosis when Casp1 was deleted. Therefore, we used Asc −/- Casp1/11 −/− (triple KO) mice and cells to address this question. Neither FlaTox-induced LDH release nor ATP loss was detected in Asc −/− Casp1/11 −/− macrophages ( Fig. 3, A and B ), indicating a complete block of pyroptosis and apoptosis. Consistently, activation of caspases was also blocked in Asc −/− Casp1/11 −/− macrophages ( Fig. 3, C to E , and fig. S6A). Fig. 3. KO of Casp1/11 and Asc completely blocks FlaTox-induced cell death and mouse death. ( A to E ) Peritoneal macrophages were treated with FlaTox for 4 hours. LDH release (A), ATP loss (B), caspase-1 (C), caspase-8 (D), and caspase-3/7 activities (E) were measured ( n = 3). ( F ) FlaTox-treated mice were monitored for survival ( n = 8). ( G ) H&E staining of small intestinal tissue of mice treated for 3 hours. Scale bars, 100 μm. ( H ) WT mice were intraperitoneally injected with FlaTox, or FlaTox + belnacasan (100 mg/kg body weight) + emricasan (20 mg/kg body weight) and were monitored for survival ( n = 6). ( I to U ) Peritoneal lavages from FlaTox-treated mice were isolated at 1 hour and 3 hours for caspase-1 (I), caspase-8 (J), caspase-3/7 (K), LDH (L), IL-1β (M), TNF (N), and IL-6 (O) ( n = 3). Peritoneal lavages at 20 min were collected for LC/MS/MS-based lipidomics (P and Q) ( n = 3). Serum was used for LDH (R), IL-1β (S), TNF (T), and IL-6 (U). n = 1 for PBS group and n = 3 for FlaTox-treated mice. ( V ) Model of death paths. ( W and X ) Mice were infected with Salmonella SL7207 and monitored for survival ( n = 10). The bacterial loads in spleen and liver were analyzed at the time of mice death ( n = 4). Graphs show means ± SE. OD 490 , optical density at 490 nm. Genetic deletion of either Casp1/11 −/− or Asc −/− had almost no effect on FlaTox-induced mouse death ( Figs. 1A and 3F ). Amazingly, Asc −/− Casp1/11 −/− mouse is completely resistant to FlaTox-induced death ( Fig. 3, F and G ), supporting the idea that cell death is the key determinant in the lethality of animal. Since the function of ASC is to mediate caspase-8 activation when Casp1 was deleted, we additionally tested whether caspase-1 inhibitor belnacasan plus pan caspase inhibitor emricasan can affect FlaTox-induced mouse death. These inhibitors indeed attenuated FlaTox-induced mouse death ( Fig. 3H ). We collected peritoneal lavages from the Asc −/− Casp1/11 −/− mice challenged with FlaTox for analyzing caspase activities, LHD release, levels of cytokines, and lipid mediators. Caspase-1, caspase-8, and caspase-3/7 activities ( Fig. 3, I to K ); LDH release ( Fig. 3L ); and IL-1β secretion ( Fig. 3M ) were found to be completely blocked. As it was observed in Gsdmd −/− and Casp1/11 −/− mice, TNF or IL-6 released into peritoneal cavity was not affected in Asc −/− Casp1/11 −/− mouse ( Fig. 3, N and O ). The generation of eicosanoids was almost completely eliminated in Asc −/− Casp1/11 −/− mouse except for PGE 2 ( Fig. 3, P and Q ). Similar to what had been observed in Nlrc4 −/− mice (fig. S1, K to N), FlaTox-induced increase of blood LDH, IL-1β, TNF, and IL-6 was blocked by deletion of Asc and Casp1/11 ( Fig. 3, R to U ). As for the increase of circulating TNF and IL-6 in FlaTox-treated mice, its association with LDH release was observed again. To this end, we can conclude that LDH release is downstream of apoptosis, and only LDH release consistently correlates with mouse death. We used RAW264.7 cells to examine the effect of Asc and Casp1 double KO on LT-induced cell death. Since RAW264.7 cells do not express ASC, RAW-asc cells were included in the experiments. We found that LT did not induce LDH release, ATP loss, activation of caspases in the cells without caspase-1, and ASC expression (fig. S7). Collectively, our data indicate that although ASC is not required for NLRC4 and NLRP1b inflammasome activation, it is required for NLRC4- and NLRP1b-mediated apoptosis in Casp1 -deficient cells ( Fig. 3V ). Thus, the (second) death path also functions in canonical inflammasomes, including those that ASC participates in but not required for their activation. As for the FlaTox-induced mice death per se, cell death dictates animal death. Salmonella Typhimurium is a type of flagellin-producing bacteria whose infection can cause host death ( 36 ). The animal death by Salmonella infection is mainly based on the bacterial dissemination in the host, and inflammasome-mediated immune responses are known to be important in limiting the bacterial burden in vivo as genetic deletion of Nlrc4 , Asc , or Casp1 leads to increased susceptibility to Salmonella infection ( 37 – 39 ). However, when the in vivo propagation of Salmonella was limited by using a propagation-defective SL7207 strain ( 40 , 41 ), the Asc −/− Casp1/11 −/− mice were significantly more resistant to Salmonella infection–caused host death in comparison with WT mice ( Fig. 3, W and X ). It appears that Salmonella propagation–caused in vivo pathological changes masked the pivotal role of cell death in animal death since the primary difference between WT and SL7207 strain is that WT but not SL7207 Salmonella propagates in mice. It seems that host responses to the process of Salmonella propagation but not the amount of bacteria loading in above experimental conditions are important for animal death as bacteria amounts in the spleen and liver had no correlation with the sensitivity to animal death ( Fig. 3X ). Inhibition of bacteria propagation in vivo is a primary therapeutic approach in treating infectious diseases, which could demask the decisive role of cell death in animal death, and inhibition of death pathways could be a clinically relevant cotreatment for lethal infections. Asc and Gsdmd double KO does not prevent mice from NLRC4 inflammasome–induced lethality because a caspase-1–dependent apoptosis pathway (the third path) is activated to arbitrate mouse death Figure 3F shows that Asc and Casp1 double KO completely blocked FlaTox-induced mouse death, but unexpectedly, double deletion of Asc and Gsdmd did not protect the animal from death ( Fig. 4A ). We obtained peritoneal cells by lavage from FlaTox-challenged mice and analyzed their caspase activities, cytokines, and LDH release. Caspase-1 activation was impaired in Asc −/− Gsdmd −/− sample at 1 hour but reached WT level at 3 hours ( Fig. 4B ). Significantly, more caspase-8 and caspase-3/7 activation was observed in Asc −/− Gsdmd −/− lavages ( Fig. 4, C and D ). LDH release in Asc −/− Gsdmd −/− lavages was not detected at early time (1 hour) but reached to a comparable level of WT lavages at 3 hours ( Fig. 4E ). Western blotting data revealed patterns of caspase cleavage that mirror results from enzymatic assays (fig. S6A). Similar to the deletion of Gsdmd ( Fig. 1D ), double deletion of Asc and Gsdmd did not affect the secretion of IL-1β into peritoneal cavity in FlaTox-treated mice ( Fig. 4F ). The production of TNF and IL-6 is not related to inflammasome activation and was not affected by double deletion of Asc and Gsdmd ( Fig. 4, G and H ). In addition, double deletion of Asc and Gsdmd did not affect the tissue damage (fig. S6B). The activation of apoptotic caspases indicates the occurrence of apoptosis, and the later LDH release is most likely caused by secondary necrosis of apoptosis. Fig. 4. Double KO of Gsdmd and Asc cannot block FlaTox-induced cell death and mouse death. ( A to H ) Mice of indicated genotypes were intraperitoneally injected with FlaTox (4 μg/g LFn-Fla body weight and 2 μg/g PA body weight) and monitored for survival (A) at indicated times ( n = 9). Peritoneal lavages were isolated by using 1 ml PBS at 1 and 3 hours after FlaTox challenge and then subjected to measurements of caspase-1 activity (B), caspase-8 activity (C), caspase-3/7 activity (D), LDH (E), IL-1β (F), TNF (G), and IL-6 (H) ( n = 3). ( I to M ) Peritoneal macrophages from mice of indicated genotypes were treated with FlaTox (2 μg/ml LFn-Fla and 2 μg/ml PA) for 4 hours. The LDH release (I) and ATP loss (J) were measured. The caspase-1 activity (K), caspase-8 activity (L), and caspase-3/7 activity (M) were measured. Graphs show means ± SE from three independent experiments. ( N ) Model of the death paths. We then analyzed the death of macrophages from those KO mice in vitro ( Fig. 4, I to M ). The cells with double deletion of Asc and Gsdmd behaved more like Gsdmd −/− other than Asc and Casp1 double KO in terms of protecting cells from death caused by NLRC4 or NLRP1b inflammasome ( Figs. 1 and 4, I to M , and fig. S6A). FlaTox-induced LDH release was blocked in Asc −/− Gsdmd −/− peritoneal macrophages until 4 hours, but ATP loss was only partially inhibited at this time point ( Fig. 4, I and J ). With the evidence of apoptotic caspase activation ( Fig. 4, L and M , and fig. S6A), we believe that the cell death mode was apoptosis. Because Asc −/− Casp1/11 −/− cells were completely protected from NLRC4- or NLRP1b-mediated death ( Fig. 3, A to E , and fig. S5), caspase-1 activation ( Fig. 4K ) observed in Asc −/− Gsdmd −/− cells must be required for their apoptotic death. The effect of Asc and Gsdmd double KO on LT-induced cell death was evaluated using RAW264.7 cells, and the data show that cells with no Asc and Gsdmd undergo apoptosis upon LT stimulation (fig. S7). It appears that caspase-1 is capable of triggering apoptosis when both Asc and Gsdmd were deleted, and we called this caspase-1–dependent apoptosis pathway the (third) death path ( Fig. 4N ). The third path is caspase-1–activated intrinsic apoptotic pathway As RAW264.7 cell is not sensitive to FlaTox, we used LT stimulation of RAW264.7 cells to study the caspase-1–dependent apoptosis. We measured LT-induced LDH release and ATP loss in RAW-asc and its Gsdmd −/− , Gsdmd −/- Casp8 −/− , and Gsdmd −/- Casp8 −/- Casp1 −/− cell lines at 4 hours after LT treatment. Deletion of Gsdmd and Casp8 inhibited LDH release ( Fig. 5A ), but significant ATP loss occurred in Gsdmd −/- Casp8 −/− cells ( Fig. 5B ), suggesting that it was caspase-1 that triggered cell death in the absence of caspase-8 and GSDMD (third death path). Gsdmd −/− Casp8 −/− Casp1 −/− cells are fully resistant to LT-induced cell death ( Fig. 5, A to F ). Caspase-3 activation should be downstream of caspase-1 since the increase of caspase-3 activity in LT-treated Gsdmd −/− Casp8 −/− cells was observed by enzymatic assay and Western blotting ( Fig. 5, C to F ). Thus, the third death pathway should be caspase-1 → caspase-3 → apoptosis. Consistently, RAW-asc Casp8 −/− Casp1 −/− cells were fully resistant to LT-induced cell death (fig. S8). Fig. 5. Caspase-1 is responsible for NLRP1b inflammasome–mediated apoptosis in Gsdmd and Casp8 double deleted cells. ( A to F ) RAW-asc cells of indicated genotypes were treated with LT (2 μg/ml LF and 2 μg/ml PA) for 4 hours. LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). Caspase-1 activity (D), caspase-8 activity (E), and caspase-3/7 activity (F) were measured. ( G and H ) RAW-asc cells of indicated genotypes were treated with LT for 90 min. ASC oligomerization was analyzed by cross-linking and Western blotting (G), and speck formation was viewed under confocal microscope (H). Graphs show means ± SE from three independent experiments. It is known that oligomerization of ASC is required for ASC to function in all subtype inflammasomes ( 1 ). Since both caspase-1 and caspase-8 can interact with ASC, we analyzed ASC oligomerization in Gsdmd −/− , Gsdmd −/- Casp8 −/− , and Gsdmd −/- Casp8 −/- Casp1 −/− RAW-asc cell lines. Cross-link and immunostaining of ASC in these different cells were performed before and after LT treatment. Results showed that LT-induced oligomerization and speck formation of ASC were not influenced by the single, double and triple deletions ( Fig. 5, G and H ). These data excluded the possibility that there was reverse signaling to affect ASC oligomerization when GSDMD, caspase-8, and caspase-1 were impaired. Caspase-1 (IL-lβ–converting enzyme) was reported to induce apoptosis in fibroblasts ( 42 ), and the cleave BH3-interacting domain death agonist (Bid) at D59 by caspase-1 could be the underlying mechanism as the cleaved Bid activates intrinsic apoptosis pathway ( 43 – 46 ). To determine whether intrinsic apoptosis plays a role in the third pathway, we deleted Apaf1 in Gsdmd/Casp8 double KO RAW-asc cells and found that it inhibited LT-induced ATP loss (apoptosis) ( Fig. 6, A to E ) but did not affect caspase-1 activation in RAW-asc Gsdmd −/- Casp8 −/− cells ( Fig. 6C ). Enzymatic assay and Western blotting confirmed abrogation of LT-induced caspase-9 and caspase-3 activation by this additional Apaf1 deletion ( Fig. 6, C to E ). Consistently, deletion of Casp9 in RAW-asc Gsdmd −/- Casp8 −/− cells also inhibited LT-mediated apoptosis ( Fig. 6, A to E ). We also measured IL-1β release by lipopolysaccharide (LPS)–primed RAW-asc cells with various gene deletions and found that LT-induced IL-1β release was partially blocked in Gsdmd −/− cells and completely eliminated in Gsdmd −/- Casp8 −/− , Gsdmd −/- Casp8 −/− Apaf1 −/− , and Gsdmd −/- Casp8 −/- Casp9 −/− cells ( Fig. 6F ). The caspase-1–dependent activation of caspase-3 and apoptosis in Asc and Gsdmd double-deficient (RAW264.7 Gsdmd −/− ) cells were blocked by genetic deletion of Casp9 (fig. S9). The modest activation of caspase-8 was also observed in Asc and Gsdmd double-deficient cells, which was abrogated by genetic deletion of Casp9 (fig. S9). This caspase-9–dependent activation of caspase-8 may be attributed to a caspase-3 feedback signal reported previously ( 47 ). Thus, caspase-1–induced apoptosis is primarily mediated by the intrinsic apoptosis pathway (apaf1 → caspase-9 → caspase-3) ( Fig. 6G ). Fig. 6. Caspase-1–initiated apoptosis during NLRP1b inflammasome activation is mediated by Apaf1–caspase-9 intrinsic apoptotic pathway. ( A to E ) RAW-asc cells of indicated genotypes were treated with LT [LF (2 μg/ml) and PA (2 μg/ml)] for 4 hours. The LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). The caspase-8 activity (D) and caspase-3/7 activity (E) were measured. ( F ) RAW-asc cells of indicated genotypes were pretreated with or without LPS (500 ng/ml) for 4 hours and then stimulated with LT for 2 hours. The IL-1β in supernatant was measured by enzyme-linked immunosorbent assay (ELISA). ( G ) Model of the death paths. Graphs show means ± SE from three independent experiments. The third death path can also be induced by the activation of NLRP3 or AIM2 inflammasome if Gsdmd and Casp8 were deleted Next, we sought to determine whether the third death path can be activated by NLRP3 and AIM2 inflammasomes. NLRP3- and AIM2-dependent inflammasome activation was induced by LPS plus nigericin or LPS plus Poly (dA:dT), respectively, in RAW-asc and its Gsdmd −/− , Gsdmd −/− Casp8 −/− , and Gsdmd −/- Casp8 −/− Casp1 −/− cells. Microscopy revealed that NLRP3 or AIM2 inflammasome activation leads to pyroptosis in RAW-asc cells, apoptosis in Gsdmd −/− and Gsdmd −/− Casp8 −/− cells, and no death of Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7A ). We also analyzed oligomerization of ASC in LPS plus nigericin or LPS plus Poly (dA:dT)–treated RAW-asc cells and found that ASC oligomerized normally in Gsdmd −/− , Gsdmd −/− Casp8 −/− , and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7B ). The IL-1β induction by Poly (dA:dT) in LPS-primed RAW-asc cells was partially blocked in Gsdmd −/− cells and completely eliminated in Gsdmd −/− Casp8 −/− and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7C ). As anticipated, caspase-1 activation was detected in all cells except Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7D ), and caspase-8 activation was only detected in Gsdmd −/− but not WT (RAW-asc), Gsdmd −/− Casp8 −/− , and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7E ). Caspase-3/7 activity was detected not only in Gsdmd −/− cells but also in Gsdmd −/− Casp8 −/− cells ( Fig. 7F ), demonstrating that the third death path (caspase-1 → caspase-3) was activated by NLRP3 and AIM2 inflammasomes when both GSDMD-mediated pyroptosis and caspase-8–mediated apoptosis were inhibited ( Fig. 7G ). These results indicate that the third death path appears to be universally used by canonical inflammasomes. We also have data that THP1 CASP1 −/− CASP8 −/− cells are fully resistant to NLRP3 and AIM2 inflammasome–mediated cell death (fig. S10), indicating that our conclusion could extend to human cells. Thus, targeting both ASC and caspase-1 or both caspase-1 and caspase-8 is a potential therapeutic way to block cell death driven by canonical inflammasome activation, which may help to prevent the lethality caused by inflammasome overactivation. Fig. 7. The third death path also applies to NLRP3 and AIM2 inflammasomes. ( A to E ) RAW-asc cells of indicated genotypes were treated with LPS (1 μg/ml) and then stimulated with nigericin (10 μM) or Poly (dA:dT) (2 μg/ml). Arrows indicate pyroptotic or apoptotic cells under microscopy (A). ASC oligomerization was measured at 90 min (B). The IL-1β in supernatant was measured by ELISA (C). Caspase-1 activity (D), caspase-8 activity (E), and caspase-3/7 activity ( F ) were measured. ( G ) Model of the death paths. Graphs show means ± SE from three independent experiments. The occurrence of cell death and subsequent cytolysis correlates with tissue damage and mouse death triggered by NLRC4 inflammasome Flagellin is a subunit protein of bacterial flagellum. FlaTox, a combination of recombinant flagellin fused to the amino-terminal domain of Bacillus anthracis lethal factor (LFn-Fla) and anthrax protective antigen (PA), was developed to selectively activate NLRC4 ( 13 ). LFn-Fla can enter cells by PA-mediated endocytosis, leading to NLRC4-dependent pyroptosis, small intestinal damage, and mouse death ( Fig. 1, A and B , and fig. S1, A to F) ( 13 , 16 , 17 , 24 ). Pyroptosis is a type of necrotic cell death with membrane disruption, which can be measured by lactate dehydrogenase (LDH) release. NLRC4-dependent pyroptosis and IL-1β release were detected in peritoneal lavage and blood of FlaTox-treated mice (fig. S1, G, H, K, and L). It was reported that extracellular flagellin induces tumor necrosis factor (TNF) and IL-6 production via an NLRC4-independent but Toll-like receptor 5 (TLR5)–dependent mechanism ( 25 ). As expected, Nlrc4 deletion only slightly reduced the TNF and IL-6 levels in peritoneal lavage from FlaTox-treated mice (fig. S1, I and J). The levels of TNF and IL-6 in the blood of FlaTox-treated Nlrc4 −/− mice were notable lower than those in wild-type (WT) mice (fig. S1, M and N). The LFn-Fla that we used can induce TNF and IL-6 production in bone marrow–derived macrophages independent of NLRC4 in the absence of PA (fig. S1, O to R). The local (peritoneal cavity) TNF and IL-6 should be induced by intraperitoneally injected FlaTox, and the circulating (blood) TNF and IL-6 ought to be controlled by other factor(s) that was directly or indirectly influenced by NLRC4 inflammasome activation. Fig. 1. Release of LDH, a cytolysis indicator, but not inflammatory cytokines or lipid mediators into peritoneal lavages is correlated with NLRC4 inflammasome–triggered mouse death. ( A ) Mice of indicated genotypes were intraperitoneally injected with FlaTox [LFn-Fla (4 μg/g body weight) and PA (2 μg/g body weight)] and monitored for survival rate ( n = 5). ( B ) Hematoxylin and eosin (H&E) staining of small intestinal tissue of mice treated as indicated for 2 hours. Scale bars, 100 μm. ( C to F ) Mice of indicated genotypes were intraperitoneally injected with FlaTox. Peritoneal lavages were isolated by using 1 ml of PBS at 1 and 2 hours after FlaTox challenge and then subjected to measurements of LDH (C), IL-1β (D), TNF (E), and IL-6 (F). n = 1 for PBS group (Ctrl) and n = 3 for FlaTox-treated group. ( G and H ) Mice of indicated genotypes were intraperitoneally injected with FlaTox for indicated times, and peritoneal lavages were collected for liquid chromatography tandem mass spectrometry (LC/MS/MS)–based lipidomics ( n = 3). ( I to K ) Mice of indicated genotypes were pretreated intraperitoneally for 30 min with of COX-1 inhibitor (SC-560) (5 μg/g body weight). Mice were intraperitoneally injected with FlaTox for indicated times, and peritoneal lavages were collected for LC/MS/MS-based lipidomics ( n = 3). Survival rate was monitored (K) ( n = 5). Since caspase-1 and GSDMD are downstream of NLRC4 to mediate pyroptosis and the release of IL-1β and IL-18, we sought to examine the effect of Gsdmd and Casp1 deletion in vivo. As we have Casp1 and Casp11 double knockout (KO) ( Casp1/11 −/− ) mice, and knowing that caspase-11 does not regulate FlaTox-induced NLRC4 inflammasome activation ( 17 ), we used Casp1/11 −/− mice in our experiments. Slightly different from the published data on FlaTox-induced pathology and body temperature drop ( 17 ), Gsdmd −/− and Casp1/11 −/− mice behaved more like WT mice in FlaTox-induced small intestinal damage and death ( Fig. 1, A and B ). It is known that deletion of Gsdmd or Casp1 blocks pyroptosis as well as IL-1β and IL-18 release ( 1 , 7 – 9 ), but we found that their deletion is quite different from Nlrc4 deletion, which totally blocks FlaTox-induced mouse death ( Fig. 1, A and B ). Thus, a more complicated mechanism is underneath the in vivo responses to FlaTox. To address the mechanism of in vivo responses to FlaTox, we analyzed FlaTox-induced pyroptosis ex vivo. We collected peritoneal lavages at 1 and 2 hours after FlaTox injection and measured the release of LDH, IL-1β, TNF, and IL-6. Deletion of Gsdmd or Casp1 effectively blocked LDH release at 1 hour. However, the level of LDH release into peritoneal cavity became comparable among Gsdmd −/− , Casp1 −/− , and WT mice at later times ( Fig. 1C ), indicating that cytolysis still occurred. It is known that in cultured cells, pyroptotic stimuli can trigger apoptosis when Casp1 or Gsdmd was deleted ( 7 , 20 , 21 ). Therefore, LDH release in Gsdmd −/− , Casp1 −/− mice at later times could be attributed to secondary necrotic lysis of the apoptotic cells or caspase-3–mediated pyroptosis via GSDME cleavage ( 26 , 27 ). However, the latter hypothesis is unlikely as no difference was observed between Gsdmd −/− mice and Gsdmd −/− Gsdme −/− mice (fig. S2, A to H). IL-1β secretion into peritoneal cavity was inhibited to different extents by deletion of Gsdmd −/− or Casp1 −/− at 1 hour, and continued inhibition was observed in Casp1 −/− but not Gsdmd −/− mice ( Fig. 1D ). The blockade of IL-1β release in Casp1 −/− mice should be resulted from failure to process IL-1β to its mature form ( 28 ), and the inhibited IL-1β release did not reduce animal death, supporting the conclusion by a previous report that flagellin-induced mouse death is independent of IL-1β and IL-18 production ( 13 ). The release of TNF and IL-6 into peritoneal cavity was not affected by gene deletion of Gsdmd or Casp1 ( Fig. 1, E and F ) because FlaTox-induced TNF and IL-6 production is NLRC4 pathway independent ( Fig. 1, E and F ). We then analyzed blood from FlaTox-treated mice and found that deletion of Gsdmd did not affect the increase of circulating LDH, IL-1β, TNF, and IL-6, and the Casp1 deletion also had no effect except for affecting IL-1β release (fig. S1, S to V). By measuring the correlations, we observed that the increase of blood TNF and IL-6 in FlaTox-treated mice is associated with LDH release but not IL-1β release. Here, it should be noted that TNF is not required for the death of FlaTox-treated mice as FlaTox was equally toxic to WT and Tnfr1 −/− mice (fig. S1W). Since inflammatory signaling lipids such as prostaglandins and leukotrienes are known to play a role in inflammasome-mediated pathological effects, we measured representatives of these lipids in peritoneal lavages at 0, 20, and 60 min after FlaTox injection. The induction of the eicosanoid lipids generated by lipoxygenase was almost completely blocked in both Gsdmd −/− and Casp1/11 −/− mice ( Fig. 1G ). Some of the cyclooxygenase (COX)–produced eicosanoids, prostaglandin D 2 (PGD 2 ), and thromboxane B2 (TXB 2 ) were significantly inhibited in Gsdmd −/− and Casp1/11 −/− mice, whereas the inhibition was slight on the production of prostaglandin E 2 (PGE 2 ) ( Fig. 1H ). We used COX inhibitor SC-560, which effectively inhibited the increase of PGE 2 and TXB 2 and, to a less extent, of PGD 2 in peritoneal lavages of FlaTox-treated mice ( Fig. 1I ), and had little or no effect on the other lipids ( Fig. 1J ). FlaTox-induced death of mice was not affected by SC-560 ( Fig. 1K ). Thus, none of these eicosanoid lipids is required for the lethality of FlaTox-treated mice ( Fig. 1, A and G to K ). Collectively, as for FlaTox-induced mouse death per se, TNF, IL-1β/18 induction, and "eicosanoid storm" seem to be dispensable, and only LDH release was found to have correlation with FlaTox-induced mouse death. Thus, cytolysis in vivo is likely the major risk of animal death. Cell death driven by NLRC4 and other canonical inflammasomes could be mediated by apoptosis when pyroptosis was blocked To understand why LDH release still occurred when pyroptotic pathway was already blocked in FlaTox-treated mice, we explored possible switches in death pathways. It was reported that ASC in NLRP3 or AIM2 complex can recruit caspase-8 to trigger apoptosis when Casp1 was deleted ( 20 , 21 ). We showed previously that Gsdmd deletion also switched NLRP3 inflammasome–mediated pyroptosis to apoptosis ( 7 ) and here that Gsdmd deletion has the same effect on cell death induced by AIM2 inflammasome (fig. S3, A to C). Caspase-1–GSDMD axis controls pyroptosis, and deletion of either Casp1 or Gsdmd switches NLRP3- and AIM2-mediated pyroptosis to apoptosis. To determine whether the switch to apoptosis also occurs in the activation of inflammasomes such as NLRC4 where ASC is not required, we used peritoneal macrophages from Gsdmd −/− mice. FlaTox induced pyroptosis in vitro in peritoneal macrophages as measured by LDH release, which was blocked by Gsdmd gene deletion ( Fig. 2A ). However, cytolysis still occurred at later times and reached a comparable level to that of WT macrophages at 8 hours ( Fig. 2A ). To understand what caused the later cytolysis, we used intracellular adenosine triphosphate (ATP) loss as a general indicator of cell death. Deletion of Gsdmd only inhibited about 20% of cell death at 4 hours ( Fig. 2B ), suggesting that cell death in a form other than pyroptosis occurred in Gsdmd −/− macrophages, and this type of cell death might be responsible for the later cytolysis. To better define the cell death pathway, we measured activities of different caspases. Gsdmd deletion did not affect FlaTox-induced autocleavage of caspase-1 as revealed by Western blotting ( Fig. 2C ). In contrast, activation of caspase-8 and caspase-3 was detected in FlaTox-treated Gsdmd −/− but not WT cells ( Fig. 2, D and E ), indicating that activation of pathways leads to apoptosis in Gsdmd −/− cells. Western blotting analysis also showed the cleavage of caspase -8 and caspase-3 in Gsdmd −/− cells ( Fig. 2C ), supporting our speculation that there is activation of an apoptotic pathway. Fig. 2. The effects of genetic deletion of Gsdmd , Caspase-1/11 , or Asc on NLRC4 inflammasome–mediated IL-1β secretion, cell death and type of cell death, and mice death. ( A to E ) Peritoneal macrophages of indicated genotypes were treated with FlaTox [LFn-Fla (2 μg/ml) and PA (2 μg/ml)] for indicated times. LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). Caspase-8 activity (D) and caspase-3/7 activity (E) in cell lysates were measured. Graphs show means ± SE from three independent experiments. ( F ) Model of the death paths. Blue color is used to indicate that this pathway is activated if the other death pathway was blocked. Note that LDH release is an indicator of pyroptosis, ATP loss is an indicator of cell death of any types, and caspase-8 and caspase-3 activations are indicators of apoptosis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. As caspase-1 is essential for NLRC4-induced pyroptosis, we used Casp1/11 −/− peritoneal macrophages to analyze the role of caspase-1 in FlaTox-induced NLRC4 inflammasome activation. Similar to the loss of Gsdmd , deletion of Casp1 effectively blocked FlaTox-induced pyroptosis ( Fig. 2A ), but we still could detect cell death ( Fig. 2B ). As expected, we cannot detect caspase-1 in caspase-1–deficient cells ( Fig. 2C ). Activation of caspase -8 and caspase-3 was observed in caspase-1–deficient cells after FlaTox treatment ( Fig. 2, C to E ), supporting our contention that when pyroptosis was blocked, cells switched to apoptosis. Collectively, we proposed that since the secondary necrotic lysis of apoptotic cells occurs in vitro, the release of LDH at later time in Fig. 2A should be caused by apoptosis. ASC is dispensable for NLRC4 inflammasome activation, but its presence can enhance NLRC4 inflammasome–mediated pyroptosis and cytokine production ( 14 , 29 , 30 ). We used peritoneal macrophages from Asc KO mice to address the role of ASC in FlaTox-induced pyroptosis. Deletion of Asc had almost no effect on FlaTox-induced pyroptosis (LDH release) ( Fig. 2A ) and apoptosis ( Fig. 2, B to E ). Autocleavage of caspase-1 was not detected in FlaTox-treated Asc KO cells, but the cleavage of GSDMD was detected ( Fig. 2C ). These data are consistent with published findings that pro–caspase-1 in inflammasome cleaves GSDMD ( 7 ). Unlike the deletion of Gsdmd or Casp1 , caspase-8 and caspase-3 activation cannot be detected in Asc -deficient cells ( Fig. 2, C to E ). These data clearly show that deletion of Asc did not alter the mechanisms of cell death in FlaTox-treated macrophages. We then explored whether the deletion of Gsdmd or Casp1 switches NLRP1b inflammasome–induced pyroptosis to apoptosis. RAW264.7 cell is a murine macrophage cell line in which NLRP1b inflammasome activation can be triggered by lethal toxin (LT; lethal factor plus PA) ( 31 – 35 ). Note that RAW264.7 cells lack ASC; RAW-asc is a RAW264.7-derived cell line containing ectopically expressed ASC ( 7 ). We confirmed that LT-induced pyroptosis in RAW264.7 and RAW-asc cells can be completely abolished by Nlrp1b deletion (fig. S4, A to C). Similar to what we observed in NLRC4 inflammasome activation, Casp1 or Gsdmd deletion blocked pyroptosis, but cells died via apoptosis instead (fig. S5, A to E). Similar data were obtained when J774, another murine cell line, was used (fig. S5, F to J). The data aforementioned support the idea that LDH release can be resulted directly from pyroptosis and indirectly from apoptosis. To explain the underlying mechanisms on the basis of the above results, we summarized pathway switches from pyroptosis to apoptosis in canonical inflammasome activation and their proposed function in vivo ( Fig. 2F ). Blockage of caspase-1–GSDMD path [named (first) path] inhibits pyroptosis and allows apoptosis instead. Inhibition of the first path blocks IL-1β/18 release and eicosanoids production. ASC is known to recruit caspase-8 [named (second) death path here] when the first path was impaired during NLRP3 or AIM2 inflammasome activation. On the basis of the data described above, we do not know at this moment which molecule(s) is responsible for caspase-8 activation in FlaTox- or LT-treated Casp1 - and/or Gsdmd -deficient cells. Nonetheless, we can conclude that apoptosis backs up pyroptosis in canonical inflammasomes, regardless of whether ASC is an indispensable or dispensable component of the given inflammasome, and not only pyroptosis but also apoptosis can be downstream of inflammasome activation to elicit cytolysis, which is likely the cause of animal death. Completely preventing cell death by simultaneously impairing ASC–caspase-8 apoptotic and caspase-1–GSDMD pyroptotic pathways inhibits NLRC4 overactivation–mediated animal death Although deletion of Asc did not affect FlaTox- or LT-induced pyroptosis, we next asked whether ASC is needed for caspase-8 activation and subsequent apoptosis when Casp1 was deleted. Therefore, we used Asc −/- Casp1/11 −/− (triple KO) mice and cells to address this question. Neither FlaTox-induced LDH release nor ATP loss was detected in Asc −/− Casp1/11 −/− macrophages ( Fig. 3, A and B ), indicating a complete block of pyroptosis and apoptosis. Consistently, activation of caspases was also blocked in Asc −/− Casp1/11 −/− macrophages ( Fig. 3, C to E , and fig. S6A). Fig. 3. KO of Casp1/11 and Asc completely blocks FlaTox-induced cell death and mouse death. ( A to E ) Peritoneal macrophages were treated with FlaTox for 4 hours. LDH release (A), ATP loss (B), caspase-1 (C), caspase-8 (D), and caspase-3/7 activities (E) were measured ( n = 3). ( F ) FlaTox-treated mice were monitored for survival ( n = 8). ( G ) H&E staining of small intestinal tissue of mice treated for 3 hours. Scale bars, 100 μm. ( H ) WT mice were intraperitoneally injected with FlaTox, or FlaTox + belnacasan (100 mg/kg body weight) + emricasan (20 mg/kg body weight) and were monitored for survival ( n = 6). ( I to U ) Peritoneal lavages from FlaTox-treated mice were isolated at 1 hour and 3 hours for caspase-1 (I), caspase-8 (J), caspase-3/7 (K), LDH (L), IL-1β (M), TNF (N), and IL-6 (O) ( n = 3). Peritoneal lavages at 20 min were collected for LC/MS/MS-based lipidomics (P and Q) ( n = 3). Serum was used for LDH (R), IL-1β (S), TNF (T), and IL-6 (U). n = 1 for PBS group and n = 3 for FlaTox-treated mice. ( V ) Model of death paths. ( W and X ) Mice were infected with Salmonella SL7207 and monitored for survival ( n = 10). The bacterial loads in spleen and liver were analyzed at the time of mice death ( n = 4). Graphs show means ± SE. OD 490 , optical density at 490 nm. Genetic deletion of either Casp1/11 −/− or Asc −/− had almost no effect on FlaTox-induced mouse death ( Figs. 1A and 3F ). Amazingly, Asc −/− Casp1/11 −/− mouse is completely resistant to FlaTox-induced death ( Fig. 3, F and G ), supporting the idea that cell death is the key determinant in the lethality of animal. Since the function of ASC is to mediate caspase-8 activation when Casp1 was deleted, we additionally tested whether caspase-1 inhibitor belnacasan plus pan caspase inhibitor emricasan can affect FlaTox-induced mouse death. These inhibitors indeed attenuated FlaTox-induced mouse death ( Fig. 3H ). We collected peritoneal lavages from the Asc −/− Casp1/11 −/− mice challenged with FlaTox for analyzing caspase activities, LHD release, levels of cytokines, and lipid mediators. Caspase-1, caspase-8, and caspase-3/7 activities ( Fig. 3, I to K ); LDH release ( Fig. 3L ); and IL-1β secretion ( Fig. 3M ) were found to be completely blocked. As it was observed in Gsdmd −/− and Casp1/11 −/− mice, TNF or IL-6 released into peritoneal cavity was not affected in Asc −/− Casp1/11 −/− mouse ( Fig. 3, N and O ). The generation of eicosanoids was almost completely eliminated in Asc −/− Casp1/11 −/− mouse except for PGE 2 ( Fig. 3, P and Q ). Similar to what had been observed in Nlrc4 −/− mice (fig. S1, K to N), FlaTox-induced increase of blood LDH, IL-1β, TNF, and IL-6 was blocked by deletion of Asc and Casp1/11 ( Fig. 3, R to U ). As for the increase of circulating TNF and IL-6 in FlaTox-treated mice, its association with LDH release was observed again. To this end, we can conclude that LDH release is downstream of apoptosis, and only LDH release consistently correlates with mouse death. We used RAW264.7 cells to examine the effect of Asc and Casp1 double KO on LT-induced cell death. Since RAW264.7 cells do not express ASC, RAW-asc cells were included in the experiments. We found that LT did not induce LDH release, ATP loss, activation of caspases in the cells without caspase-1, and ASC expression (fig. S7). Collectively, our data indicate that although ASC is not required for NLRC4 and NLRP1b inflammasome activation, it is required for NLRC4- and NLRP1b-mediated apoptosis in Casp1 -deficient cells ( Fig. 3V ). Thus, the (second) death path also functions in canonical inflammasomes, including those that ASC participates in but not required for their activation. As for the FlaTox-induced mice death per se, cell death dictates animal death. Salmonella Typhimurium is a type of flagellin-producing bacteria whose infection can cause host death ( 36 ). The animal death by Salmonella infection is mainly based on the bacterial dissemination in the host, and inflammasome-mediated immune responses are known to be important in limiting the bacterial burden in vivo as genetic deletion of Nlrc4 , Asc , or Casp1 leads to increased susceptibility to Salmonella infection ( 37 – 39 ). However, when the in vivo propagation of Salmonella was limited by using a propagation-defective SL7207 strain ( 40 , 41 ), the Asc −/− Casp1/11 −/− mice were significantly more resistant to Salmonella infection–caused host death in comparison with WT mice ( Fig. 3, W and X ). It appears that Salmonella propagation–caused in vivo pathological changes masked the pivotal role of cell death in animal death since the primary difference between WT and SL7207 strain is that WT but not SL7207 Salmonella propagates in mice. It seems that host responses to the process of Salmonella propagation but not the amount of bacteria loading in above experimental conditions are important for animal death as bacteria amounts in the spleen and liver had no correlation with the sensitivity to animal death ( Fig. 3X ). Inhibition of bacteria propagation in vivo is a primary therapeutic approach in treating infectious diseases, which could demask the decisive role of cell death in animal death, and inhibition of death pathways could be a clinically relevant cotreatment for lethal infections. Asc and Gsdmd double KO does not prevent mice from NLRC4 inflammasome–induced lethality because a caspase-1–dependent apoptosis pathway (the third path) is activated to arbitrate mouse death Figure 3F shows that Asc and Casp1 double KO completely blocked FlaTox-induced mouse death, but unexpectedly, double deletion of Asc and Gsdmd did not protect the animal from death ( Fig. 4A ). We obtained peritoneal cells by lavage from FlaTox-challenged mice and analyzed their caspase activities, cytokines, and LDH release. Caspase-1 activation was impaired in Asc −/− Gsdmd −/− sample at 1 hour but reached WT level at 3 hours ( Fig. 4B ). Significantly, more caspase-8 and caspase-3/7 activation was observed in Asc −/− Gsdmd −/− lavages ( Fig. 4, C and D ). LDH release in Asc −/− Gsdmd −/− lavages was not detected at early time (1 hour) but reached to a comparable level of WT lavages at 3 hours ( Fig. 4E ). Western blotting data revealed patterns of caspase cleavage that mirror results from enzymatic assays (fig. S6A). Similar to the deletion of Gsdmd ( Fig. 1D ), double deletion of Asc and Gsdmd did not affect the secretion of IL-1β into peritoneal cavity in FlaTox-treated mice ( Fig. 4F ). The production of TNF and IL-6 is not related to inflammasome activation and was not affected by double deletion of Asc and Gsdmd ( Fig. 4, G and H ). In addition, double deletion of Asc and Gsdmd did not affect the tissue damage (fig. S6B). The activation of apoptotic caspases indicates the occurrence of apoptosis, and the later LDH release is most likely caused by secondary necrosis of apoptosis. Fig. 4. Double KO of Gsdmd and Asc cannot block FlaTox-induced cell death and mouse death. ( A to H ) Mice of indicated genotypes were intraperitoneally injected with FlaTox (4 μg/g LFn-Fla body weight and 2 μg/g PA body weight) and monitored for survival (A) at indicated times ( n = 9). Peritoneal lavages were isolated by using 1 ml PBS at 1 and 3 hours after FlaTox challenge and then subjected to measurements of caspase-1 activity (B), caspase-8 activity (C), caspase-3/7 activity (D), LDH (E), IL-1β (F), TNF (G), and IL-6 (H) ( n = 3). ( I to M ) Peritoneal macrophages from mice of indicated genotypes were treated with FlaTox (2 μg/ml LFn-Fla and 2 μg/ml PA) for 4 hours. The LDH release (I) and ATP loss (J) were measured. The caspase-1 activity (K), caspase-8 activity (L), and caspase-3/7 activity (M) were measured. Graphs show means ± SE from three independent experiments. ( N ) Model of the death paths. We then analyzed the death of macrophages from those KO mice in vitro ( Fig. 4, I to M ). The cells with double deletion of Asc and Gsdmd behaved more like Gsdmd −/− other than Asc and Casp1 double KO in terms of protecting cells from death caused by NLRC4 or NLRP1b inflammasome ( Figs. 1 and 4, I to M , and fig. S6A). FlaTox-induced LDH release was blocked in Asc −/− Gsdmd −/− peritoneal macrophages until 4 hours, but ATP loss was only partially inhibited at this time point ( Fig. 4, I and J ). With the evidence of apoptotic caspase activation ( Fig. 4, L and M , and fig. S6A), we believe that the cell death mode was apoptosis. Because Asc −/− Casp1/11 −/− cells were completely protected from NLRC4- or NLRP1b-mediated death ( Fig. 3, A to E , and fig. S5), caspase-1 activation ( Fig. 4K ) observed in Asc −/− Gsdmd −/− cells must be required for their apoptotic death. The effect of Asc and Gsdmd double KO on LT-induced cell death was evaluated using RAW264.7 cells, and the data show that cells with no Asc and Gsdmd undergo apoptosis upon LT stimulation (fig. S7). It appears that caspase-1 is capable of triggering apoptosis when both Asc and Gsdmd were deleted, and we called this caspase-1–dependent apoptosis pathway the (third) death path ( Fig. 4N ). The third path is caspase-1–activated intrinsic apoptotic pathway As RAW264.7 cell is not sensitive to FlaTox, we used LT stimulation of RAW264.7 cells to study the caspase-1–dependent apoptosis. We measured LT-induced LDH release and ATP loss in RAW-asc and its Gsdmd −/− , Gsdmd −/- Casp8 −/− , and Gsdmd −/- Casp8 −/- Casp1 −/− cell lines at 4 hours after LT treatment. Deletion of Gsdmd and Casp8 inhibited LDH release ( Fig. 5A ), but significant ATP loss occurred in Gsdmd −/- Casp8 −/− cells ( Fig. 5B ), suggesting that it was caspase-1 that triggered cell death in the absence of caspase-8 and GSDMD (third death path). Gsdmd −/− Casp8 −/− Casp1 −/− cells are fully resistant to LT-induced cell death ( Fig. 5, A to F ). Caspase-3 activation should be downstream of caspase-1 since the increase of caspase-3 activity in LT-treated Gsdmd −/− Casp8 −/− cells was observed by enzymatic assay and Western blotting ( Fig. 5, C to F ). Thus, the third death pathway should be caspase-1 → caspase-3 → apoptosis. Consistently, RAW-asc Casp8 −/− Casp1 −/− cells were fully resistant to LT-induced cell death (fig. S8). Fig. 5. Caspase-1 is responsible for NLRP1b inflammasome–mediated apoptosis in Gsdmd and Casp8 double deleted cells. ( A to F ) RAW-asc cells of indicated genotypes were treated with LT (2 μg/ml LF and 2 μg/ml PA) for 4 hours. LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). Caspase-1 activity (D), caspase-8 activity (E), and caspase-3/7 activity (F) were measured. ( G and H ) RAW-asc cells of indicated genotypes were treated with LT for 90 min. ASC oligomerization was analyzed by cross-linking and Western blotting (G), and speck formation was viewed under confocal microscope (H). Graphs show means ± SE from three independent experiments. It is known that oligomerization of ASC is required for ASC to function in all subtype inflammasomes ( 1 ). Since both caspase-1 and caspase-8 can interact with ASC, we analyzed ASC oligomerization in Gsdmd −/− , Gsdmd −/- Casp8 −/− , and Gsdmd −/- Casp8 −/- Casp1 −/− RAW-asc cell lines. Cross-link and immunostaining of ASC in these different cells were performed before and after LT treatment. Results showed that LT-induced oligomerization and speck formation of ASC were not influenced by the single, double and triple deletions ( Fig. 5, G and H ). These data excluded the possibility that there was reverse signaling to affect ASC oligomerization when GSDMD, caspase-8, and caspase-1 were impaired. Caspase-1 (IL-lβ–converting enzyme) was reported to induce apoptosis in fibroblasts ( 42 ), and the cleave BH3-interacting domain death agonist (Bid) at D59 by caspase-1 could be the underlying mechanism as the cleaved Bid activates intrinsic apoptosis pathway ( 43 – 46 ). To determine whether intrinsic apoptosis plays a role in the third pathway, we deleted Apaf1 in Gsdmd/Casp8 double KO RAW-asc cells and found that it inhibited LT-induced ATP loss (apoptosis) ( Fig. 6, A to E ) but did not affect caspase-1 activation in RAW-asc Gsdmd −/- Casp8 −/− cells ( Fig. 6C ). Enzymatic assay and Western blotting confirmed abrogation of LT-induced caspase-9 and caspase-3 activation by this additional Apaf1 deletion ( Fig. 6, C to E ). Consistently, deletion of Casp9 in RAW-asc Gsdmd −/- Casp8 −/− cells also inhibited LT-mediated apoptosis ( Fig. 6, A to E ). We also measured IL-1β release by lipopolysaccharide (LPS)–primed RAW-asc cells with various gene deletions and found that LT-induced IL-1β release was partially blocked in Gsdmd −/− cells and completely eliminated in Gsdmd −/- Casp8 −/− , Gsdmd −/- Casp8 −/− Apaf1 −/− , and Gsdmd −/- Casp8 −/- Casp9 −/− cells ( Fig. 6F ). The caspase-1–dependent activation of caspase-3 and apoptosis in Asc and Gsdmd double-deficient (RAW264.7 Gsdmd −/− ) cells were blocked by genetic deletion of Casp9 (fig. S9). The modest activation of caspase-8 was also observed in Asc and Gsdmd double-deficient cells, which was abrogated by genetic deletion of Casp9 (fig. S9). This caspase-9–dependent activation of caspase-8 may be attributed to a caspase-3 feedback signal reported previously ( 47 ). Thus, caspase-1–induced apoptosis is primarily mediated by the intrinsic apoptosis pathway (apaf1 → caspase-9 → caspase-3) ( Fig. 6G ). Fig. 6. Caspase-1–initiated apoptosis during NLRP1b inflammasome activation is mediated by Apaf1–caspase-9 intrinsic apoptotic pathway. ( A to E ) RAW-asc cells of indicated genotypes were treated with LT [LF (2 μg/ml) and PA (2 μg/ml)] for 4 hours. The LDH release (A) and ATP loss (B) were measured. Cell lysates combined with medium were subjected to Western blot for indicated proteins (C). The caspase-8 activity (D) and caspase-3/7 activity (E) were measured. ( F ) RAW-asc cells of indicated genotypes were pretreated with or without LPS (500 ng/ml) for 4 hours and then stimulated with LT for 2 hours. The IL-1β in supernatant was measured by enzyme-linked immunosorbent assay (ELISA). ( G ) Model of the death paths. Graphs show means ± SE from three independent experiments. The third death path can also be induced by the activation of NLRP3 or AIM2 inflammasome if Gsdmd and Casp8 were deleted Next, we sought to determine whether the third death path can be activated by NLRP3 and AIM2 inflammasomes. NLRP3- and AIM2-dependent inflammasome activation was induced by LPS plus nigericin or LPS plus Poly (dA:dT), respectively, in RAW-asc and its Gsdmd −/− , Gsdmd −/− Casp8 −/− , and Gsdmd −/- Casp8 −/− Casp1 −/− cells. Microscopy revealed that NLRP3 or AIM2 inflammasome activation leads to pyroptosis in RAW-asc cells, apoptosis in Gsdmd −/− and Gsdmd −/− Casp8 −/− cells, and no death of Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7A ). We also analyzed oligomerization of ASC in LPS plus nigericin or LPS plus Poly (dA:dT)–treated RAW-asc cells and found that ASC oligomerized normally in Gsdmd −/− , Gsdmd −/− Casp8 −/− , and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7B ). The IL-1β induction by Poly (dA:dT) in LPS-primed RAW-asc cells was partially blocked in Gsdmd −/− cells and completely eliminated in Gsdmd −/− Casp8 −/− and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7C ). As anticipated, caspase-1 activation was detected in all cells except Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7D ), and caspase-8 activation was only detected in Gsdmd −/− but not WT (RAW-asc), Gsdmd −/− Casp8 −/− , and Gsdmd −/− Casp8 −/− Casp1 −/− cells ( Fig. 7E ). Caspase-3/7 activity was detected not only in Gsdmd −/− cells but also in Gsdmd −/− Casp8 −/− cells ( Fig. 7F ), demonstrating that the third death path (caspase-1 → caspase-3) was activated by NLRP3 and AIM2 inflammasomes when both GSDMD-mediated pyroptosis and caspase-8–mediated apoptosis were inhibited ( Fig. 7G ). These results indicate that the third death path appears to be universally used by canonical inflammasomes. We also have data that THP1 CASP1 −/− CASP8 −/− cells are fully resistant to NLRP3 and AIM2 inflammasome–mediated cell death (fig. S10), indicating that our conclusion could extend to human cells. Thus, targeting both ASC and caspase-1 or both caspase-1 and caspase-8 is a potential therapeutic way to block cell death driven by canonical inflammasome activation, which may help to prevent the lethality caused by inflammasome overactivation. Fig. 7. The third death path also applies to NLRP3 and AIM2 inflammasomes. ( A to E ) RAW-asc cells of indicated genotypes were treated with LPS (1 μg/ml) and then stimulated with nigericin (10 μM) or Poly (dA:dT) (2 μg/ml). Arrows indicate pyroptotic or apoptotic cells under microscopy (A). ASC oligomerization was measured at 90 min (B). The IL-1β in supernatant was measured by ELISA (C). Caspase-1 activity (D), caspase-8 activity (E), and caspase-3/7 activity ( F ) were measured. ( G ) Model of the death paths. Graphs show means ± SE from three independent experiments. DISCUSSION Inflammasome-mediated cell death on one hand contributes to the activation of immune system to control infected pathogens ( 14 – 18 ). On the other hand, cell death by itself subsidizes the pathological changes in infected animals ( 17 ). The data presented in this report demonstrate that cell death plays a decisive role in the death of mice caused by NLRC4 inflammasome in an infection-free system. In the case of live bacteria, such as S. Typhimurium infection, NLRC4-triggered cell death contributes to host death, but the outcome should be determined by multiple factors. Blocking the cell death pathways downstream of NLRC4 attenuated mice death caused by growth-defective Salmonella infection ( Fig. 3W ), but this inhibition was not observed when WT Salmonella was used, suggesting that the lethal effect of bacteria propagation is before that of cell death ( 37 – 39 ). Therefore, if propagation of infected pathogens can be blocked by pharmacological approaches, blocking inflammasome overactivation–caused cell death shall be important for preventing the host from death. A previous study showed that deletion of Casp1 and Casp8 ( Casp1 −/− Casp8 −/− Rip3 −/− mice) reduced ability of mice in controlling the propagation of orally infected S. Typhimurium ( 17 ). A recent study further showed that simultaneous genetic deletion of Casp1 , Casp11 , Casp12 , Casp8 , and Rip3 ensured the blockade of Salmonella -induced cell death and sensitized mice to the infection with a growth-attenuated Salmonella ∆ AroA strain BRD509 ( 15 ). WT mice were capable of controlling the propagation of intravenously injected BRD509 [200 colony-forming units (CFU)] within a certain level and eventually clearing the bacteria after 12 weeks, whereas the injected BRD509 in Casp1 −/− Casp11 −/− Casp12 −/− Casp8 −/− Rip3 −/− mice continued propagating, and the infected mice had to be euthanized in 4 to 5 weeks after infection because of heavy in-organ bacterial burden ( 15 ). It appears that similar to virus infections, cell death is a mechanism to eliminate intracellular bacteria such as Salmonella . In agreement with this notion, animal death caused by high-dose nonpropagative SL7207 was significantly attenuated by blocking NLRC4-mediated cell death ( Fig. 3W ). Thus, cell death should have dual roles in host defense against intracellular bacteria, and the pro-death role would be predominant if bacteria propagation has been eliminated by other means. Global induction of the production of inflammatory cytokines, the so-called cytokine storm, is believed to be the major trigger of infection/inflammation-caused animal death. The levels of FlaTox-induced cytokines such as IL-1β, TNF, and IL-6 in mice are comparable to those induced by S. Typhimurium infection, indicating the induction of cytokine storm by NLRC4 activation directly or indirectly. However, in line with a published study ( 13 ), we confirmed that there is no correlation between the levels of IL-1β, TNF, or IL-6 and death in FlaTox-treated mice ( Figs. 1 and 3 ). The increase of circulating TNF and IL-6, an NLRC4-independent response to flagellin, was associated with LDH release (cytolysis) (fig. S1, K to N and S to V, and Fig. 3, R to U ). The factors that caused the increase of circulating (blood) TNF and IL-6 in FlaTox-treated cells are very likely to be the damage-associated molecular patterns released from dead cells. We also did not find a correlation between animal death and eicosanoids production. We showed here that eicosanoids production is dependent on pyroptosis ( Fig. 1, G and H ), which is not unexpected as it was previously proposed that eicosanoid biosynthesis is stimulated by Ca 2+ influx via FlaTox-induced plasma membrane pore ( 13 ), which should be GSDMD pore based on current knowledge ( 7 – 9 ). We proposed that either "cytokine or eicosanoid storm" is not essential for FlaTox-induced animal death, or the animal death is resulted from the compound effect of a large number of different cytokines/lipids, and the commonly used measurement of cytokines cannot evaluate the effect of cytokine storm. Our data demonstrated that irreversible cell death induced by NLRC4 overactivation definitely can play a decisive role in animal death, which is consistent with a published study by Rauch et al. ( 17 ) showing that deletion of Casp1 and Casp8 , which should block all three death pathways described in this study ( Fig. 7G ), inhibited FlaTox-induced body temperature drop (an indicator of animal viability). Rauch et al. ( 17 ) also showed that Asc −/− Casp1/11 −/− and Rip3 −/− Casp1 −/− Casp8 −/− mice were fully protected from FlaTox-induced pathological changes, including hypothermia, hematocrit increase, diarrhea, intestinal epithelial cell expulsion, PGE 2 , and body temperature drop. Our data that FlaTox-induced animal death was prevented by gene deletion of Asc and Casp1/11 are in line with their finding. Cell death in vivo could be a result of cytokine storm, which, in turn, can induce cytokine production. In the situation of FlaTox-induced sterile NLRC4 activation, cell death is a direct consequence of FlaTox treatment. Pyroptosis is important for the production of some cytokines such as IL-1β and IL-18 and inflammatory lipids, and they could induce other cytokines such as TNF, but those cytokines and lipids are not essential for FlaTox-induced animal death. Bacterial infection, such as S. Typhimurium infection, is a much more complicated situation in comparison with the NLRC4 activation by FalTox, and pathway switches such as pyroptosis to caspase-8–dependent apoptosis in Salmonella -treated macrophages occur ( 15 ). The cytokine to cell death and cell death to cytokine circles could function in bacterial infections. Previous studies have revealed switches among or compensation of different cell death pathways when one of them was absent or impaired ( 7 , 8 , 15 , 20 – 22 , 48 , 49 ). In this report, we revealed the order of complementation of three distinct cell death pathways in inflammasome-mediated cell death process. It seems that cell death is destined to take place after inflammasome has been fully activated. The inevitable cell death does not necessarily promote IL-1β secretion as there was cell death in FlaTox-treated Casp1/11 −/− mice but not much IL-1β in peritoneal lavages ( Fig. 1, C and D ). IL-1β secretion is definitely not a must to reach destination by the inherent complementary cell death pathways. The preferential adoption of these cell death pathways in a specific order could be programmed for animal death since only the mutant mice in which cell death was blocked survived from FlaTox treatment ( Figs. 1, A and F , and 4A . Cell death types are not important for outcomes at animal level. It was reported that GSDMD can be inactivated by caspase-3/-7 when apoptosis occurred ( 50 ). This is the case in Casp1 −/− cells when apoptosis was induced by NLRC4, NLRP1b, NLRP3, or AIM2 inflammasome. Cleavage of GSDMD by caspase-8 was also reported to mediate pyroptosis when transforming growth factor β–activated kinase 1 was inhibited ( 51 ), although we did not detect GSDMD cleavage by caspase-8 in our experimental settings. The cleavage of GSDMD by apoptotic caspases suggests possible inhibition of pyroptosis by apoptosis. We detected activation of caspase-8 and caspase-9 when the third pathway was activated, and caspase-9 but not caspase-8 is required for the third pathway ( Fig. 6 ). Caspase-8 activation should be an associated event of caspase-9 activation and could be additive but not essential for the third path of inflammasome-induced cell death. Activation of caspase-8 by caspase-9–activated caspase-3 was reported ( 47 ), and here, it is likely the same case. A recent report showed that forced activation of caspase-1 in cells without GSDMD–induced apoptosis through the Bid → caspase-9 → caspase-3 axis ( 46 ), supporting the idea that caspase-1 can activate intrinsic apoptotic pathway when pyroptotic pathway is impaired. In addition, cleavage of Bid by caspase-8 can mediate the intrinsic apoptosis pathway ( 44 ). In the current study, we indeed observed caspase-9 activation in Casp1 -deficient conditions when NLRC4, NLRP3, AIM2, and NLRP1b inflammasomes were activated ( Fig. 2C , and figs. S3C and S5, C and H). Since FlaTox-induced mouse death is the only model available for lethality solely mediated by inflammasomes, we were unable to evaluate whether hyperactivation of the inflammasomes other than NLRC4 is lethal. Because of similarity among the subtype inflammasomes, we would predict that the conclusion of this study can be applicable to other inflammasomes. However, we shall note that injection of FlaTox in mice is an artificial approach, and unfortunately, there is no model available that can recapitulate NLRC4 activation in a more physiologically relevant setting. Efforts should be given to find relevant stimulus and animal model that can better mimic NLRC4-dependent pathological changes reported in human patients. Nonetheless, the preexistence of three death paths suggests a firm commitment to cell death after inflammasome hyperactivation, which is not important for inflammasome-mediated cytokine secretion but destined to inflammasome-caused animal death. Cell death, regardless of its type, plays a decisive role in animal death caused by NLRC4 inflammasome hyperactivation. The decisive role of cell death might be applicable to other inflammation-related lethality, such as septic shock. Cell death is a marker of the immunosuppression stage of sepsis, and this phase links to the lethality of septic individuals ( 52 , 53 ). It is highly possible that inhibition of cell death is an effective cotreatment of lethal infectious diseases when the propagation of infected pathogen can be eliminated by other therapeutic approaches. MATERIALS AND METHODS Study design The objective of this study was to investigate the relation between canonical inflammasome–mediated cell death and host death. We deleted various canonical inflammasome–related genes individually or in combination in cells and mice and analyzed their effects on cell and mouse death as well as death pathways triggered by different canonical inflammasomes. The stimulation and replicates of each experiment are presented in the figure legends. Mice All mice used were housed in a specific pathogen-free environment. WT C57BL/6J mice were originally obtained from the Jackson Laboratory. Casp1/11 −/− and Gsdmd −/− mice were used as previously described ( 7 ). Asc −/− , Nlrc4 −/− , and Gsdme −/− mice were generated by co-microinjection of in vitro translated Cas9 mRNA and guide RNA (gRNA) into the C57BL/6J zygotes. The targeting sequence in the gRNA vector was 5′-CAGACGAGCCCTTATTCAA-3′ for mouse Nlrc4 , 5′-TATGGGCGCATCCCACGCG-3′ for mouse Asc , and 5′-TTCGCCTTTCTGGACATGC-3′ for mouse Gsdme . All KO alleles have been crossed onto the C57BL/6J background. Male mice between 8 and 12 weeks of age were used for the animal experiments in the current study. All animal experiments were conducted following the guidelines for housing and care of laboratory animals and performed in accordance with regulations approved by the Animal Care and Use Committee at Xiamen University. For intraperitoneal injection of FlaTox into mice, toxin doses were PA (2 μg/g body weight) combined with LFn-Fla or LFn-Fla 3A (4 μg/g) and diluted in 250 μl of phosphate-buffered saline (PBS). The survival rate of the animals was checked every few hours. Generation of KO cell lines using CRISPR-Cas9 technique The targeting sequence in the gRNA vector was 5′-TCTCTAAAAAAGGGCCCC-3′ for mouse caspase-1 , 5′-TGCAACAGCTTCGGAGTCG-3′ for mouse Gsdmd , 5′-TATGGGCGCATCCCACGCG-3′ for mouse Asc , 5′- GTCTAGGAAGTTGACCAGC-3′ for mouse caspase-8 , 5′-GGATGAAGAGCAGCTTCTCA-3′ for mouse Nlrp1b , 5′- ATGGATCACATGATCAGTAA-3′ for mouse Apaf1 , 5′-CTGGCTTCACTCTTGCAAAG-3′ for mouse caspase-9 , 5′-ATTGACTCCGTTATTCCGAA-3′ for human CASP1 , and 5′-CCCTCAAGTTCCTGAGCCT-3′ for human CASP8 . The plasmids (vector pBOB) harboring the gene gRNA sequences and Cas9 gene were transfected into 293T in the presence of lentivirus helper plasmids, and the supernatants were collected after 48 hours. The viruses were then used to infect J774 cells, RAW264.7 cells, RAW-asc cells, or THP1 cells. KOs were confirmed by immunoblots and further confirmed by sequencing. LFn fusion protein preparation pET15b LFn-Fla (Addgene plasmid no. 84871) and pET15b LFn-Fla 3A (Addgene plasmid no. 84872) were gifts from R. Vance. The LFn-Fla and LFn-Fla 3A were expressed and purified as described in previous publications ( 13 ). The endotoxin was removed with the ToxinEraserTM Endotoxin Removal Kit (L00338, GenScript). Bacterial strains for infection studies For in vivo infection, Salmonella SL7207 (ΔaroA) was incubated with shaking at 37°C in Luria-Bertani (LB) broth for 16 hours, and then 500 μl of medium with bacteria was added to a new tube with 15 ml of LB broth shaking until an optical density at 600 nm (OD 600 ) reached 1.0. The collected Salmonella was diluted in PBS, and 5 × 10 8 CFU for Salmonella SL7207 (ΔaroA) were injected intraperitoneally in a volume of 200 μl. The number of bacteria was examined by homogenizing organs of infected mice in 5 ml of sterilized PBS. The homogenate was diluted and plated onto LB agar plates and incubated at 37°C for 16 hours. Hematoxylin and eosin staining Animals were euthanized. Small intestines were collected and fixed immediately in 4% paraformaldehyde for 24 hours. The fixed tissues were embedded in a water-free procedure. Five-micrometer sections were cut and stained with hematoxylin and eosin. Slides were analyzed on the Leica Aperio Versa 200. Eicosanoid analysis For peritoneal lavage analysis, mice were injected intraperitoneally with FlaTox. After the indicated time, mice were euthanized and injected intraperitoneally with 1 ml of cold PBS. Peritoneal lavage (600 μl) was collected, and 50 μl of which was immediately transferred to 500 μl of cold methanol for storage at −80°C. Eicosanoids and docosanoids were identified and quantified by AB SCIEX QTRAP-6500 plus. The synthetic standards were purchased from Cayman Chemical. Cell cultures and stimulation Peritoneal macrophages were isolated from mice after being injected intraperitoneally with 3% thioglycollate solution for 4 days. For NLRC4 inflammasome activation, peritoneal macrophages were treated with LFn-Fla or LFn-Fla 3A with PA for the indicated times. For NLRP1b inflammasome activation, the cells (J774, RAW264.7, or RAW-asc) were cultured with LF (2 μg/ml; no. 172C, List Biological Labs) together with PA (2 μg/ml; no. 171E, List Biological Labs) for the indicated times. LPS (L2018, Sigma-Aldrich)–primed J774 or RAW-asc cells were stimulated with 10 μM nigericin (tlrl-nig-5, InvivoGen) for NLRP3 inflammasome activation and transfected with Poly (dA:dT) (2 μg/ml; tlrl-patn-1, InvivoGen) for AIM2 inflammasome activation by using Lipofectamine LTX with Plus Reagent (15338100, Thermo Fisher Scientific). Human THP-1 cells were treated with 100 nM phorbol 12-myristate 13-acetate (PMA) for 3 hours and then cultured overnight without PMA before stimulation. Peritoneal macrophages and THP1 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS; SH30071.03, HyClone), penicillin (100 units/ml), and streptomycin (100 μg/ml). The J774, RAW264.7, and RAW-asc cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS (10099-141, Gibco), penicillin (100 units/ml), and streptomycin (100 μg/ml). Immunoblot analysis Cell lysates and culture supernatants were collected by adding 5× sample buffer [50% glycerol, 10% SDS, 5% 2-mercaptoethanol, 0.02% bromophenol blue, and 250 mM (pH 6.8) tris-HCl] for immunoblot analysis. Proteins were separated by 10 to 15% polyacrylamide gels, followed by electrophoretic transfer to polyvinylidene difluoride membranes (IPVH00010, Millipore). The membrane was then blocked by incubation with 5% bovine serum albumin before being incubated with primary antibodies. Antibodies used include caspase-1 (clone 4B4) [a gift from V. M. Dixit (Genetech, USA)], caspase-3 (9662, Cell Signaling Technology), pro–caspase-8 (4790, Cell Signaling Technology), cleaved caspase-8 (9429, Cell Signaling Technology), Apaf1 (8723, Cell Signaling Technology), caspase-9 (9508, Cell Signaling Technology), GSDMD (ab209845, Abcam), ASC (67824, Cell Signaling Technology), GSDME (ab215191, Abcam), and glyceraldehyde-3-phosphate dehydrogenase (AC002, ABclonal). Cross-linking of ASC oligomers To detect ASC oligomerization, cells were lysed with 0.5% Triton X-100 lysis buffer and then centrifuged at 6000 g at 4°C for 15 min. Supernatants were transferred to new tubes (lysates). The pellets were washed twice with tris-buffered saline and then cross-linked for 45 min at 37°C by disuccinimidyl suberate (2 mM; Thermo Fisher Scientific). The cross-linked pellets were centrifuged at 6000 g for 15 min, dissolved in SDS sample buffer, and subjected to Western blot. Confocal microscopy Mouse ASC was detected by using a rabbit-ASC monoclonal antibody (67824, Cell Signaling Technology) followed by Alexa Fluor 488 goat anti-rabbit immunoglobulin G (H + L) antibody (A11034, Invitrogen). The nuclei were stained by Hoechst (H1399, Invitrogen). The stained cells were examined under Axio Observer (Zeiss, Germany) at room temperature. Images were acquired and analyzed using Zen2012 software (Zeiss, Germany). LDH and cell viability assay Cytotoxicity was determined by measuring LDH activity in the culture medium using the Cytotoxicity LDH Assay Kit-WST (CK12-500-wells, Dojindo). For peritoneal lavage detection, 50 μl of assay reagent was added to peritoneal lavages of an equal volume isolated from mice by using 1 ml of PBS. The CellTiter-Glo Luminescent Cell Viability Assay Kit was used to determine the number of viable cells according to the manufacturer's instructions (G7571, Promega). Measurement of caspase-1, caspase-3/7, and caspase-8 activities Caspase-1, caspase-3/7, and caspase-8 activities were determined by using a caspase-Glo 1 (G9951, Promega), caspase-Glo 3/7 (G8092, Promega), or caspase-8 assay kit (G8202, Promega) according to the manufacturer's instructions. Cells were seeded in 96-well plate with white wall (Nunc). After treatment, an equal volume of caspase-Glo 1, caspase-Glo 3/7, or caspase-8 reagent was added to the cell culture medium and shaken for 30 min. Luminescent recording was performed with POLAR star Omega (BMG Labtech). For peritoneal lavage detection, 50 μl of assay reagent was added to peritoneal lavages of an equal volume isolated from mice by using 1 ml of PBS. Enzyme-linked immunosorbent assay Mouse TNF (88-7324-88, eBioscence), IL-1β (88-7013-88, eBioscence), and IL-6 (88-7064-77, eBioscence) were measured by enzyme-linked immunosorbent assay according to the manufacturer's instructions. Statistical analysis GraphPad Prism 5.0 software (GraphPad Software Inc) was used for data analysis. Data are shown as means ± SE. Survival curves were compared using log rank test (Mantel-Cox). P value less than 0.05 was considered to be statistically significant. *** P < 0.001 versus WT group and * P < 0.05 versus WT FlaTox group. Study design The objective of this study was to investigate the relation between canonical inflammasome–mediated cell death and host death. We deleted various canonical inflammasome–related genes individually or in combination in cells and mice and analyzed their effects on cell and mouse death as well as death pathways triggered by different canonical inflammasomes. The stimulation and replicates of each experiment are presented in the figure legends. Mice All mice used were housed in a specific pathogen-free environment. WT C57BL/6J mice were originally obtained from the Jackson Laboratory. Casp1/11 −/− and Gsdmd −/− mice were used as previously described ( 7 ). Asc −/− , Nlrc4 −/− , and Gsdme −/− mice were generated by co-microinjection of in vitro translated Cas9 mRNA and guide RNA (gRNA) into the C57BL/6J zygotes. The targeting sequence in the gRNA vector was 5′-CAGACGAGCCCTTATTCAA-3′ for mouse Nlrc4 , 5′-TATGGGCGCATCCCACGCG-3′ for mouse Asc , and 5′-TTCGCCTTTCTGGACATGC-3′ for mouse Gsdme . All KO alleles have been crossed onto the C57BL/6J background. Male mice between 8 and 12 weeks of age were used for the animal experiments in the current study. All animal experiments were conducted following the guidelines for housing and care of laboratory animals and performed in accordance with regulations approved by the Animal Care and Use Committee at Xiamen University. For intraperitoneal injection of FlaTox into mice, toxin doses were PA (2 μg/g body weight) combined with LFn-Fla or LFn-Fla 3A (4 μg/g) and diluted in 250 μl of phosphate-buffered saline (PBS). The survival rate of the animals was checked every few hours. Generation of KO cell lines using CRISPR-Cas9 technique The targeting sequence in the gRNA vector was 5′-TCTCTAAAAAAGGGCCCC-3′ for mouse caspase-1 , 5′-TGCAACAGCTTCGGAGTCG-3′ for mouse Gsdmd , 5′-TATGGGCGCATCCCACGCG-3′ for mouse Asc , 5′- GTCTAGGAAGTTGACCAGC-3′ for mouse caspase-8 , 5′-GGATGAAGAGCAGCTTCTCA-3′ for mouse Nlrp1b , 5′- ATGGATCACATGATCAGTAA-3′ for mouse Apaf1 , 5′-CTGGCTTCACTCTTGCAAAG-3′ for mouse caspase-9 , 5′-ATTGACTCCGTTATTCCGAA-3′ for human CASP1 , and 5′-CCCTCAAGTTCCTGAGCCT-3′ for human CASP8 . The plasmids (vector pBOB) harboring the gene gRNA sequences and Cas9 gene were transfected into 293T in the presence of lentivirus helper plasmids, and the supernatants were collected after 48 hours. The viruses were then used to infect J774 cells, RAW264.7 cells, RAW-asc cells, or THP1 cells. KOs were confirmed by immunoblots and further confirmed by sequencing. LFn fusion protein preparation pET15b LFn-Fla (Addgene plasmid no. 84871) and pET15b LFn-Fla 3A (Addgene plasmid no. 84872) were gifts from R. Vance. The LFn-Fla and LFn-Fla 3A were expressed and purified as described in previous publications ( 13 ). The endotoxin was removed with the ToxinEraserTM Endotoxin Removal Kit (L00338, GenScript). Bacterial strains for infection studies For in vivo infection, Salmonella SL7207 (ΔaroA) was incubated with shaking at 37°C in Luria-Bertani (LB) broth for 16 hours, and then 500 μl of medium with bacteria was added to a new tube with 15 ml of LB broth shaking until an optical density at 600 nm (OD 600 ) reached 1.0. The collected Salmonella was diluted in PBS, and 5 × 10 8 CFU for Salmonella SL7207 (ΔaroA) were injected intraperitoneally in a volume of 200 μl. The number of bacteria was examined by homogenizing organs of infected mice in 5 ml of sterilized PBS. The homogenate was diluted and plated onto LB agar plates and incubated at 37°C for 16 hours. Hematoxylin and eosin staining Animals were euthanized. Small intestines were collected and fixed immediately in 4% paraformaldehyde for 24 hours. The fixed tissues were embedded in a water-free procedure. Five-micrometer sections were cut and stained with hematoxylin and eosin. Slides were analyzed on the Leica Aperio Versa 200. Eicosanoid analysis For peritoneal lavage analysis, mice were injected intraperitoneally with FlaTox. After the indicated time, mice were euthanized and injected intraperitoneally with 1 ml of cold PBS. Peritoneal lavage (600 μl) was collected, and 50 μl of which was immediately transferred to 500 μl of cold methanol for storage at −80°C. Eicosanoids and docosanoids were identified and quantified by AB SCIEX QTRAP-6500 plus. The synthetic standards were purchased from Cayman Chemical. Cell cultures and stimulation Peritoneal macrophages were isolated from mice after being injected intraperitoneally with 3% thioglycollate solution for 4 days. For NLRC4 inflammasome activation, peritoneal macrophages were treated with LFn-Fla or LFn-Fla 3A with PA for the indicated times. For NLRP1b inflammasome activation, the cells (J774, RAW264.7, or RAW-asc) were cultured with LF (2 μg/ml; no. 172C, List Biological Labs) together with PA (2 μg/ml; no. 171E, List Biological Labs) for the indicated times. LPS (L2018, Sigma-Aldrich)–primed J774 or RAW-asc cells were stimulated with 10 μM nigericin (tlrl-nig-5, InvivoGen) for NLRP3 inflammasome activation and transfected with Poly (dA:dT) (2 μg/ml; tlrl-patn-1, InvivoGen) for AIM2 inflammasome activation by using Lipofectamine LTX with Plus Reagent (15338100, Thermo Fisher Scientific). Human THP-1 cells were treated with 100 nM phorbol 12-myristate 13-acetate (PMA) for 3 hours and then cultured overnight without PMA before stimulation. Peritoneal macrophages and THP1 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS; SH30071.03, HyClone), penicillin (100 units/ml), and streptomycin (100 μg/ml). The J774, RAW264.7, and RAW-asc cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS (10099-141, Gibco), penicillin (100 units/ml), and streptomycin (100 μg/ml). Immunoblot analysis Cell lysates and culture supernatants were collected by adding 5× sample buffer [50% glycerol, 10% SDS, 5% 2-mercaptoethanol, 0.02% bromophenol blue, and 250 mM (pH 6.8) tris-HCl] for immunoblot analysis. Proteins were separated by 10 to 15% polyacrylamide gels, followed by electrophoretic transfer to polyvinylidene difluoride membranes (IPVH00010, Millipore). The membrane was then blocked by incubation with 5% bovine serum albumin before being incubated with primary antibodies. Antibodies used include caspase-1 (clone 4B4) [a gift from V. M. Dixit (Genetech, USA)], caspase-3 (9662, Cell Signaling Technology), pro–caspase-8 (4790, Cell Signaling Technology), cleaved caspase-8 (9429, Cell Signaling Technology), Apaf1 (8723, Cell Signaling Technology), caspase-9 (9508, Cell Signaling Technology), GSDMD (ab209845, Abcam), ASC (67824, Cell Signaling Technology), GSDME (ab215191, Abcam), and glyceraldehyde-3-phosphate dehydrogenase (AC002, ABclonal). Cross-linking of ASC oligomers To detect ASC oligomerization, cells were lysed with 0.5% Triton X-100 lysis buffer and then centrifuged at 6000 g at 4°C for 15 min. Supernatants were transferred to new tubes (lysates). The pellets were washed twice with tris-buffered saline and then cross-linked for 45 min at 37°C by disuccinimidyl suberate (2 mM; Thermo Fisher Scientific). The cross-linked pellets were centrifuged at 6000 g for 15 min, dissolved in SDS sample buffer, and subjected to Western blot. Confocal microscopy Mouse ASC was detected by using a rabbit-ASC monoclonal antibody (67824, Cell Signaling Technology) followed by Alexa Fluor 488 goat anti-rabbit immunoglobulin G (H + L) antibody (A11034, Invitrogen). The nuclei were stained by Hoechst (H1399, Invitrogen). The stained cells were examined under Axio Observer (Zeiss, Germany) at room temperature. Images were acquired and analyzed using Zen2012 software (Zeiss, Germany). LDH and cell viability assay Cytotoxicity was determined by measuring LDH activity in the culture medium using the Cytotoxicity LDH Assay Kit-WST (CK12-500-wells, Dojindo). For peritoneal lavage detection, 50 μl of assay reagent was added to peritoneal lavages of an equal volume isolated from mice by using 1 ml of PBS. The CellTiter-Glo Luminescent Cell Viability Assay Kit was used to determine the number of viable cells according to the manufacturer's instructions (G7571, Promega). Measurement of caspase-1, caspase-3/7, and caspase-8 activities Caspase-1, caspase-3/7, and caspase-8 activities were determined by using a caspase-Glo 1 (G9951, Promega), caspase-Glo 3/7 (G8092, Promega), or caspase-8 assay kit (G8202, Promega) according to the manufacturer's instructions. Cells were seeded in 96-well plate with white wall (Nunc). After treatment, an equal volume of caspase-Glo 1, caspase-Glo 3/7, or caspase-8 reagent was added to the cell culture medium and shaken for 30 min. Luminescent recording was performed with POLAR star Omega (BMG Labtech). For peritoneal lavage detection, 50 μl of assay reagent was added to peritoneal lavages of an equal volume isolated from mice by using 1 ml of PBS. Enzyme-linked immunosorbent assay Mouse TNF (88-7324-88, eBioscence), IL-1β (88-7013-88, eBioscence), and IL-6 (88-7064-77, eBioscence) were measured by enzyme-linked immunosorbent assay according to the manufacturer's instructions. Statistical analysis GraphPad Prism 5.0 software (GraphPad Software Inc) was used for data analysis. Data are shown as means ± SE. Survival curves were compared using log rank test (Mantel-Cox). P value less than 0.05 was considered to be statistically significant. *** P < 0.001 versus WT group and * P < 0.05 versus WT FlaTox group. Funding: This work was supported by the National Natural Science Foundation of China (81788101 and 81630042 to J.H. and 31701205 to P.Z.), the National Scientific and Technological Major Project (2017ZX10202203-003 to J.H.), the National Key R&D program (2020YFA0803500 to J.H.), the 111 Project (B12001 to J.H.), the CAMS Innovation Found for Medical Science (CIFMS) (2019-I2M-5-062 to J.H.), and the U.S. National Institute of Health grant (5U19 AI100627-08 to R.J.U.). Author contributions: P.Z., Y.L., L.H., K.H., M.H., Y.W., and X.F., carried out experimental work. P.Z, Y.L., R.J.U., and J.H. designed experiments and interpreted the data. P.Z. and J.H. wrote the manuscript. J.H. conceived and supervised the study. The authors reviewed and approved the final version of the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Supplementary Materials This PDF file includes: Figs. S1 to S10 Click here for additional data file. This PDF file includes: Figs. S1 to S10 Click here for additional data file. View/request a protocol for this paper from Bio-protocol .

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3910841/

Amidate Prodrugs of 9-[2-(Phosphonomethoxy)Ethyl]Adenine as Inhibitors of Adenylate Cyclase Toxin from Bordetella pertussis

Adenylate cyclase toxin (ACT) is the key virulence factor of Bordetella pertussis that facilitates its invasion into the mammalian body. 9-[2-(Phosphonomethoxy)ethyl]adenine diphosphate (PMEApp), the active metabolite of the antiviral drug bis(POM)PMEA (adefovir dipivoxil), has been shown to inhibit ACT. The objective of this study was to evaluate six novel amidate prodrugs of PMEA, both phenyloxy phosphonamidates and phosphonodiamidates, for their ability to inhibit ACT activity in the J774A.1 macrophage cell line. The two phenyloxy phosphonamidate prodrugs exhibited greater inhibitory activity (50% inhibitory concentration [IC 50 ] = 22 and 46 nM) than the phosphonodiamidates (IC 50 = 84 to 3,960 nM). The inhibitory activity of the prodrugs correlated with their lipophilicity and the degree of their hydrolysis into free PMEA in J774A.1 cells. Although the prodrugs did not inhibit ACT as effectively as bis(POM)PMEA (IC 50 = 6 nM), they were significantly less cytotoxic. Moreover, they all reduced apoptotic effects of ACT and prevented an ACT-induced elevation of intracellular [Ca 2+ ]i. The amidate prodrugs were less susceptible to degradation in Caco-2 cells compared to bis(POM)PMEA, while they exerted good transepithelial permeability in this assay. As a consequence, a large amount of intact amidate prodrug is expected to be available to target macrophages in vivo . This feature makes nontoxic amidate prodrugs attractive candidates for further investigation as novel antimicrobial agents.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7169652/

Infectious Diseases

Infectious diseases is one of the most common cause of the visits to pediatric office and emergency department as well. Prevention of Infectious Diseases Child-Care Center Risk of acquiring infections in child-care center Poor hygiene increases the risk of young children for recurrent infections and development of antibiotic resistance . Prevention Good hand washing; wash hands with soap and water, alcohol-based antiseptic is acceptable Disinfecting environmental surfaces Frequent facility cleaning Appropriate food handling Teach children and staff to sneeze or cough into elbow (not hands) Use gloves when contacting body fluids Common organism in child-care centers: Shigella infection Transmitted from infected feces (person-to-person contact) Do : stool bacterial cultures for any symptomatic contact Know : if Shigella infections are confirmed should receive appropriate antibacterial treatment Return to child-care center ◦ If diarrhea has resolved and stool cultures are negative Nontyphoidal Salmonella species No antibiotic is required except: ◦ Infants younger than 3 months of age ◦ Immunocompromised host Infected individuals should be excluded from child care until symptoms resolve Salmonella serotype typhi Treatment is indicated for infected individuals Return to child-care center ◦ 5 years of age or younger : 48 h after antibiotic treatment ◦ Older than 5 years : 24 h after the diarrhea has resolved Other risk of infection : e.g., giardia, rotavirus, cryptosporidiosis, respiratory syncytial virus (RSV), parainfluenza virus, adeno, rhino, and corona viruses hemophilus influenza , pneumococcal, hepatitis A and, cytomegalovirus infections Prevention of Hospital and Office Infection Standard precautions are indicated in the care of all patients including: Hand hygiene before and after each patient contact Protective equipment when needed Preventive methods Alcohol-based products are preferred because of their superior activity and adherence Soap and water are preferred when hands are visibly soiled or exposed to a spore-forming organism, e.g., ( Clostridium difficile is the most common) Gloves, isolation gowns, masks, and goggles for any exposure to body fluids contaminated materials or sharps Strict aseptic technique for all invasive procedures, and for catheter care Separate well and sick children areas in the medical offices Examples of infections and agents requiring transmission-based precautions Contact precautions , e.g., RSV, C. difficile , and Staphylococcus aureus Gloves and gowns are required when there is direct patient contact Droplet precautions , e.g., Influenza, Neisseria meningitidis , and Bordetella pertussis Use of a surgical mask is required A single room is preferred Remember all office and hospital staff should receive an annual influenza immunization Airborne precautions , e.g., Mycobacterium tuberculosis , measles, and varicella (with contact precautions) Negative pressure airborne infection isolation room Room needs 6–12 air changes per hour or recirculated through a high-efficiency particulate air (HEPA) filter Tested N95 or similar sealing mask Prevention of Infection Through Breast Feeding Exclusive breastfeeding for the first 6 months is recommended by American Academy of Pediatrics (AAP) Immunologic characteristics of breast milk Postpartum colostrum contains high concentrations of antibodies and other infection-protective elements (natural immunization). The actual antibodies against specific microbial agents present in an individual woman's milk depends on her exposure and response to the particular agents. Lactoferrin : Limits bacterial growth by iron chelation. Lysozyme : Bacterial cell wall lysis. Lactalbumin : Enhance the growth Bifidobacterium and affects immune modulation. Casein : Limits adhesion of bacteria and facilitates the growth of Bifidobacterium. Carbohydrates : Enhance the growth of probiotics. Lipids : Lytic effect on many viruses and are active against Giardia as well. Absolute contraindication of breast feeding Human immunodeficiency virus 1 (HIV-1) infection (if replacement feeding is acceptable, feasible, affordable, sustainable, and safe) Human T-lymphotropic virus 1 and 2 infection (varies by country; in Japan, breastfeeding is initiated) Tuberculosis (active, untreated pulmonary tuberculosis, until effective maternal treatment for the initial 2 weeks or the infant is receiving isoniazid) Herpes simplex virus infection on a breast (until the lesions on the breast are cleared) Medical Evaluation of Internationally Adopted Children Evaluation for tuberculosis (TB) infection and purified protein derivative (PPD) testing Immunizations Written immunization record is accepted for the number of doses, interval, and appropriate age of immunization Serologic testing to determine protective antibodies: Tetanus antibodies (the test of choice) other antibodies for diphtheria, polio, and hepatitis B can be measured Pertussis titer do not reliably predict protection against infection Measles vaccine should not be administered routinely to children younger than 1 year Prevention of Vector-Borne Disease Chemoprophylaxis before travelling to endemic areas, e.g., mefloquine for malaria should be given before travelling to endemic areas Use mosquito netting during sleep in tropical areas Use protective clothing and garments Repellents , e.g., DEET (  2 months) with meningitis in combination with third generation cephalosporin Methicillin-resistant staphylococcal infection Prophylaxis before prosthetic device implantation requiring major surgery Enterally for C. difficile Acute infectious endocarditis if S. aureus is the likely cause Adverse reaction Red man syndrome, or red neck syndrome Vancomycin releases histamine that can cause pruritus, erythema of the head and neck This is a related drug infusion problem just slow down the infusion rate and premedicate the patient with diphenhydramine Ototoxicity and nephrotoxicity (follow the trough level and adjust the dose accordingly) Misuse of vancomycin cause development of resistance Indications C. difficile diarrhea (It is not systemically absorbed) S. aureus infections Clindamycin Mechanism of action Inhibit bacterial protein synthesis by binding to 50S ribosomal subunit Bacterial coverage Active against many strains of methicillin-resistant S. aureus (MRSA) Active against anaerobes Active against most staphylococcal and streptococcal infections Adverse reaction Diarrhea including C. difficile enterocolitis Macrolides, e.g., azithromycin and clarithromycin Mechanism of action Inhibit bacterial protein synthesis by binding to 50S ribosomes Azithromycin does not inhibit cytochrome P-450 as erythromycin or clarithromycin do Bacterial coverage Azithromycin is the drug of choice for pertussis, Mycoplasma and Chlamydia Adverse reaction Gastrointestinal irritation Hypertrophic pyloric stenosis if used in children less than 1 month of age Rifampin Bacterial coverage Tuberculosis Invasive H. influenzae Indications Close contacts to a child who has invasive meningococcal infection Combination with vancomycin in certain staphylococcal infections (VP shunt, osteomyelitis, endocarditis) Persistent group A streptococcal pharyngitis in combination with beta-lactam antibiotics MRSA carriage eradication attempt Fluoroquinolones, e.g., ciprofloxacin AAP recommendation of fluoroquinolones use in children If the pathogen is multidrug resistant No safe and other effective alternative Parenteral therapy is not feasible No other effective alternative oral agents Bacterial coverage UTIs caused by multidrug resistant gram negatives rods Resistant gram negative rods: P. aeruginosa Gastrointestinal and respiratory tract infection Chronic or acute osteomyelitis Adverse reaction Fluoroquinolones has no documented evidence of increased incidence of arthropathy in pediatric patient using fluoroquinolones Tetracycline Bacterial coverage Tetracycline provides coverage against tick borne organisms, e.g., (Lyme disease, Rocky Mountain spotted fever) Doxycycline and minocycline are used for acne ( Propionibacterium acnes ) Doxycycline may have MRSA coverage as well Adverse reaction Tetracyclines causes staining of dental enamels. Tetracycline is not recommended in children less than 8 years old. Tetracyclines can be used in children younger than 8 years in life threatening situations, e.g., rocky mountain spotted fever (doxycyclines is the drug of choice). Doxycycline does not cause staining of permanent teeth comparing to tetracyclines. Trimethoprim/sulfamethoxazole Bacterial coverage Pneumocystis jiroveci which is common in immunocompromised patient, e.g., HIV Urinary tract infection , treatment, and prophylaxis (drug of choice in susceptible patients) Methicillin-resistant staphylococcal infection Gastroenteritis due to salmonella, shigella, and isospora belli Burkholderia cepacia Brucella Adverse reaction Rash Neutropenia Stevens–Johnson syndrome Vancomycin Mechanism of action Inhibits bacterial cell wall synthesis by binding tightly to peptidoglycan precursors and blocking polymerization Bacterial coverage Confirmed gram positive infection in patient seriously ill or allergic to beta-lactam antibiotics Initial empiric treatment in a child (> 2 months) with meningitis in combination with third generation cephalosporin Methicillin-resistant staphylococcal infection Prophylaxis before prosthetic device implantation requiring major surgery Enterally for C. difficile Acute infectious endocarditis if S. aureus is the likely cause Adverse reaction Red man syndrome, or red neck syndrome Vancomycin releases histamine that can cause pruritus, erythema of the head and neck This is a related drug infusion problem just slow down the infusion rate and premedicate the patient with diphenhydramine Ototoxicity and nephrotoxicity (follow the trough level and adjust the dose accordingly) Misuse of vancomycin cause development of resistance Indications C. difficile diarrhea (It is not systemically absorbed) S. aureus infections Antivirals Acyclovir Mechanism of action Terminates the viral deoxyribonucleic acid (DNA) synthesis when incorporated into the viral DNA chain . Appropriate use of acyclovir Herpes simplex virus (HSV) type 1 and HSV type 2 Varicella Treatment of recurrent primary genital HSV2 or primary HSV1 mucocutaneous infections IV acyclovir is the drug of choice for treatment of HSV encephalitis Major side effect of acyclovir Acute renal failure due to precipitation in the renal tubules (proper hydration and slower infusion can minimize this problem) Nausea, vomiting, and diarrhea Valacyclovir Background Newer potent oral antiviral (Inhibits DNA polymerase; incorporates into viral DNA) Indications HSV1 HSV2 Varicella-Zoster virus (VZV) Ganciclovir Indications CMV infection Foscarnet CMV infection Acyclovir Mechanism of action Terminates the viral deoxyribonucleic acid (DNA) synthesis when incorporated into the viral DNA chain . Appropriate use of acyclovir Herpes simplex virus (HSV) type 1 and HSV type 2 Varicella Treatment of recurrent primary genital HSV2 or primary HSV1 mucocutaneous infections IV acyclovir is the drug of choice for treatment of HSV encephalitis Major side effect of acyclovir Acute renal failure due to precipitation in the renal tubules (proper hydration and slower infusion can minimize this problem) Nausea, vomiting, and diarrhea Valacyclovir Background Newer potent oral antiviral (Inhibits DNA polymerase; incorporates into viral DNA) Indications HSV1 HSV2 Varicella-Zoster virus (VZV) Ganciclovir Indications CMV infection Foscarnet CMV infection Other Antiviral Agents, Against DNA Viruses Famciclovir, valganciclovir, penciclovir, and cidofovir Nucleoside Reverse Transcriptase Inhibitors Mechanism of action These drugs inhibit replication of HIV by interfering with the reverse transcriptase enzyme Indication HIV infection Example of nucleoside reverse transcriptase inhibitors and their side effects Zidovudine (ZDV) Significant side effect; bone marrow suppression Didanosine (ddI) Significant side effects; pancreatitis and peripheral neuropathy Zalcitabine (ddC) Significant side effects; stomatitis and neuropathy Stavudine (d4T) Contraindication: ◦ Cannot be combined with ddI in pregnant women can cause fatal lactic acidosis Side effects; pancreatitis and peripheral neuropathy Abacavir Most serious side effect is FATAL hypersensitivity Nonnucleoside Reverse Transcriptase Inhibitors (NNRTI) Indication HIV infection Example of NNRTI and common side effects Efavirenz Teratogenic Nevirapine Rash Protease Inhibitors Mechanism of action Inhibit the HIV protease enzyme that involved with processing the completed virus Indication HIV infection Examples of protease inhibitors medications and the common side effects Indinavir Asymptomatic hyperlipidemia Nephrolithiasis Nelfinavir Diarrhea Saquinavir Nucleoside Reverse Transcriptase Inhibitors Mechanism of action These drugs inhibit replication of HIV by interfering with the reverse transcriptase enzyme Indication HIV infection Example of nucleoside reverse transcriptase inhibitors and their side effects Zidovudine (ZDV) Significant side effect; bone marrow suppression Didanosine (ddI) Significant side effects; pancreatitis and peripheral neuropathy Zalcitabine (ddC) Significant side effects; stomatitis and neuropathy Stavudine (d4T) Contraindication: ◦ Cannot be combined with ddI in pregnant women can cause fatal lactic acidosis Side effects; pancreatitis and peripheral neuropathy Abacavir Most serious side effect is FATAL hypersensitivity Nonnucleoside Reverse Transcriptase Inhibitors (NNRTI) Indication HIV infection Example of NNRTI and common side effects Efavirenz Teratogenic Nevirapine Rash Protease Inhibitors Mechanism of action Inhibit the HIV protease enzyme that involved with processing the completed virus Indication HIV infection Examples of protease inhibitors medications and the common side effects Indinavir Asymptomatic hyperlipidemia Nephrolithiasis Nelfinavir Diarrhea Saquinavir Antiparasites Permethrin Excellent safety profile Five percent permethrin is the drug of choice for treatment of scabies It paralyze the parasite and cause death One percent permethrin solution is effective for head lice It is not recommended in infants younger than 2 months and during pregnancy Metronidazole Mechanism of action Metronidazole is nitroimidazole bactericidal drug Indications Anaerobic bacteria Clostridium Trichomonas vaginalis Gardnerella vaginalis Treponema pallidum Oral spirochetes Helicobacter pylori Malathion It is the most effective drug in the treatment of pediculosis or head lice It has ovicidal activity Single topical application is effective in resistant cases Chloroquine Indication It is the drug of choice for malaria prophylaxis in the sensitive chloroquine regions, e.g., Central and South America Drug should be administered 1–2 weeks before travelling Adverse effect Gastrointestinal (GI) upset, headache, dizziness, blurred vision, insomnia, and pruritus Mefloquine and atovaquone/proguanil Commonly used for prophylaxis for malaria in chloroquine resistant regions, e.g., Africa and Middle east Permethrin Excellent safety profile Five percent permethrin is the drug of choice for treatment of scabies It paralyze the parasite and cause death One percent permethrin solution is effective for head lice It is not recommended in infants younger than 2 months and during pregnancy Metronidazole Mechanism of action Metronidazole is nitroimidazole bactericidal drug Indications Anaerobic bacteria Clostridium Trichomonas vaginalis Gardnerella vaginalis Treponema pallidum Oral spirochetes Helicobacter pylori Malathion It is the most effective drug in the treatment of pediculosis or head lice It has ovicidal activity Single topical application is effective in resistant cases Chloroquine Indication It is the drug of choice for malaria prophylaxis in the sensitive chloroquine regions, e.g., Central and South America Drug should be administered 1–2 weeks before travelling Adverse effect Gastrointestinal (GI) upset, headache, dizziness, blurred vision, insomnia, and pruritus Mefloquine and atovaquone/proguanil Commonly used for prophylaxis for malaria in chloroquine resistant regions, e.g., Africa and Middle east Antifungals Amphotericin B Indication Active against broad array of fungi, e.g., Candida , Aspergillus , Zygomycetes, Histoplasma , Coccidioides immitis Toxicity Febrile drug reaction Hypokalemia Hypomagnesemia Nephrotoxicity (liposomal preparation is equally effective and less nephrotoxic) Fluconazole Indications It is equally effective for treatment of invasive Candida albicans in neonates as amphotericin B Treatment of oropharyngeal or esophageal candidiasis in immunocompromised patients Treatment of vulvovaginal Candida Treatment of cryptococcal meningitis Griseofulvin It is the standard first-line therapy for tinea capitis No laboratory assessment of hepatic enzyme if used   1 year to susceptible contact Rabies Virus Background Rabies virus is a RNA virus classified in the Rhabdoviridae family Usually is transmitted by bats and carnivores, e.g., raccoon, foxes, and coyotes Clinical presentation Anxiety Dysphagia Seizures Encephalitis In most cases progress to death Prophylaxis recommendation All person bitten by, bats, carnivores, e.g., raccoon, foxes, and coyotes Domestic animals that may be infected Open wound or scratch contaminated with saliva of infected animals or human Prompt local flushing and cleaning the wound with soap and water The need for tetanus and antibiotic should be considered Passive and active immunization should be started as soon as possible Human rabies immunoglobulin (passive). Rabies vaccine (active). Both should be given together. Human rabies immunoglobulin as much as possible of the dose should be infiltrated directly to wound, the remainder of the dose should be given intramuscularly. Rabies vaccine should be given IM, the first dose immediately after exposure then repeated at days 3, 7, and 14. Arboviruses West Nile virus Dengue fever West Nile Virus Background It is the most common arbovirus identified in the USA West Nile virus is transmitted by mosquitoes Typically the spring and summer California, Colorado, and Idaho are the most common location Clinical presentation Most cases are asymptomatic May present with fever and flu-like symptoms Fever, headache, altered mental status, paresis, nerve palsies, or coma in more severe cases Diagnosis Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment Supportive Dengue Fever Background Dengue fever is an arbovirus transmitted by mosquitoes Typically the spring and summer History of travel to endemic area is the most important part to assist in the diagnosis of Dengue fever Endemic in Latin America and Puerto Rico Key West, Miami, Florida are endemic areas in the USA Clinical presentation Severe muscle, and joint pain Headache, and retro-orbital pain Nonspecific rash, nausea, vomiting, diarrhea, and respiratory symptoms It can lead to dengue shock syndrome and death Laboratory It may show leukopenia, thrombocytopenia, and modest elevation of liver enzyme Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment is supportive Enteroviruses Non-polio viruses (coxsachievirus A and B, echoviruses and enterovirus) Background More common in the summer Enteroviruses transmitted by the feco-oral route and person to person Meningitis/Encephalitis Meningitis commonly caused by echovirus Common in older children Fever, headache, photophobia, and nuchal rigidity, CSF pleocytosis Severe complications: seizure, hemiparesis, hearing loss, and mental deterioration No signs toxicity as in bacterial meningitis Best diagnostic test : CSF enterovirus PCR Herpangina Caused by Coxsackievirus type A is a subgroup of enterovirus which is a subgroup of picornavirus Sudden onset of high fever in 3–10 years of age, and can be associated with vomiting, malaise, myalgia, and backache Poor intake, drooling, sore throat, dysphagia, and dehydration may occur Oral lesions: ◦ One or more small tender papular pinpoint vescular lesions, on erythematous base on anterior pillars of the faucets, soft palate, uvula, tonsils, and tongue, then ulcerate in 3–4 days. Hand-foot-mouth disease (Fig. 5 ) Fig. 5 Hand-foot-mouth disease: a. Tender vesicles and macules on an erythematous base, and crusted vesicles on the foot and the leg. b. Multiple vesicles that erode and become surrounded by an erythematous halo in the mouth. c. Erythematous macules and vesicles on the palm Coxsackie A16 and enterovirus 71 Fever (may be present) Oral vesicles and ulcers on buccal mucosa and tongue Painful vesicles on hands and feet, it may affect the groin, and buttocks Usually last for 7–10 days Most common complication is dehydration due to odynophagia Acute hemorrhagic conjunctivitis Subconjunctival hemorrhage Swelling, redness, and tearing of the eye Resolve spontaneously within 7 days Myocarditis/pericarditis Commonly caused by Coxsackievirus B or echovirus Common symptoms; shortness of breath, chest pain, fever, and weakness Congenital and neonatal infection Can range from mild febrile infection to encephalitis and negative bacterial culture Can cause hepatic necrosis Poliovirus infection Background Polioviruses are enterovirus belong to family of Picornaviridae Clinical presentation Fever common in less than 6 years of age Aseptic meningitis Flaccid paralysis in a descending manner without reflexes The poliovirus destroys the anterior horn cells in the spinal cord Diagnosis Viral stool culture Throat swab Treatment No curative treatment Prevention Polio vaccine (IPV/OPV) Human Immunodeficiency Virus (HIV) Background HIV is RNA virus Highest infectivity due to the very high (3–4 weeks) initial viremia Nearly all patients seroconvert within 6 months of acquiring the infection Mode of transmission HIV infection is transmitted by two principal modes in the pediatric age group: Mother-to-child Transplacental transfer Exposure to maternal blood, amniotic fluid, and cervicovaginal secretions during delivery Postpartum through breastfeeding Behavioral (risk behavior in adolescent either unprotected sex or injection drugs) Clinical presentation During the "window period: Infected person has a negative HIV antibody test result, but HIV RNA testing results are usually positive Acute retroviral syndrome, characterized by: Fever, lymphadenopathy, rash, myalgia, arthralgia, headache, diarrhea, oral ulcers, leukopenia, thrombocytopenia, and transaminitis Red flags of HIV infection Thrush in apparently healthy child or adolescent Invasive candidal infections Recurrent severe infections Lymphadenopathy and/or hepatosplenomegaly Failure to thrive Parotid enlargement Diagnosis Infants born to HIV-positive mothers Most infants are normal at birth and then may develop lymphadenopathy, HSM, chronic diarrhea, failure to thrive , and oral candidiasis. Within the first 48 h, 14 days, and 4 weeks of life, 38, 93, and 96 % of infected children, respectively, have positive HIV DNA PCR results. Any positive HIV DNA PCR finding should be confirmed with follow-up HIV DNA PCR before infection is diagnosed. HIV DNA PCR testing: HIV infection can be ruled out if one of the following is true: DNA HIV PCR results are consistently negative in an infant older than 4 months in the absence of breastfeeding. Two DNA HIV PCR results obtained at least one month apart are negative in an infant older than 6 months. HIV antibody testing between 12 and 18 months of age to confirm the loss of maternal antibody is optional. Screening and diagnosis of children older than age 18 months Screening enzyme-linked immunoassay (EIA) Confirmatory test such as western blot is performed if EIA is positive Evaluation of HIV positive children CD4 percentage and absolute cell counts Plasma HIV RNA concentration (viral load) HIV genotype to assess for baseline resistance, and mutations Complete blood count with differential count Serum chemistries with liver and renal function tests Lipid profile and urinalysis For children younger than 5 years of age, CD4 percentage is the preferred test for monitoring immune status Screening for hepatitis B and C infection as well as for tuberculosis is recommended for all HIV-infected patients Treatment of HIV Triple-drug combination antiretroviral therapy effectively controls HIV infection Prevention Breastfeeding is contraindicated in HIV positive mothers All exposed infants should receive 6 weeks of ZDV Condoms and abstinence are the best forms of preventing sexual transmission of AIDS Cesarean delivery and treatment of HIV-positive mothers (specially with high viral load) decreases the risk of transmission of HIV to their infants Immunization of infants and children Immunization schedule for HIV-exposed children is the same as for their healthy peers, with only a few exceptions: ◦ Patients who have severely symptomatic illness. ◦ Patient with CD4 percentage of less than 15 % or CD4 counts of less than 200 cells/mm 3 should not receive measles-mumps-rubella (MMR), varicella vaccines or live vaccines. Annual influenza immunization is recommended for all children older than age 6 months, but only the killed vaccine. Measles Background Mode of transmission : respiratory droplets (airborne). The virus is infectious for 3–4 days before the onset of morbilliform rash and 4 days after the exanthem. Diagnosis IgM level serology (most reliable test) Antigen detection in respiratory epithelial cells Tissue by immunofluorescent method or PCR Clinical presentation Coryza Cough Conjunctivitis High fever Koplik spots Rash is erythematous maculopapular rash spread from up–down and disappear the same way Prevention Intramuscular (IM) immunoglobulin prophylaxis should be given to unimmunized child if exposed to measles infection Infants (6–12 months) should be pre-vaccinated before travelling to high risk areas, e.g., India. Children received measles vaccine before 1 year do not count and need to receive two doses of MMR after 12 months for full immunization. Infected child with measles should be placed under airborne precaution transmission and isolated for 4 days after the rash and for all duration of illness if immunocompromised. Complications Otitis media is the most common Pneumonia (common cause of death) Encephalitis Subacute sclerosing panencephalitis (SSPE) is rare and it may occur after 6–15 years Mumps Background Mumps is an acute, self-limited, systemic viral illness characterized by the swelling of one or more of the salivary glands, typically the parotid glands. The illness is caused by a specific RNA virus, known as Rubulavirus. Mode of transmission Airborne and contact to respiratory secretions Incubation period is 12–25 days Clinical presentation Symptoms in the patient's history consist mostly of fever, headache, and malaise. Within 24 h, patients may report ear pain localized near the lobe of the ear and aggravated by a chewing movement of the jaw. Unilateral or bilateral parotid swelling at least for 2 days. Complications Encephalitis and orchitis Arthritis, thyroiditis, pancreatitis, myocarditis, oophoritis (rare) Diagnosis Serology and virus isolation Prevention MMR vaccine at 1 and 4 years of age Isolation of infected individual is 9 days from the onset of parotid swelling Unimmunized children should stay at home for 26 days from the last case in school Rubella Background The name rubella is derived from a Latin term meaning "little red". Rubella is generally a benign communicable exanthematous disease. It is caused by rubella virus, which is a member of the Rubivirus genus of the family Togaviridae. Disease transmission: by droplet inhalation from the respiratory tract of an infected host. Incubation period: 14–21 days. Communicability: Patients are infectious 2 days before and 5–7 days after the rash. Clinical presentation Lymphadenopathy: Retroauricular Postauricular Posterior occipital Rash: Maculopapular erythematous rash last for 3 days Forschheimer spots; rose colored spot on soft palate Other manifestation: Pharyngitis and conjunctivitis Anorexia, headache, and malaise Low-grade fever and polyarthritis Complications Congenital rubella syndrome Cataract, salt and pepper chorioretinitis, and deafness PDA IUGR and microcephaly HSM and jaundice Blueberry muffin rash Anemia , thrombocytopenia, and leukopenia B-cell, and T-cell deficiency Metaphyseal lucencies Infant with congenital rubella may shed the virus from the nasal mucosa > 1 year to susceptible contact Rabies Virus Background Rabies virus is a RNA virus classified in the Rhabdoviridae family Usually is transmitted by bats and carnivores, e.g., raccoon, foxes, and coyotes Clinical presentation Anxiety Dysphagia Seizures Encephalitis In most cases progress to death Prophylaxis recommendation All person bitten by, bats, carnivores, e.g., raccoon, foxes, and coyotes Domestic animals that may be infected Open wound or scratch contaminated with saliva of infected animals or human Prompt local flushing and cleaning the wound with soap and water The need for tetanus and antibiotic should be considered Passive and active immunization should be started as soon as possible Human rabies immunoglobulin (passive). Rabies vaccine (active). Both should be given together. Human rabies immunoglobulin as much as possible of the dose should be infiltrated directly to wound, the remainder of the dose should be given intramuscularly. Rabies vaccine should be given IM, the first dose immediately after exposure then repeated at days 3, 7, and 14. Arboviruses West Nile virus Dengue fever West Nile Virus Background It is the most common arbovirus identified in the USA West Nile virus is transmitted by mosquitoes Typically the spring and summer California, Colorado, and Idaho are the most common location Clinical presentation Most cases are asymptomatic May present with fever and flu-like symptoms Fever, headache, altered mental status, paresis, nerve palsies, or coma in more severe cases Diagnosis Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment Supportive Dengue Fever Background Dengue fever is an arbovirus transmitted by mosquitoes Typically the spring and summer History of travel to endemic area is the most important part to assist in the diagnosis of Dengue fever Endemic in Latin America and Puerto Rico Key West, Miami, Florida are endemic areas in the USA Clinical presentation Severe muscle, and joint pain Headache, and retro-orbital pain Nonspecific rash, nausea, vomiting, diarrhea, and respiratory symptoms It can lead to dengue shock syndrome and death Laboratory It may show leukopenia, thrombocytopenia, and modest elevation of liver enzyme Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment is supportive Arboviruses West Nile virus Dengue fever West Nile Virus Background It is the most common arbovirus identified in the USA West Nile virus is transmitted by mosquitoes Typically the spring and summer California, Colorado, and Idaho are the most common location Clinical presentation Most cases are asymptomatic May present with fever and flu-like symptoms Fever, headache, altered mental status, paresis, nerve palsies, or coma in more severe cases Diagnosis Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment Supportive Dengue Fever Background Dengue fever is an arbovirus transmitted by mosquitoes Typically the spring and summer History of travel to endemic area is the most important part to assist in the diagnosis of Dengue fever Endemic in Latin America and Puerto Rico Key West, Miami, Florida are endemic areas in the USA Clinical presentation Severe muscle, and joint pain Headache, and retro-orbital pain Nonspecific rash, nausea, vomiting, diarrhea, and respiratory symptoms It can lead to dengue shock syndrome and death Laboratory It may show leukopenia, thrombocytopenia, and modest elevation of liver enzyme Fourfold rise in virus-specific serum antibodies, or positive IgM-CSF antibody titer is helpful in the diagnosis Treatment is supportive Hepatitis A Virus (HAV) Background HAV is the most common cause of viral hepatitis worldwide No known animal reservoir Mode of transmission is fecal-oral route Incubation period is 15–50 days Highest period of communicability is 1 week before and after the onset of symptoms CD8 + T cells are responsible for the destruction of infected liver cells Clinical presentation In children younger than 5 years may be asymptomatic or with just few symptoms Older children and adult may develop symptoms of acute infection which may last 2 weeks to several months Malaise, anorexia, fever, nausea, vomiting, and eventually jaundice Most of the cases generally resolve without sequelae within a few weeks Diagnosis Anti-HAV immune globulin M (IgM) in a single serum sample is a good test for current or recent infection. Prevention HAV vaccine at 12 months and booster dose at least 6 months after the initial dose. Prevention of HAV infection can be promoted by enforcing good hygiene in child care centers, with conscientious hand washing after changing diapers and before handling food. If travelling is imminent to endemic areas or the patient is immunocompromised, immunoglobulin (IG) can be administered simultaneously with vaccine. Treatment Mainly supportive Avoid acetaminophen, it can exacerbate damage to liver cells Prognosis HAV does not carry the risk of chronic infection Immunity after infection is life-long Hepatitis B Virus (HBV) Background The infection has an incubation period of 2–6 months HBV is commonly transmitted via body fluids such as blood, semen, and vaginal secretions HBV does not spread by breast feeding, kissing, hugging, sharing utensils Clinical presentation Acute self-limited hepatitis: Increase in serum transaminases and resolution of the infection within 6 months Nausea Fever Abdominal pain Jaundice, fatigue General malaise Fulminant hepatitis: Acute hepatitis associated with a change in mental status due hepatic encephalopathy Chronic hepatitis: Generally is asymptomatic in childhood, having minimal or no effect on growth and development Serum transaminase values usually are normal They can flare at any time Hepatitis B viral serology and liver functions tests HBsAg is the first serologic marker to appear and found in infected persons, its rise correlates with the acute symptoms. Anti-HBc is the single most valuable serologic marker of acute HBV infection, because it appears as early as HBsAg, and continue later in the course of the disease when HBsAg disappeared. Anti-HBs marks serologic recovery and protection; marks vaccine immunity. Both Anti HBs and Anti HBc are detected in person with resolved infection. HBeAg is present in person with active acute or chronic infection and marks infectivity. Anti-HBe marks improvement and is the goal of therapy in chronically infected patients. Remember : Alanine transaminase (AST) and aspartate aminotransferase (ALT) can be derived from muscle, you should verify that serum creatine kinase and aldolase values are within the normal range before assuming that the elevated serum AST and ALT values are hepatic in origin. Test reflecting cholestasis High-serum concentrations of gamma-glutamyl transferase High-serum alkaline phosphatase High-conjugated bilirubin Test reflecting liver failure High-prothrombin time, despite administration of vitamin K Low-serum albumin concentrations are the most useful indicators of impaired synthetic liver function HBV perinatal infection Nearly all perinatally acquired HBV infection are asymptomatic Maternal screening of all pregnant women for HBV is now standard Prophylaxis for all newborns of HBV-positive women in the first 12 h after birth: ◦ Combination of passive (IgG) and active immunization (first dose of the vaccine) followed by the complete HBV vaccine schedule Breastfeeding does not increase the risk of transmission Treatment is mainly supportive Interferon-Alpha2b and lamivudine are the current approved therapy Hepatitis C Viral Infection (HCV) Background HCV is a spherical, enveloped, single-stranded RNA virus belonging to the Flaviviridae family and Flavivirus genus Egypt had the highest number of reported infections with 22 % prevalence of HCV antibodies in persons in Egypt. Mode of transmission Infants and children The maternal-fetal route is the principal route of transmission Adults Injection during drug abuse is the most common mode of transmission Long term complication of HCV infection Chronic carrier Chronic hepatitis Hepatocellular carcinoma Testing for HCV HCV infection is investigated by measuring anti-HCV antibody and is confirmed by the detection of serum HCV RNA by PCR. Screening of infants born to HCV-infected mothers is recommended by measuring serum anti-HCV antibody at 18 months of age . Know that children with chronic hepatitis C infection should undergo periodic screening tests for hepatic complications and the treatment regimens are available. Treatment (see GI chapter for more details) Genotype 1 is the most aggressive and most resistant to antiviral therapy Genome 2 and 3 has a better response Remember : A high rate of spontaneous mutations in the viral genome is the reason for the lack of an effective vaccine . Human Papillomavirus (HPV) Background Oncogenic strain 16 and 18 are responsible for two thirds of all cervical cancers Nononcogenic HPV type 6 and 11 are responsible for > 90 % of anogenital wart Immunization Quadrivalent vaccine contains types 6, 11, 16, and 18 Bivalent vaccine contains 16 and 18 Bacterial Pathogens Gram Positive Bacteria S. aureus Background S. aureus is a well-known cause of both local and invasive infection Coagulase positive Grapelike clusters (Fig. 6 ) Fig. 6 Staphylococci in blood culture (gram stain, original magnification × 1000). The bacteria are gram-positive cocci and grow inpairs, tetrads, and clusters ( arrow ) S. aureus colonizes the nares and skin in 30–50 % of children Common staphylococcal infections: Bullous and crusted impetigo. Soft tissue or lymph node infection. If the organism seeds the bloodstream, dissemination to joints, bones, kidney, liver, muscles, lung, and heart valves may occur, causing substantial morbidity and potential mortality. S. aureus is the most common cause of osteomyelitis, except sickle cell anemia patients is usually caused by salmonella . Children with cyanotic congenital heart disease are at high risk of staphylococcal brain abscess . Children who undergo neurosurgical procedures, specially shunt revisions at high risk for staphylococcal infection. Catheters are usually associated with staphylococcal infection and must be removed if the patient develops symptoms or positive culture, and antibiotic must be started. Folliculitis/Furunculosis/Carbunculosis (Fig. 7a and b ) Fig. 7 a Furuncle: erythematous tender papulonodule with central punctum with point of fluctuant. b Folliculitis: Superficial inflammation centered around a follicle, tender to touch Background Folliculitis: superficial inflammation centered around a follicle. Furuncles: bacterial folliculitis of a single follicle that involves a deeper portion of the follicle. Carbuncle: bacterial folliculitis that involves the deeper portion of several contiguous follicles. Bacterial folliculitis most often caused by S. aureus . Hot tub folliculitis is usually caused by gram-negative bacteria (most often P. aeruginosa. It is self limited). Usually the child looks healthy and does not appear ill. Abscess (  7 days causes meningitis Treatment Ampicillin and aminoglycoside Corynebacterium diphtheriae Background Gram-positive pleomorphic bacillus It is rare due to immunization against diphtheria Clinical presentation Low-grade fever Sore throat Malaise Difficulty swallowing Bilateral cervical lymphadenopathy Grayish exudates over mucous membrane Bleeding after attempting to remove the membrane Treatment Antitoxin should be started immediately if diphtheria is suspected called equine hyperimmune antiserum IV to neutralize the toxins. Diphtheria toxins can cause myocarditis, necrosis, peripheral neuritis. Airway obstruction and neck swelling (bull neck) can occur. Know that close contact should receive single IM dose of penicillin G benzathine or oral erythromycin regardless their immunization status. Enterococcus Background Gram-positive cocci. Normal inhabitant of the gastrointestinal tract. E. faecalis and E. faecium . Most neonatal enterococcal infections are nosocomial and occur after second week of life, usually with bacteremia due to line infection or necrotizing enterocolitis (common symptoms in neonates include, fever, bradycardia, apnea, and abdominal distention). Associated infections Bacteremia in neonates Catheter associated bacteremia Endocarditis Intra-abdominal abscess UTI Antibiotics It is resistant to all cephalosporins and vancomycin as well It is susceptible to aminoglycoside and linezolid It is imperative to do sensitivity test because of increasing resistance Sensitive enterococcal sepsis or endocarditis must be treated with vancomycin, PCN, ampicillin, in addition to gentamicin Bacillus anthracis Background Large positive rods (bacilli) that cause anthrax Types of anthrax: cutaneous anthrax, pulmonic and gastrointestinal Inoculation occurs from handling contaminated substance, e.g., wool, and in the mail in cases of bioterrorism Clinical presentation Painless papules and ulcers Painless black eschar with painless swelling and induration Treatment Penicillin G or quinolones, e.g., ciprofloxacin Bacillus cereus Background It is a soil dwelling gram-positive rods, beta hemolytic bacterium. Produces gastrointestinal symptoms due enterotoxin production in vivo in the GI tract. Clinical presentation Vomiting with incubation period 1–6 h (the emetic form is commonly associated with fried rice left at room temperature) Diarrhea with incubation period 8–16 h Eye infection after traumatic eye injuries in contact lens wearers Diagnosis It is usually clinical B. cereus spores in stool Isolated toxins from suspected food items Treatment Self limited and require no antibiotics Arcanobacterium haemolyticum Background A. haemolyticum (can be mistaken with strep pharyngitis or scarlet fever) Gram positive bacillus Grows slowly as small colonies with narrow bands of hemolysis on blood-enriched agar Growth enhanced by culture on rabbit or human blood with incubation in 5 % CO 2 Clinical presentation Common in teenagers and young adults 0.5–3 % of acute pharyngitis Except for absence of palatal petechiae and strawberry tongue, the disease indistinguishable from that caused by group A Streptococcus Fever Pharyngeal exudates Cervical lymphadenopathy Scarlatiniform or maculopapular pruritic rash in 50 % of cases usually spares the palm and soles Treatment Macrolides: erythromycin or azithromycin Gram Positive Bacteria S. aureus Background S. aureus is a well-known cause of both local and invasive infection Coagulase positive Grapelike clusters (Fig. 6 ) Fig. 6 Staphylococci in blood culture (gram stain, original magnification × 1000). The bacteria are gram-positive cocci and grow inpairs, tetrads, and clusters ( arrow ) S. aureus colonizes the nares and skin in 30–50 % of children Common staphylococcal infections: Bullous and crusted impetigo. Soft tissue or lymph node infection. If the organism seeds the bloodstream, dissemination to joints, bones, kidney, liver, muscles, lung, and heart valves may occur, causing substantial morbidity and potential mortality. S. aureus is the most common cause of osteomyelitis, except sickle cell anemia patients is usually caused by salmonella . Children with cyanotic congenital heart disease are at high risk of staphylococcal brain abscess . Children who undergo neurosurgical procedures, specially shunt revisions at high risk for staphylococcal infection. Catheters are usually associated with staphylococcal infection and must be removed if the patient develops symptoms or positive culture, and antibiotic must be started. Folliculitis/Furunculosis/Carbunculosis (Fig. 7a and b ) Fig. 7 a Furuncle: erythematous tender papulonodule with central punctum with point of fluctuant. b Folliculitis: Superficial inflammation centered around a follicle, tender to touch Background Folliculitis: superficial inflammation centered around a follicle. Furuncles: bacterial folliculitis of a single follicle that involves a deeper portion of the follicle. Carbuncle: bacterial folliculitis that involves the deeper portion of several contiguous follicles. Bacterial folliculitis most often caused by S. aureus . Hot tub folliculitis is usually caused by gram-negative bacteria (most often P. aeruginosa. It is self limited). Usually the child looks healthy and does not appear ill. Abscess (  7 days causes meningitis Treatment Ampicillin and aminoglycoside Corynebacterium diphtheriae Background Gram-positive pleomorphic bacillus It is rare due to immunization against diphtheria Clinical presentation Low-grade fever Sore throat Malaise Difficulty swallowing Bilateral cervical lymphadenopathy Grayish exudates over mucous membrane Bleeding after attempting to remove the membrane Treatment Antitoxin should be started immediately if diphtheria is suspected called equine hyperimmune antiserum IV to neutralize the toxins. Diphtheria toxins can cause myocarditis, necrosis, peripheral neuritis. Airway obstruction and neck swelling (bull neck) can occur. Know that close contact should receive single IM dose of penicillin G benzathine or oral erythromycin regardless their immunization status. Enterococcus Background Gram-positive cocci. Normal inhabitant of the gastrointestinal tract. E. faecalis and E. faecium . Most neonatal enterococcal infections are nosocomial and occur after second week of life, usually with bacteremia due to line infection or necrotizing enterocolitis (common symptoms in neonates include, fever, bradycardia, apnea, and abdominal distention). Associated infections Bacteremia in neonates Catheter associated bacteremia Endocarditis Intra-abdominal abscess UTI Antibiotics It is resistant to all cephalosporins and vancomycin as well It is susceptible to aminoglycoside and linezolid It is imperative to do sensitivity test because of increasing resistance Sensitive enterococcal sepsis or endocarditis must be treated with vancomycin, PCN, ampicillin, in addition to gentamicin Bacillus anthracis Background Large positive rods (bacilli) that cause anthrax Types of anthrax: cutaneous anthrax, pulmonic and gastrointestinal Inoculation occurs from handling contaminated substance, e.g., wool, and in the mail in cases of bioterrorism Clinical presentation Painless papules and ulcers Painless black eschar with painless swelling and induration Treatment Penicillin G or quinolones, e.g., ciprofloxacin Bacillus cereus Background It is a soil dwelling gram-positive rods, beta hemolytic bacterium. Produces gastrointestinal symptoms due enterotoxin production in vivo in the GI tract. Clinical presentation Vomiting with incubation period 1–6 h (the emetic form is commonly associated with fried rice left at room temperature) Diarrhea with incubation period 8–16 h Eye infection after traumatic eye injuries in contact lens wearers Diagnosis It is usually clinical B. cereus spores in stool Isolated toxins from suspected food items Treatment Self limited and require no antibiotics Arcanobacterium haemolyticum Background A. haemolyticum (can be mistaken with strep pharyngitis or scarlet fever) Gram positive bacillus Grows slowly as small colonies with narrow bands of hemolysis on blood-enriched agar Growth enhanced by culture on rabbit or human blood with incubation in 5 % CO 2 Clinical presentation Common in teenagers and young adults 0.5–3 % of acute pharyngitis Except for absence of palatal petechiae and strawberry tongue, the disease indistinguishable from that caused by group A Streptococcus Fever Pharyngeal exudates Cervical lymphadenopathy Scarlatiniform or maculopapular pruritic rash in 50 % of cases usually spares the palm and soles Treatment Macrolides: erythromycin or azithromycin S. aureus Background S. aureus is a well-known cause of both local and invasive infection Coagulase positive Grapelike clusters (Fig. 6 ) Fig. 6 Staphylococci in blood culture (gram stain, original magnification × 1000). The bacteria are gram-positive cocci and grow inpairs, tetrads, and clusters ( arrow ) S. aureus colonizes the nares and skin in 30–50 % of children Common staphylococcal infections: Bullous and crusted impetigo. Soft tissue or lymph node infection. If the organism seeds the bloodstream, dissemination to joints, bones, kidney, liver, muscles, lung, and heart valves may occur, causing substantial morbidity and potential mortality. S. aureus is the most common cause of osteomyelitis, except sickle cell anemia patients is usually caused by salmonella . Children with cyanotic congenital heart disease are at high risk of staphylococcal brain abscess . Children who undergo neurosurgical procedures, specially shunt revisions at high risk for staphylococcal infection. Catheters are usually associated with staphylococcal infection and must be removed if the patient develops symptoms or positive culture, and antibiotic must be started. Folliculitis/Furunculosis/Carbunculosis (Fig. 7a and b ) Fig. 7 a Furuncle: erythematous tender papulonodule with central punctum with point of fluctuant. b Folliculitis: Superficial inflammation centered around a follicle, tender to touch Background Folliculitis: superficial inflammation centered around a follicle. Furuncles: bacterial folliculitis of a single follicle that involves a deeper portion of the follicle. Carbuncle: bacterial folliculitis that involves the deeper portion of several contiguous follicles. Bacterial folliculitis most often caused by S. aureus . Hot tub folliculitis is usually caused by gram-negative bacteria (most often P. aeruginosa. It is self limited). Usually the child looks healthy and does not appear ill. Abscess (  7 days causes meningitis Treatment Ampicillin and aminoglycoside Corynebacterium diphtheriae Background Gram-positive pleomorphic bacillus It is rare due to immunization against diphtheria Clinical presentation Low-grade fever Sore throat Malaise Difficulty swallowing Bilateral cervical lymphadenopathy Grayish exudates over mucous membrane Bleeding after attempting to remove the membrane Treatment Antitoxin should be started immediately if diphtheria is suspected called equine hyperimmune antiserum IV to neutralize the toxins. Diphtheria toxins can cause myocarditis, necrosis, peripheral neuritis. Airway obstruction and neck swelling (bull neck) can occur. Know that close contact should receive single IM dose of penicillin G benzathine or oral erythromycin regardless their immunization status. Enterococcus Background Gram-positive cocci. Normal inhabitant of the gastrointestinal tract. E. faecalis and E. faecium . Most neonatal enterococcal infections are nosocomial and occur after second week of life, usually with bacteremia due to line infection or necrotizing enterocolitis (common symptoms in neonates include, fever, bradycardia, apnea, and abdominal distention). Associated infections Bacteremia in neonates Catheter associated bacteremia Endocarditis Intra-abdominal abscess UTI Antibiotics It is resistant to all cephalosporins and vancomycin as well It is susceptible to aminoglycoside and linezolid It is imperative to do sensitivity test because of increasing resistance Sensitive enterococcal sepsis or endocarditis must be treated with vancomycin, PCN, ampicillin, in addition to gentamicin Bacillus anthracis Background Large positive rods (bacilli) that cause anthrax Types of anthrax: cutaneous anthrax, pulmonic and gastrointestinal Inoculation occurs from handling contaminated substance, e.g., wool, and in the mail in cases of bioterrorism Clinical presentation Painless papules and ulcers Painless black eschar with painless swelling and induration Treatment Penicillin G or quinolones, e.g., ciprofloxacin Bacillus cereus Background It is a soil dwelling gram-positive rods, beta hemolytic bacterium. Produces gastrointestinal symptoms due enterotoxin production in vivo in the GI tract. Clinical presentation Vomiting with incubation period 1–6 h (the emetic form is commonly associated with fried rice left at room temperature) Diarrhea with incubation period 8–16 h Eye infection after traumatic eye injuries in contact lens wearers Diagnosis It is usually clinical B. cereus spores in stool Isolated toxins from suspected food items Treatment Self limited and require no antibiotics Arcanobacterium haemolyticum Background A. haemolyticum (can be mistaken with strep pharyngitis or scarlet fever) Gram positive bacillus Grows slowly as small colonies with narrow bands of hemolysis on blood-enriched agar Growth enhanced by culture on rabbit or human blood with incubation in 5 % CO 2 Clinical presentation Common in teenagers and young adults 0.5–3 % of acute pharyngitis Except for absence of palatal petechiae and strawberry tongue, the disease indistinguishable from that caused by group A Streptococcus Fever Pharyngeal exudates Cervical lymphadenopathy Scarlatiniform or maculopapular pruritic rash in 50 % of cases usually spares the palm and soles Treatment Macrolides: erythromycin or azithromycin Anaerobes Clostridium botulinum Background C. botulinum is an anaerobic gram-positive rod that survives in soil and marine sediment by forming spores. Human botulism is caused by neurotoxins A, B, E, and occasionally F. Infant botulism Ingestion of honey or exposure to soils increases the risk Age between 3 weeks and 6 months Symptoms develop 3–30 days from the time of exposure Clinical presentation Constipation usually is the initial finding Feeding difficulty is a common presenting symptoms Hypotonia Increased drooling Weak cry Truncal weakness Cranial nerve palsies Generalized weakness with ventilatory failure Treatment of infant botulism Botulism immune globulin (BIG) IV should be started as early as possible if clinically suspected. No antibiotics. Foodborne botulism Background Most common source is home canned food. Symptoms develop 12–36 h after toxin ingestion. Wound botulism is similar except the incubation period between 4 and 14 days. Clinical presentation Initial symptoms: dry mouth, nausea, and diarrhea Bilateral cranial nerve palsies Eye diplopia and blurring vision Dysphagia Upper extremity weakness Respiratory dysfunction Lower extremity dysfunction Diagnosis Stool toxins detection Treatment of botulism in older patients Equine trivalent antitoxin (Type A, B, and E) Wound debridement for wound botulism is recommended Clostridium perfringens Background Gram-positive, rod shaped, anaerobic, spore forming bacterium of the genus Clostridium Spores found in raw meat and poultry Clinical presentation Sudden onset of diarrhea Crampy abdominal pain Management Resolve with 24 h No treatment is necessary Clostridium tetani Background C. tetani , an obligate anaerobic gram-positive bacillus, is the pathogen responsible for tetanus. It is nonencapsulated and form spores that are resistant to heat, desiccation, and disinfectants. Contaminated deep puncture wounds , open wounds, soil, and animals (wool) containing spores are the most common sources of this bacteria. Neonatal tetanus Contaminated umbilical cord is a common source of infection. Poor feeding (poor suck and swallowing due to muscle spasm). Constant crying Decreased movement Spasm and rigidity Generalized tetanus Trismus (lockjaw) Sardonic smile (risus sardonicus) Severe muscle spasm Opisthotonos (severe hyperextension) Laryngeal spasm can lead airway obstruction and death Tetanic seizure is severe tonic contractions with high fever Diagnosis is always clinical Treatment Human tetanus immune globulin immediately Penicillin G or metronidazole Muscle relaxants Prevention of tetanus Routine immunization with Dtap and Tdap Prevention in wound injuries guideline Tetanus vaccine + /−Tetanus immunoglobulin (TIG) Dirty wound, immunization is unknown or less than three tetanus shots: Give TIG + tetanus vaccine Dirty wound, immunized > 5 years and   10 years: Immunize, no TIG Clostridium difficile Background Gram-positive anaerobes Colonization Around 50 % of infants younger than 1 year are colonized Carriage decrease by 1–5 % by 2 years of age Risk factor: Having infected roommate or having symptomatic patient in the same ward Antibiotics , e.g., beta-lactams drugs, clindamycin, and macrolides Underlying bowel disease or surgeries Symptomatic disease is due to toxins A and B produced by the organism Clinical presentation Asymptomatic colonization is common in infants and young children Watery diarrhea Abdominal cramps Abdominal tenderness In severe cases : Systemic toxicity Bloody diarrhea Toxic megacolon, perforation or even death are complications of pseudomembranous colitis Diagnosis Documenting toxin A and B in stool (should be tested promptly or stored at 4 °C) Endoscopic finding of pseudomembranous enterocolitis Examination for occult blood is not diagnostic In young infants you must consider other causes because they are colonized Treatment Oral or IV metronidazole Oral vancomycin with or without metronidazole can be used in severe cases Oral vancomycin can be used alone in those who do not respond to metronidazole Prevention Hand washing with water and soap Know that alcohol based product are not effective in eradications of the organisms Diluted bleach solution is the best for decontamination of surfaces Limit antibiotic use Infected child should be excluded from child care facility for the duration of diarrhea Actinomycosis Background Actinomycosis is a subacute-to-chronic bacterial infection caused by filamentous, gram-positive, non acid-fast, anaerobic-to-microaerophilic bacteria. It is characterized by contagious spread, suppurative and granulomatous inflammation, and formation of multiple abscesses and sinus tracts that may discharge sulfur granules. Clinical presentation The most common clinical forms of actinomycosis are cervicofacial (i.e., lumpy jaw) usually caused by dental infection. In women, pelvic actinomycosis is possible when IUD in place. Treatment Initial therapy should include IV penicillin or ampicillin for 4–6 weeks followed by high dose of oral penicillin, clindamycin or doxycycline. Clostridium botulinum Background C. botulinum is an anaerobic gram-positive rod that survives in soil and marine sediment by forming spores. Human botulism is caused by neurotoxins A, B, E, and occasionally F. Infant botulism Ingestion of honey or exposure to soils increases the risk Age between 3 weeks and 6 months Symptoms develop 3–30 days from the time of exposure Clinical presentation Constipation usually is the initial finding Feeding difficulty is a common presenting symptoms Hypotonia Increased drooling Weak cry Truncal weakness Cranial nerve palsies Generalized weakness with ventilatory failure Treatment of infant botulism Botulism immune globulin (BIG) IV should be started as early as possible if clinically suspected. No antibiotics. Foodborne botulism Background Most common source is home canned food. Symptoms develop 12–36 h after toxin ingestion. Wound botulism is similar except the incubation period between 4 and 14 days. Clinical presentation Initial symptoms: dry mouth, nausea, and diarrhea Bilateral cranial nerve palsies Eye diplopia and blurring vision Dysphagia Upper extremity weakness Respiratory dysfunction Lower extremity dysfunction Diagnosis Stool toxins detection Treatment of botulism in older patients Equine trivalent antitoxin (Type A, B, and E) Wound debridement for wound botulism is recommended Clostridium perfringens Background Gram-positive, rod shaped, anaerobic, spore forming bacterium of the genus Clostridium Spores found in raw meat and poultry Clinical presentation Sudden onset of diarrhea Crampy abdominal pain Management Resolve with 24 h No treatment is necessary Clostridium tetani Background C. tetani , an obligate anaerobic gram-positive bacillus, is the pathogen responsible for tetanus. It is nonencapsulated and form spores that are resistant to heat, desiccation, and disinfectants. Contaminated deep puncture wounds , open wounds, soil, and animals (wool) containing spores are the most common sources of this bacteria. Neonatal tetanus Contaminated umbilical cord is a common source of infection. Poor feeding (poor suck and swallowing due to muscle spasm). Constant crying Decreased movement Spasm and rigidity Generalized tetanus Trismus (lockjaw) Sardonic smile (risus sardonicus) Severe muscle spasm Opisthotonos (severe hyperextension) Laryngeal spasm can lead airway obstruction and death Tetanic seizure is severe tonic contractions with high fever Diagnosis is always clinical Treatment Human tetanus immune globulin immediately Penicillin G or metronidazole Muscle relaxants Prevention of tetanus Routine immunization with Dtap and Tdap Prevention in wound injuries guideline Tetanus vaccine + /−Tetanus immunoglobulin (TIG) Dirty wound, immunization is unknown or less than three tetanus shots: Give TIG + tetanus vaccine Dirty wound, immunized > 5 years and   10 years: Immunize, no TIG Clostridium difficile Background Gram-positive anaerobes Colonization Around 50 % of infants younger than 1 year are colonized Carriage decrease by 1–5 % by 2 years of age Risk factor: Having infected roommate or having symptomatic patient in the same ward Antibiotics , e.g., beta-lactams drugs, clindamycin, and macrolides Underlying bowel disease or surgeries Symptomatic disease is due to toxins A and B produced by the organism Clinical presentation Asymptomatic colonization is common in infants and young children Watery diarrhea Abdominal cramps Abdominal tenderness In severe cases : Systemic toxicity Bloody diarrhea Toxic megacolon, perforation or even death are complications of pseudomembranous colitis Diagnosis Documenting toxin A and B in stool (should be tested promptly or stored at 4 °C) Endoscopic finding of pseudomembranous enterocolitis Examination for occult blood is not diagnostic In young infants you must consider other causes because they are colonized Treatment Oral or IV metronidazole Oral vancomycin with or without metronidazole can be used in severe cases Oral vancomycin can be used alone in those who do not respond to metronidazole Prevention Hand washing with water and soap Know that alcohol based product are not effective in eradications of the organisms Diluted bleach solution is the best for decontamination of surfaces Limit antibiotic use Infected child should be excluded from child care facility for the duration of diarrhea Actinomycosis Background Actinomycosis is a subacute-to-chronic bacterial infection caused by filamentous, gram-positive, non acid-fast, anaerobic-to-microaerophilic bacteria. It is characterized by contagious spread, suppurative and granulomatous inflammation, and formation of multiple abscesses and sinus tracts that may discharge sulfur granules. Clinical presentation The most common clinical forms of actinomycosis are cervicofacial (i.e., lumpy jaw) usually caused by dental infection. In women, pelvic actinomycosis is possible when IUD in place. Treatment Initial therapy should include IV penicillin or ampicillin for 4–6 weeks followed by high dose of oral penicillin, clindamycin or doxycycline. Gram Negative Bacteria Gram Negative Anaerobes Bacteroides and Fusobacterium anaerobes Causes Variety of Clinical Manifestations Depending on the Location Head and neck Retropharyngeal abscess Peritonsillar abscess Dental abscess Ludwig angina CNS Brain abscess Subdural and epidural empyema Lung Aspiration pneumonia Lung abscess Pleural empyema Abdomen Peritonitis Appendicitis Intra-abdominal abscess Skin and soft tissue Infected bite wound Necrotizing fasciitis Cellulitis Antibiotics with anaerobic activity Clindamycin Penicillin Ampicillin-sulbactam Amoxicillin-clavulanic acid Metronidazole Campylobacter species Background Campylobacter jejuni (gram-negative motile bacilli) It is one of the most common agent associated with bacterial gastroenteritis Common sources Uncooked poultry (chicken and turkey) Unpasteurized milk Dogs and cats Clinical presentation Bloody diarrhea Abdominal pain (may mimic inflammatory bowel disease in severe cases) Tenesmus Fever Diagnosis Stool culture in a selective media at temperature 42 °C incubated in gas mixture O 2 and CO 2 Azithromycin is the drug of choice Antibiotic is recommended to shorten the duration of illness and prevent relapse Chlamydophila pneumoniae Background C. pneumoniae is distinct antigenically, genetically, and morphologically from Chlamydia species Transmitted from person to another via respiratory secretion Clinical presentation Patient may be asymptomatic or mildly to moderately ill Illness is usually prolonged with cough persist for 2–6 weeks Pneumonia and pulmonary rales Acute bronchitis and bronchospasm Less commonly nonexudative pharyngitis, laryngitis, otitis media, and sinusitis Diagnosis Chest radiography; may reveal an infiltrate No reliable test to identify the organism is available Fourfold increase in immunoglobulin (Ig) G titer or IgM titer of ≥ 16 is evidence of acute infection Treatment Macrolides or tetracycline Chlamydophila psittaci Background C. psittaci is obligate intracellular bacterial pathogen. Birds are major reservoir of C. psittaci , e.g., parakeets, and parrots, also animal such as goats and cows may become infected. Clinical presentation (Psittacosis) Fever Nonproductive cough Headache Malaise Extensive interstitial pneumonia can occur Pericarditis, hepatitis, and encephalitis can occur (rare) Diagnosis Same as C. pneumonia Treatment Tetracyclines are preferred therapy except children less than 8 years of age Macrolides, e.g., azithromycin Chlamydia trachomatis Background It is the most frequently identified infectious cause of neonatal conjunctivitis; it is transmitted perinatally from infected mothers. Clinical presentation The symptoms typically develop 5–14 days after birth Conjunctival edema Hyperemia Watery-to-mucopurulent discharge A pseudomembrane may form and bloody discharge may be present if infection is prolonged Management Know that topical prophylaxis with erythromycin or silver nitrate given to all infants to prevent neonatal gonococcal conjunctivitis is ineffective against chlamydial conjunctivitis. Important : when chlamydial conjunctivitis is diagnosed in an infant, the infant's mother and her sexual partner(s) must be tested. Treatment is erythromycin PO 50 mg/kg/day in four divided doses × 14 days. Topical treatment alone is ineffective Remember : untreated infections may result in corneal and conjunctival scarring. Pneumonia due to C. trachomatis Background Small, gram-negative, obligate intracellular organisms . Transmitted to the infant from the birth canal. Generally presents as a subacute infection 2–19 weeks after birth. C. trachomatis infection may cause neonatal conjunctivitis, nasopharyngitis, otitis media, and pneumonitis. Clinical presentation Rhinorrhea, congestion, or conjunctivitis Tachypnea Staccato cough Crackles (rales) Wheezing (rare) Preterm infants may have episodes of apnea Diagnosis Chest radiography reveals infiltrates and hyperinflation Laboratory testing may reveal: Peripheral eosinophilia Elevated serum immunoglobulins A positive nasopharyngeal culture is considered diagnostic of infection Treatment Antibiotic treatment should be started presumptively on clinical grounds. Oral erythromycin for 14 days or azithromycin, 20 mg/kg/day, once daily × 3 days. If untreated, symptoms can last for months and include persistent hypoxemia. Remember: Diagnosis of chlamydial pneumonia in an infant necessitates treatment of the infant's mother and her sexual partner . Trachoma Background This disease is a chronic keratoconjunctivitis caused by the obligate intracellular bacterium C. trachomatis . Disease transmission occurs primarily between children and the women who care for them. Trachoma is the most common infectious cause of blindness worldwide. Clinical presentation Chronic follicular keratoconjunctivitis with corneal neovascularization resulting from untreated or chronic infection. Blindness occurs in up to 15 % of those infected. Trachoma rarely occurs in the USA. Diagnosis It is a clinical diagnosis and nucleic acid amplification tests (NAATs) can confirm the causative agent. The cicatricial phase has unique clinical features, which lead to definitive diagnosis in most cases. Treatment Azithromycin Neisseria gonorrhoeae (Gonococcal Infections) Background N. gonorrhoeae is a gram-negative diplococcus. Gonococcal infection is the second most common bacterial disease in the USA that is classified as a reportable and notifiable infection. It is the highest in youth, especially females between 15 and 19 years of age. The incubation period is 2–7 days. A child abuse evaluation must be performed in any prepubertal case of gonococcal infection. Neonatal conjunctivitis Conjunctivitis due to mucosal transmission during vaginal delivery. Topical antibiotics (erythromycin, silver nitrate, or tetracycline) to the eyes of a newborn within 1 h of birth can prevent the infection. Treatment is ceftriaxone 125 mg IM × 1. Gonococcal pharyngitis Genital-oral activity is the major risk Infection is asymptomatic in most cases Patients who have gonococcal pharyngitis have a significant public health impact Gonococcal pharyngitis are at risk for developing disseminated gonococcal infection (DGI) Pharyngeal infection clears spontaneously within 12 weeks Treatment is ceftriaxone 250 mg IM × 1 Gonococcal urethritis Dysuria and a mucopurulent penile discharge They may be coinfected with other sexually transmitted organisms, most commonly, C. trachomatis Positive leukocyte esterase usually seen in urine specimen Diagnosis of gonococcal urethritis Presence of intracellular diplococci in urethral discharge Treatment is ceftriaxone 250 mg IM × 1 plus azithromycin 1 g × 1 Epididymitis (gonococcus) Dysuria and a mucopurulent discharge Scrotal edema as well as scrotal, inguinal, or flank pain Urinalysis may demonstrate WBCs In most cases, this infection is transmitted sexually and may be an extension of urethritis Gonococcal proctitis Most cases of proctitis due to N. gonorrhoeae occur in homosexual males Clinical presentation Anal discharge Rectal bleeding Anorectal pain Tenesmus Constipation Disseminated gonococcal infection (DGI) DGI infection occurs in 0.5–3 % of people infected with N. gonorrhoeae DGI usually cause an asymptomatic genital infection Migratory arthritis (wrist, ankle, and knee) are the most common locations Dermatitis Tenosynovitis Fever and chills may occur Elevated white blood cell count DGI occurs more commonly in females Screening methods for infection N. gonorrhoeae and Chlamydia Culture is the gold standard for diagnosing C. trachomatis . Standard collection sites include the endocervix, male and female urethra, nasopharynx, conjunctiva, vagina, and rectum. Nucleic acid amplification tests (NAATs) amplify nucleic acid sequences specific for the organism of interest. The ease of using urine specimens, together with the high sensitivity of NAATs, has made these tests the preferred method for screening. The presence of gram-negative intracellular diplococci on microscopy suggests the diagnosis of a gonococcal infection. N. meningitidis (Meningococcal Infections) Background Aerobic gram-negative diplococcus N. meningitidis . Natural commensal organism living in the nasopharynx of humans. Children younger than 2 years of age have a nearly fivefold greater risk of contracting meningococcal disease than the general adult population. Risk of transmission; crowded living conditions, e.g., college dormitories, military barracks. Clue to clinician of invasive meningococcal infection Rash Any rash appearing in the context of a sudden febrile illness should raise concern Meningococcal rash is typically present within 24 h of any symptomatology Petechiae may be intraoral or conjunctival or be hidden in skinfolds Early rash may not be petechial True rigors Shaking chill that cannot be stopped voluntarily Prolonged (10–20 min) Neck pain Severe pain in the neck, back, or extremities May manifest in younger children as refusal to walk Meningismus : In patients older than 3 years, the classic signs of Kernig and Brudzinski may be elicited Vomiting May be associated with headache or abdominal pain without diarrhea Cushing triads: Bradycardia Hypertension Respiratory depression Purpura fulminans (meningococcemia) Aggressive spread of purpura to large areas with ischemic necrosis Sudden drops in blood pressure Acute adrenal hemorrhage (Waterhouse–Friderichsen syndrome) Diagnosis Culture of the organism from a normally sterile site is the gold standard for bacteriologic diagnosis. Cerebrospinal fluid study: CSF WBC counts are elevated in most patients who have meningitis. CSF WBC counts are low or even normal if the disease is severe and rapidly progressive. Markedly low glucose and elevated protein values are associated with the diagnosis of meningitis. All patients with meningococcal disease or meningitis must be tested for CH50 or CH100 assay (20 % of children with meningococcal disease will end having a complement deficiency). Management Know that antibiotics or fluids should not be delayed for the sake of cultures or other testing. Penicillin is effective treatment for both severe meningococcal septicemia (SMS) and meningococcal meningitis if the diagnosis is certain. Broad-spectrum antibiotics effective against N. meningitidis and other potential pathogens are indicated (e.g., ceftriaxone, cefotaxime, vancomycin). Emergency care evaluation and preferably transported via emergency medical services to allow for prompt delivery of intravenous fluids and airway management if the condition is suspected. Large isotonic fluid boluses (20 mL/kg) over the first 5 min. Inotropic/vasoactive agent such as dopamine or dobutamine. Hydrocortisone may be beneficial in children who have SMS and respond poorly to vasopressors. Prevention and indication of MCV4 (A, C,Y, and W-135) MCV4 is routinely recommended at 11–12 years of age. Unvaccinated adolescents through 18 years of age should receive a dose at the earliest opportunity. Military recruits and all college freshmen who will be living in campus dormitories. Persons who have terminal complement component deficiencies. Anatomic or functional asplenia. Note: 30 % of infections are due to serogroup B which is not covered by the vaccine. Antibiotic prophylaxis , e.g., Rifampin, ciprofloxacin, azithromycin, or ceftriaxone should be used for contacts: Child care contact Direct exposure to oral secretions of individual with meningococcal disease (such as personnel providing mouth-to-mouth resuscitation) Haemophilus influenzae Background Pleomorphic gram-negative coccobacillus . Used to be the most common cause of meningitis and serious bacteremia in children. Introduction of the H. influenzae vaccine quickly reduced the incidence of encapsulated H. influenza type b. Nontypeable strains are still responsible for a large number of mucosal infections, including conjunctivitis , otitis media, sinusitis, and bronchitis. Bacterial meningitis Peak age is less than 1 year . Mortality rate around 5 %. Common complications include: subdural empyema, brain infarct, cerebritis, ventriculitis, brain abscess , and hydrocephalus. Long-term sequelae occur in 15–30 % of survivors with sensorineural hearing loss, others include language disorders, intellectual disability (ID), and developmental delay. Dexamethasone before or with antibiotics such as ceftriaxone or cefotaxime to prevent hearing loss and neurologic sequelae. Epiglottitis H. influenzae type b (Hib) was the predominant organism (> 90 %) in pediatric epiglottitis cases (other bacteria can cause epiglottitis as well, e.g., S. pneumoniae , group A beta-hemolytic streptococci, S. aureus , and Moraxella catarrhalis . Occurs primarily in children (ages 2–7 years). The clinical triad of drooling, dysphagia, and distress is the classic presentation. Fever with associated respiratory distress or air hunger occurs in most patients. Treatment in patients with epiglottitis is directed toward relieving the airway obstruction and eradicating the infectious agent. Optimally, initial treatment is provided by a pediatric anesthesiologist and either a pediatric surgeon or a pediatric otolaryngologist. Once the airway is controlled, a pediatric intensivist is required for inpatient management. Buccal infections Buccal cellulitis previously was always caused by H. influenzae infection before the vaccine. Always associated with bacteremia if present. Present with palpable cellulitis on both checks, purplish in color and child looks very toxic. Periorbital cellulitis Previously H. influenzae was the a common cause, now pneumococcus bacteria is the most common etiology Minor trauma or insect bite of the eye lid usually associated with preseptal cellulitis due to S. aureus or a Group A Streptococcus Pyogenic arthritis H. influenzae was the most common cause of septic arthritis before Hib vaccine in children less than 2 years of age Occult bacteremia Occult bacteremia with H. influenzae will result in in 30–50 % developing meningitis or other deep, or focal infection from occult bacteremia. All occult bacteremia from H. influenzae has to be treated immediately. Pneumonia Pneumonia from H. influenzae used to cause about one third of bacterial pneumonia before Hib vaccine and usually associated with pleural effusion, positive blood culture in most of the cases . Treatment (Patient with life threatening illness) Remember : the organism produces beta lactamase which makes amoxicillin is ineffective. Cefotaxime or ceftriaxone is the antimicrobial of choice. Meropenem or chloramphenicol is another option. Amoxicillin is the drug of choice for noninvasive diseases such as otitis media or sinusitis, if amoxicillin fails, uses antibiotics against beta-lactamase-producing strains, e.g., nontypeable H. influenzae including amoxicillin/clavulanic, TMP-SMX, azithromycin, cefuroxime axetil, cefixime, and cefpodoxime. Rifampin antibiotic prophylaxis for contact with invasive H. influenzae type b infection All household who did not receive immunization Less than 4 years with incomplete immunization Younger than 12 months who did not complete primary HIB immunization Immunocompromised child Nursery school and child care center if two or more cases within 60 days Helicobacter pylori Background H. pylori is a gram-negative microaerophilic bacillus It is spiral, curved, or U-shaped and has two to six flagella at one end under microscope Transmission is fecal-oral, oral-oral from human-to human contact Diagnosis Know that AAP recommends testing only when treatment for H. pylori infection would be warranted. Endoscopy remains the gold standard for evaluating H. pylori . H. pylori stool antigen and urea breath test is a promising diagnostic tools. Serologic tests for H. pylori are unreliable marker of disease. Treatment indications Endoscopically confirmed gastric or duodenal ulcer Histologically proven gastric metaplasia Gastric mucosa-associated lymphoid lymphoma (MALT) Prior ulcer disease and current active infection First-line: 14 days treatment regimens for children generally include Clarithromycin (15 mg/kg/day divided twice a day, up to 500 mg per dose) with: Either amoxicillin (50 mg/kg/day divided BID, up to 1 g per dose) or metronidazole (20 mg/kg/day divided BID, up to 500 mg per dose) and Proton-pump inhibitor (PPI) Mycoplasma pneumonia Background M. pneumonia is the leading cause of pneumonia in school age children and young adults Infection is prevalent in person living in group setting Clinical presentation Pulmonary manifestations Nonproductive cough Chills Scattered rales Skin rash Bilateral infiltrate on chest radiograph Extrapulmonary manifestation Pharyngitis Rash Stevens–Johnson syndrome Hemolytic anemia Arthritis CNS disease (encephalitis, cranial nerve palsy (specially CNIII)) Testing for mycoplasma IgG and IgM serology or cold agglutinin Mycoplasma DNA PCR Treatment Mycoplasma lacks the cell wall and beta lactams are not effective Azithromycin is the drug of choice Pasteurella multocida Background Small gram-negative coccobacilli, it is a normal flora in number of animals, e.g., dog and cats. Dog or cat bite is a common risk. Clinical presentation Erythema, tenderness, and edema usually develop rapidly within 24 h. Infection occurs few days after the bite is usually caused by S. aureus . Treatment Clean the wound with soap and water. Treatment should cover potential pathogens, e.g., P. multocida , S. aureus , and anaerobes . Administration of antibiotic within 8–12 h of injury may decrease the risk of infection. Amoxicillin-Clavulanate is the drug of choice Ampicillin-sulbactam IV in severe cases Clindamycin and TMP-SMX is appropriate for children allergic to penicillin. Bordetella pertussis Background Pertussis is a small gram-negative coccobacillus that infects only humans. Pertussis is spread by aerosol droplets expelled while coughing or sneezing in proximity to others. Incubation period of 7–14 days. Clinical presentation Catarrhal phase Lasts from 1 to 2 weeks Mild fever Cough The cough worsens as the patient progresses to the paroxysmal phase Paroxysmal phase Lasts from 2 to 6 weeks Rapid fire or staccato cough Five to ten uninterrupted coughs occur in succession, followed by a "whoop" as the patient rapidly draws in a breath May occur several times per hour Can be associated with cyanosis, salivation, lacrimation, and posttussive emesis Despite the severe spells, patients often appear relatively well between episodes Whoop is usually absent in infants less than 6 months of age Gasping, gagging, and apnea can occur Convalescent phase Decreasing frequency and severity of the coughing episodes Lasts from weeks to months Complications of pertussis Pertussis is most severe in infants   39 °C Hemoglobinopathies, e.g., sickle cell anemia, HIV, and neoplastic diseases Immunocompromised patients at any age Typhoid fever Background Salmonella enterica, Serovar typhi ( S. typhi ) Mode of transmission Poor sanitation and overcrowding Spread by fecal-oral contamination of food or water by individuals who are carriers for S. typhi in either stool or urine Typhoid is endemic in many developing areas Clinical presentation Fever "can exceed 104 °F (40 °C)" Malaise Chills Headache, anorexia, myalgias, and dry cough may be seen Abdominal pain is common Diarrhea is more likely in children Abdominal tenderness, hepatosplenomegaly, and a coated tongue Rose spots (pink, blanchable maculopapular lesions that are 2–4 mm in diameter) are seen on the torso and abdomen Know that neonatal typhoid generally presents within 3 days of birth with fever, emesis, diarrhea, abdominal distention, pronounced hepatomegaly, jaundice , and sometimes, seizures Know that absence of abdominal or intestinal changes is not typical of typhoid Diagnosis Blood cultures are the mainstay of diagnosis Stool culture Treatment and Prognosis Treatment includes: Hydration and correction of fluid-electrolyte imbalance Antipyretics and antibiotics The choice of antibiotic as well as the route and duration depends on the host, site of infection, and sensitivities of the organism. Multidrug resistant (MDR) strains , including resistance to ampicillin and TMP-SM have emerged. IV cefotaxime or ceftriaxone for 14 days is appropriate. For severe typhoid with obtundation, stupor, coma, or shock: Two-day course of IV dexamethasone may be life-saving. Shigella Background Shigella is a gram-negative bacilli Shigella dysenteriae and Shigella flexneri usually cause bloody diarrhea Shigella sonnei and Shigella boydii usually cause watery diarrhea Ingestion of as few as 10 organism can cause diarrhea Incubation period is 2–4 days Outbreak can occur in child care centers Mode of transmission Person to person Feco-oral Ano-oral House flies Contaminated fomites Clinical presentation Range from mild diarrhea to life-threatening dysentery Fever Abdominal camps High-volume watery stools Small-volume bloody stool may follow 24–48 h later Blood-mucoid stool is a common presentation Rectal prolapse occurs in 5–8 % Complications Hemolytic-uremic syndrome Seizures Colonic perforation Toxic encephalopathy Diagnosis Stool culture is diagnostic Stool study with large number of neutrophil is suggestive but not specific Peripheral WBCs are usually elevated; bandemia is very common Treatment Antimicrobial therapy is recommended for all patient with shigellosis. Antimicrobial therapy for 5 days will shorten the duration and eradicate the organism from stool. Oral ampicillin or TMP-SMX but the resistance makes them useless of Shigella infection. Ceftriaxone, ciprofloxacin or azithromycin are usually effective. Ciprofloxacin is not recommended if less than 18 years, if there is an alternative. Daycare center Once Shigella is identified in a daycare or household, all other symptomatic individuals in these environments should be cultured for Shigella as well. Anyone found to have Shigella cannot return to daycare until the diarrhea has stopped and stool culture test is negative. Escherichia coli Background E. coli is a gram-negative, lactose fermenting, motile rod, belonging to the Enterobacteriaceae. E. coli is one of the most frequent causes of many common bacterial infections , including cholecystitis, bacteremia, cholangitis, urinary tract infection (UTI) , and traveler's diarrhea, and other clinical infections such as neonatal meningitis and pneumonia . Acute bacterial meningitis The vast majority of neonatal meningitis cases are caused by E. coli and group B streptococcal infections . Pregnant women are at a higher risk of colonization with the K1 capsular antigen strain of E. coli , which commonly observed in neonatal sepsis. Low-birth weight and a positive CSF culture result portend a poor outcome. Most survivors have subsequent neurologic or developmental abnormalities. Pneumonia E. coli respiratory tract infections are uncommon and are almost always associated with E. coli UTI . Intra-abdominal infections E. coli intra-abdominal infections often result from a perforated viscus (e.g., appendix, diverticulum) or may be associated with intra-abdominal abscess, cholecystitis, and ascending cholangitis. They can be observed in the postoperative period after anastomotic disruption. Abscesses are often polymicrobial. E. coli is one of the more common gram-negative bacilli observed together with anaerobes. Enteric infections Enterotoxigenic E. coli (ETEC) is a cause of traveler's diarrhea; TMP-SMX is the drug of choice . Enteropathogenic E. coli (EPEC) is a cause of childhood diarrhea; can be treated with TMP-SMX Enteroinvasive E. coli (EIEC) causes a Shigella -like dysentery. Enteroaggregative E. coli (EAEC) is primarily associated with persistent diarrhea in children in developing countries, and enteroadherent E. coli (EAEC) is a cause of childhood diarrhea and traveler's diarrhea in Mexico and North Africa. Enterohemorrhagic E. coli (EHEC) causes hemorrhagic colitis or hemolytic-uremic syndrome (HUS). Strains of STEC serotype O157:H7 have caused numerous outbreaks and sporadic cases of bloody diarrhea and HUS. E. coli (O157:H7) Background Gram-negative rods. It occurs in all ages. Transmitted via ingestion of contaminated food, e.g., (ground beef) or infected feces. The disease linked to eating undercooked beef, and unpasteurized milk or apple juice. Produces shiga toxins; the most virulent strain. The incidence of E. coli O157:H7 > Shigella . Clinical presentation Usually begin as nonbloody diarrhea then become bloody Severe abdominal pain is common Fever in one third of the cases May progress to hemorrhagic colitis in severe cases Hemolytic uremic syndrome (HUS) may occur Management No antibiotic is proven to be effective and no prove that antibiotic increase the risk HUS. No antibiotics are indicated. Do not use antimotility agents. UTIs The urinary tract is the most common site of E. coli infection, and more than 90 % of all uncomplicated UTIs are caused by E. coli infection . The recurrence rate after a first E. coli infection is 44 % over 12 months. E. coli UTIs are caused by uropathogenic strains of E. coli . E. coli causes a wide range of UTIs, including uncomplicated urethritis, cystitis, pyelonephritis, and urosepsis. Other miscellaneous E. coli infections : Septic arthritis. Endocarditis. Soft tissue infections especially in patients with diabetes. Yersinia enterocolitica Background Small-gram-negative coccobacillus It produces entero and endotoxins Pigs are commonly infected Ingestion of raw or improperly prepared food, such as pork (pork intestine or chitterlings), contaminated unpasteurized milk, and water Clinical presentation Blood and mucus in stool Fever Right lower quadrant pain Leukocytosis Usually confused with appendicitis Treatment No treatment for isolated intestinal infection If extraintestinal manifestation or immune compromised antibiotic is indicated Cefotaxime, TMP-SMX (if older than 2 months), or aminoglycosides Yersinia pestis Background Gram-negative coccobacillus that causes plague Wild rodents are the reservoir It is transmitted by flea or direct contact such as skinning the animals Has a high mortality rate Keyword (adenopathy and hunting) like tularemia Clinical presentation Localized lymphadenopathy "buboes" that suppurate Bubonic type can lead to pneumonic form that rapidly transmitted by coughing to others If not treated, it can lead to sepsis and death Diagnosis Lymph node aspiration or serology Treatment Gentamicin has been used successfully in the treatment of human plague Doxycycline (as dosed for anthrax) is a recommended alternative in patients who cannot take aminoglycosides or in the event of a mass casualty scenario, making parenteral therapy unachievable. Francisella tularensis Background Gram-negative pleomorphic bacillus that causes tularemia or "rabbit fever" It is found in many animals specially the rabbits Its transmitted by ticks and blood sucking flies Organism can be ingested or inhaled It is prevalent in Desert SW; Arkansas, Missouri, and Oklahoma Clinical presentation Fever, chills, myalgias, and arthralgias Irregular ulcers at the site of inoculation Lymphadenopathy that suppurate and form an ulcer Oculoglandular tularemia (Unilateral conjunctivitis, corneal ulceration) Pneumonic tularemia (Dry cough, dyspnea, and pleuritic-type chest pain) Typhoidal tularemia—Fever, chills, myalgias, malaise, and weight loss Diagnosis Serology, e.g., ELISA or PCR Treatment Gentamicin or tetracycline Prevention Avoid tick-infested areas, check cloth for ticks and use tick repellents. Avoid exposure to dead or wild mammals and wear gloves if such exposure is necessary; hands should be thoroughly washed afterwards. Rocky Mountain Spotted Fever (RMSF) Background It is a tickborne rickettsial disease Common in the Southeastern USA Caused by Rickettsia rickettsii Clinical presentation Fever Malaise Headache Abdominal pain Myalgias 3–4 days later the rash will appear Maculopapular rash start in the wrist and ankle spread centrally as well as palm and sole Rash become petechial and purpuric Laboratory ELISA or indirect fluorescent antibody detecting immunoglobulin IgM and IgG to the organism PCR is also available through CDC and prevention Treatment No need to wait to confirm the diagnosis to start treatment Tetracycline particularly doxycycline is the treatment of choice even in children less than 8 years Antibiotic is given for 5–7 days or at least 3 days after fever resolve Best outcome if the treatment started within 5 days of illness Complication Vasculitis DIC Death Ehrlichiosis Background Gram-negative cocci Transmitted by tick bite Monocytic ehrlichiosis (HME) Granulocytic ehrlichiosis (HGE) Common location Southeastern and Southcentral USA Clinical presentation Similar to RMSF but usually without rash Leukopenia Neutropenia Thrombocytopenia Hyponatremia in most of the cases Elevated liver enzymes Treatment Drug of choice is doxycycline (Table 2 ) Table 2 Difference between RMSF and ehrlichiosis Difference Rocky mountain spotted fever Ehrlichiosis Mode of transmission Tick Tick Rash Very common including palm and sole Rare Neutropenia Less common More common Thrombocytopenia Yes Yes Anemia May be present Anemia is not a feature of ehrlichiosis Hyponatremia Yes Yes Liver enzyme May be elevated Usually elevated Treatment Doxycycline Doxycycline Borrelia burgdorferi (Lyme Disease) Background Tick-borne infection caused by spirochete B. burgdorferi Transmitted by Ixodes species ticks in the nymphal stage Commonly seen in the summer. Common areas in the USA are Northeast to mid-Atlantic, e.g., Connecticut, New York, and New Jersey Early localized disease stage I Erythema migrans (pathognomonic skin lesion) either bullseye or clear center Myalgia Arthralgia Fever Early disseminated disease stage II (weeks-months later) Recurrent erythema migrans (rare) Meningitis (lymphocytic) Cranial nerve palsies, e.g., Bell palsy Peripheral neuropathy, e.g., foot drop Heart block ; first, second, or third degree heart block Late disseminated disease stage III Arthritis Oligo-migratory arthritis Remember : Lyme disease can be confused with Juvenile rheumatoid arthritis Diagnosis Erythema migrans is pathognomonic and is an early lesion and antibodies not developed yet. No need to test the patient in order to treat in the first few weeks. Serologic testing is to confirm the diagnosis in stage two or three or in atypical cases. Initial test is sensitive enzyme immunoassay assay (EIA); high false positive rate. Confirm with western blot test. Treatments Isolated Bell palsy or erythema migrans Amoxicillin if   8 years old Cardiac and neurologic complications: Ceftriaxone 75-100 mg/kg/day Treponema pallidum Background TP is spirochete mobile bacteria Mode of transmission: Sexual contact Perinatal Exposure to infected blood or tissue Clinical presentation Primary syphilis Genital chancre It is a painless papule, and then become painless ulcer, which is very contagious Secondary syphilis 2–10 weeks after the chancre heals Maculopapular rash involve the palm and sole Condyloma lata (wart like plaques around the anus or the vagina) Generalized lymphadenopathy Tertiary syphilis (symptomatic late syphilis) Cardiovascular, CNS, gummatous lesions Diagnosis Screening methods: RPR (rapid plasma reagin) and VDRL correlates with disease activity EBV infection can cause false positive results FTA-ABS confirm the diagnosis and this test remain positive for life Treatment Penicillin Doxycycline or tetracycline if allergic to penicillin Congenital syphilis (see chapter The Fetus and Newborn Infants) Leptospirosis Mode of transmission Swimming with dog or contact with fresh water contaminated with the urine of an animal that is a chronic carrier, e.g., rats. Clinical presentation Fever Headache Elevated liver enzyme Diagnosis Early blood culture, later in the disease urine culture may show the organism Treatment Penicillin or doxycycline Mycobacterium tuberculosis Background M. tuberculosis , a tubercle bacillus, is the causative agent of TB. Mycobacteria, such as M. tuberculosis , are aerobic, non spore-forming, non motile, facultative, curved intracellular rods measuring 0.2–0.5 μm by 2–4 μm. It retains many stains after decolorization with acid-alcohol, which is the basis of the acid-fast stains used for pathologic identification. TB is transmitted most commonly via airborne spread. Kissing, shaking hand, and sharing food do not spread the infection. TB is unlikely to spread from child to another child   15 mg/day of prednisone for 1 month) Children   90 %) in pediatric epiglottitis cases (other bacteria can cause epiglottitis as well, e.g., S. pneumoniae , group A beta-hemolytic streptococci, S. aureus , and Moraxella catarrhalis . Occurs primarily in children (ages 2–7 years). The clinical triad of drooling, dysphagia, and distress is the classic presentation. Fever with associated respiratory distress or air hunger occurs in most patients. Treatment in patients with epiglottitis is directed toward relieving the airway obstruction and eradicating the infectious agent. Optimally, initial treatment is provided by a pediatric anesthesiologist and either a pediatric surgeon or a pediatric otolaryngologist. Once the airway is controlled, a pediatric intensivist is required for inpatient management. Buccal infections Buccal cellulitis previously was always caused by H. influenzae infection before the vaccine. Always associated with bacteremia if present. Present with palpable cellulitis on both checks, purplish in color and child looks very toxic. Periorbital cellulitis Previously H. influenzae was the a common cause, now pneumococcus bacteria is the most common etiology Minor trauma or insect bite of the eye lid usually associated with preseptal cellulitis due to S. aureus or a Group A Streptococcus Pyogenic arthritis H. influenzae was the most common cause of septic arthritis before Hib vaccine in children less than 2 years of age Occult bacteremia Occult bacteremia with H. influenzae will result in in 30–50 % developing meningitis or other deep, or focal infection from occult bacteremia. All occult bacteremia from H. influenzae has to be treated immediately. Pneumonia Pneumonia from H. influenzae used to cause about one third of bacterial pneumonia before Hib vaccine and usually associated with pleural effusion, positive blood culture in most of the cases . Treatment (Patient with life threatening illness) Remember : the organism produces beta lactamase which makes amoxicillin is ineffective. Cefotaxime or ceftriaxone is the antimicrobial of choice. Meropenem or chloramphenicol is another option. Amoxicillin is the drug of choice for noninvasive diseases such as otitis media or sinusitis, if amoxicillin fails, uses antibiotics against beta-lactamase-producing strains, e.g., nontypeable H. influenzae including amoxicillin/clavulanic, TMP-SMX, azithromycin, cefuroxime axetil, cefixime, and cefpodoxime. Rifampin antibiotic prophylaxis for contact with invasive H. influenzae type b infection All household who did not receive immunization Less than 4 years with incomplete immunization Younger than 12 months who did not complete primary HIB immunization Immunocompromised child Nursery school and child care center if two or more cases within 60 days Helicobacter pylori Background H. pylori is a gram-negative microaerophilic bacillus It is spiral, curved, or U-shaped and has two to six flagella at one end under microscope Transmission is fecal-oral, oral-oral from human-to human contact Diagnosis Know that AAP recommends testing only when treatment for H. pylori infection would be warranted. Endoscopy remains the gold standard for evaluating H. pylori . H. pylori stool antigen and urea breath test is a promising diagnostic tools. Serologic tests for H. pylori are unreliable marker of disease. Treatment indications Endoscopically confirmed gastric or duodenal ulcer Histologically proven gastric metaplasia Gastric mucosa-associated lymphoid lymphoma (MALT) Prior ulcer disease and current active infection First-line: 14 days treatment regimens for children generally include Clarithromycin (15 mg/kg/day divided twice a day, up to 500 mg per dose) with: Either amoxicillin (50 mg/kg/day divided BID, up to 1 g per dose) or metronidazole (20 mg/kg/day divided BID, up to 500 mg per dose) and Proton-pump inhibitor (PPI) Mycoplasma pneumonia Background M. pneumonia is the leading cause of pneumonia in school age children and young adults Infection is prevalent in person living in group setting Clinical presentation Pulmonary manifestations Nonproductive cough Chills Scattered rales Skin rash Bilateral infiltrate on chest radiograph Extrapulmonary manifestation Pharyngitis Rash Stevens–Johnson syndrome Hemolytic anemia Arthritis CNS disease (encephalitis, cranial nerve palsy (specially CNIII)) Testing for mycoplasma IgG and IgM serology or cold agglutinin Mycoplasma DNA PCR Treatment Mycoplasma lacks the cell wall and beta lactams are not effective Azithromycin is the drug of choice Pasteurella multocida Background Small gram-negative coccobacilli, it is a normal flora in number of animals, e.g., dog and cats. Dog or cat bite is a common risk. Clinical presentation Erythema, tenderness, and edema usually develop rapidly within 24 h. Infection occurs few days after the bite is usually caused by S. aureus . Treatment Clean the wound with soap and water. Treatment should cover potential pathogens, e.g., P. multocida , S. aureus , and anaerobes . Administration of antibiotic within 8–12 h of injury may decrease the risk of infection. Amoxicillin-Clavulanate is the drug of choice Ampicillin-sulbactam IV in severe cases Clindamycin and TMP-SMX is appropriate for children allergic to penicillin. Bordetella pertussis Background Pertussis is a small gram-negative coccobacillus that infects only humans. Pertussis is spread by aerosol droplets expelled while coughing or sneezing in proximity to others. Incubation period of 7–14 days. Clinical presentation Catarrhal phase Lasts from 1 to 2 weeks Mild fever Cough The cough worsens as the patient progresses to the paroxysmal phase Paroxysmal phase Lasts from 2 to 6 weeks Rapid fire or staccato cough Five to ten uninterrupted coughs occur in succession, followed by a "whoop" as the patient rapidly draws in a breath May occur several times per hour Can be associated with cyanosis, salivation, lacrimation, and posttussive emesis Despite the severe spells, patients often appear relatively well between episodes Whoop is usually absent in infants less than 6 months of age Gasping, gagging, and apnea can occur Convalescent phase Decreasing frequency and severity of the coughing episodes Lasts from weeks to months Complications of pertussis Pertussis is most severe in infants   39 °C Hemoglobinopathies, e.g., sickle cell anemia, HIV, and neoplastic diseases Immunocompromised patients at any age Typhoid fever Background Salmonella enterica, Serovar typhi ( S. typhi ) Mode of transmission Poor sanitation and overcrowding Spread by fecal-oral contamination of food or water by individuals who are carriers for S. typhi in either stool or urine Typhoid is endemic in many developing areas Clinical presentation Fever "can exceed 104 °F (40 °C)" Malaise Chills Headache, anorexia, myalgias, and dry cough may be seen Abdominal pain is common Diarrhea is more likely in children Abdominal tenderness, hepatosplenomegaly, and a coated tongue Rose spots (pink, blanchable maculopapular lesions that are 2–4 mm in diameter) are seen on the torso and abdomen Know that neonatal typhoid generally presents within 3 days of birth with fever, emesis, diarrhea, abdominal distention, pronounced hepatomegaly, jaundice , and sometimes, seizures Know that absence of abdominal or intestinal changes is not typical of typhoid Diagnosis Blood cultures are the mainstay of diagnosis Stool culture Treatment and Prognosis Treatment includes: Hydration and correction of fluid-electrolyte imbalance Antipyretics and antibiotics The choice of antibiotic as well as the route and duration depends on the host, site of infection, and sensitivities of the organism. Multidrug resistant (MDR) strains , including resistance to ampicillin and TMP-SM have emerged. IV cefotaxime or ceftriaxone for 14 days is appropriate. For severe typhoid with obtundation, stupor, coma, or shock: Two-day course of IV dexamethasone may be life-saving. Shigella Background Shigella is a gram-negative bacilli Shigella dysenteriae and Shigella flexneri usually cause bloody diarrhea Shigella sonnei and Shigella boydii usually cause watery diarrhea Ingestion of as few as 10 organism can cause diarrhea Incubation period is 2–4 days Outbreak can occur in child care centers Mode of transmission Person to person Feco-oral Ano-oral House flies Contaminated fomites Clinical presentation Range from mild diarrhea to life-threatening dysentery Fever Abdominal camps High-volume watery stools Small-volume bloody stool may follow 24–48 h later Blood-mucoid stool is a common presentation Rectal prolapse occurs in 5–8 % Complications Hemolytic-uremic syndrome Seizures Colonic perforation Toxic encephalopathy Diagnosis Stool culture is diagnostic Stool study with large number of neutrophil is suggestive but not specific Peripheral WBCs are usually elevated; bandemia is very common Treatment Antimicrobial therapy is recommended for all patient with shigellosis. Antimicrobial therapy for 5 days will shorten the duration and eradicate the organism from stool. Oral ampicillin or TMP-SMX but the resistance makes them useless of Shigella infection. Ceftriaxone, ciprofloxacin or azithromycin are usually effective. Ciprofloxacin is not recommended if less than 18 years, if there is an alternative. Daycare center Once Shigella is identified in a daycare or household, all other symptomatic individuals in these environments should be cultured for Shigella as well. Anyone found to have Shigella cannot return to daycare until the diarrhea has stopped and stool culture test is negative. Escherichia coli Background E. coli is a gram-negative, lactose fermenting, motile rod, belonging to the Enterobacteriaceae. E. coli is one of the most frequent causes of many common bacterial infections , including cholecystitis, bacteremia, cholangitis, urinary tract infection (UTI) , and traveler's diarrhea, and other clinical infections such as neonatal meningitis and pneumonia . Acute bacterial meningitis The vast majority of neonatal meningitis cases are caused by E. coli and group B streptococcal infections . Pregnant women are at a higher risk of colonization with the K1 capsular antigen strain of E. coli , which commonly observed in neonatal sepsis. Low-birth weight and a positive CSF culture result portend a poor outcome. Most survivors have subsequent neurologic or developmental abnormalities. Pneumonia E. coli respiratory tract infections are uncommon and are almost always associated with E. coli UTI . Intra-abdominal infections E. coli intra-abdominal infections often result from a perforated viscus (e.g., appendix, diverticulum) or may be associated with intra-abdominal abscess, cholecystitis, and ascending cholangitis. They can be observed in the postoperative period after anastomotic disruption. Abscesses are often polymicrobial. E. coli is one of the more common gram-negative bacilli observed together with anaerobes. Enteric infections Enterotoxigenic E. coli (ETEC) is a cause of traveler's diarrhea; TMP-SMX is the drug of choice . Enteropathogenic E. coli (EPEC) is a cause of childhood diarrhea; can be treated with TMP-SMX Enteroinvasive E. coli (EIEC) causes a Shigella -like dysentery. Enteroaggregative E. coli (EAEC) is primarily associated with persistent diarrhea in children in developing countries, and enteroadherent E. coli (EAEC) is a cause of childhood diarrhea and traveler's diarrhea in Mexico and North Africa. Enterohemorrhagic E. coli (EHEC) causes hemorrhagic colitis or hemolytic-uremic syndrome (HUS). Strains of STEC serotype O157:H7 have caused numerous outbreaks and sporadic cases of bloody diarrhea and HUS. E. coli (O157:H7) Background Gram-negative rods. It occurs in all ages. Transmitted via ingestion of contaminated food, e.g., (ground beef) or infected feces. The disease linked to eating undercooked beef, and unpasteurized milk or apple juice. Produces shiga toxins; the most virulent strain. The incidence of E. coli O157:H7 > Shigella . Clinical presentation Usually begin as nonbloody diarrhea then become bloody Severe abdominal pain is common Fever in one third of the cases May progress to hemorrhagic colitis in severe cases Hemolytic uremic syndrome (HUS) may occur Management No antibiotic is proven to be effective and no prove that antibiotic increase the risk HUS. No antibiotics are indicated. Do not use antimotility agents. UTIs The urinary tract is the most common site of E. coli infection, and more than 90 % of all uncomplicated UTIs are caused by E. coli infection . The recurrence rate after a first E. coli infection is 44 % over 12 months. E. coli UTIs are caused by uropathogenic strains of E. coli . E. coli causes a wide range of UTIs, including uncomplicated urethritis, cystitis, pyelonephritis, and urosepsis. Other miscellaneous E. coli infections : Septic arthritis. Endocarditis. Soft tissue infections especially in patients with diabetes. Yersinia enterocolitica Background Small-gram-negative coccobacillus It produces entero and endotoxins Pigs are commonly infected Ingestion of raw or improperly prepared food, such as pork (pork intestine or chitterlings), contaminated unpasteurized milk, and water Clinical presentation Blood and mucus in stool Fever Right lower quadrant pain Leukocytosis Usually confused with appendicitis Treatment No treatment for isolated intestinal infection If extraintestinal manifestation or immune compromised antibiotic is indicated Cefotaxime, TMP-SMX (if older than 2 months), or aminoglycosides Yersinia pestis Background Gram-negative coccobacillus that causes plague Wild rodents are the reservoir It is transmitted by flea or direct contact such as skinning the animals Has a high mortality rate Keyword (adenopathy and hunting) like tularemia Clinical presentation Localized lymphadenopathy "buboes" that suppurate Bubonic type can lead to pneumonic form that rapidly transmitted by coughing to others If not treated, it can lead to sepsis and death Diagnosis Lymph node aspiration or serology Treatment Gentamicin has been used successfully in the treatment of human plague Doxycycline (as dosed for anthrax) is a recommended alternative in patients who cannot take aminoglycosides or in the event of a mass casualty scenario, making parenteral therapy unachievable. Francisella tularensis Background Gram-negative pleomorphic bacillus that causes tularemia or "rabbit fever" It is found in many animals specially the rabbits Its transmitted by ticks and blood sucking flies Organism can be ingested or inhaled It is prevalent in Desert SW; Arkansas, Missouri, and Oklahoma Clinical presentation Fever, chills, myalgias, and arthralgias Irregular ulcers at the site of inoculation Lymphadenopathy that suppurate and form an ulcer Oculoglandular tularemia (Unilateral conjunctivitis, corneal ulceration) Pneumonic tularemia (Dry cough, dyspnea, and pleuritic-type chest pain) Typhoidal tularemia—Fever, chills, myalgias, malaise, and weight loss Diagnosis Serology, e.g., ELISA or PCR Treatment Gentamicin or tetracycline Prevention Avoid tick-infested areas, check cloth for ticks and use tick repellents. Avoid exposure to dead or wild mammals and wear gloves if such exposure is necessary; hands should be thoroughly washed afterwards. Rocky Mountain Spotted Fever (RMSF) Background It is a tickborne rickettsial disease Common in the Southeastern USA Caused by Rickettsia rickettsii Clinical presentation Fever Malaise Headache Abdominal pain Myalgias 3–4 days later the rash will appear Maculopapular rash start in the wrist and ankle spread centrally as well as palm and sole Rash become petechial and purpuric Laboratory ELISA or indirect fluorescent antibody detecting immunoglobulin IgM and IgG to the organism PCR is also available through CDC and prevention Treatment No need to wait to confirm the diagnosis to start treatment Tetracycline particularly doxycycline is the treatment of choice even in children less than 8 years Antibiotic is given for 5–7 days or at least 3 days after fever resolve Best outcome if the treatment started within 5 days of illness Complication Vasculitis DIC Death Ehrlichiosis Background Gram-negative cocci Transmitted by tick bite Monocytic ehrlichiosis (HME) Granulocytic ehrlichiosis (HGE) Common location Southeastern and Southcentral USA Clinical presentation Similar to RMSF but usually without rash Leukopenia Neutropenia Thrombocytopenia Hyponatremia in most of the cases Elevated liver enzymes Treatment Drug of choice is doxycycline (Table 2 ) Table 2 Difference between RMSF and ehrlichiosis Difference Rocky mountain spotted fever Ehrlichiosis Mode of transmission Tick Tick Rash Very common including palm and sole Rare Neutropenia Less common More common Thrombocytopenia Yes Yes Anemia May be present Anemia is not a feature of ehrlichiosis Hyponatremia Yes Yes Liver enzyme May be elevated Usually elevated Treatment Doxycycline Doxycycline Borrelia burgdorferi (Lyme Disease) Background Tick-borne infection caused by spirochete B. burgdorferi Transmitted by Ixodes species ticks in the nymphal stage Commonly seen in the summer. Common areas in the USA are Northeast to mid-Atlantic, e.g., Connecticut, New York, and New Jersey Early localized disease stage I Erythema migrans (pathognomonic skin lesion) either bullseye or clear center Myalgia Arthralgia Fever Early disseminated disease stage II (weeks-months later) Recurrent erythema migrans (rare) Meningitis (lymphocytic) Cranial nerve palsies, e.g., Bell palsy Peripheral neuropathy, e.g., foot drop Heart block ; first, second, or third degree heart block Late disseminated disease stage III Arthritis Oligo-migratory arthritis Remember : Lyme disease can be confused with Juvenile rheumatoid arthritis Diagnosis Erythema migrans is pathognomonic and is an early lesion and antibodies not developed yet. No need to test the patient in order to treat in the first few weeks. Serologic testing is to confirm the diagnosis in stage two or three or in atypical cases. Initial test is sensitive enzyme immunoassay assay (EIA); high false positive rate. Confirm with western blot test. Treatments Isolated Bell palsy or erythema migrans Amoxicillin if   8 years old Cardiac and neurologic complications: Ceftriaxone 75-100 mg/kg/day Treponema pallidum Background TP is spirochete mobile bacteria Mode of transmission: Sexual contact Perinatal Exposure to infected blood or tissue Clinical presentation Primary syphilis Genital chancre It is a painless papule, and then become painless ulcer, which is very contagious Secondary syphilis 2–10 weeks after the chancre heals Maculopapular rash involve the palm and sole Condyloma lata (wart like plaques around the anus or the vagina) Generalized lymphadenopathy Tertiary syphilis (symptomatic late syphilis) Cardiovascular, CNS, gummatous lesions Diagnosis Screening methods: RPR (rapid plasma reagin) and VDRL correlates with disease activity EBV infection can cause false positive results FTA-ABS confirm the diagnosis and this test remain positive for life Treatment Penicillin Doxycycline or tetracycline if allergic to penicillin Congenital syphilis (see chapter The Fetus and Newborn Infants) Leptospirosis Mode of transmission Swimming with dog or contact with fresh water contaminated with the urine of an animal that is a chronic carrier, e.g., rats. Clinical presentation Fever Headache Elevated liver enzyme Diagnosis Early blood culture, later in the disease urine culture may show the organism Treatment Penicillin or doxycycline Mycobacterium tuberculosis Background M. tuberculosis , a tubercle bacillus, is the causative agent of TB. Mycobacteria, such as M. tuberculosis , are aerobic, non spore-forming, non motile, facultative, curved intracellular rods measuring 0.2–0.5 μm by 2–4 μm. It retains many stains after decolorization with acid-alcohol, which is the basis of the acid-fast stains used for pathologic identification. TB is transmitted most commonly via airborne spread. Kissing, shaking hand, and sharing food do not spread the infection. TB is unlikely to spread from child to another child   15 mg/day of prednisone for 1 month) Children   90 %) in pediatric epiglottitis cases (other bacteria can cause epiglottitis as well, e.g., S. pneumoniae , group A beta-hemolytic streptococci, S. aureus , and Moraxella catarrhalis . Occurs primarily in children (ages 2–7 years). The clinical triad of drooling, dysphagia, and distress is the classic presentation. Fever with associated respiratory distress or air hunger occurs in most patients. Treatment in patients with epiglottitis is directed toward relieving the airway obstruction and eradicating the infectious agent. Optimally, initial treatment is provided by a pediatric anesthesiologist and either a pediatric surgeon or a pediatric otolaryngologist. Once the airway is controlled, a pediatric intensivist is required for inpatient management. Buccal infections Buccal cellulitis previously was always caused by H. influenzae infection before the vaccine. Always associated with bacteremia if present. Present with palpable cellulitis on both checks, purplish in color and child looks very toxic. Periorbital cellulitis Previously H. influenzae was the a common cause, now pneumococcus bacteria is the most common etiology Minor trauma or insect bite of the eye lid usually associated with preseptal cellulitis due to S. aureus or a Group A Streptococcus Pyogenic arthritis H. influenzae was the most common cause of septic arthritis before Hib vaccine in children less than 2 years of age Occult bacteremia Occult bacteremia with H. influenzae will result in in 30–50 % developing meningitis or other deep, or focal infection from occult bacteremia. All occult bacteremia from H. influenzae has to be treated immediately. Pneumonia Pneumonia from H. influenzae used to cause about one third of bacterial pneumonia before Hib vaccine and usually associated with pleural effusion, positive blood culture in most of the cases . Treatment (Patient with life threatening illness) Remember : the organism produces beta lactamase which makes amoxicillin is ineffective. Cefotaxime or ceftriaxone is the antimicrobial of choice. Meropenem or chloramphenicol is another option. Amoxicillin is the drug of choice for noninvasive diseases such as otitis media or sinusitis, if amoxicillin fails, uses antibiotics against beta-lactamase-producing strains, e.g., nontypeable H. influenzae including amoxicillin/clavulanic, TMP-SMX, azithromycin, cefuroxime axetil, cefixime, and cefpodoxime. Rifampin antibiotic prophylaxis for contact with invasive H. influenzae type b infection All household who did not receive immunization Less than 4 years with incomplete immunization Younger than 12 months who did not complete primary HIB immunization Immunocompromised child Nursery school and child care center if two or more cases within 60 days Helicobacter pylori Background H. pylori is a gram-negative microaerophilic bacillus It is spiral, curved, or U-shaped and has two to six flagella at one end under microscope Transmission is fecal-oral, oral-oral from human-to human contact Diagnosis Know that AAP recommends testing only when treatment for H. pylori infection would be warranted. Endoscopy remains the gold standard for evaluating H. pylori . H. pylori stool antigen and urea breath test is a promising diagnostic tools. Serologic tests for H. pylori are unreliable marker of disease. Treatment indications Endoscopically confirmed gastric or duodenal ulcer Histologically proven gastric metaplasia Gastric mucosa-associated lymphoid lymphoma (MALT) Prior ulcer disease and current active infection First-line: 14 days treatment regimens for children generally include Clarithromycin (15 mg/kg/day divided twice a day, up to 500 mg per dose) with: Either amoxicillin (50 mg/kg/day divided BID, up to 1 g per dose) or metronidazole (20 mg/kg/day divided BID, up to 500 mg per dose) and Proton-pump inhibitor (PPI) Mycoplasma pneumonia Background M. pneumonia is the leading cause of pneumonia in school age children and young adults Infection is prevalent in person living in group setting Clinical presentation Pulmonary manifestations Nonproductive cough Chills Scattered rales Skin rash Bilateral infiltrate on chest radiograph Extrapulmonary manifestation Pharyngitis Rash Stevens–Johnson syndrome Hemolytic anemia Arthritis CNS disease (encephalitis, cranial nerve palsy (specially CNIII)) Testing for mycoplasma IgG and IgM serology or cold agglutinin Mycoplasma DNA PCR Treatment Mycoplasma lacks the cell wall and beta lactams are not effective Azithromycin is the drug of choice Pasteurella multocida Background Small gram-negative coccobacilli, it is a normal flora in number of animals, e.g., dog and cats. Dog or cat bite is a common risk. Clinical presentation Erythema, tenderness, and edema usually develop rapidly within 24 h. Infection occurs few days after the bite is usually caused by S. aureus . Treatment Clean the wound with soap and water. Treatment should cover potential pathogens, e.g., P. multocida , S. aureus , and anaerobes . Administration of antibiotic within 8–12 h of injury may decrease the risk of infection. Amoxicillin-Clavulanate is the drug of choice Ampicillin-sulbactam IV in severe cases Clindamycin and TMP-SMX is appropriate for children allergic to penicillin. Bordetella pertussis Background Pertussis is a small gram-negative coccobacillus that infects only humans. Pertussis is spread by aerosol droplets expelled while coughing or sneezing in proximity to others. Incubation period of 7–14 days. Clinical presentation Catarrhal phase Lasts from 1 to 2 weeks Mild fever Cough The cough worsens as the patient progresses to the paroxysmal phase Paroxysmal phase Lasts from 2 to 6 weeks Rapid fire or staccato cough Five to ten uninterrupted coughs occur in succession, followed by a "whoop" as the patient rapidly draws in a breath May occur several times per hour Can be associated with cyanosis, salivation, lacrimation, and posttussive emesis Despite the severe spells, patients often appear relatively well between episodes Whoop is usually absent in infants less than 6 months of age Gasping, gagging, and apnea can occur Convalescent phase Decreasing frequency and severity of the coughing episodes Lasts from weeks to months Complications of pertussis Pertussis is most severe in infants   39 °C Hemoglobinopathies, e.g., sickle cell anemia, HIV, and neoplastic diseases Immunocompromised patients at any age Typhoid fever Background Salmonella enterica, Serovar typhi ( S. typhi ) Mode of transmission Poor sanitation and overcrowding Spread by fecal-oral contamination of food or water by individuals who are carriers for S. typhi in either stool or urine Typhoid is endemic in many developing areas Clinical presentation Fever "can exceed 104 °F (40 °C)" Malaise Chills Headache, anorexia, myalgias, and dry cough may be seen Abdominal pain is common Diarrhea is more likely in children Abdominal tenderness, hepatosplenomegaly, and a coated tongue Rose spots (pink, blanchable maculopapular lesions that are 2–4 mm in diameter) are seen on the torso and abdomen Know that neonatal typhoid generally presents within 3 days of birth with fever, emesis, diarrhea, abdominal distention, pronounced hepatomegaly, jaundice , and sometimes, seizures Know that absence of abdominal or intestinal changes is not typical of typhoid Diagnosis Blood cultures are the mainstay of diagnosis Stool culture Treatment and Prognosis Treatment includes: Hydration and correction of fluid-electrolyte imbalance Antipyretics and antibiotics The choice of antibiotic as well as the route and duration depends on the host, site of infection, and sensitivities of the organism. Multidrug resistant (MDR) strains , including resistance to ampicillin and TMP-SM have emerged. IV cefotaxime or ceftriaxone for 14 days is appropriate. For severe typhoid with obtundation, stupor, coma, or shock: Two-day course of IV dexamethasone may be life-saving. Typhoid fever Background Salmonella enterica, Serovar typhi ( S. typhi ) Mode of transmission Poor sanitation and overcrowding Spread by fecal-oral contamination of food or water by individuals who are carriers for S. typhi in either stool or urine Typhoid is endemic in many developing areas Clinical presentation Fever "can exceed 104 °F (40 °C)" Malaise Chills Headache, anorexia, myalgias, and dry cough may be seen Abdominal pain is common Diarrhea is more likely in children Abdominal tenderness, hepatosplenomegaly, and a coated tongue Rose spots (pink, blanchable maculopapular lesions that are 2–4 mm in diameter) are seen on the torso and abdomen Know that neonatal typhoid generally presents within 3 days of birth with fever, emesis, diarrhea, abdominal distention, pronounced hepatomegaly, jaundice , and sometimes, seizures Know that absence of abdominal or intestinal changes is not typical of typhoid Diagnosis Blood cultures are the mainstay of diagnosis Stool culture Treatment and Prognosis Treatment includes: Hydration and correction of fluid-electrolyte imbalance Antipyretics and antibiotics The choice of antibiotic as well as the route and duration depends on the host, site of infection, and sensitivities of the organism. Multidrug resistant (MDR) strains , including resistance to ampicillin and TMP-SM have emerged. IV cefotaxime or ceftriaxone for 14 days is appropriate. For severe typhoid with obtundation, stupor, coma, or shock: Two-day course of IV dexamethasone may be life-saving. Shigella Background Shigella is a gram-negative bacilli Shigella dysenteriae and Shigella flexneri usually cause bloody diarrhea Shigella sonnei and Shigella boydii usually cause watery diarrhea Ingestion of as few as 10 organism can cause diarrhea Incubation period is 2–4 days Outbreak can occur in child care centers Mode of transmission Person to person Feco-oral Ano-oral House flies Contaminated fomites Clinical presentation Range from mild diarrhea to life-threatening dysentery Fever Abdominal camps High-volume watery stools Small-volume bloody stool may follow 24–48 h later Blood-mucoid stool is a common presentation Rectal prolapse occurs in 5–8 % Complications Hemolytic-uremic syndrome Seizures Colonic perforation Toxic encephalopathy Diagnosis Stool culture is diagnostic Stool study with large number of neutrophil is suggestive but not specific Peripheral WBCs are usually elevated; bandemia is very common Treatment Antimicrobial therapy is recommended for all patient with shigellosis. Antimicrobial therapy for 5 days will shorten the duration and eradicate the organism from stool. Oral ampicillin or TMP-SMX but the resistance makes them useless of Shigella infection. Ceftriaxone, ciprofloxacin or azithromycin are usually effective. Ciprofloxacin is not recommended if less than 18 years, if there is an alternative. Daycare center Once Shigella is identified in a daycare or household, all other symptomatic individuals in these environments should be cultured for Shigella as well. Anyone found to have Shigella cannot return to daycare until the diarrhea has stopped and stool culture test is negative. Escherichia coli Background E. coli is a gram-negative, lactose fermenting, motile rod, belonging to the Enterobacteriaceae. E. coli is one of the most frequent causes of many common bacterial infections , including cholecystitis, bacteremia, cholangitis, urinary tract infection (UTI) , and traveler's diarrhea, and other clinical infections such as neonatal meningitis and pneumonia . Acute bacterial meningitis The vast majority of neonatal meningitis cases are caused by E. coli and group B streptococcal infections . Pregnant women are at a higher risk of colonization with the K1 capsular antigen strain of E. coli , which commonly observed in neonatal sepsis. Low-birth weight and a positive CSF culture result portend a poor outcome. Most survivors have subsequent neurologic or developmental abnormalities. Pneumonia E. coli respiratory tract infections are uncommon and are almost always associated with E. coli UTI . Intra-abdominal infections E. coli intra-abdominal infections often result from a perforated viscus (e.g., appendix, diverticulum) or may be associated with intra-abdominal abscess, cholecystitis, and ascending cholangitis. They can be observed in the postoperative period after anastomotic disruption. Abscesses are often polymicrobial. E. coli is one of the more common gram-negative bacilli observed together with anaerobes. Enteric infections Enterotoxigenic E. coli (ETEC) is a cause of traveler's diarrhea; TMP-SMX is the drug of choice . Enteropathogenic E. coli (EPEC) is a cause of childhood diarrhea; can be treated with TMP-SMX Enteroinvasive E. coli (EIEC) causes a Shigella -like dysentery. Enteroaggregative E. coli (EAEC) is primarily associated with persistent diarrhea in children in developing countries, and enteroadherent E. coli (EAEC) is a cause of childhood diarrhea and traveler's diarrhea in Mexico and North Africa. Enterohemorrhagic E. coli (EHEC) causes hemorrhagic colitis or hemolytic-uremic syndrome (HUS). Strains of STEC serotype O157:H7 have caused numerous outbreaks and sporadic cases of bloody diarrhea and HUS. E. coli (O157:H7) Background Gram-negative rods. It occurs in all ages. Transmitted via ingestion of contaminated food, e.g., (ground beef) or infected feces. The disease linked to eating undercooked beef, and unpasteurized milk or apple juice. Produces shiga toxins; the most virulent strain. The incidence of E. coli O157:H7 > Shigella . Clinical presentation Usually begin as nonbloody diarrhea then become bloody Severe abdominal pain is common Fever in one third of the cases May progress to hemorrhagic colitis in severe cases Hemolytic uremic syndrome (HUS) may occur Management No antibiotic is proven to be effective and no prove that antibiotic increase the risk HUS. No antibiotics are indicated. Do not use antimotility agents. UTIs The urinary tract is the most common site of E. coli infection, and more than 90 % of all uncomplicated UTIs are caused by E. coli infection . The recurrence rate after a first E. coli infection is 44 % over 12 months. E. coli UTIs are caused by uropathogenic strains of E. coli . E. coli causes a wide range of UTIs, including uncomplicated urethritis, cystitis, pyelonephritis, and urosepsis. Other miscellaneous E. coli infections : Septic arthritis. Endocarditis. Soft tissue infections especially in patients with diabetes. Yersinia enterocolitica Background Small-gram-negative coccobacillus It produces entero and endotoxins Pigs are commonly infected Ingestion of raw or improperly prepared food, such as pork (pork intestine or chitterlings), contaminated unpasteurized milk, and water Clinical presentation Blood and mucus in stool Fever Right lower quadrant pain Leukocytosis Usually confused with appendicitis Treatment No treatment for isolated intestinal infection If extraintestinal manifestation or immune compromised antibiotic is indicated Cefotaxime, TMP-SMX (if older than 2 months), or aminoglycosides Yersinia pestis Background Gram-negative coccobacillus that causes plague Wild rodents are the reservoir It is transmitted by flea or direct contact such as skinning the animals Has a high mortality rate Keyword (adenopathy and hunting) like tularemia Clinical presentation Localized lymphadenopathy "buboes" that suppurate Bubonic type can lead to pneumonic form that rapidly transmitted by coughing to others If not treated, it can lead to sepsis and death Diagnosis Lymph node aspiration or serology Treatment Gentamicin has been used successfully in the treatment of human plague Doxycycline (as dosed for anthrax) is a recommended alternative in patients who cannot take aminoglycosides or in the event of a mass casualty scenario, making parenteral therapy unachievable. Francisella tularensis Background Gram-negative pleomorphic bacillus that causes tularemia or "rabbit fever" It is found in many animals specially the rabbits Its transmitted by ticks and blood sucking flies Organism can be ingested or inhaled It is prevalent in Desert SW; Arkansas, Missouri, and Oklahoma Clinical presentation Fever, chills, myalgias, and arthralgias Irregular ulcers at the site of inoculation Lymphadenopathy that suppurate and form an ulcer Oculoglandular tularemia (Unilateral conjunctivitis, corneal ulceration) Pneumonic tularemia (Dry cough, dyspnea, and pleuritic-type chest pain) Typhoidal tularemia—Fever, chills, myalgias, malaise, and weight loss Diagnosis Serology, e.g., ELISA or PCR Treatment Gentamicin or tetracycline Prevention Avoid tick-infested areas, check cloth for ticks and use tick repellents. Avoid exposure to dead or wild mammals and wear gloves if such exposure is necessary; hands should be thoroughly washed afterwards. Rocky Mountain Spotted Fever (RMSF) Background It is a tickborne rickettsial disease Common in the Southeastern USA Caused by Rickettsia rickettsii Clinical presentation Fever Malaise Headache Abdominal pain Myalgias 3–4 days later the rash will appear Maculopapular rash start in the wrist and ankle spread centrally as well as palm and sole Rash become petechial and purpuric Laboratory ELISA or indirect fluorescent antibody detecting immunoglobulin IgM and IgG to the organism PCR is also available through CDC and prevention Treatment No need to wait to confirm the diagnosis to start treatment Tetracycline particularly doxycycline is the treatment of choice even in children less than 8 years Antibiotic is given for 5–7 days or at least 3 days after fever resolve Best outcome if the treatment started within 5 days of illness Complication Vasculitis DIC Death Ehrlichiosis Background Gram-negative cocci Transmitted by tick bite Monocytic ehrlichiosis (HME) Granulocytic ehrlichiosis (HGE) Common location Southeastern and Southcentral USA Clinical presentation Similar to RMSF but usually without rash Leukopenia Neutropenia Thrombocytopenia Hyponatremia in most of the cases Elevated liver enzymes Treatment Drug of choice is doxycycline (Table 2 ) Table 2 Difference between RMSF and ehrlichiosis Difference Rocky mountain spotted fever Ehrlichiosis Mode of transmission Tick Tick Rash Very common including palm and sole Rare Neutropenia Less common More common Thrombocytopenia Yes Yes Anemia May be present Anemia is not a feature of ehrlichiosis Hyponatremia Yes Yes Liver enzyme May be elevated Usually elevated Treatment Doxycycline Doxycycline Borrelia burgdorferi (Lyme Disease) Background Tick-borne infection caused by spirochete B. burgdorferi Transmitted by Ixodes species ticks in the nymphal stage Commonly seen in the summer. Common areas in the USA are Northeast to mid-Atlantic, e.g., Connecticut, New York, and New Jersey Early localized disease stage I Erythema migrans (pathognomonic skin lesion) either bullseye or clear center Myalgia Arthralgia Fever Early disseminated disease stage II (weeks-months later) Recurrent erythema migrans (rare) Meningitis (lymphocytic) Cranial nerve palsies, e.g., Bell palsy Peripheral neuropathy, e.g., foot drop Heart block ; first, second, or third degree heart block Late disseminated disease stage III Arthritis Oligo-migratory arthritis Remember : Lyme disease can be confused with Juvenile rheumatoid arthritis Diagnosis Erythema migrans is pathognomonic and is an early lesion and antibodies not developed yet. No need to test the patient in order to treat in the first few weeks. Serologic testing is to confirm the diagnosis in stage two or three or in atypical cases. Initial test is sensitive enzyme immunoassay assay (EIA); high false positive rate. Confirm with western blot test. Treatments Isolated Bell palsy or erythema migrans Amoxicillin if   8 years old Cardiac and neurologic complications: Ceftriaxone 75-100 mg/kg/day Rocky Mountain Spotted Fever (RMSF) Background It is a tickborne rickettsial disease Common in the Southeastern USA Caused by Rickettsia rickettsii Clinical presentation Fever Malaise Headache Abdominal pain Myalgias 3–4 days later the rash will appear Maculopapular rash start in the wrist and ankle spread centrally as well as palm and sole Rash become petechial and purpuric Laboratory ELISA or indirect fluorescent antibody detecting immunoglobulin IgM and IgG to the organism PCR is also available through CDC and prevention Treatment No need to wait to confirm the diagnosis to start treatment Tetracycline particularly doxycycline is the treatment of choice even in children less than 8 years Antibiotic is given for 5–7 days or at least 3 days after fever resolve Best outcome if the treatment started within 5 days of illness Complication Vasculitis DIC Death Ehrlichiosis Background Gram-negative cocci Transmitted by tick bite Monocytic ehrlichiosis (HME) Granulocytic ehrlichiosis (HGE) Common location Southeastern and Southcentral USA Clinical presentation Similar to RMSF but usually without rash Leukopenia Neutropenia Thrombocytopenia Hyponatremia in most of the cases Elevated liver enzymes Treatment Drug of choice is doxycycline (Table 2 ) Table 2 Difference between RMSF and ehrlichiosis Difference Rocky mountain spotted fever Ehrlichiosis Mode of transmission Tick Tick Rash Very common including palm and sole Rare Neutropenia Less common More common Thrombocytopenia Yes Yes Anemia May be present Anemia is not a feature of ehrlichiosis Hyponatremia Yes Yes Liver enzyme May be elevated Usually elevated Treatment Doxycycline Doxycycline Borrelia burgdorferi (Lyme Disease) Background Tick-borne infection caused by spirochete B. burgdorferi Transmitted by Ixodes species ticks in the nymphal stage Commonly seen in the summer. Common areas in the USA are Northeast to mid-Atlantic, e.g., Connecticut, New York, and New Jersey Early localized disease stage I Erythema migrans (pathognomonic skin lesion) either bullseye or clear center Myalgia Arthralgia Fever Early disseminated disease stage II (weeks-months later) Recurrent erythema migrans (rare) Meningitis (lymphocytic) Cranial nerve palsies, e.g., Bell palsy Peripheral neuropathy, e.g., foot drop Heart block ; first, second, or third degree heart block Late disseminated disease stage III Arthritis Oligo-migratory arthritis Remember : Lyme disease can be confused with Juvenile rheumatoid arthritis Diagnosis Erythema migrans is pathognomonic and is an early lesion and antibodies not developed yet. No need to test the patient in order to treat in the first few weeks. Serologic testing is to confirm the diagnosis in stage two or three or in atypical cases. Initial test is sensitive enzyme immunoassay assay (EIA); high false positive rate. Confirm with western blot test. Treatments Isolated Bell palsy or erythema migrans Amoxicillin if   8 years old Cardiac and neurologic complications: Ceftriaxone 75-100 mg/kg/day Treponema pallidum Background TP is spirochete mobile bacteria Mode of transmission: Sexual contact Perinatal Exposure to infected blood or tissue Clinical presentation Primary syphilis Genital chancre It is a painless papule, and then become painless ulcer, which is very contagious Secondary syphilis 2–10 weeks after the chancre heals Maculopapular rash involve the palm and sole Condyloma lata (wart like plaques around the anus or the vagina) Generalized lymphadenopathy Tertiary syphilis (symptomatic late syphilis) Cardiovascular, CNS, gummatous lesions Diagnosis Screening methods: RPR (rapid plasma reagin) and VDRL correlates with disease activity EBV infection can cause false positive results FTA-ABS confirm the diagnosis and this test remain positive for life Treatment Penicillin Doxycycline or tetracycline if allergic to penicillin Congenital syphilis (see chapter The Fetus and Newborn Infants) Leptospirosis Mode of transmission Swimming with dog or contact with fresh water contaminated with the urine of an animal that is a chronic carrier, e.g., rats. Clinical presentation Fever Headache Elevated liver enzyme Diagnosis Early blood culture, later in the disease urine culture may show the organism Treatment Penicillin or doxycycline Leptospirosis Mode of transmission Swimming with dog or contact with fresh water contaminated with the urine of an animal that is a chronic carrier, e.g., rats. Clinical presentation Fever Headache Elevated liver enzyme Diagnosis Early blood culture, later in the disease urine culture may show the organism Treatment Penicillin or doxycycline Mycobacterium tuberculosis Background M. tuberculosis , a tubercle bacillus, is the causative agent of TB. Mycobacteria, such as M. tuberculosis , are aerobic, non spore-forming, non motile, facultative, curved intracellular rods measuring 0.2–0.5 μm by 2–4 μm. It retains many stains after decolorization with acid-alcohol, which is the basis of the acid-fast stains used for pathologic identification. TB is transmitted most commonly via airborne spread. Kissing, shaking hand, and sharing food do not spread the infection. TB is unlikely to spread from child to another child   15 mg/day of prednisone for 1 month) Children   5 % Chest radiography may show consolidations and hilar lymphadenopathy Treatment Amphotericin B in Severe disseminated disease Fluconazole for CNS infections Blastomyces Blastomyces causes illness similar to Histoplasma and Coccidioides It is seen in Arkansas and Wisconsin hunters and loggers Outbreak occurred in kids visited Wisconsin lodge and beaver dam Blastomyces may disseminate to the skin and cause crusted skin lesions Bone lesion more common with blastomycosis Itraconazole or amphotericin B is the treatment of choice depending on the severity Sporotrichosis schenckii Common in florists Symptoms may take from 7 to 30 day after inoculation Present with painless papule at the site of inoculation then ulcerates Extracutaneous manifestation may occur Itraconazole is the drug of choice Saturated solution K iodide, is much less costly and still recommended as an alternative treatment Candida Species Candida albicans is the most commonly isolated species, and cause infections (Candidiasis or thrush). Systemic infections of blood stream and major organs (invasive candidiasis or candidemia, particularly in immunocompromised patients. Candida appears as budding yeast cells and pseudohyphae (Fig. 15 ). Fig. 15 Candida albicans in blood culture (gram stain, original magnification × 1000). Budding yeast cells (blastoconidia, black arrow ) and pseudohyphae ( white arrow ) Oral Thrush Background Common is the first 6 postnatal months Possibly due to infants' immunologic immaturity Infection sources Contaminated bottle nipples, pacifier, or dropper, e.g., vitamin dropper. Infected mother's nipples (although the incidence is high in formula fed infants). Maternal vaginal colonization with Candida . Recognize Recurrent or persistent oral thrush beyond 6–12 months raises the concern of immunodeficiency, especially if associated with failure to thrive or hepatosplenomegaly. Risk of infection Use of inhaled steroid without adequate rinsing afterward or oral antibiotics can cause oral thrush. Poorly controlled diabetes in adult can cause candida infection however is not associated with gestational diabetes. Clinical presentation Infant may have trouble feeding in severe cases. Tiny focal white area that enlarge to white patches on oral mucosa (Fig. 16 ). Fig. 16 Thrush: Tiny focal white areas that enlarge to white patches on oral mucosa, it was difficult to remove the white spots with the tongue blade If scraped with a tongue blade, lesions are difficult to remove and leave behind an inflamed base that may be painful and may bleed. Examine the patient with diaper dermatitis for oral lesions. Treatment Oral nystatin. Once-daily oral fluconazole is superior to oral nystatin for resistant thrush and effective candidal diaper dermatitis. Candidal Diaper Dermatitis Clinical presentation Lesions consist of beefy-red plaques, often with scalloped borders. Satellite papules and pustules may be observed surrounding the plaques (Fig. 17 ). Fig. 17 Candidal diaper rash: lesions consist of beefy-red plaques, with satellite papules Maceration is often present, especially in intertriginous areas. Treatment Once-daily oral fluconazole is superior to oral nystatin for resistant thrush and effective candidal diaper dermatitis. Topical clotrimazole if resistant to topical nystatin. Vulvovaginitis Background Common in pubertal and adolescent girls Risk factors Oral antibiotics Oral contraceptive Pregnancy Poor hygiene Diabetes Clinical presentation Vulvar/vaginal erythema, and itching White, cottage cheese like vaginal discharge Treatment Topical nystatin or clotrimazole Single dose of oral fluconazole Candidal Infections in Neonates Background Very low-birth weight Prolonged venous catheter (obtain culture from the catheter) Treatment Remove the catheter Parenteral amphotericin (lipid-complex formulation (less nephrotoxic)) Monitor for hypokalemia Aspergillus Background Aspergillus species is ubiquitous molds found in organic matter. Most common species affect the human is Aspergillus fumigatus and Aspergillus niger . Mode of transmission Inhalation of fungus spores Clinical presentation Underlying asthma or cystic fibrosis May presents with fever and pulmonary infiltrates not responsive to antibiotics (allergic bronchopulmonary aspergillosis) Patient may cough mucous plug Underlying preexisting cavities, e.g., TB, sarcoidosis, or CF Aspergilloma or fungal ball, it may cause hemoptysis Allergic fungal sinusitis Present with purulent discharge and unilateral opacity Immunocompromised patient may present: Fever, cough, dyspnea, pleuritic chest pain, and hemoptysis Diagnosis Elevated IgE level Deterioration of lung function and increase in sputum production in chronically ill patients, e.g., CF or asthma Peripheral eosinophilia Sputum culture Bronchoalveolar lavage (BAL) CT scan Treatment of allergic pulmonary aspergillosis Oral steroids Cryptococcosis Background Infection with the encapsulated yeast Cryptococcus neoformans can result in harmless colonization of the airways It can also lead to meningitis or disseminated disease, especially in persons with defective cell-mediated immunity. Cryptococcosis represents a major life-threatening fungal infection in patients with severe HIV infection and may also complicate organ transplantation, reticuloendothelial malignancy, corticosteroid treatment, or sarcoidosis. Clinical presentation Severity of symptoms and presentation depends on the immune status and the affected organs Pulmonary ; cough, pleuritic chest pain, fever, dyspnea, weight loss, and malaise Meningitis ; headache, lethargy, confusion, seizures, and coma Skin ; papules, pustules, nodules, ulcers, or draining sinuses Diagnosis Cutaneous lesions: Biopsy with fungal stains and cultures. Blood: Fungal culture, cryptococcal serology, and cryptococcal antigen testing. Cerebrospinal fluid: India ink smear , fungal culture, and cryptococcal antigen testing. In AIDS patients with cryptococcal pneumonia , culture of bronchoalveolar lavage washings. Treatment for cryptococcal meningitis Amphotericin B, and flucytosine for 2 weeks Flucytosine speeds clearance of viable yeast from CSF but is potentially toxic, especially in patients with renal dysfunction Then fluconazole for 8-10 weeks Malassezia furfur Overview Can cause tinea versicolor (see skin disorders) Can cause neonatal infection in NICU babies receiving TPN with lipids NICU babies with M. furfur may present with fever, bilateral interstitial infiltrates, and increased WBCs M. furfur requires olive oil overlay to grow Management of infection in Neonates Removal of catheters Stop lipid infusion Start amphotericin B or fluconazole Histoplasmosis Background Endemic areas: Ohio, Missouri, and Mississippi River valleys Mode of transmission Inhalation of spores from birds excreta or contaminated soil No person to person transmission Clinical presentation Flu like symptoms Pulmonary infiltrates Hilar lymphadenopathy with or without calcifications Erythema nodosum In younger children may develop progressive disseminated histoplasmosis Treatment Amphotericin B Coccidioides (Coccidioidomycosis) Background Endemic areas California, Arizona, New Mexico, and Texas Mode of transmission Inhalation of airborne spores Clinical presentation Most cases are asymptomatic Fever Cough Weight loss (common) Fatigue Shortness of breath Chills Erythema nodosum Night sweat Mild respiratory distress or respiratory failure in severe cases Diagnosis Culture and DNA probe is the most definitive method for the diagnosis High index of suspicion is important in patient who travelled or underlying medical conditions Elevated ESR Lymphocytosis and monocytosis Eosinophilia > 5 % Chest radiography may show consolidations and hilar lymphadenopathy Treatment Amphotericin B in Severe disseminated disease Fluconazole for CNS infections Blastomyces Blastomyces causes illness similar to Histoplasma and Coccidioides It is seen in Arkansas and Wisconsin hunters and loggers Outbreak occurred in kids visited Wisconsin lodge and beaver dam Blastomyces may disseminate to the skin and cause crusted skin lesions Bone lesion more common with blastomycosis Itraconazole or amphotericin B is the treatment of choice depending on the severity Sporotrichosis schenckii Common in florists Symptoms may take from 7 to 30 day after inoculation Present with painless papule at the site of inoculation then ulcerates Extracutaneous manifestation may occur Itraconazole is the drug of choice Saturated solution K iodide, is much less costly and still recommended as an alternative treatment Candidal Diaper Dermatitis Clinical presentation Lesions consist of beefy-red plaques, often with scalloped borders. Satellite papules and pustules may be observed surrounding the plaques (Fig. 17 ). Fig. 17 Candidal diaper rash: lesions consist of beefy-red plaques, with satellite papules Maceration is often present, especially in intertriginous areas. Treatment Once-daily oral fluconazole is superior to oral nystatin for resistant thrush and effective candidal diaper dermatitis. Topical clotrimazole if resistant to topical nystatin. Vulvovaginitis Background Common in pubertal and adolescent girls Risk factors Oral antibiotics Oral contraceptive Pregnancy Poor hygiene Diabetes Clinical presentation Vulvar/vaginal erythema, and itching White, cottage cheese like vaginal discharge Treatment Topical nystatin or clotrimazole Single dose of oral fluconazole Candidal Infections in Neonates Background Very low-birth weight Prolonged venous catheter (obtain culture from the catheter) Treatment Remove the catheter Parenteral amphotericin (lipid-complex formulation (less nephrotoxic)) Monitor for hypokalemia Aspergillus Background Aspergillus species is ubiquitous molds found in organic matter. Most common species affect the human is Aspergillus fumigatus and Aspergillus niger . Mode of transmission Inhalation of fungus spores Clinical presentation Underlying asthma or cystic fibrosis May presents with fever and pulmonary infiltrates not responsive to antibiotics (allergic bronchopulmonary aspergillosis) Patient may cough mucous plug Underlying preexisting cavities, e.g., TB, sarcoidosis, or CF Aspergilloma or fungal ball, it may cause hemoptysis Allergic fungal sinusitis Present with purulent discharge and unilateral opacity Immunocompromised patient may present: Fever, cough, dyspnea, pleuritic chest pain, and hemoptysis Diagnosis Elevated IgE level Deterioration of lung function and increase in sputum production in chronically ill patients, e.g., CF or asthma Peripheral eosinophilia Sputum culture Bronchoalveolar lavage (BAL) CT scan Treatment of allergic pulmonary aspergillosis Oral steroids Cryptococcosis Background Infection with the encapsulated yeast Cryptococcus neoformans can result in harmless colonization of the airways It can also lead to meningitis or disseminated disease, especially in persons with defective cell-mediated immunity. Cryptococcosis represents a major life-threatening fungal infection in patients with severe HIV infection and may also complicate organ transplantation, reticuloendothelial malignancy, corticosteroid treatment, or sarcoidosis. Clinical presentation Severity of symptoms and presentation depends on the immune status and the affected organs Pulmonary ; cough, pleuritic chest pain, fever, dyspnea, weight loss, and malaise Meningitis ; headache, lethargy, confusion, seizures, and coma Skin ; papules, pustules, nodules, ulcers, or draining sinuses Diagnosis Cutaneous lesions: Biopsy with fungal stains and cultures. Blood: Fungal culture, cryptococcal serology, and cryptococcal antigen testing. Cerebrospinal fluid: India ink smear , fungal culture, and cryptococcal antigen testing. In AIDS patients with cryptococcal pneumonia , culture of bronchoalveolar lavage washings. Treatment for cryptococcal meningitis Amphotericin B, and flucytosine for 2 weeks Flucytosine speeds clearance of viable yeast from CSF but is potentially toxic, especially in patients with renal dysfunction Then fluconazole for 8-10 weeks Malassezia furfur Overview Can cause tinea versicolor (see skin disorders) Can cause neonatal infection in NICU babies receiving TPN with lipids NICU babies with M. furfur may present with fever, bilateral interstitial infiltrates, and increased WBCs M. furfur requires olive oil overlay to grow Management of infection in Neonates Removal of catheters Stop lipid infusion Start amphotericin B or fluconazole Histoplasmosis Background Endemic areas: Ohio, Missouri, and Mississippi River valleys Mode of transmission Inhalation of spores from birds excreta or contaminated soil No person to person transmission Clinical presentation Flu like symptoms Pulmonary infiltrates Hilar lymphadenopathy with or without calcifications Erythema nodosum In younger children may develop progressive disseminated histoplasmosis Treatment Amphotericin B Histoplasmosis Background Endemic areas: Ohio, Missouri, and Mississippi River valleys Mode of transmission Inhalation of spores from birds excreta or contaminated soil No person to person transmission Clinical presentation Flu like symptoms Pulmonary infiltrates Hilar lymphadenopathy with or without calcifications Erythema nodosum In younger children may develop progressive disseminated histoplasmosis Treatment Amphotericin B Coccidioides (Coccidioidomycosis) Background Endemic areas California, Arizona, New Mexico, and Texas Mode of transmission Inhalation of airborne spores Clinical presentation Most cases are asymptomatic Fever Cough Weight loss (common) Fatigue Shortness of breath Chills Erythema nodosum Night sweat Mild respiratory distress or respiratory failure in severe cases Diagnosis Culture and DNA probe is the most definitive method for the diagnosis High index of suspicion is important in patient who travelled or underlying medical conditions Elevated ESR Lymphocytosis and monocytosis Eosinophilia > 5 % Chest radiography may show consolidations and hilar lymphadenopathy Treatment Amphotericin B in Severe disseminated disease Fluconazole for CNS infections Blastomyces Blastomyces causes illness similar to Histoplasma and Coccidioides It is seen in Arkansas and Wisconsin hunters and loggers Outbreak occurred in kids visited Wisconsin lodge and beaver dam Blastomyces may disseminate to the skin and cause crusted skin lesions Bone lesion more common with blastomycosis Itraconazole or amphotericin B is the treatment of choice depending on the severity Sporotrichosis schenckii Common in florists Symptoms may take from 7 to 30 day after inoculation Present with painless papule at the site of inoculation then ulcerates Extracutaneous manifestation may occur Itraconazole is the drug of choice Saturated solution K iodide, is much less costly and still recommended as an alternative treatment Sporotrichosis schenckii Common in florists Symptoms may take from 7 to 30 day after inoculation Present with painless papule at the site of inoculation then ulcerates Extracutaneous manifestation may occur Itraconazole is the drug of choice Saturated solution K iodide, is much less costly and still recommended as an alternative treatment Protozoa Giardia lamblia (Giardiasis) Background Giardiasis is an infection of the small intestine caused by the flagellated protozoan Giardia intestinalis . Mode of transmission Travelers and hikers who drink water contaminated with stool from infected animals such as beavers, muskrats, and sheep. Outbreaks also may occur from sewage contamination of water supplies. Unprotected anal sex also is a source of transmission. Child care centers from fecal-oral transmission. Food-associated outbreaks may occur. Clinical presentation Most infections remaining asymptomatic Watery diarrhea with abdominal cramping Nausea Vomiting Weight loss Flatulence Diagnosed Microscopic examination of the stool for cysts or by antigen detection Treatment Indicated for all symptomatic patients. Metronidazole, a single dose of tinidazole, or nitazoxanide for 3 days. Immunocompromised patients, e.g., AIDS at increased risk for chronic giardiasis and treatment failure. Entamoeba histolytica Background Amebiasis is caused by pathogenic species of Entamoeba Mode of transmission Fecal-oral route Travel to high-risk area, e.g., Mexico Clinical presentation Can be asymptomatic Amebic dysentery or colitis Bloody diarrhea with mucus Tenesmus Hepatic abscess Fever Abdominal pain Tender enlarged liver Elevated liver enzymes Elevated ESR Diagnosis Stool microscopic examination Stool antigen Serum antibody Ultrasound if liver abscess is suspected Treatment Symptomatic cases. Metronidazole followed by paromomycin or iodoquinol to eradicate colonization. Asymptomatic amebiasis in non endemic areas should be treated with a luminal agent (iodoquinol, paromomycin, or diloxanide furoate) to eradicate infection. Amebic liver abscess can be cured without drainage and even by 1 dose of metronidazole. Cryptosporidiosis Background Cryptosporidiosis, caused by Cryptosporidium protozoa Transmitted via feco-oral route; child care centers, and swimming pools Clinical presentation Diarrhea Chronic diarrhea in immunodeficient patients Treatment Many immunocompetent patients who have cryptosporidiosis have self-limited disease and do not require therapy A 3-day course of nitazoxanide: To reduce the duration and transmission of diarrhea in children older than 1 year of age No swimming pool for at least 2 weeks after the diarrhea stopped Toxoplasma gondii (Toxoplasmosis) Background Obligate intracellular protozoa Mode of transmission Ingestion of contaminated raw or uncooked meat Cats excreta Organ transplants Transplacental to fetus causes congenital toxoplasmosis (see chapter Fetus and Newborns) Clinical presentation Most cases are asymptomatic Fever Malaise Rash Myalgia Cervical lymphadenopathy (most common sign) Brain abscess (test for HIV) Chorioretinitis usually present years later (mostly congenital) Diagnosis Head CT: ring-enhanced lesion Toxoplasma IgM antibodies PCR Treatment Pyrimethamine plus sulfadiazine and folic acid Lifelong therapy in HIV patients Pneumocystis jiroveci (Carinii) Background Unicellular fungi that do not respond to antifungal treatment Mode of transmission is unknown Commonly seen in immunocompromised patients, e.g., HIV patients Clinical presentation Subacute diffuse pneumonitis Dyspnea Tachycardia Oxygen desaturation Nonproductive cough Fever Diagnosis Chest radiography Bilateral diffuse interstitial disease Low CD4 Bronchoalveolar lavage Lung biopsy Treatment TMP-SMX IV pentamidine in severe cases Prophylaxis in immunocompromised patients TMP-SMX Plasmodium (Malaria) Background Intracellular protozoa Transmitted by mosquito bites in endemic area, e.g., south Africa Plasmodium falciparum Most severe Symptoms develop within a month from returning from endemic area Most common cause of congenital malaria Complications Cerebral malaria Pulmonary edema Severe anemia Renal failure Shock Treatment Chloroquine sensitive ◦ Chloroquine Chloroquine resistant: ◦ Quinine plus doxycycline or clindamycin ◦ Or atovaquone-proguanil ◦ Or mefloquine Severe cases ◦ Quinidine gluconate IV plus doxycycline or clindamycin Plasmodium malariae , P. vivax , and P. ovale Periodicity of symptoms Nephrotic syndrome- P. malariae (most benign form) Hypersplenism and splenic rupture- P. vivax and P. ovale Treatment ◦ Chloroquine plus primaquine for P. vivax , and P. ovale ◦ Chloroquine phosphate for P. malaria Clinical presentation of malaria History of travelling to endemic areas in the past years Paroxysmal fever, sweat and rigors Pallor and jaundice Headache and myalgia Abdominal pain Vomiting and diarrhea In severe cases Change in mental status Hepatosplenomegaly Anemia Thrombocytopenia Hypotension Hypoglycemia Hyperkalemia Respiratory distress Diagnosis RBCs smear Prevention Travelling to chloroquine resistant areas, e.g., South Africa Atovaquone-proguanil 2 weeks before and 4 weeks after or Doxycycline (> 8 years old) Mefloquine (safe for pregnant) Travelling to chloroquine sensitive areas, e.g., South America Chloroquine 2 weeks before and 4 weeks after or Atovaquone-proguanil or Mefloquine Giardia lamblia (Giardiasis) Background Giardiasis is an infection of the small intestine caused by the flagellated protozoan Giardia intestinalis . Mode of transmission Travelers and hikers who drink water contaminated with stool from infected animals such as beavers, muskrats, and sheep. Outbreaks also may occur from sewage contamination of water supplies. Unprotected anal sex also is a source of transmission. Child care centers from fecal-oral transmission. Food-associated outbreaks may occur. Clinical presentation Most infections remaining asymptomatic Watery diarrhea with abdominal cramping Nausea Vomiting Weight loss Flatulence Diagnosed Microscopic examination of the stool for cysts or by antigen detection Treatment Indicated for all symptomatic patients. Metronidazole, a single dose of tinidazole, or nitazoxanide for 3 days. Immunocompromised patients, e.g., AIDS at increased risk for chronic giardiasis and treatment failure. Entamoeba histolytica Background Amebiasis is caused by pathogenic species of Entamoeba Mode of transmission Fecal-oral route Travel to high-risk area, e.g., Mexico Clinical presentation Can be asymptomatic Amebic dysentery or colitis Bloody diarrhea with mucus Tenesmus Hepatic abscess Fever Abdominal pain Tender enlarged liver Elevated liver enzymes Elevated ESR Diagnosis Stool microscopic examination Stool antigen Serum antibody Ultrasound if liver abscess is suspected Treatment Symptomatic cases. Metronidazole followed by paromomycin or iodoquinol to eradicate colonization. Asymptomatic amebiasis in non endemic areas should be treated with a luminal agent (iodoquinol, paromomycin, or diloxanide furoate) to eradicate infection. Amebic liver abscess can be cured without drainage and even by 1 dose of metronidazole. Cryptosporidiosis Background Cryptosporidiosis, caused by Cryptosporidium protozoa Transmitted via feco-oral route; child care centers, and swimming pools Clinical presentation Diarrhea Chronic diarrhea in immunodeficient patients Treatment Many immunocompetent patients who have cryptosporidiosis have self-limited disease and do not require therapy A 3-day course of nitazoxanide: To reduce the duration and transmission of diarrhea in children older than 1 year of age No swimming pool for at least 2 weeks after the diarrhea stopped Cryptosporidiosis Background Cryptosporidiosis, caused by Cryptosporidium protozoa Transmitted via feco-oral route; child care centers, and swimming pools Clinical presentation Diarrhea Chronic diarrhea in immunodeficient patients Treatment Many immunocompetent patients who have cryptosporidiosis have self-limited disease and do not require therapy A 3-day course of nitazoxanide: To reduce the duration and transmission of diarrhea in children older than 1 year of age No swimming pool for at least 2 weeks after the diarrhea stopped Toxoplasma gondii (Toxoplasmosis) Background Obligate intracellular protozoa Mode of transmission Ingestion of contaminated raw or uncooked meat Cats excreta Organ transplants Transplacental to fetus causes congenital toxoplasmosis (see chapter Fetus and Newborns) Clinical presentation Most cases are asymptomatic Fever Malaise Rash Myalgia Cervical lymphadenopathy (most common sign) Brain abscess (test for HIV) Chorioretinitis usually present years later (mostly congenital) Diagnosis Head CT: ring-enhanced lesion Toxoplasma IgM antibodies PCR Treatment Pyrimethamine plus sulfadiazine and folic acid Lifelong therapy in HIV patients Pneumocystis jiroveci (Carinii) Background Unicellular fungi that do not respond to antifungal treatment Mode of transmission is unknown Commonly seen in immunocompromised patients, e.g., HIV patients Clinical presentation Subacute diffuse pneumonitis Dyspnea Tachycardia Oxygen desaturation Nonproductive cough Fever Diagnosis Chest radiography Bilateral diffuse interstitial disease Low CD4 Bronchoalveolar lavage Lung biopsy Treatment TMP-SMX IV pentamidine in severe cases Prophylaxis in immunocompromised patients TMP-SMX Plasmodium (Malaria) Background Intracellular protozoa Transmitted by mosquito bites in endemic area, e.g., south Africa Plasmodium falciparum Most severe Symptoms develop within a month from returning from endemic area Most common cause of congenital malaria Complications Cerebral malaria Pulmonary edema Severe anemia Renal failure Shock Treatment Chloroquine sensitive ◦ Chloroquine Chloroquine resistant: ◦ Quinine plus doxycycline or clindamycin ◦ Or atovaquone-proguanil ◦ Or mefloquine Severe cases ◦ Quinidine gluconate IV plus doxycycline or clindamycin Plasmodium malariae , P. vivax , and P. ovale Periodicity of symptoms Nephrotic syndrome- P. malariae (most benign form) Hypersplenism and splenic rupture- P. vivax and P. ovale Treatment ◦ Chloroquine plus primaquine for P. vivax , and P. ovale ◦ Chloroquine phosphate for P. malaria Clinical presentation of malaria History of travelling to endemic areas in the past years Paroxysmal fever, sweat and rigors Pallor and jaundice Headache and myalgia Abdominal pain Vomiting and diarrhea In severe cases Change in mental status Hepatosplenomegaly Anemia Thrombocytopenia Hypotension Hypoglycemia Hyperkalemia Respiratory distress Diagnosis RBCs smear Prevention Travelling to chloroquine resistant areas, e.g., South Africa Atovaquone-proguanil 2 weeks before and 4 weeks after or Doxycycline (> 8 years old) Mefloquine (safe for pregnant) Travelling to chloroquine sensitive areas, e.g., South America Chloroquine 2 weeks before and 4 weeks after or Atovaquone-proguanil or Mefloquine Helminthic Organisms Enterobius vermicularis (Pinworm) Mode of transmission From one person to another via feco-oral route Eggs survive up to 3 weeks and are ingested from finger nails, bedding, and toys Autoinfection Clinical presentation Anal and vulvar itching (more at night) Enuresis Diagnosis Visualizing the adult worm at night on the perineum Transparent tape collected over three consecutive mornings under microscope low power Treatment Albendazole Ascaris lumbricoides (Ascariasis) Mode of transmission Ingestion of eggs from contaminated soil (feco-oral) Clinical presentation Most patient are asymptomatic Nonspecific abdominal pain or discomfort Intestinal obstruction (large number of worms) Due to larvae migration to the liver and lung: Obstructive jaundice Peritonitis Cough (Loeffler's syndrome) Diagnosis Seeing the ova on microscopic stool examination Seeing the adult worm itself Treatment Albendazole or pyrantel pamoate Necator americanus (Hookworm) or Ancylostoma duodenale Background Found in rural, tropical and subtropical locales Mode of transmission Skin penetration of larvae from soil contaminated by human feces Can cause itchiness and burning sensation May be ingested as well Can cause pharyngitis and gastroenteritis Clinical presentation (blood sucker worm from the intestine) Failure to thrive Short stature Anemia due to chronic blood loss Diagnosis Finding the eggs stool (may take 5–10 weeks after infection) Treatment Albendazole Trichuriasis (Whipworms) It is due to infection of large intestine with Trichuris trichiura . More common in the Southern USA. Transmitted to human by ingesting eggs. Usually asymptomatic if only few worms. Can cause fever, abdominal pain, weight loss, blood in stool and rectal prolapse. Presence of eggs in stool is diagnostic. Treatment is mebendazole. Trichinosis (Trichinella spiralis) Trichinella spiralis is usually found in pork. Symptoms depend on the worm location. After ingestion the eggs hatch, larvae invade the duodenum, and causes abdominal symptoms. Larvae penetrate, reach bloodstream, end in muscular tissue and causes muscle pain. If the larvae reach the heart can cause myocarditis. Ocular involvement; presence of chemosis, periorbital edema, and eosinophilia usually suggest the diagnosis. Diagnosis is confirmed by rising titers. Strongyloides stercoralis S. stercoralis is common in certain areas of the USA. In the USA this infection is common in Kentucky and Tennessee. It is the only helminthic organism replicates in the body with autoinfection, and the infection may persist for decades. Can cause pulmonary symptoms with eosinophilia and GI symptoms as well. It is potentially fetal in immunosuppressed patients. Diagnosis of serial stool studies for larvae not the eggs. Treatment is ivermectin or thiabendazole. Toxocariasis Toxocara canis and Toxocara catis can cause visceral larva migrans. It is transmitted to human by ingesting soil contaminated with dog or cat excreta. In human larva do not develop into adult worms but rather migrate through the host tissue; causing eosinophilia. Treatment is albendazole or mebendazole. Cestodes (Platyhelminthes) Platyhelminthes include cestodes (tapeworms) and trematodes (flukes). Cestodes are flatworms (tapeworms).The pork tapeworm. Taenia solium, present in two different ways. If the cysticerci are ingested, taeniasis develops and tape worm grows in the intestine. If contaminated food with eggs is ingested, the patient will develop cysticercosis. Cysticerci go in CNS and the eyes and do nothing until they die. Diagnosis of neurocysticercosis must be considered in the patients with new onset seizures and history of travelling to or immigration from Mexico, Central or South America or who is a household from these areas. Trematodes (Platyhelminthes) Trematodes or flukes. Clonorchis sinensis is the Chinese liver fluke. Schistosoma haematobium infects the bladder and cause urinary symptoms . Schistosoma mansoni is a fluke found in Africa, the Middle East, and South America . Schistosoma japonicum is found in Asia . Most serious complications of Schistosomiasis is cirrhosis with esophageal varices. Treatment is praziquantel Fever Without Focus Febrile Neonate Background It is difficult to distinguish between a serious bacterial infection and self limited viral illness in this age group. Neonates who have fever and do not appear ill have a 7 % risk of having a serious bacterial infection. Serious bacterial infections include occult bacteremia, meningitis, pneumonia , osteomyelitis, septic arthritis, enteritis, and UTI . Late onset neonatal bacterial diseases, e.g., group B Streptococci , E. coli , and Listeria monocytogenes and perinatal herpes (HSV) infection. If the neonate has fever recorded at home by reliable parents, the patient should be treated as febrile neonate. If excessive clothing and blanket falsely elevating the temperature, the excessive covering should be removed and retake the temperature in 15–30 min. Management All febrile neonates must be hospitalized. Full sepsis evaluation including blood, urine, CSF should be cultured. Child should receive empirical antibiotics such as cefotaxime and ampicillin. Acyclovir should be included if HSV infection is suspected. CSF studies should include cell count, glucose, and protein level, Gram stain, cultures; HSV, and enterovirus PCR should be considered. Stool culture and CXR may be included. Fever in 1–3 Months Infants Background Large majority of the children with fever without localizing signs in 1–3 months age group likely viral syndrome. Most viral diseases has distinct seasonal pattern unlike bacteria, e.g., respiratory syncytial virus, and influenza more common during winter and enterovirus infection more common during summer and fall. Management Ill appearing (toxic) febrile infants ≤ 3 months: Require prompt hospitalization, immediate parenteral antibiotics after blood and CSF cultures are obtained. Well appearing infants 1–3 months who is previously healthy with no evidence of focus of infection: WBCs count of 5000–15,000 cells/µL, an absolute band count of ≤ 1500 cells/µL, and normal urinalysis, and negative culture (blood and urine) results are unlikely to have a serious bacterial infection. The decision to obtain CSF studies in the well appearing 1–3 months old infant depends on the decision to administer empirical antibiotics. If close observation without antibiotics planned, a lumbar puncture may be deferred. Fever in 3–36 Months of Age Background Approximately 30 % of febrile children in the 3–36 months age group have no localizing signs of infection. Viral infections are the cause of the vast majority of fevers in this population. Risk factors indicating probability of occult bacteremia Temperature ≥ 39 °C, WBC count ≥ 15,000/µL, elevated absolute neutrophil count, bands, ESR and CRP. The risk of bacteremia and/or pneumonia or pyelonephritis, among infants 3–36 months of age increases as temperature (specially > 40 °C) and WBCs count (specially > 25,000) increases. Management Toxic appearing febrile children 3–36 months of age who do not have focal infection should be hospitalized, and prompt institution of parenteral antibiotics after blood, urine and CSF cultures are obtained (full sepsis evaluation). For nontoxic appearing infants who have temperature   6 years of age often have respiratory or genitourinary tract infection, localized infection (abscess, osteomyelitis), JIA, or rarely leukemia. Adolescent patients more likely to have TB, inflammatory bowel disease , autoimmune process or lymphoma in addition to the causes of FUO in younger children. Exposure to wild or domestic animals, and zoonotic infection. History of pica should be elicited; ingestion of dirt is a particularly important due to infection with Toxocara canis or Toxoplasma gondii . Physical examination is essential to find any physical clues to underlying diagnosis, e.g., lymphadenopathy, rash, joint swelling, etc. Laboratory it is determined on case-by-case bases. ESR > 30 mm/h indicates inflammation and need further evaluation. ESR > 100 mm/h suggests tuberculosis, Kawasaki disease, malignancy or autoimmune disease. Low ESR does not eliminate the possibility of infection. CRP is another acute phase reactant that is elevated and returns to normal more rapidly than ESR. Cultures, serologic studies, imaging studies and biopsies depending on each case. Treatment The ultimate treatment of FUO is tailored to the underlying diagnosis. Empirical trials of antimicrobial agents may be dangerous and obscure the diagnosis of infective endocarditis, meningitis, parameningeal infection, and osteomyelitis. Antipyretics for fever and relief of symptoms. Enterobius vermicularis (Pinworm) Mode of transmission From one person to another via feco-oral route Eggs survive up to 3 weeks and are ingested from finger nails, bedding, and toys Autoinfection Clinical presentation Anal and vulvar itching (more at night) Enuresis Diagnosis Visualizing the adult worm at night on the perineum Transparent tape collected over three consecutive mornings under microscope low power Treatment Albendazole Ascaris lumbricoides (Ascariasis) Mode of transmission Ingestion of eggs from contaminated soil (feco-oral) Clinical presentation Most patient are asymptomatic Nonspecific abdominal pain or discomfort Intestinal obstruction (large number of worms) Due to larvae migration to the liver and lung: Obstructive jaundice Peritonitis Cough (Loeffler's syndrome) Diagnosis Seeing the ova on microscopic stool examination Seeing the adult worm itself Treatment Albendazole or pyrantel pamoate Necator americanus (Hookworm) or Ancylostoma duodenale Background Found in rural, tropical and subtropical locales Mode of transmission Skin penetration of larvae from soil contaminated by human feces Can cause itchiness and burning sensation May be ingested as well Can cause pharyngitis and gastroenteritis Clinical presentation (blood sucker worm from the intestine) Failure to thrive Short stature Anemia due to chronic blood loss Diagnosis Finding the eggs stool (may take 5–10 weeks after infection) Treatment Albendazole Trichuriasis (Whipworms) It is due to infection of large intestine with Trichuris trichiura . More common in the Southern USA. Transmitted to human by ingesting eggs. Usually asymptomatic if only few worms. Can cause fever, abdominal pain, weight loss, blood in stool and rectal prolapse. Presence of eggs in stool is diagnostic. Treatment is mebendazole. Trichinosis (Trichinella spiralis) Trichinella spiralis is usually found in pork. Symptoms depend on the worm location. After ingestion the eggs hatch, larvae invade the duodenum, and causes abdominal symptoms. Larvae penetrate, reach bloodstream, end in muscular tissue and causes muscle pain. If the larvae reach the heart can cause myocarditis. Ocular involvement; presence of chemosis, periorbital edema, and eosinophilia usually suggest the diagnosis. Diagnosis is confirmed by rising titers. Strongyloides stercoralis S. stercoralis is common in certain areas of the USA. In the USA this infection is common in Kentucky and Tennessee. It is the only helminthic organism replicates in the body with autoinfection, and the infection may persist for decades. Can cause pulmonary symptoms with eosinophilia and GI symptoms as well. It is potentially fetal in immunosuppressed patients. Diagnosis of serial stool studies for larvae not the eggs. Treatment is ivermectin or thiabendazole. Toxocariasis Toxocara canis and Toxocara catis can cause visceral larva migrans. It is transmitted to human by ingesting soil contaminated with dog or cat excreta. In human larva do not develop into adult worms but rather migrate through the host tissue; causing eosinophilia. Treatment is albendazole or mebendazole. Cestodes (Platyhelminthes) Platyhelminthes include cestodes (tapeworms) and trematodes (flukes). Cestodes are flatworms (tapeworms).The pork tapeworm. Taenia solium, present in two different ways. If the cysticerci are ingested, taeniasis develops and tape worm grows in the intestine. If contaminated food with eggs is ingested, the patient will develop cysticercosis. Cysticerci go in CNS and the eyes and do nothing until they die. Diagnosis of neurocysticercosis must be considered in the patients with new onset seizures and history of travelling to or immigration from Mexico, Central or South America or who is a household from these areas. Trematodes (Platyhelminthes) Trematodes or flukes. Clonorchis sinensis is the Chinese liver fluke. Schistosoma haematobium infects the bladder and cause urinary symptoms . Schistosoma mansoni is a fluke found in Africa, the Middle East, and South America . Schistosoma japonicum is found in Asia . Most serious complications of Schistosomiasis is cirrhosis with esophageal varices. Treatment is praziquantel Fever Without Focus Febrile Neonate Background It is difficult to distinguish between a serious bacterial infection and self limited viral illness in this age group. Neonates who have fever and do not appear ill have a 7 % risk of having a serious bacterial infection. Serious bacterial infections include occult bacteremia, meningitis, pneumonia , osteomyelitis, septic arthritis, enteritis, and UTI . Late onset neonatal bacterial diseases, e.g., group B Streptococci , E. coli , and Listeria monocytogenes and perinatal herpes (HSV) infection. If the neonate has fever recorded at home by reliable parents, the patient should be treated as febrile neonate. If excessive clothing and blanket falsely elevating the temperature, the excessive covering should be removed and retake the temperature in 15–30 min. Management All febrile neonates must be hospitalized. Full sepsis evaluation including blood, urine, CSF should be cultured. Child should receive empirical antibiotics such as cefotaxime and ampicillin. Acyclovir should be included if HSV infection is suspected. CSF studies should include cell count, glucose, and protein level, Gram stain, cultures; HSV, and enterovirus PCR should be considered. Stool culture and CXR may be included. Fever in 1–3 Months Infants Background Large majority of the children with fever without localizing signs in 1–3 months age group likely viral syndrome. Most viral diseases has distinct seasonal pattern unlike bacteria, e.g., respiratory syncytial virus, and influenza more common during winter and enterovirus infection more common during summer and fall. Management Ill appearing (toxic) febrile infants ≤ 3 months: Require prompt hospitalization, immediate parenteral antibiotics after blood and CSF cultures are obtained. Well appearing infants 1–3 months who is previously healthy with no evidence of focus of infection: WBCs count of 5000–15,000 cells/µL, an absolute band count of ≤ 1500 cells/µL, and normal urinalysis, and negative culture (blood and urine) results are unlikely to have a serious bacterial infection. The decision to obtain CSF studies in the well appearing 1–3 months old infant depends on the decision to administer empirical antibiotics. If close observation without antibiotics planned, a lumbar puncture may be deferred. Fever in 3–36 Months of Age Background Approximately 30 % of febrile children in the 3–36 months age group have no localizing signs of infection. Viral infections are the cause of the vast majority of fevers in this population. Risk factors indicating probability of occult bacteremia Temperature ≥ 39 °C, WBC count ≥ 15,000/µL, elevated absolute neutrophil count, bands, ESR and CRP. The risk of bacteremia and/or pneumonia or pyelonephritis, among infants 3–36 months of age increases as temperature (specially > 40 °C) and WBCs count (specially > 25,000) increases. Management Toxic appearing febrile children 3–36 months of age who do not have focal infection should be hospitalized, and prompt institution of parenteral antibiotics after blood, urine and CSF cultures are obtained (full sepsis evaluation). For nontoxic appearing infants who have temperature   6 years of age often have respiratory or genitourinary tract infection, localized infection (abscess, osteomyelitis), JIA, or rarely leukemia. Adolescent patients more likely to have TB, inflammatory bowel disease , autoimmune process or lymphoma in addition to the causes of FUO in younger children. Exposure to wild or domestic animals, and zoonotic infection. History of pica should be elicited; ingestion of dirt is a particularly important due to infection with Toxocara canis or Toxoplasma gondii . Physical examination is essential to find any physical clues to underlying diagnosis, e.g., lymphadenopathy, rash, joint swelling, etc. Laboratory it is determined on case-by-case bases. ESR > 30 mm/h indicates inflammation and need further evaluation. ESR > 100 mm/h suggests tuberculosis, Kawasaki disease, malignancy or autoimmune disease. Low ESR does not eliminate the possibility of infection. CRP is another acute phase reactant that is elevated and returns to normal more rapidly than ESR. Cultures, serologic studies, imaging studies and biopsies depending on each case. Treatment The ultimate treatment of FUO is tailored to the underlying diagnosis. Empirical trials of antimicrobial agents may be dangerous and obscure the diagnosis of infective endocarditis, meningitis, parameningeal infection, and osteomyelitis. Antipyretics for fever and relief of symptoms. Febrile Neonate Background It is difficult to distinguish between a serious bacterial infection and self limited viral illness in this age group. Neonates who have fever and do not appear ill have a 7 % risk of having a serious bacterial infection. Serious bacterial infections include occult bacteremia, meningitis, pneumonia , osteomyelitis, septic arthritis, enteritis, and UTI . Late onset neonatal bacterial diseases, e.g., group B Streptococci , E. coli , and Listeria monocytogenes and perinatal herpes (HSV) infection. If the neonate has fever recorded at home by reliable parents, the patient should be treated as febrile neonate. If excessive clothing and blanket falsely elevating the temperature, the excessive covering should be removed and retake the temperature in 15–30 min. Management All febrile neonates must be hospitalized. Full sepsis evaluation including blood, urine, CSF should be cultured. Child should receive empirical antibiotics such as cefotaxime and ampicillin. Acyclovir should be included if HSV infection is suspected. CSF studies should include cell count, glucose, and protein level, Gram stain, cultures; HSV, and enterovirus PCR should be considered. Stool culture and CXR may be included. Fever in 1–3 Months Infants Background Large majority of the children with fever without localizing signs in 1–3 months age group likely viral syndrome. Most viral diseases has distinct seasonal pattern unlike bacteria, e.g., respiratory syncytial virus, and influenza more common during winter and enterovirus infection more common during summer and fall. Management Ill appearing (toxic) febrile infants ≤ 3 months: Require prompt hospitalization, immediate parenteral antibiotics after blood and CSF cultures are obtained. Well appearing infants 1–3 months who is previously healthy with no evidence of focus of infection: WBCs count of 5000–15,000 cells/µL, an absolute band count of ≤ 1500 cells/µL, and normal urinalysis, and negative culture (blood and urine) results are unlikely to have a serious bacterial infection. The decision to obtain CSF studies in the well appearing 1–3 months old infant depends on the decision to administer empirical antibiotics. If close observation without antibiotics planned, a lumbar puncture may be deferred. Fever in 3–36 Months of Age Background Approximately 30 % of febrile children in the 3–36 months age group have no localizing signs of infection. Viral infections are the cause of the vast majority of fevers in this population. Risk factors indicating probability of occult bacteremia Temperature ≥ 39 °C, WBC count ≥ 15,000/µL, elevated absolute neutrophil count, bands, ESR and CRP. The risk of bacteremia and/or pneumonia or pyelonephritis, among infants 3–36 months of age increases as temperature (specially > 40 °C) and WBCs count (specially > 25,000) increases. Management Toxic appearing febrile children 3–36 months of age who do not have focal infection should be hospitalized, and prompt institution of parenteral antibiotics after blood, urine and CSF cultures are obtained (full sepsis evaluation). For nontoxic appearing infants who have temperature   6 years of age often have respiratory or genitourinary tract infection, localized infection (abscess, osteomyelitis), JIA, or rarely leukemia. Adolescent patients more likely to have TB, inflammatory bowel disease , autoimmune process or lymphoma in addition to the causes of FUO in younger children. Exposure to wild or domestic animals, and zoonotic infection. History of pica should be elicited; ingestion of dirt is a particularly important due to infection with Toxocara canis or Toxoplasma gondii . Physical examination is essential to find any physical clues to underlying diagnosis, e.g., lymphadenopathy, rash, joint swelling, etc. Laboratory it is determined on case-by-case bases. ESR > 30 mm/h indicates inflammation and need further evaluation. ESR > 100 mm/h suggests tuberculosis, Kawasaki disease, malignancy or autoimmune disease. Low ESR does not eliminate the possibility of infection. CRP is another acute phase reactant that is elevated and returns to normal more rapidly than ESR. Cultures, serologic studies, imaging studies and biopsies depending on each case. Treatment The ultimate treatment of FUO is tailored to the underlying diagnosis. Empirical trials of antimicrobial agents may be dangerous and obscure the diagnosis of infective endocarditis, meningitis, parameningeal infection, and osteomyelitis. Antipyretics for fever and relief of symptoms. Central Nervous System (CNS) Infections Encephalitis Definition Inflammation of the brain Causes Viral, e.g., West Nile virus and herpesvirus (most common) Bacteria, e.g., Mycoplasma , tertiary syphilis Noninfectious, e.g., autoimmune Prion protein Parasitic Fungal Acute cerebellar ataxia Ataxia Nystagmus Cerebellar dysarthria Epidemiology WNV remains the most commonly encountered arboviral encephalitis agent. California encephalitis viruses have the greatest proportion of pediatric symptomatic infections (88 % of cases). Eastern equine encephalitis has the highest overall mortality rate of 42 %. The importance of local epidemiological information and seasonality cannot be ignored. Enteroviruses are most often seen in spring and summer. Arthropod-borne illnesses, in the summer and fall. Clinical presentation Altered mental status Seizures Weakness Sensory disturbances Nonepileptic movement disorders Young children in absence of identifiable cause may present with: Somnolence Disinterest in feeding Weak suck and irritability Loss of head control Abnormal eye movements Further clinical clues: Fever (either acutely or in the 1–4 week interval before the onset of symptoms) Meningeal irritation Any child presenting with uncharacteristic behavior that is persistent and disproportionate to environmental and situational factors Initial evaluation of the patient include: Seasonal presentation. History of immunosuppression. Travel history. Recent local epidemiological information. Presence of focal neurologic symptoms or deficits. Investigation Complete blood count. Complete metabolic panel. Urinalysis. MRI or CT scan for intracranial pressure. EEG. Enteroviral infections can produce a sepsis-like syndrome with more remarkable hematologic abnormalities. Neonatal HSV infections sometimes produce hepatic function abnormalities and disseminated intravascular coagulation. SIADH. Lumbar puncture if normal pressure. Cerebrospinal spinal fluid study: The lumbar puncture is the single most utilized test for the diagnosis of encephalitis . Increased opening pressure. Normal or elevated protein concentration. Normal glucose level. Pleocytosis, polymorphonuclear leukocytes and then converts to lymphocytic in many viral cases. Monocytic, predominance may show with progression of the disease. Hemorrhagic pleocytosis with HSV. Atypical lymphocytes with EBV. Mononuclear leukocytes with echovirus or varicella-zoster infection. PCR amplification of viral DNA. Pleocytosis tends to be less dramatic in parainfectious encephalitis or acute cerebellar ataxia. Fourfold rise in titer, especially immunoglobulin M, against a suspected agent is most often considered diagnostic. Intravenous acyclovir while waiting for lumbar puncture, or while waiting for laboratory results, including HSV PCR. Intracranial hypertension conservative measures Head elevation Hyperventilation Fluid restriction Mannitol is used on a limited basis Treatment of seizure Benzodiazepines (midazolam, lorazepam, diazepam) in the beginning followed by loading dose of fosphenytoin, or Phenobarbital. Meningitis Neonatal Streptococcal Meningitis GBS remains the predominant neonatal meningitis pathogen . Early-onset disease , infants typically manifest with signs suggestive of sepsis, often with pneumonia, but less commonly with meningitis. Late-onset disease ; the typical infant who has late-onset disease is 3–4 weeks of age and presents with meningitis or bacteremia. Neonatal Gram-negative Meningitis Gram-negative bacillary meningitis is rare and E. coli being the most commonly isolated pathogen. Other gram-negative neonatal meningitis pathogens such as Citrobacter koseri , Enterobacter sakazakii , and Serratia marcescens . Neonatal Herpes Simplex (HSV) Infection HSV in the newborn can present as isolated skin or mucous membrane lesions, encephalitis, or a disseminated process. HSV infection occurs most commonly in infants born to mothers who have active primary infection. Frequently no maternal history or clinical evidence is available to alert the practitioner to this diagnosis. The incubation period is 2 days to 2 weeks, and most infants who develop HSV CNS infection are 2–3 weeks of age . Neonatal Listeria meningitis Common sources: Unpasteurized milk Soft cheeses Prepared ready-to-eat meats Undercooked poultry Unwashed raw vegetables Can precipitate abortion and preterm delivery. Septic appearance in the neonate is typical in cases of early onset. Papular truncal rash has been identified. S. pneumoniae Pneumococcus is the leading pathogen causing bacterial meningitis in infants and young children in developed countries. N. meningitidis Meningococcal disease generally occurs in otherwise healthy individuals and often has a fulminant presentation with high fatality rates. Aseptic meningitis Enteroviruses virus infection is the most common . B. burgdorferi in mid-Atlantic states. Vasculitis in the setting of systemic lupus erythematosus or Kawasaki disease. Drug-induced such as ibuprofen, and IV immunoglobulin Other Causes of Meningitis M. tuberculosis B. burgdorferi Rickettsia rickettsii Clinical Manifestations of Meningitis Infants younger than 1 month of age who have viral or bacterial meningitis Fever Hypothermia Lethargy Irritability Poor feeding Signs and symptoms of increased intracranial pressure and meningeal inflammation Vomiting Apnea Seizures also can occur Older children and adolescents often experience Malaise Myalgia Headache Photophobia Neck stiffness Anorexia Nausea. Physical Examination Altered levels of consciousness can present as irritability, somnolence, lethargy, or coma Intracranial pressure include: Papilledema. Diplopia. Unilateral or bilateral dilated pupil. Poorly reactive pupils. Bulging fontanelle in infants. Head circumference always should be obtained, especially in those who have an open fontanelle. Meningismus is suggestive of meningeal irritation. Kernig sign: The patient lies supine and the thigh is flexed at a right angle to the trunk. If knee extension from this position elicits pain, the Kernig sign is positive. Brudzinski sign: The patient lies supine and flexes his or her neck. A positive sign occurs if the patient also reflexively flexes the lower extremities, typically at the knees. Absence of Kernig and Brudzinski signs does not exclude meningitis. Exanthems typical for enterovirus, borreliosis (erythema migrans), and invasive meningococcal or pneumococcal disease (petechiae and purpura) may be present. Diagnosis All children who are suspected of having meningitis should have their CSF examined unless lumbar puncture is contraindicated. Contraindications of lumbar puncture include: Focal neurologic deficits. Signs of increased intracranial pressure. Uncorrected coagulopathy. Cardiopulmonary compromise. Computed tomography (CT) scan is performed before lumbar puncture if any signs of ICP. CSF finding of Bacterial meningitis (Table 5 ). Glucose concentration usually is less than one half of the measured serum value. Protein value often is greater than 1.0 g/dL (10 g/L). WBC often greater than 1.0 × 10 3 /mcL (1.0 × 10 9 /L), with a predominance of polymorphonuclear leukocytes. Gram stain is extremely helpful if positive. CSF culture remains the gold standard for diagnosing bacterial meningitis. CSF finding viral meningitis WBC count of 0.05–0.5 × 10 3 /mcL (0.05–0.5 × 10 9 /L). Neutrophil predominance is common early in the course of infection, shifting to lymphocytic predominance quickly during the illness. Glucose and protein concentrations frequently are normal, although the protein value can be slightly elevated. Gram stain is universally negative. In cases of enteroviral meningitis, enteroviral PCR can confirm the diagnosis. Tuberculous meningitis, epidemiologic clue, high protein and lymphocytosis. SIADH and hyponatremia commonly occur in bacterial meningitis. Leukopenia, thrombocytopenia, and coagulopathy may be present in meningococcal and rickettsial infection. Table 5 Cerebrospinal fluid analysis (Adapted from Wubbel L, McCracken GH. Pediatr Rev. 1998) Glucose (mg/dL) Protein (g/L) White blood cell (× 10 3 /mcL) Differential count Gram stain Healthy newborn 30–120 30–150 100 > 1.0 = 1000 > 50 PMNs Often > 90 % Enteroviral meningitis > 1/2 serum 40–60 0.05–0.5 > 50 % PMNs early 48 h Negative Lyme meningitis > 1/2 serum 0.05–0.5 Predominance of lymphocytes and monocytes Negative Tuberculous meningitis 100 0.05–0.5 Predominance of lymphocyte Negative This table is just a guide and should not be used in isolation without clinical correlation because overlap between values in each of these categories is significant PMN polymorphonuclear leukocytes. Management Therapy should not be delayed if CNS infection is suspected. Appropriate antimicrobials are required in bacterial meningitis, HSV encephalitis, Lyme meningitis, tuberculous meningitis, and rickettsial infection, and in all cases, timely diagnosis and correct antimicrobial choice are critical. If the practitioner cannot perform a lumbar puncture or there are contraindications to CSF examination, a blood culture should be obtained and antibiotics administered promptly. Drug choice and duration For infants Ampicillin (300 mg/kg/day divided every 6 h) and cefotaxime (200–300 mg/kg/day divided every 6 h) is appropriate. Acyclovir (60 mg/kg/day divided every 8 h) should be added if HSV infection is a concern. Vancomycin (60 mg/kg/day given every 6 h) should be added, if the Gram stain suggests pneumococcus. Children older than 2 months of age Vancomycin (60 mg/kg/day divided every 6 h) plus ceftriaxone (100 mg/kg/day given in one dose or divided into two doses) or cefotaxime (200–300 mg/kg/day divided every 6 h) should be used for empiric coverage. Once culture and susceptibility data are available, definitive therapy can be selected. HSV meningitis Neonatal HSV CNS infection typically is treated with IV acyclovir (60 mg/kg/day divided every 8 h) for 21 days. The dosing for non-neonates is 30 mg/kg/day divided every 8 h IV for 14–21 days. Follow-up CSF HSV DNA PCR should be evaluated at day 21 and the course of therapy extended if the result still is positive. Corticosteroids in bacterial meningitis Adjunctive treatment has reduced rates of mortality, severe hearing loss, and neurologic sequelae significantly in adults who have community-acquired bacterial meningitis. For children beyond the neonatal age groups, available data suggest that the use of adjunctive corticosteroids may be beneficial for Hib meningitis and could be considered in cases of pneumococcal meningitis. The dose of dexamethasone for bacterial meningitis is 0.6 mg/kg/day divided into four doses and administered IV for 4 days. The first dose should be given before or concurrently with antibiotics. Care of the child exposed to meningitis Meningococcal and Hib disease create an increased risk for secondary infection in contacts. Rifampin generally is the drug of choice for chemoprophylaxis in children. Prognosis Intellectual deficits (intelligence quotient   100 > 1.0 = 1000 > 50 PMNs Often > 90 % Enteroviral meningitis > 1/2 serum 40–60 0.05–0.5 > 50 % PMNs early 48 h Negative Lyme meningitis > 1/2 serum 0.05–0.5 Predominance of lymphocytes and monocytes Negative Tuberculous meningitis 100 0.05–0.5 Predominance of lymphocyte Negative This table is just a guide and should not be used in isolation without clinical correlation because overlap between values in each of these categories is significant PMN polymorphonuclear leukocytes. Management Therapy should not be delayed if CNS infection is suspected. Appropriate antimicrobials are required in bacterial meningitis, HSV encephalitis, Lyme meningitis, tuberculous meningitis, and rickettsial infection, and in all cases, timely diagnosis and correct antimicrobial choice are critical. If the practitioner cannot perform a lumbar puncture or there are contraindications to CSF examination, a blood culture should be obtained and antibiotics administered promptly. Drug choice and duration For infants Ampicillin (300 mg/kg/day divided every 6 h) and cefotaxime (200–300 mg/kg/day divided every 6 h) is appropriate. Acyclovir (60 mg/kg/day divided every 8 h) should be added if HSV infection is a concern. Vancomycin (60 mg/kg/day given every 6 h) should be added, if the Gram stain suggests pneumococcus. Children older than 2 months of age Vancomycin (60 mg/kg/day divided every 6 h) plus ceftriaxone (100 mg/kg/day given in one dose or divided into two doses) or cefotaxime (200–300 mg/kg/day divided every 6 h) should be used for empiric coverage. Once culture and susceptibility data are available, definitive therapy can be selected. HSV meningitis Neonatal HSV CNS infection typically is treated with IV acyclovir (60 mg/kg/day divided every 8 h) for 21 days. The dosing for non-neonates is 30 mg/kg/day divided every 8 h IV for 14–21 days. Follow-up CSF HSV DNA PCR should be evaluated at day 21 and the course of therapy extended if the result still is positive. Corticosteroids in bacterial meningitis Adjunctive treatment has reduced rates of mortality, severe hearing loss, and neurologic sequelae significantly in adults who have community-acquired bacterial meningitis. For children beyond the neonatal age groups, available data suggest that the use of adjunctive corticosteroids may be beneficial for Hib meningitis and could be considered in cases of pneumococcal meningitis. The dose of dexamethasone for bacterial meningitis is 0.6 mg/kg/day divided into four doses and administered IV for 4 days. The first dose should be given before or concurrently with antibiotics. Care of the child exposed to meningitis Meningococcal and Hib disease create an increased risk for secondary infection in contacts. Rifampin generally is the drug of choice for chemoprophylaxis in children. Prognosis Intellectual deficits (intelligence quotient   100 > 1.0 = 1000 > 50 PMNs Often > 90 % Enteroviral meningitis > 1/2 serum 40–60 0.05–0.5 > 50 % PMNs early 48 h Negative Lyme meningitis > 1/2 serum 0.05–0.5 Predominance of lymphocytes and monocytes Negative Tuberculous meningitis 100 0.05–0.5 Predominance of lymphocyte Negative This table is just a guide and should not be used in isolation without clinical correlation because overlap between values in each of these categories is significant PMN polymorphonuclear leukocytes. Management Therapy should not be delayed if CNS infection is suspected. Appropriate antimicrobials are required in bacterial meningitis, HSV encephalitis, Lyme meningitis, tuberculous meningitis, and rickettsial infection, and in all cases, timely diagnosis and correct antimicrobial choice are critical. If the practitioner cannot perform a lumbar puncture or there are contraindications to CSF examination, a blood culture should be obtained and antibiotics administered promptly. Drug choice and duration For infants Ampicillin (300 mg/kg/day divided every 6 h) and cefotaxime (200–300 mg/kg/day divided every 6 h) is appropriate. Acyclovir (60 mg/kg/day divided every 8 h) should be added if HSV infection is a concern. Vancomycin (60 mg/kg/day given every 6 h) should be added, if the Gram stain suggests pneumococcus. Children older than 2 months of age Vancomycin (60 mg/kg/day divided every 6 h) plus ceftriaxone (100 mg/kg/day given in one dose or divided into two doses) or cefotaxime (200–300 mg/kg/day divided every 6 h) should be used for empiric coverage. Once culture and susceptibility data are available, definitive therapy can be selected. HSV meningitis Neonatal HSV CNS infection typically is treated with IV acyclovir (60 mg/kg/day divided every 8 h) for 21 days. The dosing for non-neonates is 30 mg/kg/day divided every 8 h IV for 14–21 days. Follow-up CSF HSV DNA PCR should be evaluated at day 21 and the course of therapy extended if the result still is positive. Corticosteroids in bacterial meningitis Adjunctive treatment has reduced rates of mortality, severe hearing loss, and neurologic sequelae significantly in adults who have community-acquired bacterial meningitis. For children beyond the neonatal age groups, available data suggest that the use of adjunctive corticosteroids may be beneficial for Hib meningitis and could be considered in cases of pneumococcal meningitis. The dose of dexamethasone for bacterial meningitis is 0.6 mg/kg/day divided into four doses and administered IV for 4 days. The first dose should be given before or concurrently with antibiotics. Care of the child exposed to meningitis Meningococcal and Hib disease create an increased risk for secondary infection in contacts. Rifampin generally is the drug of choice for chemoprophylaxis in children. Prognosis Intellectual deficits (intelligence quotient   < 70), hydrocephalus, spasticity, blindness, and severe hearing loss are the most common sequelae. Hearing loss occurs in approximately 30 % of patients, can be unilateral or bilateral, and is more common in pneumococcal than meningococcal meningitis. Brain Abscess Causes of brain abscess Chronic otitis media Paranasal sinus infection Mastoiditis Head injury S. aureus Metastatic spread, e.g., endocarditis Right-to-left cardiac or pulmonary shunts, especially in the presence of cyanotic congenital heart disease Clinical presentation Headache (most common) May be throbbing Worsen with changes in posture or Valsalva maneuver Drowsiness Confusion Vomiting Drowsiness, and coma Hemiparesis Papilledema Frontal lobe abscesses Apathy, memory deficits Personality change Mental slowing Cerebellar abscesses Nystagmus Defective conjugate eye movements to that side Ataxia Hypotonia Laboratory diagnosis Little in the laboratory investigation of patients who have brain abscesses is specific to the diagnosis except for culture of the purulent material and antibiotic sensitivity of the responsible organism. Neuroimaging CT scan of the brain Ill-defined Low-density change within the parenchyma Enhancement occurs following administration of contrast material Classic ring-enhancing lesion with surrounding edema Calcification is common in abscesses in neonates Magnetic resonance imaging (MRI) Antimicrobial therapy For abscesses arising as a result of sinusitis in which streptococci are the most likely organisms, penicillin or cefotaxime and metronidazole. Chronic otitis media or mastoiditis often is associated with P. aeruginosa and Enterobacteriaceae, antibiotics to treat abscesses secondary to these infections should include penicillin, metronidazole, and a third-generation cephalosporin. Metastatic abscesses require a regimen based on the likely site of primary infection. S. aureus commonly is isolated in abscess following trauma. Surgical intervention Provide a specimen of purulent material for bacteriologic analysis and antibiotic sensitivity testing. Remove purulent material, thereby lowering intracranial pressure and decreasing the mass effect of the abscess. Decompress and irrigate the ventricular system and debride the abscess in the event of its rupture into the ventricular system .

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3045091/

Strategies for the discovery of therapeutic Aptamers

Importance of the field Therapeutic aptamers are synthetic, structured oligonucleotides that bind to a very broad range of targets with high affinity and specificity. They are an emerging class of targeting ligand that show great promise for treating a number of diseases. A series of aptamers currently in various stages of clinical development highlights the potential of aptamers for therapeutic applications. Area covered in this review This review will cover in vitro selection of oligonucleotide ligands, called aptamers, from a combinatorial library using the Systematic Evolution of Ligands by Exponential Enrichment (SELEX) process as well as the other known strategies for finding aptamers against various targets. What the reader will gain Readers will gain an understanding of the highly useful strategies for successful aptamer discovery. They may also be able combine two or more of the presented strategies for their aptamer discovery projects. Take home message Although many processes are available for discovering aptamers, it is not trivial to discover an aptamer candidate that is ready to move toward pharmaceutical drug development. It is also apparent that there have been relatively few therapeutic advances and clinical trials undertaken due to the small number of companies that participate in aptamer development. Importance of the field Therapeutic aptamers are synthetic, structured oligonucleotides that bind to a very broad range of targets with high affinity and specificity. They are an emerging class of targeting ligand that show great promise for treating a number of diseases. A series of aptamers currently in various stages of clinical development highlights the potential of aptamers for therapeutic applications. Area covered in this review This review will cover in vitro selection of oligonucleotide ligands, called aptamers, from a combinatorial library using the Systematic Evolution of Ligands by Exponential Enrichment (SELEX) process as well as the other known strategies for finding aptamers against various targets. What the reader will gain Readers will gain an understanding of the highly useful strategies for successful aptamer discovery. They may also be able combine two or more of the presented strategies for their aptamer discovery projects. Take home message Although many processes are available for discovering aptamers, it is not trivial to discover an aptamer candidate that is ready to move toward pharmaceutical drug development. It is also apparent that there have been relatively few therapeutic advances and clinical trials undertaken due to the small number of companies that participate in aptamer development.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8545222/

Inclusion of Infants and Neonates in Pediatric Orphan Product Approvals

The Orphan Drug Act (ODA) of 1983 was enacted to provide financial incentives to drug sponsors to develop therapies for rare diseases. Although this act increased the number of orphan products approved, there are still a limited number of products available for the pediatric population because orphan drug products are exempt from the Pediatric Research Equity Act. The objectives of this study were (i) to evaluate the pediatric orphan drug studies submitted to the US Food and Drug Administration (FDA) in the period of 2007–2018 and (ii) to examine whether orphan drug products were fully labeled with a pediatric indication in infants and neonates. Out of the 468 indications evaluated, 171 (37%) were FDA-labeled for use in the pediatric population. Labeling for the 12 to < 18 years age group was most common (98%). Fifty-two percent of FDA-labeled pediatric indications included the newborn to < 2 years of age group. In this newborn to < 2 years age group, the indication was labeled without pivotal clinical trials in 43% of the programs. Of the 60 new indications not labeled down to birth, 50% were found to have an age of onset and diagnosis that occurs earlier than the age approved for use of the product for that indication. In summary, although the ODA has been successful in improving pediatric access to medications for rare diseases, our analysis identified the incomplete labeling for pediatric patients under 2 years of age. Strategies to include the birth to < 2 years old group of pediatric patients in orphan drug development programs should be explored.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10749683/

Application of Artificial Intelligence in the Management of Drinking Water: A Narrative Review

Waterborne illnesses are a significant concern worldwide. The management of water resources can be facilitated by artificial intelligence (AI) with the help of data analytics, regression models, and algorithms. Achieving the Sustainable Development Goals (SDGs) of the 2030 Agenda for Sustainable Development of the United Nations depends on understanding, communicating, and measuring the value of water and incorporating it into decision-making. Various barriers are used from the source to the consumer to prevent microbiological contamination of drinking water sources or reduce contamination to levels safe for human health. Infrastructure development and capacity-building policies should be integrated with guidelines on applying AI to problems relating to water to ensure good development outcomes. Communities can live healthily with such technology if they can provide clean, economical, and sustainable water to the ecosystem as a whole. Quick and accurate identification of waterborne pathogens in drinking and recreational water sources is essential for treating and controlling the spread of water-related diseases, especially in resource-constrained situations. To ensure successful development outcomes, policies on infrastructure development and capacity building should be combined with those on applying AI to water-related problems. The primary focus of this study is the use of AI in managing drinking water and preventing waterborne illness. Introduction and background Water is the natural elixir of life, as the saying goes. Life must have access to clean water to survive. Nonetheless, we discover that waterborne infections cause devastation in sizable areas of developing and undeveloped countries. According to the WHO, 3.6 million people worldwide die from waterborne illnesses, with children accounting for around 2.2 million fatalities. Waterborne diseases are illnesses brought on by drinking water infected with pathogens like dangerous bacteria, viruses, protozoa, etc [ 1 ]. These harmful microorganisms contaminate water due to improper sanitation practices, and sewage spills into drinking water sources. Both governmental and non-governmental organizations (NGOs) have made every effort to raise the drinking water standard. But for most people, access to safe drinking water remains a pipe dream. To successfully address this worldwide threat, optimal solutions built on cutting-edge deep learning and machine learning developments can be deployed [ 2 ]. The leading causes of water pollution are bacteria, viruses, parasites, insecticides, pharmaceuticals, plastics, feces, radioactive materials, fertilizers, and pesticides. These compounds are frequently invisible contaminants since they do not permanently alter the color of water [ 3 ]. The bacteria known as Anthrax bacilli, found in tanning wastes, are the most dangerous water contaminants. They all cause water contamination, make it hazardous to drink, and if it is utilized, cause illnesses transmitted through water [ 4 ]. Waterborne infections are a big concern on a global scale. Around three million people globally die from water-related diseases, with 1.2 million fatalities occurring in children, according to a United Nations report [ 1 ]. Clean water is a vital component of our health. The WHO estimates drinking contaminated water can result in cholera, dysentery, typhoid, polio infections, and 485,000 annual diarrheal fatalities. Identifying pathogens is consequently crucial for treating and preventing aquatic illnesses. A ground-breaking development in object detection has been made possible by the advent of machine learning and deep learning, both powered by AI [ 2 ]. The current state of water resources makes it clear that better management is required. Recognizing, assessing, and elaborating on the value of water and incorporating it into decision-making is crucial for managing water resources sustainably and reasonably and achieving the Sustainable Development Goals (SDGs) of the 2030 Agenda for Sustainable Development of the United Nations [ 5 ]. An essential prerequisite for our survival is adequate water quality [ 6 ]. The likelihood of an outbreak of waterborne disease is increased in disaster-affected areas if clean water supplies are destroyed or combined with contaminated water. A nation like Haiti, where a sizable portion of the population lacks access to fresh water and even the most basic sanitation services, would suffer more in the aftermath of a disaster. Humanitarian relief and managing epidemic diseases would benefit significantly from early warning of waterborne diseases like cholera. Finding the appropriate characteristics to predict a future epidemic better is not easy in disease forecasting [ 7 ]. Methodology Literature searches for articles published between 2000 and 2023 were made using Google Scholar, Pub Med, and other databases. The terms "safe water," "waterborne diseases," "machine learning," and "future" were also utilized. The inclusion criteria were publications between 2000 and 2023, English-language publications, and articles discussing using AI in all situations to manage waterborne infections and drinking water. We excluded duplicates, abstracts, works written in languages other than English, unpublished works, and materials that didn't have much to do with AI and the management of clean, safe drinking water. Review Need for clean and safe drinking water A global requirement for clean water, sanitary conditions, and robust aquatic ecosystems is the treatment of water and wastewater [ 8 , 9 ]. Furthermore, having access to high-quality water is necessary for having a sustainable economy [ 10 ]. The health effects of natural disasters are one of their most frequent side effects, and they are directly tied to the water and sanitation issues that develop after such events [ 11 ]. The physical health of displaced persons is at risk both during and after crisis scenarios because they have fewer opportunities to practice good personal hygiene and have enough access to fresh water. As a result, the likelihood of an increase in waterborne diseases in a post-disaster situation is higher due to inadequate water, sanitation, and hygiene (WASH) conditions [ 12 ]. Some of the most common waterborne pathogens and diseases caused by them are given below. Bacteria like Escherichia coli, Yersinia enterocolitica, Leptospira spp., Campylobacter jejuni, Salmonella spp., Vibrio cholerae, Salmonella typhi, and Shigella spp. cause a variety of water-borne illnesses, from septicemia and gastroenteritis to specific ailments like yersiniosis, leptospirosis, reactive arthritis, salmonellosis, cholera, typhoid fever, and bacillary dysentery. Meningitis, infectious hepatitis, gastroenteritis, heart defects, and hepatitis A and E are caused by viruses such as rotavirus, coronaviruses, and enteroviruses. Diseases like amebiasis, cryptosporidiosis, and giardiasis, which cause fevers and diarrhea, are caused by protozoa, such as Entamoeba histolytica, Cryptosporidium, Microsporidia, and Giardia lamblia. Helminths like Necator americanus, Taenia spp., Enterobius vermicularis, Ascaris lumbricoides, and Ancylostoma spp. can cause a variety of worm-related waterborne diseases, including necatoriasis, taeniasis, enterobiasis, ascariasis, and ancylostomiasis. Effective prevention and treatment techniques are essential for public health programs, and understanding these organisms and the diseases they are associated with is imperative [ 13 ]. Natural disasters frequently result in food insecurity and malnutrition due to the destruction of agricultural regions and crops, which raises the risk of cholera and other outbreaks of diarrheal diseases [ 14 ]. The primary means of transmission of cholera is through the intake of water and food contaminated with Vibrio cholera, the disease's causative agent. The condition is common in places with poor sanitary and water infrastructure [ 15 , 16 ]. Over 100,000-150,000 people worldwide die from cholera each year [ 17 , 18 ]. To lessen consumers' risk from contaminated water, automatic anomaly detection monitoring is crucial in water utilities' distribution systems. Two significant issues and occurrences in the water quality anomaly identification domain are unbalanced class distribution and missing data. An overestimated classification accuracy can be produced by learning algorithms in an unbalanced dataset due to a bias favoring the majority class at the expense of the minority class. The effectiveness of learning algorithms in real-world water quality anomaly detection situations is significantly hampered by these two issues. So, in order to improve performance, they must be properly studied and handled. AI The aim of AI is to develop methods and equations that, when included with a system, allow the system to handle problems in a manner that is similar to a human [ 19 ]. AI is the ability of a system to carry out a set of tasks in a manner that is similar to how humans carry them out [ 20 , 21 ]. Any intelligent system needs to be able to learn in order to acquire new skills through experiments, store knowledge, and apply knowledge to solve issues. The ability of an AI system to learn, adapt, and gradually forget outdated or irrelevant information to enhance future scenarios is its key advantage. The foundation of AI is its ability to learn and its purported robustness. Because of this, we think there aren't any practical alternatives to AI techniques for creating autonomous systems that will help people. The main contribution of AI is to reduce the need for human participation and assistance in routine control and adaptability operations. Human intervention in artificially intelligent systems should become less active and humans should have more of an observer role [ 22 ]. According to an IBM (International Business Machines Corporation) business policy document [ 23 ], all facets and technologies of our lives are developing toward making a "smart planet." The goal of developing the field of AI is to create autonomous, intelligent systems. ChatGPT is one of the most sophisticated AI systems as of 2023; it is a component of strong big language models, specifically a sizable neural network trained on zetta-bytes of internet-based text data. It can comprehend spoken language and produce responses that resemble those of humans. Just to clarify, ChatGPT is only one illustration of the cutting-edge AI systems that are now in use. DeepMind, an Alphabet Inc. subsidiary, created an AI called AlphaGo. In 2016, AlphaGo made headlines after defeating Lee Sedol, the most intelligent Go player in the world, in a five-game match. An old Chinese board game, AlphaGo, has straightforward rules but is exceedingly difficult to play without human intuition. According to DeepMind researchers, their AI is all-purpose, which means it can do much more than just play Go. Engineers utilized it to address various issues, such as protein folding and controlling Google data centers' cooling systems. IBM created the AI system known as Watson. It was initially created as a question-answering chatbot based on neural networks and advanced natural language processing. Additionally, Watson excels in healthcare applications since it can foretell the likelihood of skin malignancies based solely on a person's photograph. It can more accurately identify a variety of ailments like cancer, cardiovascular disease, heart disease, etc., and also makes drug recommendations. Several hospitals and health facilities throughout the world make use of Watson's skills [ 24 ]. The application of AI is widespread now. Accenture and Frontier Economics' investigation also shows that the technology has much more potential. According to the analysis, due to artificial intelligence, industrialized countries' productivity growth will improve by up to 40% by 2035 [ 25 ]. Role of artificial intelligence in safe water supply A discipline of computer science that deals with the simulation of intelligent behavior in computers or a machine's ability to resemble intelligent human behavior is known as AI [ 26 ]. AI or machine learning is primarily used to make decisions about providing efficient water supply. These tasks include maximizing the information and data water utilities have access to improve service delivery, capital investment optimization, and operating cost reduction, including social and environmental externalities. Water utilities frequently adopt business practices from other industries, particularly those in the energy industry, without fully comprehending the underlying presumptions and repercussions of doing so [ 27 ]. Coastal distribution and seasonal climate dynamics have been connected in earlier research to the pathogenic Vibrio cholera bacteria that cause human cholera illness. Numerous possibilities exist for developing cholera-risk apps for the environment that use remote-sensed significant climate parameters and random forest classifiers. More research on the present random forest model and its primary climate variables is based on cholera surveillance datasets in additional coastal areas affected by the outbreak to determine the method's relevance and effectiveness for cholera forecasting systems. Numerous significant outbreaks have been caused by decreasing water quality or failing drinking water infrastructure. These problems can be detected beforehand by a real-time drinking water quality monitoring system, which can then notify operators and prompt them to take the necessary action. Although often used for this purpose, Supervisory Control and Data Acquisition (SCADA) has several disadvantages, such as issues with sensor scalability, a lack of predictive capability, and increased effort for operators owing to the constant onslaught of pointless alarms. AI can facilitate the management of water resources with data analytics, regression models, and algorithms. These cutting-edge technologies make it easier to design efficient water networks and systems. AI makes it feasible to construct water facilities and assess the quality of the water supply. Governmental agencies and water managers can employ artificial intelligence to create an intelligent water infrastructure to manage water efficiently and adapt to the environment. These environmentally friendly and economically advantageous technologies will be able to fully use all available water management alternatives and foresee possible risks [ 28 ]. Over the world, significant outbreaks caused by water infrastructure failure or water quality degradation have often occurred. These problems can be detected beforehand by a real-time drinking water quality monitoring system, which can also notify operators to take the necessary action. Despite being commonly used for this purpose, SCADA has many limitations, including sensor scalability challenges, a lack of predictive capability, and an increased burden for operators bombarded with unneeded warnings. Cloud Internet of Things (IoT), AI and Soft Computing (AI and SC), and other technologies can reduce operator dependency and improve system operations [ 29 ]. Water contamination is the fundamental cause of many diseases in the world. Sensors must be employed to gauge the water's quality to stop the spread of waterborne diseases. The connected works still have problems with communication, mobility, accuracy, and scalability. A new SCADA system that incorporates IoT technologies was proposed in a real-time study for monitoring water quality. Using an Arduino Atmega 368 (Arduino Corporation, Somerville, MA, United States) and a Global System for Mobile Communication (GSM) module, it intends to detect water contamination, pipeline breaches, and automatic measurements of parameters (such as temperature sensor, flow sensor, and color sensor) in real time. The system is used in Tirunelveli, a significant city in Tamilnadu, India, to automatically capture sensor data from pressure, pH level, and energy sensors. The SCADA system now has more sensors at a lower cost. The outcomes demonstrate that the suggested approach performs better than those already in use and yields superior results. SCADA with GSM connectivity collects precise real-time sensor values for flow, temperature, color, and turbidity [ 30 ]. Safe water AI is one of the creative ways that AI is being used to achieve SDGs linked to water quality. Convolution Neural Network (CNN) and IoT technologies, developed in the USA, allow for real-time analysis and identification of pollutants like bacteria, even without an internet connection. The system is made out of cheap commercial off-the-shelf parts. The clean water AI package costs USD 500 right now. Further price reductions are anticipated as AI technology advances and is adopted by more individuals [ 31 ]. By 2050, 70% of the world's population is expected to live in cities. Uncontrolled urbanization can lead to cities that worsen poverty, inequality, informal settlements/communities, pollution, and unemployment. It can also encroach on bio-diverse areas and productive agricultural fields, releasing unchecked pollutants into vulnerable water supplies. Contrarily, multi-level governance, and integrated regional and urban planning can preserve and improve water resources, storage, and retention while encouraging investment in climate-resilient infrastructure, supporting stormwater management and disaster reduction, and boosting the blue economy [ 32 ]. Scientists and engineers can now reduce the inaccuracy associated with a system or particle's geometry or size by using AIs. The method that is most frequently used to accomplish this is to train an AI model using data that comes from systems whose behavior is already well understood. These methods are beneficial for nano-materials because it can frequently be challenging to reproduce the various effects and phenomena observed in materials like graphene. This program has a great deal of potential. In fact, it promises to incorporate machine learning into production methods, which would catalyze the advancement of both AI and nanotechnologies in the future [ 33 ]. There is an increased demand for fresh water in many crowded, growing cities worldwide, and planners are unsure how to meet this demand in the future. Communities may use such technology to breathe easily by extending the concept of affordable, clean, and sustainable water to the ecosystem as a whole. Understanding the primary environmental problems the world is currently facing and taking into account potential solutions from emerging nanotechnologies can help achieve sustainability in clean water [ 34 ]. There are now numerous ways to clean drinking water, including chemical processes that release toxins into the liquid media, killing cyanobacteria and causing cell lysis to reduce the microbial burden [ 35 ]. Water sector evolution due to AI The water sector is embracing AI, which powers machine learning-based intelligent operations that maximize resource consumption and operational budgets for businesses. 1. To bring in intelligent infrastructure solutions, water and wastewater operations will invest in technology over the next ten years; 2. By lowering energy costs, optimizing the use of chemicals for treatment, and facilitating proactive asset maintenance, AI will result in considerable operating expenses (OpEx) reductions in water and wastewater operations; 3. AI will forecast emergencies, learn from them faster, and identify trends that might point to an impending break event. As a result, notifications will get better over time; 4. AI will offer powerful decision-making intelligence to help operators make critical decisions without having to evaluate complex variables independently. AI empowers operators with intelligent recommendations and machine learning-driven decision-making, whether it's controlling the operation of pumps, calculating chemical dosages, or selecting whether to maintain assets; 5. AI will optimize pump runtimes to ensure that energy is only used when necessary to reduce energy consumption for water and wastewater operations; 6. AI will maintain clean water at an affordable cost for both public and private use. To ensure that effluence criteria are followed, and compliance fines are avoided, AI learns from the distinctive features of your site; 7. AI will simplify data integrity by processing this heterogeneous data to make it clear, valuable, safe, and the basis for high-fidelity recommendations; 8. AI will run genuinely intelligent water systems. Organizations can seek data-driven, innovative management of water systems due to the deployment of AI. As a result, water management will be dependable, long-lasting, and affordable [ 36 ]. Table 1 lists several uses of artificial intelligence (AI) in the management of water resources, each system having its unique advantages. Table 1 AI-based models for water optimization Model Name Description Benefits Reservoir Operation Optimization [ 37 ] AI models analyze historical data, rainfall patterns, and water demand to optimize reservoir operations, maximizing water storage and release timing. Ensures efficient water storage and release, improves flood control, and optimizes hydropower generation. Water Allocation Optimization [ 38 ] AI algorithms consider water demand, availability, and environmental constraints to optimize water allocation schemes, minimizing conflicts and maximizing efficiency. Promotes equitable water distribution, reduces user conflicts, and improves water-use efficiency. Drought Forecasting and Mitigation [ 39 ] AI-based models analyze rainfall patterns, soil moisture levels, and climate data to predict and mitigate drought events. Enables early warning systems, improves drought preparedness, and facilitates proactive water management. Integrated Water Resource Management [ 40 ] AI-driven systems integrate data from multiple sources, including weather forecasts, river flows, and water usage, to provide holistic water resource management strategies. It enhances decision-making processes, optimizes water allocation, and supports sustainable water management. Environmental Impact Assessment [ 41 ] AI models assess the environmental impact of water resource management activities, considering water quality, ecosystem health, and habitat preservation factors. Facilitates environmentally sustainable practices, protects ecosystems, and preserves biodiversity. Real-Time Water Monitoring and Control [ 42 ] AI-powered sensors and data analytics enable real-time monitoring of water parameters and automatic control of water systems, optimizing water allocation and usage in response to changing conditions. It improves operational efficiency, reduces water losses, and facilitates adaptive water management. Stakeholder Engagement and Decision Support [ 43 ] AI-based tools facilitate stakeholder engagement and provide decision support by analyzing data, generating insights, and fostering collaborative water resource management approaches. Enhances communication and cooperation among stakeholders and supports evidence-based decision-making. Conclusions To prevent microbiological contamination of drinking water sources or to reduce contamination to levels safe for human health, a variety of barriers are used from the start to when the water is used by the consumer. The precautions that promote safety include preserving and maintaining treated water quality, managing distribution systems (whether piped or otherwise), and protecting water resources. The best management strategy puts less emphasis on using treatment technologies to eliminate pathogens and more on preventing or minimizing their entrance into water sources. Rapid and sensitive identification of waterborne pathogens in drinking and recreational water sources is essential for the treatment and control of the spread of water-related diseases, especially in resource-constrained situations. Localization is necessary before AI models, methods, and technology are implemented. The opportunities and problems associated with water differ by nation and place due to varying levels of infrastructure accessibility and implementation capabilities to address the issues and potential provided by AI. Before applying AI to address water-related problems, it is essential to conduct baseline studies to assess the implementation capabilities, return on investment, and impact of the intervention. Before implementing AI in the water sector, policymakers should thoroughly examine social, economic, and cultural issues. Support is required for AI for water-based activities to produce good development outcomes. To ensure successful development outcomes, infrastructure development, and capacity-building policies should be linked with guidelines on using AI for water-related challenges. Skills development is necessary for all parties interested in water, and capacity development programs must consider this. These measures for increasing capacity should promote interdisciplinary and cross-disciplinary study and research. The basic requirements for computation, energy, the generation of data, and storage should be considered while developing infrastructure policies. It's crucial to arrange these policies correctly.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2834293/

Potent inhibitors of furin and furin-like proprotein convertases containing decarboxylated P1 arginine mimetics

Furin belongs to the family of proprotein convertases (PCs) and is involved in numerous normal physiological and pathogenic processes, such as viral propagation, bacterial toxin activation, cancer and metastasis. Furin and related furin-like PCs cleave their substrates at characteristic multibasic consensus sequences, preferentially after an arginine residue. By incorporation of decarboxylated arginine mimetics in P1 position of substrate analogue peptidic inhibitors we could identify highly potent furin inhibitors. The most potent compound, phenylacetyl-Arg-Val-Arg-4-amidinobenzylamide ( 15 ), inhibits furin with a K i -value of 0.81 nM and has also comparable affinity to other PCs like PC1/3, PACE4, and PC5/6, whereas PC2 and PC7 or trypsin-like serine proteases were poorly affected. In fowl plague virus (influenza A, H7N1)-infected MDCK cells inhibitor 15 reduced proteolytic hemagglutinin cleavage and was able to reduce virus propagation in a long term infection test. Molecular modelling revealed several key interactions of the 4-amidinobenzylamide residue in the S1 pocket of furin contributing to the excellent affinity of these inhibitors.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9306250/

Intensive care unit versus high-dependency care unit admission on mortality in patients with septic shock: a retrospective cohort study using Japanese claims data

Background Septic shock is a common and life-threatening condition that requires intensive care. Intensive care units (ICUs) in Japan are classified into ICUs and high-dependency care units (HDUs), depending on presence of full-time certified intensivists and the number of assigned nurses. Compared with other developed countries, there are fewer intensive care beds and certified intensivists in Japan; therefore, non-intensivists often treat patients with septic shock in HDUs. It is unknown where we should treat patients with septic shock because no studies have compared the clinical outcomes between ICU and HDU treatment. This study aimed to elucidate which units should admit patients with septic shock by comparing mortality data and resource use between ICU and HDU admissions. Methods In this retrospective cohort study, we used a nationwide Japanese administrative database to identify adult patients with septic shock who were admitted to ICUs or HDUs between January 2010 and February 2021. The patients were divided into two groups, based on admittance to ICU or HDU on the day of hospitalization. The primary outcome was 30-day all-cause mortality adjusted for covariates using Cox regression analyses; the secondary outcomes were the length of ICU or HDU stay and length of hospital stay. Results Of the 10,818 eligible hospitalizations for septic shock, 6584 were in the ICU group, and 4234 were in the HDU group. Cox regression analyses revealed that patients admitted to the ICUs had lower 30-day mortality (adjusted hazard ratio: 0.89; 95% confidence interval: 0.83–0.96; P = 0.005). Linear regression analyses showed no significant difference in hospital length of stay or ICU or HDU length of stay. Conclusions An association was observed between ICU admission and lower 30-day mortality in patients with septic shock. These findings could provide essential insights for building a more appropriate treatment system. Background Septic shock is a common and life-threatening condition that requires intensive care. Intensive care units (ICUs) in Japan are classified into ICUs and high-dependency care units (HDUs), depending on presence of full-time certified intensivists and the number of assigned nurses. Compared with other developed countries, there are fewer intensive care beds and certified intensivists in Japan; therefore, non-intensivists often treat patients with septic shock in HDUs. It is unknown where we should treat patients with septic shock because no studies have compared the clinical outcomes between ICU and HDU treatment. This study aimed to elucidate which units should admit patients with septic shock by comparing mortality data and resource use between ICU and HDU admissions. Methods In this retrospective cohort study, we used a nationwide Japanese administrative database to identify adult patients with septic shock who were admitted to ICUs or HDUs between January 2010 and February 2021. The patients were divided into two groups, based on admittance to ICU or HDU on the day of hospitalization. The primary outcome was 30-day all-cause mortality adjusted for covariates using Cox regression analyses; the secondary outcomes were the length of ICU or HDU stay and length of hospital stay. Results Of the 10,818 eligible hospitalizations for septic shock, 6584 were in the ICU group, and 4234 were in the HDU group. Cox regression analyses revealed that patients admitted to the ICUs had lower 30-day mortality (adjusted hazard ratio: 0.89; 95% confidence interval: 0.83–0.96; P = 0.005). Linear regression analyses showed no significant difference in hospital length of stay or ICU or HDU length of stay. Conclusions An association was observed between ICU admission and lower 30-day mortality in patients with septic shock. These findings could provide essential insights for building a more appropriate treatment system. Background Sepsis is a common and life-threatening disease with high mortality of 12.5–15%, so its disease burden is enormous [ 1 , 2 ]. It is estimated that 47–50 million people worldwide, including children and those in developing countries, suffer from sepsis annually, and at least 11 million people die from sepsis [ 3 ]. It is also reported that more than 10,000 people die from sepsis every year in Japan [ 4 ]. Septic shock is defined as sepsis associated with circulatory and cellular metabolic abnormalities, and patients with septic shock have high hospital mortality [ 5 ]. Besides appropriate antimicrobial therapy and source control with drainage or surgery, intensive care with fluid resuscitation and vasopressors is the essential treatment strategy recommended by international guidelines for septic shock [ 6 , 7 ]. In contrast, there is no clear recommendation or consensus on where to treat patients with septic shock. "Sepsis Treatment System" was first mentioned in the Japanese Clinical Practice Guidelines 2020 (J-SSCG 2020), which recommends that patients with sepsis who do not respond to initial fluid resuscitation be managed in units where intensive care can be provided [ 8 ]. However, no reports compare the clinical outcomes of patients admitted to intensive care units (ICUs) versus non-ICU settings; therefore, it is difficult to appropriately define the appropriate "units". Furthermore, although the overall number of beds per population is larger, there are several problems with intensive care in Japan, such as a smaller number of certified intensivists and intensive care beds than in other developed countries [ 9 ]. Therefore, even in the case of critical illnesses, such as septic shock, a certain number of patients are treated by non-intensivists in non-ICU settings in Japan. This study aimed to investigate the practice pattern for patients with septic shock in Japan and to elucidate which units should admit patients by comparing the mortality of ICU admission with high-dependency care unit (HDU) admission. Methods Overview of ICU and HDU system in Japan There are three major categories of acute hospital beds in Japan depending on the patient-to-nurse ratio: ICUs, HDUs, and general wards (Table 1 ). Compared with HDUs, ICUs require more nurses and space, resulting in more expensive charges per admission. ICUs were further divided into two categories. ICUs for which "ICU management fee 1" can be charged are located in large, well-equipped hospitals such as university hospitals. They have higher standards for full-time staff and facilities: two or more certified intensivists, certified nurses, and certified clinical engineers. Conversely, the other category of ICUs requires a full-time physician, but not necessarily certified intensivists. HDUs and general wards do not require a full-time physician. In Japan, large hospitals often have both ICU and HDU, while middle-sized hospitals often have only ICU or HDU. Smaller community-based hospitals often do not have intensive care units. The Japanese government's insurance policy limits ICU admission to critically ill patients, such as those with loss of consciousness, respiratory failure, and shock. However, the actual decision of admission to either ICUs or HDUs depends on the medical system of each region, availability of beds, and judgment of the attending physician. Therefore, critically ill patients in Japan are often admitted to HDUs in Japan. Table 1 Categories of acute hospital beds in Japan Charges for admission Patient–nurse ratio Criteria ICUs ICU management fee 1 2:1 Full-time staff (two or more experienced certified intensivists, certified nurses, and clinical engineers) ICU management fee 2 2:1 Full-time staff (two or more experienced certified intensivists, certified nurses, and clinical engineers) Emergency and critical care unit management fee 2 2:1 Full-time physician (not necessary intensivists) Emergency and critical care unit management fee 4 2:1 Full-time physician (not necessary intensivists) ICU management fee 3 2:1 Full-time physician (not necessary intensivists) ICU management fee 4 2:1 Full-time physician (not necessary intensivists) HDUs Emergency and critical care unit management fee 1 4:1 Full-time physician (not necessary intensivists) Emergency and critical care unit management fee 3 4:1 Full-time physician (not necessary intensivists) High care unit management fee 1 4:1 No need for full-time physician High care unit management fee 2 5:1 No need for full-time physician General wards 7:1 No need for full-time physician Study design and data source We conducted a retrospective cohort study using the diagnostic procedure combination (DPC) database provided by Medical Data Vision Co., Ltd. (MDV; Tokyo, Japan) (MDV). This database has been used in previous epidemiological studies [ 10 , 11 ]. DPC is a payment system for acute hospital inpatients, in which provider reimbursement is calculated based on a per-diem fee according to the diagnosis category [ 12 ]. The MDV database is fully anonymized and includes more than 35 million inpatient data points from 438 acute care hospitals, which account for approximately 25% of all hospitals that have opted for DPC (as of the end of April 2021). The database contains demographic data, medical and pharmacy claims data, clinical diagnoses, and medical procedures. The clinical diagnoses were recorded using the International Classification of Diseases, 10th revision (ICD-10) codes. Medical procedures were recorded using Japanese classification codes and medical billing codes. Unfortunately, this database does not include physiological data such as vital signs or information about hospitals in which each patient was hospitalized. The study protocol was approved by the Ethics Committee of Kyoto University Graduate School and Faculty of Medicine (R2653). Study participants We identified patients with septic shock who were ≥ 18 years and admitted to ICUs or HDUs for intensive care on the day of hospitalization between January 2010 and February 2021. In this study, we defined patients who met the following criteria as having septic shock. First, patients with the diagnosis of both infection (ICD-10 codes A039, A021, A047, A207, A217, A227, A239, A241, A267, A280, A282, A327, A392, A393, A394, A400, A401, A402, A403, A408, A409, A410, A411, A412, A413, A414, A415, A418, A419, A427, B007, B377, J189, J440, N390) and organ dysfunction (ICD-10 codes J960, J969, J80, R092, R570, R571, R578, R579, I951, I959, N170, N171, N172, N178, N179, K720, K729, K763, F050, F059, G931, G934, G938, D695, D696, D65) were identified using the ICD-10 codes that matched the ICD-9 codes used in the previous validation study [ 13 ] (Table 2 ). These diagnoses were identified from the database as main diagnosis, admission diagnosis, diagnosis with the first or second highest medical costs, or comorbidity at admission. Second, patients in whom both intravenous antibiotics and noradrenaline were used on the day of hospitalization. If a patient was hospitalized more than once, we counted each hospitalization as a single hospitalization. Table 2 ICD-10 codes used for inclusion criteria Infection A039 Shigellosis, unspecified A021 Salmonella sepsis A047 Enterocolitis due to Clostridium difficile A207 Septicemic plague A217 Generalized tularemia A227 Anthrax sepsis A239 Brucellosis, unspecified A241 Acute and fulminating melioidosis A267 Erysipelothrix sepsis A280 Pasteurellosis A282 Extraintestinal yersiniosis A327 Listerial sepsis A392 Acute meningococcemia A393 Chronic meningococcemia A394 Meningococcemia, unspecified A400 Sepsis due to Streptococcus , group A A401 Sepsis due to Streptococcus , group B A402 Sepsis due to Streptococcus , group D A403 Sepsis due to Streptococcus pneumoniae A408 Other streptococcal sepsis A409 Streptococcal sepsis, unspecified A410 Sepsis due to Staphylococcus aureus A411 Other sepsis A412 Sepsis due to unspecified Staphylococcus A413 Sepsis due to Haemophilus influenzae A414 Sepsis due to anaerobes A415 Sepsis due to other Gram-negative organisms A418 Other specified sepsis A419 Sepsis, unspecified, includes: septicemia A427 Actinomycotic sepsis B007 Disseminated herpes viral disease, includes herpes viral sepsis B377 Candidal sepsis J189 Pneumonia, unspecified organism J440 Chronic obstructive pulmonary disease with acute lower respiratory infection N390 Urinary tract infection, site not specified Organ dysfunction Respiratory J960 Acute respiratory failure J969 Respiratory failure, unspecified J80 Diseases of bronchus, not elsewhere classified R092 Respiratory arrest Cardiovascular R570 Cardiogenic shock R571 Hypovolemic shock R578 Other shock R579 Shock, unspecified I951 Orthostatic hypotension I959 Hypotension, unspecified Renal N170 Acute renal failure with tubular necrosis N171 Acute renal failure with acute cortical necrosis N172 Acute renal failure with medullary necrosis N178 Other acute renal failure N179 Acute renal failure, unspecified Neurological K720 Acute and subacute hepatic failure K729 Hepatic failure, unspecified K763 Infarction of liver F050 Delirium not superimposed on dementia, so described F059 Delirium, unspecified G931 Anoxic brain damage, not elsewhere classified G934 Encephalopathy, unspecified G938 Metabolic encephalopathy Hematological D695 Secondary thrombocytopenia D696 Thrombocytopenia, unspecified D65 Disseminated intravascular coagulation (defibrination syndrome) Patients who died within 24 h after hospitalization were excluded because they were probably so severely ill they would have died, regardless of the unit type. Patients with the following diseases and procedures at the time of hospitalization were also excluded because they may be incorrectly included by the criteria above: patients complicated with congestive heart failure (ICD-10 code I509), complicated with severe acute pancreatitis (ICD-10 code K859), and patients who underwent the following procedures: percutaneous coronary intervention, coronary artery bypass grafting, valve replacement, valvuloplasty, transcatheter aortic valve implantation, operation for aortic aneurysm or dissection, or endovascular aortic repair. Exposure and comparison We defined patients admitted to the ICUs on the day of hospitalization as the exposure group and those admitted to HDUs on the day of hospitalization as the comparison group. We identified admission to the ICUs using Japanese claims codes (classification codes A3002, A3004, A3011, A3012, A3013, A3014) or the HDUs (classification codes A3001, A30011, A3003, A3004, A301-21, A301-22, A301-24). The claims codes present in both groups (classification code A3004) were further distinguished using the accompanying medical billing codes. We excluded the management fee for severe burns from A3003, A3004, A3012, and A3014 to exclude patients with severe burns. If patients with claims codes of both units were identified on the day of hospitalization, we considered them to belong to the first unit group where the initial location they were admitted prior to transfer. Outcomes The primary outcome of this study was 30-day all-cause mortality. We identified patient deaths using the discharge outcomes recorded in the DPC database. The secondary outcomes were the length of ICU or HDU stay, length of hospital stay, discharge destination, and Barthel index at discharge. The Barthel index (BI) was calculated based on the activities of daily living (ADL) scores recorded in the DPC database. The cumulative BI score ranges from 0 to 100 points, with 0 indicating complete dependence in activities of daily living and 100 indicating complete independence. Covariates The covariates for adjusting confounding factors were age, sex, Charlson comorbidity index [ 14 ], admission year, ambulance use, emergency charge, admission from the nursing home, and facility information, such as teaching hospital and number of hospital beds. The following procedures and treatments performed on the day of hospitalization were also identified from the database: emergency surgery or drainage procedures performed for infectious source control, mechanical ventilation, continuous renal replacement therapy, polymyxin B-immobilized fiber column direct hemoperfusion (PMX-DHP), venoatrial extracorporeal membrane oxygenation (VA-ECMO), use of two or more vasoactive agents (dopamine, noradrenaline, dobutamine, epinephrine, and vasopressin), blood transfusion (red blood cells, platelets, fresh frozen plasma), albumin preparations infusion, sedative drugs, narcotic drugs, recombinant thrombomodulin, antithrombin III preparations, low-dose glucocorticoids, and intravenous immunoglobulin. Each patient's infection source was identified using ICD-10 codes recorded at admission, combined with emergency surgery or drainage procedures performed. Statistical analysis Categorical and ordinal variables were summarized using numbers and percentages. If normally distributed continuous variables were summarized using mean and standard deviation, or median and interquartile range if not normally distributed. We compared 30-day mortality between the ICU and HDU groups using the Kaplan–Meier method and log-rank test and estimated the hazard ratio using multivariable Cox proportional hazard models, adjusting for the covariates mentioned above. Patients who were transferred to other hospitals and discharged within 30 days of hospitalization were censored. The survival period was calculated from the date of hospitalization to the date of death from any cause within 30 days. Secondary outcomes were analyzed using a logistic regression model to evaluate the association between ICU admission and in-hospital mortality. A linear regression model was used to assess the length of ICU (or HDU) stay, length of hospital stay, and Barthel index on discharge. We adjusted all secondary outcomes for the same covariates as those in the survival analysis. Subgroup analyses were performed for age, procedures performed on the day of hospitalization, and the source of infection. Sensitivity analyses were performed for limited populations as follows: (a) population which include the patients who met the exclusion criteria; (b) population with ICD-9 codes for infection and organ dysfunction from a previous study that did not match ICD-10 codes, supplemented with the corresponding ICD-10 codes (supplement ICD-9 codes: A41.50, A41.51, A41.52, A41.58 with ICD-10 code: A498, supplement ICD-9 code: R572 with ICD-10 code: A419); (c) admission to hospitals with both ICUs and HDUs; (d) 14-day all-cause mortality, and (e) in-hospital mortality. We conducted sensitivity analyses by changing the definition of exposure and comparison (f): (1) "ICU management fee 1" vs. "ICU management fee 3" and "Emergency and critical care unit management fee 2" to examine whether "ICU management fee 1" had better outcomes in ICUs; (2) "ICU management fee 1" vs. "Emergency and critical care unit management fee 1", to compare outcomes for the most resource-rich ICUs and HDUs, respectively; (3) "ICU management fee 3" and "Emergency and critical care unit management fee 2" vs. "Emergency and critical care unit management fee 1", to compare outcomes in more resource-poor ICUs to those in the most resource-rich HDUs. We also performed propensity score matching analyses using the nearest neighbor matching (g): (1) caliper width of 0.1 of the standard deviation; (2) caliper width of 0.2 of the standard deviation. A multivariable logistic regression model using all the covariates same as the primary analysis was employed to compute the propensity scores for patients admitted to the ICUs on the day of hospitalization. The statistical significance level was set at a two-tailed p < 0.05, and all statistical analyses were conducted using SAS ver. 9.4 (SAS Institute Inc., Cary, NC, USA). Overview of ICU and HDU system in Japan There are three major categories of acute hospital beds in Japan depending on the patient-to-nurse ratio: ICUs, HDUs, and general wards (Table 1 ). Compared with HDUs, ICUs require more nurses and space, resulting in more expensive charges per admission. ICUs were further divided into two categories. ICUs for which "ICU management fee 1" can be charged are located in large, well-equipped hospitals such as university hospitals. They have higher standards for full-time staff and facilities: two or more certified intensivists, certified nurses, and certified clinical engineers. Conversely, the other category of ICUs requires a full-time physician, but not necessarily certified intensivists. HDUs and general wards do not require a full-time physician. In Japan, large hospitals often have both ICU and HDU, while middle-sized hospitals often have only ICU or HDU. Smaller community-based hospitals often do not have intensive care units. The Japanese government's insurance policy limits ICU admission to critically ill patients, such as those with loss of consciousness, respiratory failure, and shock. However, the actual decision of admission to either ICUs or HDUs depends on the medical system of each region, availability of beds, and judgment of the attending physician. Therefore, critically ill patients in Japan are often admitted to HDUs in Japan. Table 1 Categories of acute hospital beds in Japan Charges for admission Patient–nurse ratio Criteria ICUs ICU management fee 1 2:1 Full-time staff (two or more experienced certified intensivists, certified nurses, and clinical engineers) ICU management fee 2 2:1 Full-time staff (two or more experienced certified intensivists, certified nurses, and clinical engineers) Emergency and critical care unit management fee 2 2:1 Full-time physician (not necessary intensivists) Emergency and critical care unit management fee 4 2:1 Full-time physician (not necessary intensivists) ICU management fee 3 2:1 Full-time physician (not necessary intensivists) ICU management fee 4 2:1 Full-time physician (not necessary intensivists) HDUs Emergency and critical care unit management fee 1 4:1 Full-time physician (not necessary intensivists) Emergency and critical care unit management fee 3 4:1 Full-time physician (not necessary intensivists) High care unit management fee 1 4:1 No need for full-time physician High care unit management fee 2 5:1 No need for full-time physician General wards 7:1 No need for full-time physician Study design and data source We conducted a retrospective cohort study using the diagnostic procedure combination (DPC) database provided by Medical Data Vision Co., Ltd. (MDV; Tokyo, Japan) (MDV). This database has been used in previous epidemiological studies [ 10 , 11 ]. DPC is a payment system for acute hospital inpatients, in which provider reimbursement is calculated based on a per-diem fee according to the diagnosis category [ 12 ]. The MDV database is fully anonymized and includes more than 35 million inpatient data points from 438 acute care hospitals, which account for approximately 25% of all hospitals that have opted for DPC (as of the end of April 2021). The database contains demographic data, medical and pharmacy claims data, clinical diagnoses, and medical procedures. The clinical diagnoses were recorded using the International Classification of Diseases, 10th revision (ICD-10) codes. Medical procedures were recorded using Japanese classification codes and medical billing codes. Unfortunately, this database does not include physiological data such as vital signs or information about hospitals in which each patient was hospitalized. The study protocol was approved by the Ethics Committee of Kyoto University Graduate School and Faculty of Medicine (R2653). Study participants We identified patients with septic shock who were ≥ 18 years and admitted to ICUs or HDUs for intensive care on the day of hospitalization between January 2010 and February 2021. In this study, we defined patients who met the following criteria as having septic shock. First, patients with the diagnosis of both infection (ICD-10 codes A039, A021, A047, A207, A217, A227, A239, A241, A267, A280, A282, A327, A392, A393, A394, A400, A401, A402, A403, A408, A409, A410, A411, A412, A413, A414, A415, A418, A419, A427, B007, B377, J189, J440, N390) and organ dysfunction (ICD-10 codes J960, J969, J80, R092, R570, R571, R578, R579, I951, I959, N170, N171, N172, N178, N179, K720, K729, K763, F050, F059, G931, G934, G938, D695, D696, D65) were identified using the ICD-10 codes that matched the ICD-9 codes used in the previous validation study [ 13 ] (Table 2 ). These diagnoses were identified from the database as main diagnosis, admission diagnosis, diagnosis with the first or second highest medical costs, or comorbidity at admission. Second, patients in whom both intravenous antibiotics and noradrenaline were used on the day of hospitalization. If a patient was hospitalized more than once, we counted each hospitalization as a single hospitalization. Table 2 ICD-10 codes used for inclusion criteria Infection A039 Shigellosis, unspecified A021 Salmonella sepsis A047 Enterocolitis due to Clostridium difficile A207 Septicemic plague A217 Generalized tularemia A227 Anthrax sepsis A239 Brucellosis, unspecified A241 Acute and fulminating melioidosis A267 Erysipelothrix sepsis A280 Pasteurellosis A282 Extraintestinal yersiniosis A327 Listerial sepsis A392 Acute meningococcemia A393 Chronic meningococcemia A394 Meningococcemia, unspecified A400 Sepsis due to Streptococcus , group A A401 Sepsis due to Streptococcus , group B A402 Sepsis due to Streptococcus , group D A403 Sepsis due to Streptococcus pneumoniae A408 Other streptococcal sepsis A409 Streptococcal sepsis, unspecified A410 Sepsis due to Staphylococcus aureus A411 Other sepsis A412 Sepsis due to unspecified Staphylococcus A413 Sepsis due to Haemophilus influenzae A414 Sepsis due to anaerobes A415 Sepsis due to other Gram-negative organisms A418 Other specified sepsis A419 Sepsis, unspecified, includes: septicemia A427 Actinomycotic sepsis B007 Disseminated herpes viral disease, includes herpes viral sepsis B377 Candidal sepsis J189 Pneumonia, unspecified organism J440 Chronic obstructive pulmonary disease with acute lower respiratory infection N390 Urinary tract infection, site not specified Organ dysfunction Respiratory J960 Acute respiratory failure J969 Respiratory failure, unspecified J80 Diseases of bronchus, not elsewhere classified R092 Respiratory arrest Cardiovascular R570 Cardiogenic shock R571 Hypovolemic shock R578 Other shock R579 Shock, unspecified I951 Orthostatic hypotension I959 Hypotension, unspecified Renal N170 Acute renal failure with tubular necrosis N171 Acute renal failure with acute cortical necrosis N172 Acute renal failure with medullary necrosis N178 Other acute renal failure N179 Acute renal failure, unspecified Neurological K720 Acute and subacute hepatic failure K729 Hepatic failure, unspecified K763 Infarction of liver F050 Delirium not superimposed on dementia, so described F059 Delirium, unspecified G931 Anoxic brain damage, not elsewhere classified G934 Encephalopathy, unspecified G938 Metabolic encephalopathy Hematological D695 Secondary thrombocytopenia D696 Thrombocytopenia, unspecified D65 Disseminated intravascular coagulation (defibrination syndrome) Patients who died within 24 h after hospitalization were excluded because they were probably so severely ill they would have died, regardless of the unit type. Patients with the following diseases and procedures at the time of hospitalization were also excluded because they may be incorrectly included by the criteria above: patients complicated with congestive heart failure (ICD-10 code I509), complicated with severe acute pancreatitis (ICD-10 code K859), and patients who underwent the following procedures: percutaneous coronary intervention, coronary artery bypass grafting, valve replacement, valvuloplasty, transcatheter aortic valve implantation, operation for aortic aneurysm or dissection, or endovascular aortic repair. Exposure and comparison We defined patients admitted to the ICUs on the day of hospitalization as the exposure group and those admitted to HDUs on the day of hospitalization as the comparison group. We identified admission to the ICUs using Japanese claims codes (classification codes A3002, A3004, A3011, A3012, A3013, A3014) or the HDUs (classification codes A3001, A30011, A3003, A3004, A301-21, A301-22, A301-24). The claims codes present in both groups (classification code A3004) were further distinguished using the accompanying medical billing codes. We excluded the management fee for severe burns from A3003, A3004, A3012, and A3014 to exclude patients with severe burns. If patients with claims codes of both units were identified on the day of hospitalization, we considered them to belong to the first unit group where the initial location they were admitted prior to transfer. Outcomes The primary outcome of this study was 30-day all-cause mortality. We identified patient deaths using the discharge outcomes recorded in the DPC database. The secondary outcomes were the length of ICU or HDU stay, length of hospital stay, discharge destination, and Barthel index at discharge. The Barthel index (BI) was calculated based on the activities of daily living (ADL) scores recorded in the DPC database. The cumulative BI score ranges from 0 to 100 points, with 0 indicating complete dependence in activities of daily living and 100 indicating complete independence. Covariates The covariates for adjusting confounding factors were age, sex, Charlson comorbidity index [ 14 ], admission year, ambulance use, emergency charge, admission from the nursing home, and facility information, such as teaching hospital and number of hospital beds. The following procedures and treatments performed on the day of hospitalization were also identified from the database: emergency surgery or drainage procedures performed for infectious source control, mechanical ventilation, continuous renal replacement therapy, polymyxin B-immobilized fiber column direct hemoperfusion (PMX-DHP), venoatrial extracorporeal membrane oxygenation (VA-ECMO), use of two or more vasoactive agents (dopamine, noradrenaline, dobutamine, epinephrine, and vasopressin), blood transfusion (red blood cells, platelets, fresh frozen plasma), albumin preparations infusion, sedative drugs, narcotic drugs, recombinant thrombomodulin, antithrombin III preparations, low-dose glucocorticoids, and intravenous immunoglobulin. Each patient's infection source was identified using ICD-10 codes recorded at admission, combined with emergency surgery or drainage procedures performed. Statistical analysis Categorical and ordinal variables were summarized using numbers and percentages. If normally distributed continuous variables were summarized using mean and standard deviation, or median and interquartile range if not normally distributed. We compared 30-day mortality between the ICU and HDU groups using the Kaplan–Meier method and log-rank test and estimated the hazard ratio using multivariable Cox proportional hazard models, adjusting for the covariates mentioned above. Patients who were transferred to other hospitals and discharged within 30 days of hospitalization were censored. The survival period was calculated from the date of hospitalization to the date of death from any cause within 30 days. Secondary outcomes were analyzed using a logistic regression model to evaluate the association between ICU admission and in-hospital mortality. A linear regression model was used to assess the length of ICU (or HDU) stay, length of hospital stay, and Barthel index on discharge. We adjusted all secondary outcomes for the same covariates as those in the survival analysis. Subgroup analyses were performed for age, procedures performed on the day of hospitalization, and the source of infection. Sensitivity analyses were performed for limited populations as follows: (a) population which include the patients who met the exclusion criteria; (b) population with ICD-9 codes for infection and organ dysfunction from a previous study that did not match ICD-10 codes, supplemented with the corresponding ICD-10 codes (supplement ICD-9 codes: A41.50, A41.51, A41.52, A41.58 with ICD-10 code: A498, supplement ICD-9 code: R572 with ICD-10 code: A419); (c) admission to hospitals with both ICUs and HDUs; (d) 14-day all-cause mortality, and (e) in-hospital mortality. We conducted sensitivity analyses by changing the definition of exposure and comparison (f): (1) "ICU management fee 1" vs. "ICU management fee 3" and "Emergency and critical care unit management fee 2" to examine whether "ICU management fee 1" had better outcomes in ICUs; (2) "ICU management fee 1" vs. "Emergency and critical care unit management fee 1", to compare outcomes for the most resource-rich ICUs and HDUs, respectively; (3) "ICU management fee 3" and "Emergency and critical care unit management fee 2" vs. "Emergency and critical care unit management fee 1", to compare outcomes in more resource-poor ICUs to those in the most resource-rich HDUs. We also performed propensity score matching analyses using the nearest neighbor matching (g): (1) caliper width of 0.1 of the standard deviation; (2) caliper width of 0.2 of the standard deviation. A multivariable logistic regression model using all the covariates same as the primary analysis was employed to compute the propensity scores for patients admitted to the ICUs on the day of hospitalization. The statistical significance level was set at a two-tailed p < 0.05, and all statistical analyses were conducted using SAS ver. 9.4 (SAS Institute Inc., Cary, NC, USA). Results Overall, 11,699 hospitalizations of patients with septic shock and admissions to ICUs or HDUs were identified between April 2008 and February 2021 (Fig. 1 ). Of these, 881 hospitalizations met the exclusion criteria, and 10,818 eligible hospitalizations of 10,754 patients were included in the analysis. Of the included hospitalizations, 6584 (60.9%) were in the ICU group, and 4234 (39.1%) were in the HDU group. No patient was admitted to both ICU and HDU on the day of hospitalization. The baseline characteristics of the patients are presented in Table 3 . Table 4 shows the treatments performed on the day of hospitalization. Although the baseline characteristics of both groups were similar, patients in the ICU group were more likely to receive intravenous drugs and interventions, such as mechanical ventilation, catheterization, or abdominal surgery. Fig. 1 Study flow diagram Table 3 Baseline characteristics of patients Overall ICU HDU n = 10,818 n = 6584 (60.9%) n = 4234 (39.1%) Age (years), median (IQR) 76.0 (67.0–84.0) 75.0 (66.0–82.0) 78.0 (69.0–85.0) Male sex, n (%) 6165 (57.0) 3813 (57.9) 2352 (55.6) BMI a (kg/m 2 ), median (IQR) 21.3 (18.6–24.3) 21.5 (18.7–24.5) 21.1(18.3–24.0) Charlson comorbidity index 0, n (%) 8013 (74.1) 4857 (73.8) 3156 (74.5) 1, n (%) 1949 (18.0) 1181 (17.9) 768 (18.1) ≤ 2, n (%) 856 (7.9) 546 (8.3) 310 (7.3) Source of infection Bacteremia/sepsis, n (%) 5228 (48.3) 3182 (48.3) 2046 (48.3) Respiratory, n (%) 1635 (15.1) 971 (14.8) 664 (15.7) Gastrointestinal, n (%) 1739 (16.1) 1293 (19.6) 446 (10.5) Urinary tract, n (%) 1136 (10.5) 526 (8.0) 610 (14.4) Hepatobiliary, n (%) 848 (7.8) 451 (6.9) 397 (9.4) Skin/soft tissue, n (%) 174 (1.6) 123 (1.9) 51 (1.2) Admission From home, n (%) 7265 (67.2) 4345 (66.0) 2920 (69.0) From other hospital, n (%) 1793 (16.6) 1195 (18.2) 598 (14.1) From nursing home, n (%) 852 (7.9) 374 (5.7) 478 (11.3) Admission year 2008–2012, n (%) 344 (3.2) 256 (3.9) 88 (2.1) 2013–2017, n (%) 4879 (45.1) 3117 (47.3) 1762 (41.6) 2018–2021, n (%) 5595 (51.7) 3211 (48.8) 2384 (56.3) Ambulance use, n (%) 8870 (82.1) 5478 (83.4) 3392 (80.2) Emergency charge, n (%) 5170 (47.8) 3145 (47.8) 2025 (47.8) Hospital beds ≤ 199, n (%) 112 (1.0) 58 (0.9) 54 (1.3) 200–499, n (%) 4796 (44.3) 2583 (39.2) 2213 (52.3) ≤ 500, n (%) 5910 (54.6) 3943 (59.9) 1967 (46.5) Teaching hospital, n (%) 10,156 (93.9) 6233 (94.7) 3923 (92.7) IQR interquartile range, BMI body mass index a The number of patients missing BMI: ICU 649, HDU 543 Table 4 Treatment performed on the day of hospitalization Overall ICU HDU n = 10,818 n = 6584 (60.9%) n = 4234 (39.1%) Vasoactive agents Dopamine, n (%) 2168 (20.0) 1413 (21.5%) 755 (17.8) Adrenaline, n (%) 982 (9.1) 720 (10.9%) 262 (6.2) Dobutamine, n (%) 868 (8.0) 649 (9.9%) 219 (5.2) Vasopressin, n (%) 1875 (17.3) 1479 (22.5%) 396 (9.4) ≥2 drugs, n (%) 4459 (41.2) 3146 (47.8%) 1313 (31.0) ≥3 drugs, n (%) 1193 (11.0) 924 (14.0%) 269 (6.4) Transfusion Red blood cell, n (%) 1805 (16.7) 1404 (21.3) 401 (9.5) Platelet, n (%) 667 (6.2) 521 (7.9) 146 (3.5) Fresh frozen plasma, n (%) 1581 (14.6) 1300 (19.7) 281 (6.6) Albumin, n (%) 3226 (29.8) 2403 (36.5) 823 (19.4) Globulin, n (%) 1511 (14.0) 1105 (16.8) 406 (9.6) Recombinant thrombomodulin, n (%) 2169 (20.1) 1281 (19.5) 888 (21.0) Antithrombin III, n (%) 1108 (10.2) 827 (12.6) 281 (6.6) Hydrocortisone, n (%) 2706 (25.0) 1922 (29.2) 784 (18.5) Sedative drugs, n (%) 5796 (53.6) 4325 (65.7) 1471 (34.7) Narcotic drugs, n (%) 4562 (42.2) 3481 (52.9) 1081 (25.5) Prophylaxis of gastrointestinal ulcer, n (%) 6327 (58.5) 4470 (67.9) 1857 (43.9) Prophylaxis of deep vein thrombosis, n (%) 3916 (36.2) 2762 (42.0) 1154 (27.3) Rehabilitation, n (%) 516 (4.8) 469 (7.1) 47 (1.1) Mechanical ventilation, n (%) 4049 (37.4) 2986 (45.4) 1063 (25.1) CRRT, n (%) 1691 (15.6) 1351 (20.5) 340 (8.0) PMX-DHP, n (%) 1025 (9.5) 746 (11.3) 279 (6.6) VA-ECMO/IABP, n (%) 200 (1.9) 172 (2.6) 28 (0.7) Central venous catheter, n (%) 7087 (65.5) 5028 (76.4) 2059 (48.6) Arterial line, n (%) 6979 (64.5) 5159 (78.4) 1820 (43.0) Urinary catheter, n (%) 8237 (76.1) 5109 (77.6) 3128 (73.9) Nasogastric tube, n (%) 4449 (41.1) 3406 (51.7) 1043 (24.6) Blood culture test, n (%) 9571 (88.5) 5799 (88.1) 3772 (89.1) Antibiotics Penicillin, n (%) 3459 (32.0) 2087 (31.7) 1372 (32.4) Cephalosporin, n (%) 2566 (23.7) 1554 (23.6) 1012 (23.9) Carbapenem, n (%) 5896 (54.5) 3683 (55.9) 2213 (52.3) Quinolone, n (%) 500 (4.6) 361 (5.5) 139 (3.3) Anti-MRSA, n (%) 1484 (13.7) 1102 (16.7) 382 (9.0) Aminoglycoside, n (%) 208 (1.9) 119 (1.8) 89 (2.1) Metronidazole, n (%) 116 (1.1) 90 (1.4) 26 (0.6) Anti-fungal, n (%) 180 (1.7) 137 (2.1) 43 (1.0) Anti-viral, n (%) 248 (2.3) 176 (2.7) 72 (1.7) Drainage Endoscopic, n (%) 357 (3.3) 178 (2.7) 179 (4.2) Percutaneous, n (%) 214 (2.0) 127 (1.9) 87 (2.1) Urinary, n (%) 754 (7.0) 388 (5.9) 366 (8.8) Surgery Abdominal surgery, n (%) 1583 (14.6) 1207 (18.3) 376 (8.9) Limb surgery, n (%) 92 (0.9) 67 (1.0) 25 (0.6) CRRT continuous renal replacement therapy, PMX-DHP polymyxin B immobilized fiber column direct hemoperfusion, VA-ECMO venoatrial extracorporeal membrane oxygenation, IABP intra-aortic balloon pumping Figure 2 shows the Kaplan–Meier survival plots for both groups. Cox regression analyses revealed that patients in the ICU group had lower 30-day mortality (adjusted hazard ratio: 0.89; 95% confidence interval 0.83–0.96; p = 0.005) (Table 5 ). The secondary outcomes are presented in Table 5 . The incidence proportion of all-cause in-hospital death was approximately 30% in both groups. Logistic regression analysis showed that patients in the ICU group had lower in-hospital mortality than patients in the HDU group (adjusted odds ratio: 0.82; 95% confidence interval: 0.75–0.90; p < 0.001). Linear regression analyses showed no significant difference in hospital length of stay and ICU or HDU length of stay. Fig. 2 Survival analysis Table 5 Primary outcome and secondary outcomes Overall ICU HDU Point estimates 95% CI P value Primary outcome 30-day mortality 2602 (24.0) 1576 (23.9) 1026 (24.2) 0.89 a 0.83–0.96 0.005 Secondary outcomes In-hospital death 3308 (30.6) 2041 (31.0) 1267 (29.9) 0.82 b 0.75–0.90 < 0.001 Hospital length of stay, days 25.0 (13.0–46.0) 26.0 (14.0–48.0) 22.0 (12.0–43.0) − 0.31 c − 1.92 to 1.28 0.69 ICU or HDU length of stay, days 6.0(3.0–13.0) 7.0(4.0–14.0) 5.0 (3.0–10.0) 0.11 c − 0.09 to 0.31 0.29 Discharge to home 3504 (32.4) 2036 (30.9) 1468 (34.7) 1.03 b 0.94–1.14 0.42 Discharge to other hospitals 3501 (32.4) 2289 (34.8) 1212 (28.6) 1.20 b 1.09–1.31 < 0.001 Discharge to nursing home 1310 (12.1) 812 (12.3) 498 (11.8) 1.12 b 0.95–1.31 0.15 Barthel index on discharge § 50.0 (0.0–100.0) 50.0 (0.0–100.0) 45.0 (0.0–100.0) 2.32 c 0.12–4.53 0.038 Data are presented as number of events (%) or mean (IQR) IQR interquartile range, CI confidence interval, CRRT continuous renal replacement therapy, PMX-DHP polymyxin B immobilized fiber column direct hemoperfusion, VA-ECMO venoatrial extracorporeal membrane oxygenation, IABP intra-aortic balloon pumping § The number of patients missing Barthel index: ICU 2499, HDU 1529 a Adjusted HR adjusted for age, sex, Charlson comorbidity index, admission year, ambulance use, teaching hospitals, emergency charge, hospital beds, patients from nursing home, source of infection, drainage, surgery, mechanical ventilation, CRRT, PMX-DHP, VA-ECMO, catecholamines, vasopressin, use of two or more catecholamines, transfusions (red blood cell, platelet, fresh frozen plasma), albumin, globulin, sedatives drugs, opioids drugs, recombinant thrombomodulin, antithrombin III, and hydrocortisone b Adjusted odds ratio adjusted for the same covariates as * c Regression coefficient adjusted for the same covariates as * The results of subgroup and sensitivity analyses are shown in Table 6 . Subgroup analyses showed that patients admitted to the ICU with VA-ECMO or with hepatobiliary diseases had significantly lower mortality. Sensitivity analyses performed in a population with different inclusion criteria also showed results consistent with those of the primary analysis. Fourteen-day and in-hospital all-cause mortality in Cox regression analyses were also lower in the ICU group. Cox regression analyses on propensity score-matched populations with different calipers also showed lower 30-day mortality in the ICU group. Table 6 Subgroup analysis and sensitivity analysis 30-day mortality Adjusted HR 95% CI P value Number of events/number of patients (%) e Overall ICU HDU Subgroup analysis Age, years 0.71 § < 65 399/2149 (18.5) 282/1479 (19.0) 117/670 (17.4) 0.92 a 0.74–1.13 0.440 65–74 607/2657 (22.8) 376/1679 (22.3) 231/978 (23.6) 0.83 a 0.71–0.98 0.029 75–84 889/3679 (24.1) 540/2213 (24.4) 349/1466 (23.8) 0.93 a 0.82–1.06 0.33 ≥ 85 707/2333 (30.3) 378/1213 (31.1) 329/1120 (29.3) 0.92 a 0.79–1.06 0.27 Procedures Mechanical ventilation 1266/4049 (31.2) 896/2986 (30.0) 370/1063 (34.8) 0.95 b 0.85–1.07 0.44 CRRT 500/1691 (29.5) 403/1351 (29.8) 97/340 (28.5) 1.08 b 0.89–1.33 0.44 PMX 257/1025 (25.0) 180/746 (24.1) 77/279 (27.6) 0.91 b 0.7–1.19 0.52 VA-ECMO/IABP 67/200 (33.5) 51/172 (29.6) 16/28 (57.1) 0.35 b 0.17–0.69 0.002 Source of infection Respiratory disease 442/1635 (27.0) 239/971 (24.6) 203/664 (30.5) 0.86 b 0.71–1.03 0.1 Urinary tract disease 102/1136 (8.9) 50/526 (9.5) 52/610 (8.5) 1.06 b 0.72–1.57 0.75 Gastrointestinal disease 379/1739 (21.7) 278/1293 (21.5) 101/446 (22.6) 1.10 b 0.88–1.36 0.38 Hepatobiliary disease 128/848 (15.0) 61/451 (13.5) 67/397 (16.8) 0.68 b 0.47–0.99 0.046 Skin/soft tissue 34/174 (19.5) 26/123 (21.1) 8/51 (15.6) 1.63 b 0.72–3.71 0.23 Sensitivity analysis (a) population which include the patients who met the exclusion criteria 2859/11699 (24.4) 1759/7218 (24.3) 1100/4481 (24.5) 0.9 c 0.84–0.97 0.008 (b) ICD-9 codes from the previous study supplemented with the corresponding ICD-10 codes 3204/13816 (23.1) 1965/8448 (23.2) 1239/5368 (23.0) 0.92 c 0.86–0.98 0.02 (c) Hospital with ICUs and HDUs 1997/8311 (24.0) 1395/5798 (24.0) 602/2513 (23.9) 0.86 c 0.78–0.95 0.002 (d) 14-day mortality 1800/10818 (16.6) 1068/6584 (16.2) 732/4234 (17.2) 0.88 c 0.82–0.95 0.002 (e) In-hospital mortality 3308/10818 (30.5) 2041/6584 (31.0) 1267/4234 (29.9) 0.89 c 0.83–0.96 0.005 (f) changing the definition of exposure and comparison d (1) 892/3539 (25.2) 167/671 (24.8) 725/2868 (25.2) 1.02 a 0.86–1.21 0.78 (2) 589/2407 (24.4) 167/671 (24.8) 422/1736 (24.3) 0.86 a 0.71–1.04 0.14 (3) 1147/4604 (24.9) 725/2868 (25.2) 422/1736 (24.3) 0.88 a 0.77–1.00 0.052 (g) propensity score-matched population (1) caliper width of 0.1 of SD 1598/6788 (23.5) 746/3394 (21.9) 852/3394 (25.1) 0.89 0.82–0.97 0.013 (2) caliper width of 0.2 of SD 1765/7432 (23.7) 840/3716 (22.6) 925/3716 (24.8) 0.91 0.84–0.99 0.03 CI confidence interval, SD standard deviation, CRRT continuous renal replacement therapy, PMX-DHP polymyxin B immobilized fiber column direct hemoperfusion, VA-ECMO venoatrial extracorporeal membrane oxygenation; IABP, intra-aortic balloon pumping § P for interaction a Adjusted for sex, Charlson comorbidity index, admission year, ambulance use, teaching hospitals, emergency charge, hospital beds, patients from nursing home, source of infection, drainage, surgery, blood culture test, urinary chemistry test, mechanical ventilation, CRRT, PMX-DHP, VA-ECMO, catecholamines, vasopressin, use of two or more catecholamines, transfusions (red blood cell, platelet, fresh frozen plasma), albumin, globulin, sedatives drugs, opioids drugs, recombinant thrombomodulin, antithrombin III, and hydrocortisone b Adjusted for age and the same covariates as *. Among these, procedure and source of infection that fell into each subgroup were excluded from the covariates c Adjusted for age and the same covariates as * d (1) "ICU management fee 1" vs. "ICU management fee 3" and "Emergency and critical care unit management fee 2", (2) "ICU management fee 1" vs. "Emergency and critical care unit management fee 1", (3) "ICU management fee 3" and "Emergency and critical care unit management fee 2" vs. "Emergency and critical care unit management fee 1" e The "number of events" indicates deaths within 30 days of hospitalization, except for "14-day mortality", which indicates deaths within 14 days Discussion In this study, we investigated the association between ICU admission and 30-day mortality in patients with septic shock. In this large-scale cohort study in Japan, approximately 40% of the patients with septic shock were admitted to HDUs. Compared with HDU admission, ICU admission showed high-frequency administration of intravenous drugs, blood transfusions, and blood products. Moreover, the frequency of mechanical ventilation, renal replacement therapy, and device placement was also higher in the ICU group. Survival analysis showed that ICU admission had lower 30-day mortality than HDU admission. Subgroup analyses of patients who underwent VA-ECMO and those with hepatobiliary infections showed that ICU admission results in better outcomes than HDU admission. Sensitivity analyses also showed results consistent with the primary analysis. These results provide valuable insights into the treatment locations of patients with septic shock. International guidelines for septic shock are based on studies that focus on ICU treatment, while treatment outside ICUs has not been well evaluated [ 6 ]. Therefore, there is only limited available evidence regarding where patients with septic shock should be treated. To the best of our knowledge, this is the first study to compare the outcomes of patients with septic shock in ICUs and HDUs. Like our study, some observational studies using the Japanese DPC database have compared the clinical outcomes of ICUs and HDUs. Miki et al. reported that patients with acute myocardial infarction who were admitted to ICUs showed lower 30-day mortality than those admitted to non-ICUs [ 15 ]. Iwashita et al. also reported that the in-hospital mortality of patients who underwent mechanical ventilation in ICUs was lower than that of patients who underwent mechanical ventilation in HDUs [ 16 ]. In their study's subgroup analysis of patients with sepsis, even though the ICU group had more patients requiring renal replacement therapy and device placement, suggesting that the severity of patients was higher in the ICU group, in-hospital mortality was also lower in the ICU group. In contrast, a recent propensity score-matched analysis by Ohbe et al. found that in-hospital mortality did not differ for patients with acute heart failure admitted to ICUs from that of patients admitted to high-dependency care units [ 17 ]. The reason for the better outcomes in the ICU group in the two previous studies and our study is unclear. However, it has been reported that patients treated in certified ICUs have better outcomes than those treated in non-certified ICUs and that high sepsis bundle adherence is associated with improved survival [ 18 , 19 ]. Additionally, our sensitivity analysis by intensive care unit category showed that patients outcomes were better in the units with more nurses. Therefore, we can speculate that several factors influence these results: the number of intensivists and certified nurses, especially nursing staff, and the quality of care provided in the ICUs, including prevention of complications and high sepsis bundle compliance. Based on the results of our study, it seems that patients with septic shock should be admitted to ICUs for intensive care. However, there are currently not enough intensivists and intensive care beds in Japan. Although the number of beds is rapidly increasing in response to the COVID-19 pandemic, the Japanese Society of Intensive Care Medicine reported that there were 7015 ICU beds and 13,003 HDU beds nationwide in 2020, with an overall intensive care bed count of approximately 15.9 beds per 100,000 population [ 20 ]. In contrast, the number of intensive care beds per 100,000 population in the United States and Germany was 34.7 (as of 2009) and 29.2 (as of 2010), respectively [ 21 , 22 ]. Moreover, Japan has only 2115 certified intensivists (as of April 1, 2021) compared with approximately 12,000 certified intensivists in the United States. Further, intensivists in Japan tend to be unevenly distributed in some urban areas [ 20 , 23 ]. Therefore, it is not feasible to treat all patients with sepsis in the ICUs. To improve clinical outcomes in HDUs, we propose the following interventions for non-intensivists and non-certified nurses involved in intensive care: disseminate standardized treatments described in the guidelines, provide education using off-the-job training like Fundamental Critical Care Support (FCCS), and provide medical support systems such as tele-ICU. Subsequently, building a treatment system in each region should allow for better determination of which patients should be treated in the ICUs. Our study had several limitations. First, some critical data were unavailable owing to the nature of the database designed for billing purposes. There is no information on the specialty of attending physicians, hospital volume, or commonly used severity scores based on physiological and laboratory parameters such as the Sequential Organ Failure Assessment (SOFA) score or Acute Physiology and Chronic Health Evaluation (APACHE) II score. These factors may be unmeasured confounders and influence the outcomes. In open ICUs, non-intensivists often provide treatment, while closed ICUs exist only in large hospitals, such as university hospitals. The database does not provide information on whether each patient was admitted to either open or closed ICUs. As mentioned above, some ICUs are staffed by certified intensivists; therefore, patients in closed ICUs are more likely to be treated by specialists. However, in other ICUs and HDUs, patients are more likely to be treated by nonspecialists. Furthermore, it has been reported that the hospital case sepsis volume influences the outcomes [ 24 ]. Since our database does not include information on where each patient was hospitalized, we do not have access to the annual number of cases. Compared to hospitals with only HDUs, case volume in hospitals with ICUs can be expected to be higher. In such cases, the effect of ICU admission on outcomes may be overestimated. Although we adjusted the severity with demographic data and the treatments performed at the time of hospitalization based on previous studies [ 17 , 25 ], there may have been a difference in severity caused by these unmeasured confounders. Second, misclassification of the included patients may have occurred due to uncertainty in the validity of the adopted ICD-10 codes as our eligibility criteria. These codes were partially modified versions of the ICD-9 codes used in a previous validation study, revealing that the sensitivity and specificity for severe sepsis (former criteria) in ICUs were 65% and 88%, respectively. Our study attempted to improve the validity by adding intravenous antibiotics and vasoactive agents to the inclusion criteria. Additionally, we performed several sensitivity analyses using different inclusion criteria and confirmed that the results were consistent. However, since there are no reports on diagnostic accuracy, the external validity of this study on the Japanese DPC database is unknown. The remaining uncertainty in the inclusion criteria is a major limitation of this study. Third, there may be confounding by indication in selecting the treatment location. As previously described, large Japanese hospitals often have ICUs and HDUs. Whether patients are admitted to ICUs or HDUs is often determined by the admission rules of each hospital. For example, patients from the emergency room are admitted to the ICU, whereas patients after emergency surgery are admitted to the HDU. Although the database does not include information on hospitals where each patient was hospitalized, we obtained additional information from the MDV on whether the hospital had both ICU and HDU after 2018. This additional information is somewhat inaccurate because units may be converted during the observation period, resulting in the misclassification of ICUs and HDUs. However, a sensitivity analysis was performed on a limited population hospitalized in a hospital with both ICUs and HDUs, and the results were consistent with the primary analysis. Fourth, there is a problem regarding non-informative censoring. In Japan, acute care hospitals are required to decrease hospital lengths of stay for bed control reasons, so patients whose conditions have stabilized are often transferred to skilled nursing and rehabilitation centers, particularly from inpatient units within large Japanese hospitals. Logistic regression analysis showed that patients in the ICU group were more frequently transferred to other hospitals than patients in the HDU group. This larger number of censored cases transferred to other hospitals in the ICU group may have missed deaths after transfer and resulted in an underestimation of mortality for the ICU group. Finally, extrapolation of results should be performed with caution. The overall in-hospital mortality in both groups was approximately 30%, which is higher than the mortality reported in a previous study of patients with sepsis in Japan but consistent with that reported in a systematic review of patients with septic shock [ 26 , 27 ]. We suspect this is because our inclusion criteria required the use of intravenous vasoactive agents to identify patients with shock and thus included patients with higher severity. Therefore, we suggest that the results of our study can only be extrapolated to critically ill patients with septic shock. Conclusions In this retrospective cohort study, ICU admission for patients with septic shock was associated with lower mortality than HDU admission. Further investigations are required to develop an optimal sepsis treatment system.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3586164/

From molecules to dynamic biological communities

Microbial ecology is flourishing, and in the process, is making contributions to how the ecology and biology of large organisms is understood. Ongoing advances in sequencing technology and computational methods have enabled the collection and analysis of vast amounts of molecular data from diverse biological communities. While early studies focused on cataloguing microbial biodiversity in environments ranging from simple marine ecosystems to complex soil ecologies, more recent research is concerned with community functions and their dynamics over time. Models and concepts from traditional ecology have been used to generate new insight into microbial communities, and novel system-level models developed to explain and predict microbial interactions. The process of moving from molecular inventories to functional understanding is complex and challenging, and never more so than when many thousands of dynamic interactions are the phenomena of interest. We outline the process of how epistemic transitions are made from producing catalogues of molecules to achieving functional and predictive insight, and show how those insights not only revolutionize what is known about biological systems but also about how to do biology itself. Examples will be drawn primarily from analyses of different human microbiota, which are the microbial consortia found in and on areas of the human body, and their associated microbiomes (the genes of those communities). Molecular knowledge of these microbiomes is transforming microbiological knowledge, as well as broader aspects of human biology, health and disease. Introduction: the revolution in DNA sequencing provides new insight into a range of microbial phenomena Microbial ecology used to be a small and specialized field that struggled to identify more than a tiny proportion of the Earth's microbial biodiversity. Part of the problem was due to the prevalence of pure-culture methods, in which microorganisms had to be removed from their natural environments (which included communities of other organisms) and cultured in laboratories. Recent advances in molecular techniques, sequencing technologies and computational methods have enabled researchers to explore the microbial world at unprecedented levels, with a focus on the natural habitats of microorganisms. The combination of these advances has so far produced remarkable insight into the role of microorganisms in human health and their powerful effects on the natural world, while at the same time developing novel evidence about the evolution and diversification of life on Earth. In this article, we discuss how these advances have allowed researchers to create new lines of inquiry, we summarize important biological and philosophical results from recent publications, and we discuss how our improved understanding of microbial ecology may affect our lives in the coming years. The last decade has seen a transformation and democratization of DNA sequencing (Shendure and Ji 2008 ). High-throughput sequencing, of the type necessary to characterize the complex microbial communities that inhabit our bodies, used to be the exclusive province of a few large sequencing centers: only research groups with access to substantial resources could engage in sequencing projects. Now, a benchtop machine that fits in an individual investigator's laboratory can produce billions of 100-nucleotide sequences per month. For comparison, a bacterial genome from the gut is typically about three million nucleotides and the human genome is about three billion nucleotides. However, the number of bacterial genomes that inhabit a human implies that they contribute far more genes than does our human genome (Turnbaugh et al. 2007 ). Playing music from a digital file once required a high-end workstation but can now be performed on a handheld device because transistors can now be packed more densely onto a microchip. In the same way, characterizing the types (e.g., the strains, species or phyla) of microbes present in a given sample (the microbiota) or the genes present in these microbes (the microbiome) are problems that can be addressed with a fixed amount of sequencing that is rapidly becoming cheaper and more accessible. These transformations in sequencing technology have correspondingly changed what it means to undertake a sequencing project. When sequences were very expensive (in the late 1970s and early 1980s), it was a substantial accomplishment to sequence even one gene from one species. Correspondingly, the focus was on identifying genes that acted as the best phylogenetic markers. These were short fragments of sequences from which inferences about the patterns of evolution were likely to match the inferred patterns of evolution of the corresponding species. These markers therefore provided efficient readouts of evolutionary history while minimizing sequencing costs. For example, ribosomal RNA genes, which play essential structural and catalytic roles in the ribosome and are thought to be almost exclusively vertically transmitted (Lawrence 1999 ; Amann et al. 1995 ), have been especially useful for reconstructing phylogenetic trees, including phylogenetic trees of organisms that have not been isolated in pure culture (Pace 1997 ). Initial studies focused on the 5S rRNA gene (Woese and Fox 1977 ), although expansion to longer rRNA genes, notably the small subunit rRNA, has allowed substantially greater phylogenetic resolution (Lane et al. 1985 ; Winker and Woese 1991 ). Here we describe several conceptual changes deeper sequencing has led to already, and will refine in the future. From catalogs to robust, reproducible community patterns The initial focus on cataloging the rRNA genes in individual species allowed phylogenies of known taxonomic groups to be reconstructed. This work provided the framework for our initial understanding that life on Earth falls into at least three distinct lineages: the Archaea, the Bacteria, and the Eukarya (initially described as the archaebacteria, the eubacteria, and the urkaryotes, respectively) (Woese and Fox 1977 ). These findings, which focused on sequencing DNA from known species, were soon complemented by a radical idea: that these phylogenetic marker genes could be isolated from unknown species via bulk DNA extraction directly from the environment. This technique, pioneered by the Pace lab (Pace et al. 1986 ), allowed researchers to start building catalogs of the known and unknown organisms, the DNA of which was present in any given environment. As the cost of sequencing DNA declined, the focus on sequencing single marker genes such as the 16S rRNA gene expanded to include shotgun metagenomic surveys, in which total DNA extracted from a sample is fragmented and sequenced. Both approaches are widely employed today. Marker-gene surveys are used to investigate the microbiota of a sample, and metagenomic surveys are used to investigate both the microbiota and the microbiome of a sample. These two views of microbial communities can yield different findings, because functional genes are frequently transferred horizontally (i.e., between different lineages). In contrast, rRNA genes are almost always transferred vertically. However, several recent studies have shown similar patterns emerging from studies involving both types of data (Turnbaugh et al. 2009a ; Fierer et al. 2012b ; Harris et al. 2013 ). The 26 years of sequencing since Pace's first community sequencing efforts have revealed a picture of 85+ phyla within the bacteria alone, and in some cases as many as 15 new candidate phyla have been detected in a single study (Ley et al. 2006 ; Harris et al. 2013 ). The bacterial and archaeal census has been estimated to reach as many as 10 6 –10 9 species (Schloss and Handelsman 2004 ), when calculated using sequence similarity criteria. Robust patterns of microbial community composition have now been observed, in a wide range of host-associated and free-living contexts. For example, human body sites are highly distinct from one another and highly diverse among individuals (Costello et al. 2009 ; HMP-Consortium 2012 ). Although any two humans are >99 % identical in their genome composition (Venter et al. 2001 ), there are no species-level OTUs (operational taxonomic units) shared across the gut microbial communities of all humans (Yatsunenko et al. 2012 ). This lack of shared OTUs suggests that many of the phenotypic differences that we see between humans may arise from differences in our microbiota, rather than differences in our genomes. We suspect that this observation will drive many advances in medicine in the coming years. For example, lean and obese individuals differ systematically in their gut microbial communities (Ley et al. 2006 ; Turnbaugh et al. 2009a ; Knights et al. 2011 ) but much less so in their genomic composition. Obesity can be identified 90 % of the time using the bacteria in the feces alone (Knights et al. 2011 ), but with only 58 % accuracy from variations in the genomes of different individuals (Sandholt et al. 2010 ). Similarly surprising insights have arisen in environmental microbiology. For example, pH has been found to be the main driver of microbial communities in soil (Lauber et al. 2009 ; Rousk et al. 2010 ; Chu et al. 2010 ; Fierer et al. 2012a ), and salinity plays a crucial role in structuring both free-living bacterial and archaeal communities across many environments (Lozupone and Knight 2007 ; Caporaso et al. 2011b ; Tamames et al. 2010 ; Auguet et al. 2010 ). These patterns can be striking: for example, seasonal patterns in marine water microbial diversity are highly reproducible in the Western English Channel (Gilbert et al. 2012 ), with the same organisms dominating microbial communities in different seasons annually. However, most of the organisms present in any given season are found even at just a single time-point if more sequences (millions rather than thousands) are collected from the sample (Caporaso et al. 2012 ). These results suggest that seasonal differences do not arise from the presence or absence of community members, but rather from variations in the abundance of organisms that are always present. This finding reinforces the point that much of what we think we know about the microbial world may be limited by the amount of sequencing that it is cost-effective to perform. The work to catalog Earth's microbial diversity has thus produced a compendium of rich and detailed observations, and efforts such as the Earth Microbiome Project (Gilbert et al. 2010 ; Knight et al. 2012 ) will round out our encyclopedia of our microbial world. But cataloging alone is insufficient: a list of the species present in a rainforest, for example, speaks little to the interactions, functions or potential of the organisms so listed. The problem with phylogenetic marker gene surveys, such as the 16S rRNA gene sequencing projects described above, is that they tell us the 'who', without the 'how', thus failing to answer the most pressing questions. For instance, how can an organism live at pH 0 (Edwards et al. 2000 ), and what can such capacities teach us about the potential for pollution mitigation or for life on other planets? Endeavors such as the Genomic Encyclopedia of Bacteria and Archaea (GEBA) (Wu et al. 2009 ) perform whole-genome sequencing on organisms that are as phylogenetically divergent as possible from previously sequenced organisms. Even a small amount of this phylogenetically targeted genome sequencing provides novel gene discovery that greatly outpaces gene discovery from organisms chosen arbitrarily or at random. Targeted sequencing can inform us about the reproducibility of the evolutionary process among organisms from different lineages that adapt to similar environments. For example, comparative genomics based on whole-genome data, and linked to rich evolutionary history and detailed environmental information (derived from marker gene databases and marker gene surveys, respectively), can offer insights into which types of biochemical or regulatory functions are necessary to survive in a given environment. These results enable an understanding of the systems biology of microbial communities, which can ultimately be applied to engineer microbial communities to treat disease, generate electricity, or clean up hazardous waste sites. However, marker gene surveys still improve our understanding of microbial ecology and enable novel findings and technological applications. We will focus on this technique for the remainder of the paper to show how this is the case. How do we know which microbes are present? A key problem with studies of the microbiome lies in determining which organisms are present. All stages of the process, including DNA extraction, amplification of specific target genes, clustering of sequences, and identification of taxonomic group are prone to both error and bias (Hamady and Knight 2009 ). As the number of sequences involved in a given study has grown, reliance on advanced computational methods has increased (Gonzalez and Knight 2012 ). However, the algorithm that is chosen can have large impacts both on beliefs about what organisms are present in a given environment (Liu et al. 2008 ) and how many kinds of organisms are present (Kunin et al. 2010 ; Quince et al. 2009 ). Even defining kinds of organisms is complicated at the microbial level. In lieu of a robust definition of a microbial species (Cohan 2002 ), the percentage of sequence identity of a marker gene is often used to define operational taxonomic units or OTUs. For example, most 16S rRNA gene-based studies treat a cluster of sequence fragment 'reads' (the output of a DNA sequencing instrument, and thus the typical observation in studies of microbial communities) that are >97 % identical to one another as members of the same OTU. 97 % identity is treated as a proxy for species-level groupings of organisms, although this definition is known to be problematic for several reasons. One is that the rate of evolution of the 16S rRNA gene differs among taxonomic lineages, so the same number of differences in the sequence may represent different times since divergence from a most recent common ancestor. The choice of algorithm for assigning sequences to OTUs can also have a large impact on which sequences are clustered into the same OTU and on how many OTUs are observed in a study. For example, it is not clear whether a 97 % sequence identity threshold means that each sequence added to an OTU must be 97 % similar to all other sequences in the OTU cluster, or whether each sequence should be 97 % similar to the sequence that defines the center of the cluster (i.e. the cluster centroid) (Schloss and Handelsman 2005 ; Schloss and Westcott 2011 ). Because neither laboratory nor computational protocols are standardized, reported differences among studies often stem from differences in methodologies rather than from differences in the underlying biology. And because techniques for performing meta-analyses of microbiome data are still only emerging, it is often difficult to standardize a reanalysis, and comparisons of results across studies and especially among laboratories must be performed with caution. Modern marker-gene-based studies often investigate the composition of microbial communities at the OTU level, due to difficulties in relating counts of short DNA sequence fragments to named species. Although short reads of sequences (100–400 bases is currently typical, depending on sequencing platform) from the genomes of well-studied organisms can often be assigned at least to the family level, and sometimes at the genus or species level, many sequences cannot confidently be assigned to known named taxonomic groups. The limitation here is primarily the amount of information present in short reads of marker genes for differentiating closely related taxa. Figure One shows that when working with the most informative region of the 16S rRNA gene for broad analyses of bacterial and archaeal communities, the fraction of reads that can be assigned to taxonomic groups increases as expected with the length of the sequence. In real-world experiments (as opposed to the simulation presented in Fig. 1 ) this effect is exacerbated by PCR and sequencing biases and errors. Fig. 1 Relationship between sequencing read-length and our ability to classify sequences using the RDP Classifier, a popular taxonomic assignment method based on oligonucleotide frequencies (Wang et al. 2007 ). Simulated sequences were generated from 16S genes to represent the complete sequence between the 515F/806R primers (the "full amplicon") or shorter 150 or 100 base pair reads from the 515f forward primer Our inability to assign detailed taxonomy to short reads is often not important for many of the questions that are interesting to address at the community level. Phylogenetic diversity calculations allow us to determine the relative similarity of microbial communities, using similarity of the fragment of the marker gene as a proxy for the relatedness of the organisms represented by those marker genes. Although in principle horizontal gene transfer, the movement of genes among different genomes, could obscure the phylogenetic pattern, in practice the difference in gene content between two organisms closely tracks the differences in marker genes such as the 16S rRNA gene (Zaneveld et al. 2010 ; Konstantinidis and Tiedje 2005 ). However, there are cases in which genomes with identical 16S rRNA genes have markedly different properties (e.g., Bacillus cereus, a harmless soil bacterium, and Bacillus anthracis , the causative agent of anthrax, are almost indistinguishable except for a plasmid that confers pathogenicity (Ivanova et al. 2003 )). Additionally, our conclusions are limited by our depth of sequencing (i.e., the number of marker gene sequence reads collected from a sample). A study that collects 1,000 sequences per sample will miss species that are only present at an abundance of one cell in a million. These limitations to knowledge are widely appreciated by specialists, but are often omitted in popular accounts and in descriptions for non-specialists. Is there a core human microbiome? Our initial expectations of the microbial diversity living within and on human beings were limited and biased because relatively few microbes can be grown in culture (Rappé and Giovannoni 2003 ; Staley and Konopka 1985 ) and because many phylogenetically and functionally distinct kinds of microbes are difficult to distinguish by morphological or biochemical characteristics. For instance, Escherichia coli was believed to be a common and abundant gut microorganism inhabiting most members of the human population. However, culture-independent surveys based on 16S rRNA gene sequencing and/or shotgun metagenomic sequencing (in which all the DNA from a given community is extracted and analyzed) typically find it at less than 1 % abundance in the gut of healthy adults (Eckburg et al. 2005 ; Turnbaugh et al. 2009a ; Costello et al. 2009 ; Qin et al. 2010 ). The scientific and medical community sought to determine the "core" microbiome of humans at the level of microbial species shared by everyone (Turnbaugh et al. 2007 ). Surprisingly, such a core does not seem to exist at the level of species; instead what appears to be shared are microbial functions (Turnbaugh et al. 2009a ; Qin et al. 2010 ). One suggestion is that there might be a few types of common but only partially overlapping (or perhaps non-overlapping) microbial communities. One study found just three "enterotypes" or types of gut bacterial communities in human populations (Arumugam et al. 2011 ), although this simplistic picture appears not to be true when additional subjects and populations are considered (Wu et al. 2011 ; MacDonald et al. 2012 ; Jeffery et al. 2012 ; Claesson et al. 2012 ; Yatsunenko et al. 2012 ; HMP-Consortium 2012 ). However, the idea that human gut microbial communities might be classified into just a few types is conceptually appealing and has received much media attention (Brandon 2011 ; Yong 2012 ; Zimmer 2011 ), so debate on this topic is likely to continue. The microbial diversity revealed due to improvements in culture-independent techniques, in part due to the vast decrease in sequencing costs noted above, has been remarkable. There are no shared OTUs across the gut communities of all humans, even at a depth of coverage of one million sequences per sample (HMP-Consortium 2012 ). This unexpected finding has given rise to the idea of microbes as personal identification markers (Fierer et al. 2010 ). In addition, because monozygotic twins differ in their microbiota (Turnbaugh et al. 2009a ; Yatsunenko et al. 2012 ), it could be argued that our microbiota are more personally unique than our own genomes. In some sense, whether or not there is a core microbiome is a purely definitional issue. Finding a core depends on the level at which sequences are aggregated (grouping together more similar or more distantly related groups of organism, for example), the abundance threshold that may be set deliberately or may be intrinsically limited by technology or study design (for example, if only 1,000 sequences per sample are collected, organisms that are as rare as one in a million microbes will be missed), and the fraction of individuals that a taxon must appear into be considered "core" (for example, the MetaHIT consortium used a 50 % threshold (Qin et al. 2010 )). Some kind of core can always be defined. A more productive research direction is to ask whether there are systematic differences among the microbial communities of every human that can be correlated with the physiological state of each individual. Microbial community states associated with disease Much attention has focused on testing whether differences in microbial diversity correlate with physiological states, especially disease states. For example, Ruth Ley, Peter Turnbaugh and colleagues in the laboratory of Jeffrey Gordon embarked on an exploration of changes in the microbiota associated with obesity in different mouse models. This seminal work revealed robust differences in the gut communities of these mice compared with lean mice, both in the case of genetically induced obesity in the ob/ob leptin model (Ley et al. 2005 ) and in diet-induced obesity (Turnbaugh et al. 2008 ). Remarkably, increased adiposity was transmissible to genetically normal mice on a standard, calorie-controlled diet by transferring these microbial communities from the obese mice to the normal mice (Turnbaugh et al. 2006 , 2008 ). The major taxonomic difference between the mice microbiota was the relative abundance of the phyla Bacteroidetes and Firmicutes. This finding has been shown to hold for human hosts as well (Ley et al. 2006 ), although the same pattern has not been replicated in all human studies (Duncan et al. 2008 ; Schwiertz et al. 2010 ). As mentioned above, we can now predict—based on the microbial community composition alone—whether an individual is obese or lean at 90 % accuracy (Knights et al. 2011 ) while predictions based on host genomic markers perform little better than chance (Sandholt et al. 2010 ). Interestingly, these predictions work best when the microbes are classified into broad groups. Clustering the sequences into groups at the 80 % sequence identity level (corresponding approximately to bacterial phyla) actually works better than clustering the sequences into groups at the 97 % sequence identity level (corresponding approximately to bacterial species) for classifying people as lean or obese. These more detailed analyses at the species-proxy level do, however, provide better resolution when classifying multiple samples from the same site (Knights et al. 2011 ). A possible explanation for the improved predictability using phylum-level classification could be that differences in biochemical pathways are differentially represented across phyla but conserved across OTUs within phyla. These biochemical pathways are the primary features that differentiate obese from lean individuals. Models trained on data that are too specific (i.e., clustered at 97 % identity rather than a lower percent identity) are prone to overfitting, and have reduced predictive capacity. But it is important to bear in mind that the phylogenetic levels at which bacteria are associated with particular states may vary considerably, depending on the ecology of the particular phenotype or disease. Recent large-scale endeavors, such as the Human Microbiome Project (NIH 2012 ), the American Gut (Human-Food-Project 2012 ) and the Personal Genome Project (Personal-Genome-Project 2012 ) are opening up new opportunities for analysis because they are building a base of healthy microbiomic data against which disease states (collected by some of these projects) can be contrasted. This is important because of the breadth of diseases associated with the microbiome. Disease states that have been found to be associated with features of the microbiome include inflammatory bowel disease (Frank et al. 2007 ; Michail et al. 2012 ), wasting diseases (Gordon et al. 2012 ), obesity (Kallus and Brandt 2012 ), halitosis (Kazor et al. 2003 ), dental caries (Yang et al. 2012 ), and perhaps even autism (Finegold et al. 2010 ). The gut microbiome appears to be causal for certain disease states, and is not just a biomarker. Causality can be inferred when, for example, fecal transplantation (and thus microbiota inoculation) in human subjects is used successfully to treat inflammatory bowel disease (IBD—primarily ulcerative colitis) (Landy et al. 2011 ) and insulin sensitivity associated with metabolic syndrome (Vrieze et al. 2012 ). These results indicate that gut microbes play an active role in these disease states and are not merely effects of the host's condition. It is possible that in the not-to-distant future a microbiome sample will become a normal component of a health checkup. Microbiome analyses may be used to diagnose disease and could provide possible avenues for the prevention of disease through predictive tests. As we mentioned above, molecular samples from microbial communities may track or predict disease states better than does the human genome. Changes in the microbiome over time Microbial ecology shares similarities with traditional ecology, yet there are some important differences. In the ecology of macroorganisms, it is often possible to observe interactions directly, such as predation or competition for resources. Such observations are much more difficult in the microbial world, and ecological interactions must often be inferred from statistical variations in sequence data instead. Species definitions, although notoriously problematic even for macroorganisms, are even more difficult in microbes, and operational definitions based on similarities in DNA sequences must be used instead (as already discussed). Additionally, the cost of DNA sequencing posed a barrier until recently to collecting the detailed time-series and spatial datasets that are necessary for ecological modeling in microorganisms. However, some aspects of microbial ecology are substantially easier than in large-organism ecology. For example, the reliance on DNA sequence data means that with advances in technology, even a deep sampling of the population (millions of individuals) can be performed rapidly, and observation biases are likely to be less profound than when attempting to glimpse rare and elusive insects or mammals. The ability to collect large-scale information about microbial populations is likely to allow classical ecological models to be applied to the microbial world far more effectively than has been possible in macroecology, because more types of microbes can be simultaneously observed with large population sizes and with replicated sampling. Ecological principles offer more than just ways to stratify the human population (e.g., by disease state). At infancy, our microbial populations go through remarkable changes in structure prior to reaching a resemblance to most adult communities. Inoculation is not necessarily from our mothers, and is substantially influenced by delivery mode. Microbial communities of children delivered vaginally initially tend to resemble their mother's vaginal communities, while the microbial communities of children delivered by C-section initially tend to resemble human skin communities. Skin inoculations may be obtained from the mother, the medical staff involved in the delivery, or hospital surroundings (many of which harbor communities resembling human skin) (Biasucci et al. 2010 ; Dominguez-Bello et al. 2010 ). Stabilization of the microbiota of human children occurs around the third year of life (Yatsunenko et al. 2012 ), but routine disruptions, adjustments and fluctuations appear to be normal in healthy individuals (Costello et al. 2009 ; Caporaso et al. 2011a ). While in general, the intra-individual microbiome variation is less than inter-individual, the amount of variability over long time periods (Caporaso et al. 2011a ) gives rise to the idea of microbial "weather" in which microbial communities react to dietary and health conditions (even as they causally affect them). This phenomenon may be especially important in determining the health of the elderly (Claesson et al. 2012 ). A revelatory aspect to studies of the microbiome is that classical ecological models and datasets previously only obtainable for a few economically important systems, such as fisheries, are now testable on the microbial scale because of the ability to assess simultaneously the relative abundance of thousands of species in thousands of samples (Gonzalez et al. 2011 ). However, this move towards accounts of microbial communities in terms of alternative stable states and dynamical systems (Costello et al. 2012 ; Lozupone et al. 2012 ; Gajer et al. 2012 ) is not entirely without peril. In the absence of theories of underlying causes, defining the number and boundaries of these states can be technique-dependent and implicitly theory-laden in ways difficult to identify—especially by investigators who are not specialists in the relevant mathematical techniques. With the availability of larger datasets and the ability to track communities over time, key ecological concepts such as resilience, alternative stable states, predator–prey cycling, and bottom-up versus top-down regulation of ecosystems will be increasingly important. However, it is equally important not to forget the lessons learned from past applications of these techniques, especially in traditional ecological modeling. For example, it has been known for almost four decades that Lotka-Volterra predator–prey dynamics with time lags produce patterns that would appear as completely uncorrelated between two species that in fact do interact deterministically (Fig. 2 ) (Holling 1973 ). However, this fact is routinely ignored in network analyses that seek to find connections among organisms by building a network in which nodes correspond to organisms, and edges correspond to pairs of organisms that are correlated. Correlation is usually assessed by determining whether the abundances of two taxa are correlated across a set of samples, typically using the Pearson correlation coefficient that assumes that all interactions are linear. In other words, taxa are linked if their correlation coefficient exceeds an arbitrary researcher-defined threshold. These networks are often used to find groups of organisms that "co-occur", presumably because of shared environmental preferences or because of mutualistic ecological interactions. Hence these network methods, which often rely on linear correlations among organisms to detect relationships (Qin et al. 2010 ; Steele et al. 2011 ; Barberán et al. 2012 ), would incorrectly assert organisms to lack ecological connections even when these connections are fully deterministic. This happens simply because the inference procedure requires an understanding of the time-evolution of the system in order to find these causal links. Fig. 2 Predator-prey dynamics for two species X and Y lead to a scatterplot (relating sampled species abundances) that is interpretable when successive time-points are connected ( a ). If, however, the information about time were not included ( b ), these dynamics would appear uncorrelated because when X is high, Y can be either high or low, and vice versa. Thus, even in a completely deterministic system, it is impossible to tell whether two species interact with each another simply by examining multiple samples in which both are present. However, this technique is widely used in practice despite its limitations. Figure adapted from (Holling 1973 ) The analysis of time-series in microbial ecology has also been limited because the performance of standard signal processing methods is degraded with uneven sampling periods and small numbers of data points (Moller-Levet et al. 2003 ; Mason 1978 ; Mallat 1989 ). Such degradations have historically been common in microbial ecology datasets due to the cost of obtaining the data. However, we have already obtained valuable information about the temporal dynamics of a few microbial communities, such as the assembly of an infant's gut microbiome and its transition towards a healthy human adult gut microbiome (Koenig et al. 2011 ). In the few cases in which even sampling has been performed or can be assumed, techniques exist to detect abrupt disruptions (Beltran et al. 1994 ; Mallat and Zhong 1992 ). In these contexts, such disruptions could mean one of the interventions that has been shown to have large effects in mice or humans such as diet change (Turnbaugh et al. 2009b ) or antibiotic administration (Dethlefsen et al. 2008 ; Dethlefsen and Relman 2011 ). Therefore, as in disease surveillance, choosing a specific analytical approach (for example co-occurrence analysis, clustering analysis, and control systems analysis) depends to a large extent on whether the goal is to monitor a trend, detect an outbreak or provide general awareness of the possibility of change (Robertson and Nelson 2010 ). Conclusions and outlook Overall, the ability to collect far larger amounts of sequence data has led to much broader and deeper characterizations of the human microbiome and microbial communities in other habitats, especially when linked to rich contextual information about the provenance and status of each sample (Knight et al. 2012 ). In particular, the increased use of time-series studies (enabled by the decline in the cost of sequencing) allows us to apply for the first time a wide range of ecological models to the microbial world. Perturbation experiments are especially important for understanding how microbial communities change and for understanding groups of species that change together and interact in complex ways. However, this expanded body of ecological data introduces substantial epistemic issues, especially in regard to how data are interpreted via models and concepts. For example, the definition of OTUs at both the organism and the gene level (e.g. in the construction of "gene catalogs" (Qin et al. 2010 )) is in many respects a return to phenetic methods, which have been criticized due to their lack of theoretical justification and their instability when more data are added (de Quieroz and Good 1997 ). The methodological principle of clustering sequences at some threshold before analysis is also not well grounded theoretically. One example would be if a single nucleotide change in the 16S rRNA gene of a single species distinguished exactly lean from obese humans, or co-varied perfectly with disease severity in IBD. Such findings would be of enormous importance yet would be missed completely by current techniques. Similarly, we know that because of factors such as horizontal gene transfer, gene- and taxon-level analysis will not map precisely on to each another, yet the data to perform such analysis and the theoretical framework for reconciling differences is at this point largely lacking. Some of the solutions to these problems are being sought in large-scale projects such as the Earth Microbiome Project (Gilbert et al. 2010 ; Knight et al. 2012 ). These research consortia are working towards understand relationships among microbial processes across different systems and timescales. They will be especially important for identifying which theoretical constructs across different scales and levels of analysis are especially useful both for understanding and predicting microbial community responses. And as this article has made clear, the availability of large datasets and the development of new methods with which to analyze them have already produced dramatic changes in how the microbial world is understood, and its relationship to the rest of the biological world. As the many human microbiome studies discussed above show, microbial ecology—especially molecular microbial ecology, even at its relatively crude stage of development—is transforming how human biology itself is understood. This transformation, which we expect to occur not just in human biology but in traditional ecology and biology more broadly, will raise philosophical issues that require the attention of scientists and philosophers. We have indicated just some of these issues, dealing with the units of analysis and the causal powers associated with them, and how imperfect methods and models become more refined and effective in the process of inquiry. Philosophy of biology itself can learn a great deal from these recent and future developments in microbial ecology, as other papers in this special issue demonstrate.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3128999/

Detrimental effects of anti-apoptotic treatments in spinal cord injury

Long term functional impairments due to spinal cord injury (SCI) in the rat result from secondary apoptotic death regulated, in part, by SCI-induced decreases in protein levels of the anti-apoptotic protein Bcl-x L . We have shown that exogenous administration of Bcl-x L spares neurons 24h after SCI. However, long term effects of chronic application of Bcl-x L have not been characterized. To counteract SCI-induced decreases in Bcl-x L and resulting apoptosis, we used the TAT protein transduction domain fused to the Bcl-x L protein (Tat-Bcl-x L ), or its anti-apoptotic domain BH4 (Tat-BH4). We used intrathecal delivery of Tat-Bcl-x L , or Tat-BH4, into injured spinal cords for 24h or 7 days, and apoptosis, neuronal death and locomotor recovery were assessed up to 2 months after injury. Both, Tat-Bcl-x L and Tat-BH4, significantly decreased SCI-induced apoptosis in thoracic segments containing the site of injury (T10) at 24h or 7 days after SCI. However, the 7 day delivery of Tat-Bcl-x L , or Tat-BH4, also induced a significant impairment of locomotor recovery that lasted beyond the drug delivery time. We found that the 7 day administration of Tat-Bcl-x L , or Tat-BH4, significantly increased non-apoptotic neuronal loss and robustly augmented microglia/macrophage activation. These results indicate that the anti-apoptotic treatment targeting Bcl-x L shifts neuronal apoptosis to necrosis, increases the inflammatory response and impairs locomotor recovery. Our results suggest that a combinatorial treatment consisting of anti-apoptotic and anti-inflammatory agents may be necessary to achieve tissue preservation and significant improvement in functional recovery after SCI.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2841224/

Prolyl 4-hydroxylase

Posttranslational modifications can cause profound changes in protein function. Typically, these modifications are reversible, and thus provide a biochemical on–off switch. In contrast, proline residues are the substrates for an irreversible reaction that is the most common posttranslational modification in humans. This reaction, which is catalyzed by prolyl 4-hydroxylase (P4H), yields (2 S, 4 R )-4-hydroxyproline (Hyp). The protein substrates for P4Hs are diverse. Likewise, the biological consequences of prolyl hydroxylation vary widely, and include altering protein conformation and protein–protein interactions, and enabling further modification. The best known role for Hyp is in stabilizing the collagen triple helix. Hyp is also found in proteins with collagen-like domains, as well as elastin, conotoxins, and argonaute 2. A prolyl hydroxylase domain protein acts on the hypoxia inducible factor α, which plays a key role in sensing molecular oxygen, and could act on inhibitory κB kinase and RNA polymerase II. P4Hs are not unique to animals, being found in plants and microbes as well. Here, we review the enzymic catalysts of prolyl hydroxylation, along with the chemical and biochemical consequences of this subtle but abundant posttranslational modification.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8874468/

A Comparison with Adverse Events Following Immunization Associated with Sabin-Strains and Salk-Strains Inactivated Polio Vaccines in Zhejiang Province, China

Objectives: One dose of Sabin-strains inactivated polio vaccine (IPV) was introduced into the Chinese immunization program on 1 May 2016. This study aimed to evaluate the safety of Sabin-strains IPV and provide a comparison with conventional Salk-strains IPV. Methods: Adverse events following immunization (AEFI) records associated with Sabin-strains IPV and Salk-strains IPV were extracted from the national AEFI surveillance system (NAEFISS) from 1 May 2016 to 31 December 2020. The vaccination information on Sabin-strains IPV and Salk-strains IPV during the same period was obtained from the Zhejiang provincial immunization information system. Reporting rates of AEFI were calculated by age, city, severity of AEFI, categories of AEFI, and reaction categories and were compared between Sabin-strains IPV and Salk-strains IPV. Results: In total, 3,861,758 doses of Sabin-strains IPV and 1,018,604 doses of Salk-strains IPV were administered during the study period. The overall AEFI reporting rate for Sabin-strains IPV (3.96/10,000 doses) was significantly lower than that for Salk-strains IPV (5.03/10,000 doses) due to the reporting rate of the minor vaccine product-related reaction following Sabin-strains IPV was significantly lower than that for Salk-strains IPV (2.76/10,000 doses vs. 3.83/10,000 doses). The most frequently reported symptoms/signs were fever, induration/swelling, and rash/urticaria. The most frequently reported serious AEFI with a causal relationship was febrile convulsion, with the reporting rates of 0.10/10,000 doses for Sabin-strains IPV and 0.08/10,000 doses for Salk-strains IPV. No significant difference was found in the reporting rates of the other serious AEFI between the two types of IPV. Conclusion: Most of the AEFI following Sabin-strains IPV and Salk-strains IPV were mild and common adverse reactions. The reporting rate of serious AEFI was not significantly different between Sabin-strains IPV and Salk-strains IPV. Sabin-strains IPV had a favorable safety profile and could be widely used. 1. Introduction The strategy of switching from the trivalent oral live attenuated polio vaccine (tOPV) to the bivalent OPV (bOPV), withdrawing the type II strain, was recommended by the Polio Eradication and Endgame Strategic Plan 2013–2018. Simultaneously, the World Health Organization (WHO) also recommended that OPV-using countries should arrange at least one dose of inactivated polio vaccine (IPV) into the routine polio vaccine schedule to prevent the possible outbreaks of polio virus type II or the circulation of vaccine-derived polio virus type II [ 1 ]. Based on these strategies, China switched from using tOPV to using bOPV in the routine immunization program and introduced one dose of IPV into the polio vaccine schedule on 1 May 2016, which was synchronized with more than 150 other countries using OPV in their national immunization programs [ 2 ]. The new Chinese polio vaccine schedule included one IPV dose given at 2 months of age, followed by three bOPV doses at 3, 4, and 48 months of age [ 3 ]. The China drug administration approved two types of IPV [ 3 ]. One was manufactured from Salk-strains and had been available as a voluntary vaccine subject to payment since 2009. Another was manufactured from Sabin-strains and was licensed in 2015. Salk-strains IPV or Salk-strains containing combination vaccines have been widely used in the USA and other developed countries and their safety and effectiveness have been proved [ 4 , 5 ]. However, Salk-strains IPV could not be widely used in many developing countries due to its high biosafety risks during production and the limited production capacity and its relevant high cost [ 6 ]. WHO encouraged the research and development of new types of IPV using less virulent virus strains, such as Sabin-strains IPV [ 7 ]. Up to date, only China and Japan license and use Sabin-strains IPV. Immunogenicity and safety of Sabin-strains IPV and the Sabin-strains IPV-OPV sequential schedule have been evaluated by clinical trials [ 8 , 9 , 10 ]. However, the clinical trials were not powered sufficiently to detect the infrequent or rare adverse events following immunization (AEFI) due to the limited sample size. Continuous assessment of the safety analysis of the post-licensure vaccines can provide a tool to evaluate the benefit–risk profiles of a specific vaccine which cannot be evaluated in pre-licensure clinical trials. Additionally, post-licensure AEFI surveillance enables the detection of signals that will generate hypotheses, as well as the identification and rectification of gaps in this program to strengthen the routine vaccination [ 11 , 12 ]. Before the switch of polio vaccine schedule, only a few pilot studies (replaced the first dose of tOPV with Sabin-strains IPV) reported the results on the post-licensure AEFI surveillance of Sabin-strains IPV [ 13 , 14 , 15 ]. We also know that limited data were available regarding the safety of Sabin-strains IPV after its licensure, and consequently, during its large-scale use. Continuous surveillance on the safety of Sabin-strains IPV based on a large population helped to capture the rare AEFI or the unexpected AEFI, which could evaluate the safety profile of Sabin-strains IPV in a comprehensive manner. Our study aimed to assess the safety of Sabin-strains IPV. Specifically, all AEFI associated with Sabin-strains IPV in Zhejiang province were summarized from the national passive surveillance data collected in the first 4.5 years (from 1 May 2016 to 31 December 2020) after the introduction of Sabin-strains IPV into the routine immunization program, with a comparison with the Salk-strains IPV during the same period. This study provides a reference for the widespread use of Salk-strains IPV in the developing countries at the endgame of polio eradication. 2. Materials and Methods 2.1. Study Area Zhejiang is a developed province with a large population of 70 million people in eastern areas of China. Of the total population, 7.23% children were under 7 years of age. Zhejiang province launched the EPI in 1978 with four vaccines and it continued to increase the number of vaccines up to 11 to date, with the administration of 20 million doses of vaccines each year. 2.2. Product Information The poliovirus strains used by Salk-IPV are wild-type polioviruses including Mahoney (type 1), MEF-1 (type 2), and Saukett (type 3). These virus strains are inoculated in Vero cells and inactivated with formalin. Every dose of Salk-IPV contains the D-antigens of 40DU, 8DU, and 32DU for type 1, 2, and 3, respectively. Other ingredients include 2-phenoxyethanol, ethanol, methanol, sodium hydroxide, hydrochloric acid, neomycin, streptomycin, and polymyxin B. The final product is packaged with a prefilled single syringe and is valid for 36 months if it is stored under the condition of 2–8 °C. The poliovirus strains used by Sabin-IPV are Sabin virus type 1 and type 2 and Pfizer (type 3). These virus strains are inoculated in Vero cells and inactivated with formalin. Every dose of Sabin-IPV contains the D-antigens of 30DU, 32DU, and 45DU for type 1, 2, and 3, respectively. Other ingredients include 2-phenoxyethanol, sodium hydroxide, hydrochloric acid, neomycin, and streptomycin. The final product is packaged in a Cilin bottle as one dose per vial and is valid for 24 months if it is stored under the condition of 2–8 °C. 2.3. Polio Vaccines Schedule in Zhejiang Province Since 1 May 2016, one dose of Sabin-strains IPV has been included in the Zhejiang province routine immunization program to be given at 2 months of age. On 1 November 2019, two doses of Sabin-strains IPV were included to be given at 2 months and 3 months of age, respectively. The Sabin-strains IPV was funded by the government and provided to all eligible children. However, every eligible child also had a supplementary option to choose the self-paid Salk-strains IPV to replace the two doses of Sabin-strains IPV and even the following two doses of bOPV. The alternative schedule of the Salk-strains IPV included four doses at 2, 3, 4, and 18 months of age. 2.4. National Adverse Events following Immunization Surveillance System The national adverse events following immunization surveillance system (NAEFISS) was an official immunization safety surveillance system that was established by the Chinese Center for Disease Control and Prevention based on WHO guidance [ 16 ]. NAEFISS aims to detect new, unusual, or rare AEFI, to evaluate the safety of newly licensed vaccines, to identify potential risk factors for AEFI, to monitor increases in known AEFI, to determine the possible reporting clusters, and to provide a reliable safety monitoring system. Zhejiang province joined the NAEFISS and started systematic surveillance in 2009. NAEFISS was upgraded in 2012 for adding variables of the case reporting form and rules of data logic verification. 2.5. AEFI Reporting and Investigation Procedures The national AEFI guidance was released in 2010 [ 16 ]. An AEFI case is defined as a reaction or an event occurring after vaccine administration that is suspected to be related to the vaccination. According to the national AEFI guidance, healthcare facilities, vaccination clinics, CDC at each administrative level, adverse drug reaction monitoring agencies, and vaccine manufacturers are required to report AEFI cases. Additionally, public members or guardians or parents can also notify any of the authorized reporters mentioned above to report an AEFI case. After identifying an AEFI case, all the above-authorized reporters should report it to the vaccination clinic or the county-level CDC within the jurisdiction. The vaccination clinic or the county-level CDC then completes an "AEFI Case Reporting Card" and submits the data to NAEFISS. The investigation is required for all AEFI, except for the non-serious vaccine product-related reaction with a clear diagnosis (e.g., fever, erythema, swelling, and induration at the injection site). If necessary, an AEFI case is investigated by the county-level CDC within the jurisdiction. In case of deaths, serious AEFI, AEFI clusters, and AEFI of significant public concern that are suspected to be related to vaccination, prefectural or provincial within the jurisdiction expert committees (comprising pediatricians, physicians, epidemiologists, pharmacists, vaccine researchers, etc.) are responsible for further investigation and causality assessment. The reporting form or the investigation collects the information on the vaccinated individual, storage and transportation of vaccines, vaccine administration, and the AEFI itself. Signs and symptoms of AEFI are coded using the International Classification of Diseases (version 10.0, ICD-10), a clinically validated, internationally standardized terminology. A single AEFI report may be assigned more than one term and be referred to more than one suspected vaccine. 2.6. AEFI Category An AEFI is any adverse medical occurrence that follows immunization, but which does not necessarily have a causal relationship with vaccination. According to the cause-specific classification of AEFI from WHO [ 17 ], AEFI cases are divided into five types: (1) vaccine product-related reaction (including non-serious reaction and serious reaction); (2) vaccination error (including errors in vaccine handling, errors in vaccine prescribing or non-adherence to recommendations for use, and errors in administration); (3) vaccine quality defect-related reaction; (4) coincidental event; and (5) anxiety reaction. According to the national AEFI guidance [ 16 ], a serious AEFI is defined as an event that results in death, is life-threatening, requires hospitalization or prolongs the existing hospitalization, results in persistent or significant disability, causes a congenital anomaly or birth defect, or requires intervention to prevent permanent impairment or damage. For example, serious AEFI include but are not limited to allergic shock, allergic laryngeal edema, allergic purpura, thrombocytopenic purpura, Arthus reaction, febrile convulsion, epilepsy, brachial neuritis, polyneuritis, Guillain–Barre syndrome, encephalopathy, encephalitis and meningitis, syncope, etc. 2.7. Data Source The vaccination records of Sabin-strains IPV and Salk-strains IPV (standalone only, combination vaccines were not included) administered from 1 May 2016 to 31 December 2020 in Zhejiang province were extracted from the Zhejiang provincial immunization information system (ZJIIS), which was established and maintained by Zhejiang Provincial Center for Disease Control and Prevention (ZJCDC) to collect and manage the information on vaccine procurement and distribution, vaccination clinic information, and cold chain temperature information. The function of ZJIIS can be found elsewhere [ 18 ]. Each recipient's vaccination information is submitted to ZJIIS by vaccination staff. AEFI records following Sabin-strains IPV and Salk-strains IPV administered from 1 May 2016 to 31 December 2020 in Zhejiang province were obtained from NAEFISS on 1 April 2021 to account for data lags due to the correction and cleaning. 2.8. Outcome and Data Analysis A database of the AEFI records associated with Sabin-strains IPV and Salk-strains IPV administration during the study period was organized as an Excel file (Microsoft Office Excel 2020). The characteristics of AEFI records were summarized by gender of case, causal category, type of reporter, severity, city, patient age, interval of AEFI onset (from vaccination date (day 0) to onset of first symptoms), and symptoms/signs/diagnoses. Each AEFI record might list several symptoms, signs, and/or diagnoses, but only the main symptom or the most serious diagnosis was included. The reporting rates of the general AEFI and the serious AEFI were calculated for Sabin-strains IPV and Salk-strains IPV, respectively. The reporting rate was calculated by dividing the number of AEFI reports by the relevant vaccine doses administered. We graphically depicted monthly reporting rates during the study period to display its trends. Some of the serious AEFI following Sabin-IPV and Salk-IPV were summarized by the confirmed causal relationship. The Pearson chi-square test or Fisher's exact test was used to examine differences in reporting rate between Sabin-strains IPV and Salk-strains IPV. If at least one theoretical frequency of the cross-table was 0.05), vaccination errors (0.01/10,000 doses vs. 0.01/10,000 doses, p > 0.05), and coincidental events (0.65/10,000 doses vs. 0.70/10,000 doses, p > 0.05). No AEFI reports were classified as anxiety-related reactions. A trend of a higher reporting rate was observed in the subsequent doses (e.g., the reporting rate of the fourth dose of Sabin-strains IPV was 7.88/10,000 doses and was 6.04/10,000 doses for the fourth dose of Salk-strains IPV, p < 0.05). The majority reports came from healthcare providers (97.91% for Sabin-strains IPV and 96.88% for Salk-strains IPV). The reporting rate onset within 48 h after Sabin-strains IPV vaccination was 1.94/10,000 doses, while it was significantly higher for Salk-strains IPV (2.97/10,000 doses). The reporting rate of AEFI following Sabin-strains IPV administered standalone was 2.48/10,000 doses, while it was significantly lower for Salk-strains IPV (0.46/10,000 doses). For both two types of IPV, the top three most frequently reported symptoms and/or signs were fever (1.56/10,000 doses for Sabin-strains IPV and 2.03/10,000 doses for Salk-strains IPV), induration/swelling (1.01/10,000 doses for Sabin-strains IPV and 1.52/10,000 doses for Salk-strains IPV), and rash/urticaria (0.95/10,000 doses for Sabin-strains IPV and 1.00/10,000 doses for Salk-strains IPV). Fever and induration/swelling were reported at a significantly higher rate following Salk-strains IPV administration compared to Sabin-strains IPV ( p < 0.05). No significant difference was observed in the reporting rate for other symptoms and/or signs between the two types of IPV ( Table 2 ). After the causality assessment, the most frequently reported serious clinical diagnosis was febrile convulsion (0.10/10,000 doses for Sabin-strains IPV and 0.08/10,000 doses for Salk-strains IPV). Thirty AEFI cases were classified as related to vaccination with Sabin-strains IPV, including six allergic purpura cases (reporting rate 0.02/10,000 doses), eight thrombocytopenia cases (reporting rate 0.02/10,000 doses), four angioedema cases (reporting rate 0.01/10,000 doses), and eleven febrile convulsion cases (reporting rate 0.10/10,000 doses). Nine AEFI cases were classified as related to vaccination with Salk-strains IPV, including two allergic purpura cases (reporting rate 0.02/10,000 doses), two thrombocytopenia cases (reporting rate 0.02/10,000 doses), two angioedema cases (reporting rate 0.02/10,000 doses), and three febrile convulsion cases (reporting rate 0.08/10,000 doses). The other serious AEFI associated with two types of IPV were identified as coincidental events. No statistically significant difference was found in the reporting rates of the other serious AEFI with the causal relationship between the two types of IPV ( Table 3 ). 4. Discussion This study presented an evaluation of AEFI of the domestic Sabin-strains IPV through the national passive surveillance system on AEFI in Zhejiang province, with a time span of 4.5 years after the switch of polio vaccination strategy. Furthermore, this study also compared the reporting rates of AEFI following the new licensed Sabin-strains IPV to those of the conventional Salk-strains IPV. The overall AEFI reporting rates for Sabin-strains IPV and Salk-strains IPV were 3.96 and 5.04 per 10,000 doses, respectively. Compared with the previous AEFI surveillance data of the domestic Sabin-strains IPV, the overall AEFI reporting rate of Sabin-strains IPV observed in this study was much lower than those from two pilot studies. For example, Fu [ 13 ] reported the AEFI rate of Sabin-strains IPV was 16.89/10,000 doses through the passive surveillance and Shi [ 14 ] found the AEFI rate of Sabin-strains IPV was 246.43/10,000 doses through the active surveillance. There were two possible explanations for this difference. One was that the active AEFI surveillance was carried out during the pilot period. The sensitivity of active monitoring was much higher than that of the routine passive surveillance. Another was that a new vaccine would attract a high social concern of its safety as reporters could have been more likely to report more AEFI cases at the very beginning of its use. Similar rises in reporting rates following the vaccine introduction had been reported as a phenomenon of the Weber effect [ 19 ]. It described an increase in the reporting rates of adverse events occurring in the time period shortly after the drug marketing, followed by a stabilization. Except for the two reports mentioned above, the overall AEFI reporting rate for Sabin-strains IPV in this study is higher than that reported by Wang [ 15 ] (2.69/10,000 doses), which was based on the passive surveillance. We assumed that it might be associated with the disparities in the study design, such as reporting sensitivity, sample size, and observation period. Salk-strains IPV has been widely used in the developed countries usually in the form of a combination vaccine (e.g., DTaP-Hib/IPV) [ 20 , 21 ]. Only a few AEFI surveillance reports were associated with the vaccination of standalone Salk-strains IPV. The overall AEFI reporting rate of the standalone Salk-strains IPV ranged from 0.01 to 0.11 per 10,000 doses in some developed countries [ 22 , 23 , 24 ]. However, the AEFI reporting rate of Salk-strains IPV in this study was much higher than those from the previous reports. We assumed that the variation in the AEFI reporting rate following Salk-strains IPV in different countries might be explained by the variable reporting requirements, case definitions, and surveillance sensitivity. The AEFI reporting rate for Salk-strains IPV observed in our study was also higher than that reported by Li [ 25 ] (1.6/10,000 doses) in the analysis of the nationwide AEFI surveillance through the NAEFISS. However, this report used the number of Salk-strains IPV distribution as the denominator for calculating the reporting rate of AEFI. It would underestimate the AEFI reporting rate as the number of distributed doses was always greater than the actual number of administered doses. Furthermore, the surveillance sensitivity in Zhejiang province was higher than that of the other regions of China and the national average level [ 25 ]. In our study, we found the AEFI reporting rate of Salk-strains IPV was slightly higher than that of Sabin-strains IPV. Further analysis indicated that the difference between two types of IPV was mainly due to the significantly higher reporting rate of the minor vaccine product-related reactions for Salk-strains IPV. The differences between two types of IPV in the reporting rate of severe vaccine product-related reactions, vaccination errors, and coincidental events were not significant. Consistent with the previous reports, we found the most common AEFI cases for two types of IPV were classified as the minor vaccine product-related reactions, which were relatively minor, mild, and self-limited [ 13 , 14 , 15 ]. Fever was the most frequent symptom and/or sign observed in AEFI cases associated with both Sabin-strains IPV and Salk-strains IPV. It was in line with the results from surveillance reports and clinical trials [ 13 , 14 , 15 , 25 ]. One phase III clinical trial of Sabin-strains IPV found that the incidence of fever was more frequent in the Sabin-strains IPV group than that in the Salk-strains IPV group [ 18 ]. On the contrary, our result was opposite to these two pieces of evidence mentioned above, with a higher reporting rate observed in Salk-strains IPV. The specific reason for this phenomenon was yet unknown, but it might be related to different production processes. Induration/swelling at injection site was the second most frequently reported symptom associated with both two types of IPV, which was similar to the results from several previous reports as the most common local symptom [ 18 , 26 ]. Both two types of IPV have the aluminum adjuvant, which had been reported that it could induce the redness and induration [ 27 , 28 ]. We also found the reporting rate of induration/swelling was significantly higher following Salk-strains IPV. It might be explained by the different dosages of adjuvant or the different syringes and needles used in two types of IPV vaccination. Pre-filled syringes produced by Becton, Dickinson, and Company (BD) were used to inject Salk-IPV, while common syringes and needles were used to inject Sabin-IPV. However, it needs to be further investigated. Rash/urticaria, which was a common clinical manifestation of the hypersensitivity reaction, was the third most frequently reported AEFI following two types of IPV. It was consistent with the results from other AEFI surveillance studies [ 29 , 30 ]. The main reason would be due to the allergy of the recipient to the components of the vaccine, such as the proteins. After the causality assessment, the proportion of serious AEFI that had a causal relationship was very small in both two types of IPV, as well as the relevant reporting rates. Generally, the reporting rate of the serious AEFI was very low associated with Sabin-strains IPV and Salk-strains IPV. Sabin-strains IPV had a safety profile consistent with that of Salk-strains IPV [ 7 ]. Febrile convulsion was the most frequently reported serious AEFI that was found to be related to both Sabin-strains IPV and Salk-strains IPV vaccination. A febrile convulsion is defined as a seizure occurring in a child 6 months to 5 years old that is accompanied by a fever (≥38 °C) without central nervous system infection [ 31 ]. Febrile convulsion can be induced under the stress of a fever and vaccination has been found to be the second most common cause. However, controversy on the association between vaccination and febrile convulsion still exists [ 32 ]. Another finding for the causality assessment of the serious AEFI reports was that the consistency for Sabin-IPV was not as good as that for Salk-IPV. The possible explanation was that the reporting sensitivity of the serious AEFI following Sabin-IPV was higher due to the doubts on safety of the new vaccines from the public. However, most of the reported serious AEFI following Sabin-IPV were considered as coincidental events after rigorous evaluation on causal relationships. There were still several limitations regarding this study. First, NAEFISS, as a passive surveillance system, has some inherent disadvantages, such as reporting bias (under-reporting or over-reporting), inconsistency and completeness of reports, and a lack of control groups. Second, the description of mild signs and symptoms without medical checks mainly relied on the recollections of parents or caregivers. This information was not verified by health professionals. Third, the standard definition of AEFI diagnosis, which was created by the Brighton Collaboration [ 33 ] and widely used internationally, was not applied in our study as the national AEFI guidance did not use the definitions of AEFI from the Brighton Collaboration. There would be obstacles for comparing our findings to that from other studies. 5. Conclusions This study indicated that most of the AEFI following Sabin-strains IPV and Salk-strains IPV were mild and common adverse reactions. The reporting rate of serious AEFI that had a causal relationship with IPV was very low and no difference was observed between Sabin-strains IPV and Salk-strains IPV. The reporting rate of the minor vaccine product-related reactions following Salk-strains IPV was slightly higher but within a reasonable range. Therefore, we assumed that the Sabin-strains IPV had a favorable safety profile that was similar to the Salk-strains IPV. A polio-free world needs the improved safety levels in vaccine production facilities. As such, the Sabin-strains IPV could be widely used, which could contribute to achieving the polio eradication target in developing countries and maintaining a polio-free world.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9535169/

Unravelling the reservoirs for colonisation of infants with Campylobacter spp. in rural Ethiopia: protocol for a longitudinal study during a global pandemic and political tensions

Introduction Undernutrition is an underlying cause of mortality in children under five (CU5) years of age. Animal-source foods have been shown to decrease malnutrition in CU5. Livestock are important reservoirs for Campylobacter bacteria, which are recognised as risk factors for child malnutrition. Increasing livestock production may be beneficial for improving nutrition of children but these benefits may be negated by increased exposure to Campylobacter and research is needed to evaluate the complex pathways of Campylobacter exposure and infection applicable to low-income and middle-income countries. We aim to identify reservoirs of infection with Campylobacter spp. of infants in rural Eastern Ethiopia and evaluate interactions with child health (environmental enteric dysfunction and stunting) in the context of their sociodemographic environment. Methods and analysis This longitudinal study involves 115 infants who are followed from birth to 12 months of age and are selected randomly from 10 kebeles of Haramaya woreda, East Hararghe zone, Oromia region, Ethiopia. Questionnaire-based information is obtained on demographics, livelihoods, wealth, health, nutrition and women empowerment; animal ownership/management and diseases; and water, sanitation and hygiene. Faecal samples are collected from infants, mothers, siblings and livestock, drinking water and soil. These samples are analysed by a range of phenotypic and genotypic microbiological methods to characterise the genetic structure of the Campylobacter population in each of these reservoirs, which will support inference about the main sources of exposure for infants. Ethics and dissemination Ethical approval was obtained from the University of Florida Internal Review Board (IRB201903141), the Haramaya University Institutional Health Research Ethics Committee (COHMS/1010/3796/20) and the Ethiopia National Research Ethics Review Committee (SM/14.1/1059/20). Written informed consent is obtained from all participating households. Research findings will be disseminated to stakeholders through conferences and peer-reviewed journals and through the Feed the Future Innovation Lab for Livestock Systems. Introduction Undernutrition is an underlying cause of mortality in children under five (CU5) years of age. Animal-source foods have been shown to decrease malnutrition in CU5. Livestock are important reservoirs for Campylobacter bacteria, which are recognised as risk factors for child malnutrition. Increasing livestock production may be beneficial for improving nutrition of children but these benefits may be negated by increased exposure to Campylobacter and research is needed to evaluate the complex pathways of Campylobacter exposure and infection applicable to low-income and middle-income countries. We aim to identify reservoirs of infection with Campylobacter spp. of infants in rural Eastern Ethiopia and evaluate interactions with child health (environmental enteric dysfunction and stunting) in the context of their sociodemographic environment. Methods and analysis This longitudinal study involves 115 infants who are followed from birth to 12 months of age and are selected randomly from 10 kebeles of Haramaya woreda, East Hararghe zone, Oromia region, Ethiopia. Questionnaire-based information is obtained on demographics, livelihoods, wealth, health, nutrition and women empowerment; animal ownership/management and diseases; and water, sanitation and hygiene. Faecal samples are collected from infants, mothers, siblings and livestock, drinking water and soil. These samples are analysed by a range of phenotypic and genotypic microbiological methods to characterise the genetic structure of the Campylobacter population in each of these reservoirs, which will support inference about the main sources of exposure for infants. Ethics and dissemination Ethical approval was obtained from the University of Florida Internal Review Board (IRB201903141), the Haramaya University Institutional Health Research Ethics Committee (COHMS/1010/3796/20) and the Ethiopia National Research Ethics Review Committee (SM/14.1/1059/20). Written informed consent is obtained from all participating households. Research findings will be disseminated to stakeholders through conferences and peer-reviewed journals and through the Feed the Future Innovation Lab for Livestock Systems. Introduction Undernutrition underlies 45% of worldwide under-five mortality. 1 In Eastern Africa, 5.3% of children under five (CU5) are stunted (a measure of chronic undernutrition, defined as length or height for age Z score< −2) and 34.5% are wasted (a measure of recent undernutrition, defined as length/height for weight Z score< −2). 2 Stunting increases the risks of intermittent illness, reduction of vaccine effectiveness, and suboptimal intellectual development in CU5, and over time is associated with lower income and increased morbidity and mortality. 3–5 Wasting is a more acute state and is associated with higher mortality risk. 6 Pathways by which agriculture can alleviate child undernutrition include production of animal-source foods (ASF), household income and women's empowerment. 7 Demand for ASF in Africa is growing rapidly. The livestock master plan (LMP) of Ethiopia aims to increase livestock production for both economic development and the nutritional needs of its growing population. 8 9 In Ethiopia, the LMP is implemented through commercialised and smallholder family production systems. 10 Smallholder systems are often implemented by farmers in low-resource rural settings, 10 where livestock waste management and water, sanitation and hygiene (WaSH) practices are likely to be deficient. As illustrated in the renowned 'F-diagram', the absence of such measures is likely to foster contamination by animal/human faeces, which serve as a reservoir of enteric pathogens. 11 The infection of enteric pathogens in CU5 is not only associated with symptomatic illness (eg, diarrhoea); importantly, asymptomatic enteric infection may lead to environmental enteric dysfunction (EED). 12 With the realisation that global child undernutrition cannot be solely attributed to deficient diets and diarrhoea, researchers have hypothesised that EED may be the key mediator between environmental exposures to enteric pathogens and undernutrition, calling for expansion of the long-established UNICEF framework of child undernutrition to incorporate poor gut health as an immediate cause of undernutrition. 13–15 In recent years, two randomised controlled trials found no significant effect of traditional WaSH interventions on stunting. 16 Researchers subsequently called for 'transformative WaSH' to eradicate household faecal contamination. Without tackling environmental contamination (animal waste, soil, poor food hygiene), the negative findings on child growth persisted in a recent intervention implementing this transformative approach, despite finding a partial protective effect against EED. 17 These findings emphasise the necessity of targeting exposure to animal faeces and zoonotic enteric pathogens through effective control measures in future interventions, particularly where livestock farming contributes significantly to rural livelihoods. A recent risk-benefit analysis of smallholder livestock production summarised a variety of enteric pathogens originating from livestock and associated with EED and undernutrition. Mainly, asymptomatic infections of children with Campylobacter spp were found to be associated with EED outcomes of epithelial damage, inflammation and increased permeability of the gut, as well as growth faltering and reduced weight gain. 13 A study in Peru found that asymptomatic infection of Campylobacter had a stronger effect size on slowing weight gain than its symptomatic counterpart. 18 In the Etiology, Risk Factors, and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health (MAL-ED) study, Campylobacter spp in child stools detected by Enzyme-Linked Immuno Assay (ELISA, identifying all species in the genus) had a nearly twofold higher effect on growth faltering than the thermotolerant species C. jejuni / C. coli detected by molecular methods, underscoring the potentially important role of non-thermotolerant Campylobacter spp. in undernutrition. 12 19 C. jejuni / C. coli are the most common causal agents of bacterial enteritis globally, 20 and their animal reservoirs include livestock animals. 21 The epidemiology and reservoirs of non-thermotolerant Campylobacter (eg, C. fetus , C. hyointestinalis , Candidatus Campylobacter infans ) are less well understood. 22 We conducted a cross-sectional study of smallholder farming households in Haramaya woreda (district) in rural eastern Ethiopia to measure the prevalence of Campylobacter spp., EED and undernutrition in children aged 10–16 months old and related risk factors. 23 A combination of poor child diet, inadequate WaSH conditions and poor livestock waste management potentially accounted for a high prevalence of (predominantly asymptomatic) Campylobacter spp colonisation, EED and undernutrition. Although the study was powered for prevalence estimation, current breastfeeding status of the child and child consumption of ASF were both significantly associated with Campylobacter spp. colonisation, while access to improved drinking water was protective against EED. Notably, we found a diversity of Campylobacter spp. colonised the children's gut, with non-thermotolerant species occurring more frequently and in higher abundance than thermotolerant species. 24 Although previous studies investigated the health impact and reservoirs of non-thermotolerant species, none of them targeted low-income and middle-income country settings and CU5 as study population. 25–27 The findings from our cross-sectional study underscore the need for attribution research examining various livestock species and other potential reservoirs for Campylobacter colonisation of infants. This paper presents a longitudinal study protocol and adaptations that have been implemented due to political unrest and the COVID-19 pandemic. The objectives of the longitudinal study are: To assess the prevalence, species composition, and genomic diversity of thermotolerant and non-thermotolerant Campylobacter spp. in infants, adults, livestock and other reservoirs in the Haramaya woreda. To determine the attribution of Campylobacter infections in infants to humans, livestock and other reservoirs (ie, drinking water, soil) based on the genetic population structure of Campylobacter spp. circulating in these reservoirs. To assess the associations among the presence of Campylobacter spp. and the nutritional status (i.e., stunting) of infants in relation to their socioeconomic environment. Methods and analysis Study setting The study is conducted in the rural Haramaya woreda, East Hararghe zone, Oromia Region, Ethiopia ( figure 1 ). Haramaya woreda, at an altitude of 1400–2340 m above sea level, has 36 rural kebeles (the smallest administrative unit in Ethiopia) and three urban kebeles. Khat, vegetables and fruits are important cash crops. Figure 1 Geographic location of study area. Right panels from top to bottom: Ethiopia, East Hararghe zone, Oromiya region with Haramaya woreda indicated in purple. Left panel: Haramaya woreda with study sites indicated in purple. Haramaya University (HU) campus is indicated in orange. The urban centre of Harar borders the woreda to the East. AM, Amuma; AW, Adele Walta; BG, Biftu Geda; BK, Bachake; DA, Damota; GC, Gobe Challa; IO, Ifa Oromia; KR, Kuro; NG, Nageya; QD, Qerensa Dereba. According to the 2016 Ethiopian Demographic and Health Survey, stunting in Haramaya is higher at 45.8% than the national average of 38%. 28 Haramaya University (HU) has established a Health and Demographic Surveillance Site (HDSS) in 12 kebeles in the Haramaya woreda. 29 Overall, 10 of these 12 kebeles ( figure 1 ) will provide the source population for the study. Two kebeles were excluded because of small population size and proximity to the urban centre of Harar. The total population of the 10 kebeles is 92 900, with population numbers per kebele ranging between 4900 and 14 300 ( table 1 ). Table 1 Population size and number of participating households in selected kebeles Kebele Population Participants Kebele Population Participants Adele Walta (AW) 5100 11 Gobe Challa (GC) 14 300 12 Amuma (AM) 7400 12 Ifa Oromia (IO) 12 200 11 Bachake (BK) 4900 9 Kuro (KR) 11 700 12 Biftu Geda (BG) 11 700 12 Nageya (NG) 11 500 12 Damota (DA) 6600 12 Qerensa Dereba (QD) 7500 12 Source: Haramaya University Health and Demographic Surveillance Site (unpublished). Sample size and power The study is powered for prevalence estimation. A sample of 100 infants or animals allows estimation of a 50% prevalence with precision 10% at 95% confidence, and a power of 80%. We allowed for 20% attrition and aimed to enrol a sample of 120 newborn infants. Overall, 100 pure cultures of each thermotolerant and non-thermotolerant Campylobacter spp. per reservoir (infants, mothers, siblings, four livestock species (cattle, goat, sheep, chicken), water, soil) are analysed by whole-genome sequencing, 1400 in total. This power calculation is based on Smid et al , 30 who showed that the precision of attribution is reduced if less than 100 isolates per animal reservoir are used. Study population A birth registry has been developed, which leverages the biannual update of the HDSS by cross-tabulation of expected birthdates in the selected kebeles based on the date of the mother's last menstrual period and the estimated month of pregnancy at the time of the interview. Data on expected deliveries were updated every month, and one DHSS data collector was assigned per kebele to record actual births. Newborn children were randomly selected with the aim to include 12 infants per kebele in the first month after birth. Families were eligible for participation if they had no plans to move out of the Haramaya woreda within 6 months; the mother had resided in the woreda for at least 3 months during pregnancy; and the mother was over 16 years of age when giving birth. Infants were excluded if the birth weight was <2500 g, if the infant or mother required extended stay (more than 4 days) in the hospital after birth, or if the infant had visible congenital abnormality or known serious medical illness or enteropathy, diagnosed by a medical doctor. Enrolment started in December 2020 and was completed by June 2021; 115 infants have successfully been enrolled ( table 1 ). Written informed consent was obtained from all participating households (husband and wife) using a form in the local language (Afan Oromo). Infant health measurements and interviews Infants are followed up from birth to 1 year of age. At enrolment and every 3 months anthropometric measurements (age in days, recumbent length, weight, mid-upper arm circumference (after 6 months of age) and head circumference) are collected. At the end of the follow-up period (age 12–14 months), EED is detected by a combination of the lactulose absorption test 31 and analysis for faecal myeloperoxidase (MPO) using a commercially available ELISA MPO RUO assay (Alpco, Salem, New Hampshire, USA). Collection and management of biological samples Stool samples are obtained every 4 weeks from all infants and biannually from all mothers, siblings and livestock (one sample from chickens, cattle, goats and sheep per household). Samples from the environment (three soil samples using bootsocks and one sample of drinking water per household) are also collected biannually. Mothers are provided modified disposable diapers with a clean plastic sheet, sterilised by UV for 10 min and an ice box with ice pack on the day before sample collection. The mother is asked to fit the modified diaper to the infant in the early morning and, after the infant has defecated, to wrap and place the diaper and contents in the ice box. When appropriate, the mother collects a sample of stool from the youngest sibling, as above, and of her own stool in a screw cap bottle. Stool samples are transferred to sterile prelabelled whirl-pack bags, which are transported to the lab in ice boxes. Samples are transported to a dedicated laboratory at the HU main campus ( figure 1 ) in an ice box within a maximum of 6 hours. Samples for nucleic acid extraction and sequencing are transferred to a nucleic acid stabilising reagent. On arrival in the laboratory, remaining faecal samples are distributed over barcoded tubes and frozen at – 80°C, partly with addition of 15% v/v glycerol. Samples for MPO analysis are distributed in barcoded tubes in the field, immediately flash-frozen in liquid nitrogen and stored at – 80°C immediately on arrival in the laboratory. Data collection and management All data are collected by trained personnel employed by HU who are proficient in the local language (Afan Oromo), knowledgeable of the local cultural background, and have appropriate scientific backgrounds (health sciences, veterinary sciences and social sciences). Household questionnaires on demographics; livelihoods; wealth; animal ownership, management and disease; WaSH; infant health and nutrition; and women's empowerment are presented to mothers and fathers two times during the study. Mothers answer a short questionnaire on infant health, vaccinations, breastfeeding practices, antibiotic use and diets during monthly collection of the stool samples. Families who decide to discontinue participation are presented with an exit interview. All data are collected on tablets using the REDCap mobile app and uploaded to a REDCap database, hosted at the University of Florida. The REDCap codebook is available as online supplemental file 1 . 10.1136/bmjopen-2022-061311.supp1 Supplementary data Detection, quantification, isolation and characterisation of Campylobacter spp in human stools, livestock faeces and environmental samples Molecular detection of Campylobacter spp Genomic DNA is extracted from the field samples using commercial kits. Detection and quantification of Campylobacter is performed using genus-specific Taqman real-time PCR 32 and species-specific Sybr Green real-time PCR. 32–36 Stool samples with a Ct value lower than 35 and 1 Ct value below the negative controls (sterile water, Salmonella enterica ssp enterica serotype Typhimurium strain LT2 and Escherichia coli genomic DNA) are considered positive for Campylobacter . Detection of Campylobacter spp. by culture Samples positive for Campylobacter by qPCR and a random selection of 10% of negative samples are selected for detection of thermophilic and non-thermophilic Campylobacter spp. For thermophilic Campylobacter spp, decimal dilutions are plated on CHROMagar Campylobacter (DRG International, Springfield, New Jersey USA) and incubated for up to 48 hours at 42°C in microaerophilic conditions (85% nitrogen, 10% carbon dioxide, 5% oxygen) using anaerobic jars and GasPak EZ Campy Container System Sachets. For non-thermophilic Campylobacter spp, decimal dilutions are plated on Columbia agar supplemented with 5% defibrinated sheep blood, Skirrow supplement (2 µL/mL), amphotericin B (5 µg/mL), cefoperazone (8 µg/mL) and Campylobacter growth supplement. The plates are incubated at 37°C for up to 72 hours in anaerobic conditions using anaerobic jars and GasPak anaerobic sachets. If no characteristic growth is observed, samples are enriched in Preston broth (faeces, soil samples for thermophilic Campylobacter spp) or Bolton broth (otherwise), incubated and plated as above. A random selection of up to five presumptively positive colonies from each sample and assay are confirmed by real-time PCR using genus-specific primers. 37 Colonies confirmed as Campylobacter spp. are stored in glycerol at −80°C and shipped to the US for speciation using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and whole-genome sequence analysis of confirmed isolates. Metagenomic sequencing Metagenomic sequencing will be used to complement detection and genetic characterisation of Campylobacter sp in children, mothers, siblings and livestock by with a culture-independent method providing sequences of (genes from) the dominant Campylobacter sp. These data will also serve to validate qPCR results and are expected to provide information on infection with other enteric pathogens. We aim to apply 16s rRNA sequencing 38 to infant stool samples after birth and at 4, and 12 months as well as at the time of EED measurement, and to all available stool samples from siblings and mothers. Shotgun sequencing 24 will be applied to a random selection of DNA extracts from human and livestock samples from 45 households. Data analysis Objective 1 The prevalence, species composition and genomic diversity of thermotolerant and non-thermotolerant Campylobacter spp. in infants, adults, livestock and other reservoirs in the Haramaya woreda will be characterised by descriptive analysis; identification of species and clonal complexes using legacy, core genome and/or whole-genome MLST: phylogenetic analysis, 39 determination of index of diversity, linkage equilibrium and rarefaction analysis. 40 Further, the virulome and resistome of the Campylobacter isolates will be analysed using comprehensive virulence factors 41 and antibiotic resistance gene databases. 42 In addition, a functional profiling of the sequenced genome will be predicted. 43 Specificity and sensitivity of qPCR and culture methods for detecting Campylobacter at the genus and biological group levels will be evaluated using Bayesian modelling of performance characteristics in the absence of a gold standard. 44–46 Metagenomic sequencing data will be analysed by analysed by latest versions of QIIME V.2, 47 CCMetagen 48 and IDseq pipelines. 49 We will fit a Bayesian two-state Markov model for the temporal process of colonisation and clearance of Campylobacter in each infant. The daily transition probabilities between colonisation and clearance will be regressed on time, demographic variables, household characteristics and most recent Campylobacter prevalence among samples from livestock, environment and other household members (sampled biannually). Spatiotemporal effects will be analysed using boosted regression trees 50 to model the influence of regional environmental variables on the distribution of infections of infants with the Campylobacter genus as primary outcome variable. Depending on results, this analysis will also be performed for one or more dominant species. To explore the spatial heterogeneity of effects of covariates on Campylobacter infections geographically weighted logistic regression models 51 will be used. Objective 2 The primary outcome variable for this objective is the proportion of Campylobacter infections in infants to that is attributed to mothers, siblings, livestock and environmental reservoirs based on the genetic population structure of Campylobacter sp circulating in these reservoirs. Source attribution will be based on modified Hald 52 and asymmetric island 53 models using R packages sourceR 54 and islandR ( https://github.com/jmarshallnz/islandR ). We will also explore the use of whole-genome 55 56 and metagenomic sequencing data 57 for attribution. Objective 3 Data on sociodemographics, livelihoods, economics, environmental health and sanitation will be analysed descriptively, characterising the context within which the study is occurring. Infant health, including breastfeeding practices, infant diet, vaccination status, use of antibiotics and presence of fever and diarrhoea, will also be described. Women's empowerment in agriculture will be assessed using five domains of empowerment generated by the Women's Empowerment in Agriculture Index (WEAI). Within this context, the primary outcome variables (EED biomarkers and nutritional outcome data, including length-for-age Z score and stunting status) will be examined for their associations with the cumulative burden of Campylobacter using generalised linear models and generalised estimating equations, respectively. We will adjust for potential confounders including sex, socioeconomic status and kebele residence, and include potential covariates such as breast feeding, infant health, household WaSH conditions, animal ownership, WEAI and food security. To better estimate the cumulative burden of Campylobacter we will extract data from the Bayesian Markov model that simulates the underlying daily presence/absence status of any Campylobacter sp in symptomatic or asymptomatic children. We will also perform exploratory analysis of associations between the gut microbiome of infants and infection with Campylobacter spp, EED biomarkers and stunting. Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. Adaptive project management The project site was deliberately selected based on the strength of potential partnerships and the realities of local populations; namely, the high rates of malnutrition among rural smallholder farmers, whose livelihoods employ a mixed livestock/crop system. While it was predictable that this type of environment would present significant logistical, technical and sociocultural constraints to research, it was also essential to examine the fundamental research questions. However, from the onset, the project has experienced additional unexpected challenges. The strength of the research partnership and continuous adaptation to the scientific protocol and data collection procedures have allowed the research to advance, despite changes in the local context. Numerous unforeseen challenges have presented, perhaps most significantly the COVID-19 pandemic. The project start was delayed from April to December 2020. Extensive COVID-19 safety protocols were developed and continuously adapted to allow the research to advance but to prevent COVID-19 transmission among team members and with the study population. Availability and work/transport dynamics of field and laboratory staff were limited by these necessary measures; examples include weekly screening for COVID-19 symptoms of all team members and associated quarantine procedures, and a policy of having only three field team members as passengers (masked) in a vehicle with the masked driver, with windows down. International travel was completely stopped for over a year. Laboratory analyses were also affected by COVID-19 related global supply chain challenges and prolonged customs clearance procedures, resulting in late delivery of critical equipment and supplies. In addition to COVID-19, other unforeseeable challenges for the field work included the escalating conflict in Tigray, which starting in November 2020 and resulted in a declaration of a national State of Emergency in November 2021, and unusually heavy rains from August to October 2021 resulting in deteriorating road conditions and collapsing bridges. Other more typical setbacks that nonetheless required additional adaptation included reduced availability of the study population during Ramadan (May 2021), personal leave among team members (eg, multiple pregnancies, death of an infant, illness) and concerns about political insecurity surrounding general elections in June 2021. Consequent to these challenges, the originally planned follow-up period of 18 months was reduced to 12 months. Further adjustments included the timing and distribution of field team members in communities for data collection, number of samples collected per household, number of anthropometric measurements collected per infant. The protocol in this manuscript includes all changes made until the moment of writing (December 2021), with the caveat that because of increased travel times due to checkpoints and roadblocks, planned EED measurements and end-line long interviews are not yet possible and will be attempted later. Communication, both between the research team and participants, as well as among members of the research team, has been critical to the adaptive management of the project. Interaction with the study population was facilitated by a Community Advisory Board (CAB), established during formative research and made up of representatives of the participating kebeles. 58 The CAB met prior to the start of field work and facilitated the research teams' interactions with local communities. Local CAB members were informed before the team travelled to field sites and were invited into and facilitated critical research-community member communication. This included discussions between the project manager and families who had concerns about ongoing participation in the project. Realtime information from the CAGED project field teams was instrumental in adapting the project to local norms, expectations and realities, many of which shifted and were reestablished during the COVID-19 pandemic. The field teams shared formal and informal data using a variety of communication platforms. Weekly face-to-face meetings were held for all field teams to discussion and debate approaches to challenges as they occurred; these were followed same day by Zoom-based conversation with Principal Investigators to finalise changes to the weekly plan. The broader team relied heavily on WhatsApp threads for day-to-day communication, field updates or other rapid communication. Zoom-based calls were also used for trainings that targeted the field team, with content ranging from research protocol updates to refresher training on data collection methods, to the importance of COVID-19 vaccines. Data collection has only proceeded because of the research team's existing partnership, the cultural competency of local team members, and a collective openness to learning and adaptability. In near constant conversation with researchers on the ground about how stakeholders within the study areas were receiving, reacting and engaging the research programme, the team has effectively adapted the scientific protocol from its original design, without jeopardising its scientific integrity, to answer the research questions. This exercise in adaptive management was facilitated by established trust, transparent communication, local project ownership and a firm commitment to the project's success. Study setting The study is conducted in the rural Haramaya woreda, East Hararghe zone, Oromia Region, Ethiopia ( figure 1 ). Haramaya woreda, at an altitude of 1400–2340 m above sea level, has 36 rural kebeles (the smallest administrative unit in Ethiopia) and three urban kebeles. Khat, vegetables and fruits are important cash crops. Figure 1 Geographic location of study area. Right panels from top to bottom: Ethiopia, East Hararghe zone, Oromiya region with Haramaya woreda indicated in purple. Left panel: Haramaya woreda with study sites indicated in purple. Haramaya University (HU) campus is indicated in orange. The urban centre of Harar borders the woreda to the East. AM, Amuma; AW, Adele Walta; BG, Biftu Geda; BK, Bachake; DA, Damota; GC, Gobe Challa; IO, Ifa Oromia; KR, Kuro; NG, Nageya; QD, Qerensa Dereba. According to the 2016 Ethiopian Demographic and Health Survey, stunting in Haramaya is higher at 45.8% than the national average of 38%. 28 Haramaya University (HU) has established a Health and Demographic Surveillance Site (HDSS) in 12 kebeles in the Haramaya woreda. 29 Overall, 10 of these 12 kebeles ( figure 1 ) will provide the source population for the study. Two kebeles were excluded because of small population size and proximity to the urban centre of Harar. The total population of the 10 kebeles is 92 900, with population numbers per kebele ranging between 4900 and 14 300 ( table 1 ). Table 1 Population size and number of participating households in selected kebeles Kebele Population Participants Kebele Population Participants Adele Walta (AW) 5100 11 Gobe Challa (GC) 14 300 12 Amuma (AM) 7400 12 Ifa Oromia (IO) 12 200 11 Bachake (BK) 4900 9 Kuro (KR) 11 700 12 Biftu Geda (BG) 11 700 12 Nageya (NG) 11 500 12 Damota (DA) 6600 12 Qerensa Dereba (QD) 7500 12 Source: Haramaya University Health and Demographic Surveillance Site (unpublished). Sample size and power The study is powered for prevalence estimation. A sample of 100 infants or animals allows estimation of a 50% prevalence with precision 10% at 95% confidence, and a power of 80%. We allowed for 20% attrition and aimed to enrol a sample of 120 newborn infants. Overall, 100 pure cultures of each thermotolerant and non-thermotolerant Campylobacter spp. per reservoir (infants, mothers, siblings, four livestock species (cattle, goat, sheep, chicken), water, soil) are analysed by whole-genome sequencing, 1400 in total. This power calculation is based on Smid et al , 30 who showed that the precision of attribution is reduced if less than 100 isolates per animal reservoir are used. Study population A birth registry has been developed, which leverages the biannual update of the HDSS by cross-tabulation of expected birthdates in the selected kebeles based on the date of the mother's last menstrual period and the estimated month of pregnancy at the time of the interview. Data on expected deliveries were updated every month, and one DHSS data collector was assigned per kebele to record actual births. Newborn children were randomly selected with the aim to include 12 infants per kebele in the first month after birth. Families were eligible for participation if they had no plans to move out of the Haramaya woreda within 6 months; the mother had resided in the woreda for at least 3 months during pregnancy; and the mother was over 16 years of age when giving birth. Infants were excluded if the birth weight was <2500 g, if the infant or mother required extended stay (more than 4 days) in the hospital after birth, or if the infant had visible congenital abnormality or known serious medical illness or enteropathy, diagnosed by a medical doctor. Enrolment started in December 2020 and was completed by June 2021; 115 infants have successfully been enrolled ( table 1 ). Written informed consent was obtained from all participating households (husband and wife) using a form in the local language (Afan Oromo). Infant health measurements and interviews Infants are followed up from birth to 1 year of age. At enrolment and every 3 months anthropometric measurements (age in days, recumbent length, weight, mid-upper arm circumference (after 6 months of age) and head circumference) are collected. At the end of the follow-up period (age 12–14 months), EED is detected by a combination of the lactulose absorption test 31 and analysis for faecal myeloperoxidase (MPO) using a commercially available ELISA MPO RUO assay (Alpco, Salem, New Hampshire, USA). Collection and management of biological samples Stool samples are obtained every 4 weeks from all infants and biannually from all mothers, siblings and livestock (one sample from chickens, cattle, goats and sheep per household). Samples from the environment (three soil samples using bootsocks and one sample of drinking water per household) are also collected biannually. Mothers are provided modified disposable diapers with a clean plastic sheet, sterilised by UV for 10 min and an ice box with ice pack on the day before sample collection. The mother is asked to fit the modified diaper to the infant in the early morning and, after the infant has defecated, to wrap and place the diaper and contents in the ice box. When appropriate, the mother collects a sample of stool from the youngest sibling, as above, and of her own stool in a screw cap bottle. Stool samples are transferred to sterile prelabelled whirl-pack bags, which are transported to the lab in ice boxes. Samples are transported to a dedicated laboratory at the HU main campus ( figure 1 ) in an ice box within a maximum of 6 hours. Samples for nucleic acid extraction and sequencing are transferred to a nucleic acid stabilising reagent. On arrival in the laboratory, remaining faecal samples are distributed over barcoded tubes and frozen at – 80°C, partly with addition of 15% v/v glycerol. Samples for MPO analysis are distributed in barcoded tubes in the field, immediately flash-frozen in liquid nitrogen and stored at – 80°C immediately on arrival in the laboratory. Data collection and management All data are collected by trained personnel employed by HU who are proficient in the local language (Afan Oromo), knowledgeable of the local cultural background, and have appropriate scientific backgrounds (health sciences, veterinary sciences and social sciences). Household questionnaires on demographics; livelihoods; wealth; animal ownership, management and disease; WaSH; infant health and nutrition; and women's empowerment are presented to mothers and fathers two times during the study. Mothers answer a short questionnaire on infant health, vaccinations, breastfeeding practices, antibiotic use and diets during monthly collection of the stool samples. Families who decide to discontinue participation are presented with an exit interview. All data are collected on tablets using the REDCap mobile app and uploaded to a REDCap database, hosted at the University of Florida. The REDCap codebook is available as online supplemental file 1 . 10.1136/bmjopen-2022-061311.supp1 Supplementary data Detection, quantification, isolation and characterisation of Campylobacter spp in human stools, livestock faeces and environmental samples Molecular detection of Campylobacter spp Genomic DNA is extracted from the field samples using commercial kits. Detection and quantification of Campylobacter is performed using genus-specific Taqman real-time PCR 32 and species-specific Sybr Green real-time PCR. 32–36 Stool samples with a Ct value lower than 35 and 1 Ct value below the negative controls (sterile water, Salmonella enterica ssp enterica serotype Typhimurium strain LT2 and Escherichia coli genomic DNA) are considered positive for Campylobacter . Detection of Campylobacter spp. by culture Samples positive for Campylobacter by qPCR and a random selection of 10% of negative samples are selected for detection of thermophilic and non-thermophilic Campylobacter spp. For thermophilic Campylobacter spp, decimal dilutions are plated on CHROMagar Campylobacter (DRG International, Springfield, New Jersey USA) and incubated for up to 48 hours at 42°C in microaerophilic conditions (85% nitrogen, 10% carbon dioxide, 5% oxygen) using anaerobic jars and GasPak EZ Campy Container System Sachets. For non-thermophilic Campylobacter spp, decimal dilutions are plated on Columbia agar supplemented with 5% defibrinated sheep blood, Skirrow supplement (2 µL/mL), amphotericin B (5 µg/mL), cefoperazone (8 µg/mL) and Campylobacter growth supplement. The plates are incubated at 37°C for up to 72 hours in anaerobic conditions using anaerobic jars and GasPak anaerobic sachets. If no characteristic growth is observed, samples are enriched in Preston broth (faeces, soil samples for thermophilic Campylobacter spp) or Bolton broth (otherwise), incubated and plated as above. A random selection of up to five presumptively positive colonies from each sample and assay are confirmed by real-time PCR using genus-specific primers. 37 Colonies confirmed as Campylobacter spp. are stored in glycerol at −80°C and shipped to the US for speciation using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and whole-genome sequence analysis of confirmed isolates. Metagenomic sequencing Metagenomic sequencing will be used to complement detection and genetic characterisation of Campylobacter sp in children, mothers, siblings and livestock by with a culture-independent method providing sequences of (genes from) the dominant Campylobacter sp. These data will also serve to validate qPCR results and are expected to provide information on infection with other enteric pathogens. We aim to apply 16s rRNA sequencing 38 to infant stool samples after birth and at 4, and 12 months as well as at the time of EED measurement, and to all available stool samples from siblings and mothers. Shotgun sequencing 24 will be applied to a random selection of DNA extracts from human and livestock samples from 45 households. Molecular detection of Campylobacter spp Genomic DNA is extracted from the field samples using commercial kits. Detection and quantification of Campylobacter is performed using genus-specific Taqman real-time PCR 32 and species-specific Sybr Green real-time PCR. 32–36 Stool samples with a Ct value lower than 35 and 1 Ct value below the negative controls (sterile water, Salmonella enterica ssp enterica serotype Typhimurium strain LT2 and Escherichia coli genomic DNA) are considered positive for Campylobacter . Detection of Campylobacter spp. by culture Samples positive for Campylobacter by qPCR and a random selection of 10% of negative samples are selected for detection of thermophilic and non-thermophilic Campylobacter spp. For thermophilic Campylobacter spp, decimal dilutions are plated on CHROMagar Campylobacter (DRG International, Springfield, New Jersey USA) and incubated for up to 48 hours at 42°C in microaerophilic conditions (85% nitrogen, 10% carbon dioxide, 5% oxygen) using anaerobic jars and GasPak EZ Campy Container System Sachets. For non-thermophilic Campylobacter spp, decimal dilutions are plated on Columbia agar supplemented with 5% defibrinated sheep blood, Skirrow supplement (2 µL/mL), amphotericin B (5 µg/mL), cefoperazone (8 µg/mL) and Campylobacter growth supplement. The plates are incubated at 37°C for up to 72 hours in anaerobic conditions using anaerobic jars and GasPak anaerobic sachets. If no characteristic growth is observed, samples are enriched in Preston broth (faeces, soil samples for thermophilic Campylobacter spp) or Bolton broth (otherwise), incubated and plated as above. A random selection of up to five presumptively positive colonies from each sample and assay are confirmed by real-time PCR using genus-specific primers. 37 Colonies confirmed as Campylobacter spp. are stored in glycerol at −80°C and shipped to the US for speciation using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and whole-genome sequence analysis of confirmed isolates. Metagenomic sequencing Metagenomic sequencing will be used to complement detection and genetic characterisation of Campylobacter sp in children, mothers, siblings and livestock by with a culture-independent method providing sequences of (genes from) the dominant Campylobacter sp. These data will also serve to validate qPCR results and are expected to provide information on infection with other enteric pathogens. We aim to apply 16s rRNA sequencing 38 to infant stool samples after birth and at 4, and 12 months as well as at the time of EED measurement, and to all available stool samples from siblings and mothers. Shotgun sequencing 24 will be applied to a random selection of DNA extracts from human and livestock samples from 45 households. Data analysis Objective 1 The prevalence, species composition and genomic diversity of thermotolerant and non-thermotolerant Campylobacter spp. in infants, adults, livestock and other reservoirs in the Haramaya woreda will be characterised by descriptive analysis; identification of species and clonal complexes using legacy, core genome and/or whole-genome MLST: phylogenetic analysis, 39 determination of index of diversity, linkage equilibrium and rarefaction analysis. 40 Further, the virulome and resistome of the Campylobacter isolates will be analysed using comprehensive virulence factors 41 and antibiotic resistance gene databases. 42 In addition, a functional profiling of the sequenced genome will be predicted. 43 Specificity and sensitivity of qPCR and culture methods for detecting Campylobacter at the genus and biological group levels will be evaluated using Bayesian modelling of performance characteristics in the absence of a gold standard. 44–46 Metagenomic sequencing data will be analysed by analysed by latest versions of QIIME V.2, 47 CCMetagen 48 and IDseq pipelines. 49 We will fit a Bayesian two-state Markov model for the temporal process of colonisation and clearance of Campylobacter in each infant. The daily transition probabilities between colonisation and clearance will be regressed on time, demographic variables, household characteristics and most recent Campylobacter prevalence among samples from livestock, environment and other household members (sampled biannually). Spatiotemporal effects will be analysed using boosted regression trees 50 to model the influence of regional environmental variables on the distribution of infections of infants with the Campylobacter genus as primary outcome variable. Depending on results, this analysis will also be performed for one or more dominant species. To explore the spatial heterogeneity of effects of covariates on Campylobacter infections geographically weighted logistic regression models 51 will be used. Objective 2 The primary outcome variable for this objective is the proportion of Campylobacter infections in infants to that is attributed to mothers, siblings, livestock and environmental reservoirs based on the genetic population structure of Campylobacter sp circulating in these reservoirs. Source attribution will be based on modified Hald 52 and asymmetric island 53 models using R packages sourceR 54 and islandR ( https://github.com/jmarshallnz/islandR ). We will also explore the use of whole-genome 55 56 and metagenomic sequencing data 57 for attribution. Objective 3 Data on sociodemographics, livelihoods, economics, environmental health and sanitation will be analysed descriptively, characterising the context within which the study is occurring. Infant health, including breastfeeding practices, infant diet, vaccination status, use of antibiotics and presence of fever and diarrhoea, will also be described. Women's empowerment in agriculture will be assessed using five domains of empowerment generated by the Women's Empowerment in Agriculture Index (WEAI). Within this context, the primary outcome variables (EED biomarkers and nutritional outcome data, including length-for-age Z score and stunting status) will be examined for their associations with the cumulative burden of Campylobacter using generalised linear models and generalised estimating equations, respectively. We will adjust for potential confounders including sex, socioeconomic status and kebele residence, and include potential covariates such as breast feeding, infant health, household WaSH conditions, animal ownership, WEAI and food security. To better estimate the cumulative burden of Campylobacter we will extract data from the Bayesian Markov model that simulates the underlying daily presence/absence status of any Campylobacter sp in symptomatic or asymptomatic children. We will also perform exploratory analysis of associations between the gut microbiome of infants and infection with Campylobacter spp, EED biomarkers and stunting. Objective 1 The prevalence, species composition and genomic diversity of thermotolerant and non-thermotolerant Campylobacter spp. in infants, adults, livestock and other reservoirs in the Haramaya woreda will be characterised by descriptive analysis; identification of species and clonal complexes using legacy, core genome and/or whole-genome MLST: phylogenetic analysis, 39 determination of index of diversity, linkage equilibrium and rarefaction analysis. 40 Further, the virulome and resistome of the Campylobacter isolates will be analysed using comprehensive virulence factors 41 and antibiotic resistance gene databases. 42 In addition, a functional profiling of the sequenced genome will be predicted. 43 Specificity and sensitivity of qPCR and culture methods for detecting Campylobacter at the genus and biological group levels will be evaluated using Bayesian modelling of performance characteristics in the absence of a gold standard. 44–46 Metagenomic sequencing data will be analysed by analysed by latest versions of QIIME V.2, 47 CCMetagen 48 and IDseq pipelines. 49 We will fit a Bayesian two-state Markov model for the temporal process of colonisation and clearance of Campylobacter in each infant. The daily transition probabilities between colonisation and clearance will be regressed on time, demographic variables, household characteristics and most recent Campylobacter prevalence among samples from livestock, environment and other household members (sampled biannually). Spatiotemporal effects will be analysed using boosted regression trees 50 to model the influence of regional environmental variables on the distribution of infections of infants with the Campylobacter genus as primary outcome variable. Depending on results, this analysis will also be performed for one or more dominant species. To explore the spatial heterogeneity of effects of covariates on Campylobacter infections geographically weighted logistic regression models 51 will be used. Objective 2 The primary outcome variable for this objective is the proportion of Campylobacter infections in infants to that is attributed to mothers, siblings, livestock and environmental reservoirs based on the genetic population structure of Campylobacter sp circulating in these reservoirs. Source attribution will be based on modified Hald 52 and asymmetric island 53 models using R packages sourceR 54 and islandR ( https://github.com/jmarshallnz/islandR ). We will also explore the use of whole-genome 55 56 and metagenomic sequencing data 57 for attribution. Objective 3 Data on sociodemographics, livelihoods, economics, environmental health and sanitation will be analysed descriptively, characterising the context within which the study is occurring. Infant health, including breastfeeding practices, infant diet, vaccination status, use of antibiotics and presence of fever and diarrhoea, will also be described. Women's empowerment in agriculture will be assessed using five domains of empowerment generated by the Women's Empowerment in Agriculture Index (WEAI). Within this context, the primary outcome variables (EED biomarkers and nutritional outcome data, including length-for-age Z score and stunting status) will be examined for their associations with the cumulative burden of Campylobacter using generalised linear models and generalised estimating equations, respectively. We will adjust for potential confounders including sex, socioeconomic status and kebele residence, and include potential covariates such as breast feeding, infant health, household WaSH conditions, animal ownership, WEAI and food security. To better estimate the cumulative burden of Campylobacter we will extract data from the Bayesian Markov model that simulates the underlying daily presence/absence status of any Campylobacter sp in symptomatic or asymptomatic children. We will also perform exploratory analysis of associations between the gut microbiome of infants and infection with Campylobacter spp, EED biomarkers and stunting. Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. Adaptive project management The project site was deliberately selected based on the strength of potential partnerships and the realities of local populations; namely, the high rates of malnutrition among rural smallholder farmers, whose livelihoods employ a mixed livestock/crop system. While it was predictable that this type of environment would present significant logistical, technical and sociocultural constraints to research, it was also essential to examine the fundamental research questions. However, from the onset, the project has experienced additional unexpected challenges. The strength of the research partnership and continuous adaptation to the scientific protocol and data collection procedures have allowed the research to advance, despite changes in the local context. Numerous unforeseen challenges have presented, perhaps most significantly the COVID-19 pandemic. The project start was delayed from April to December 2020. Extensive COVID-19 safety protocols were developed and continuously adapted to allow the research to advance but to prevent COVID-19 transmission among team members and with the study population. Availability and work/transport dynamics of field and laboratory staff were limited by these necessary measures; examples include weekly screening for COVID-19 symptoms of all team members and associated quarantine procedures, and a policy of having only three field team members as passengers (masked) in a vehicle with the masked driver, with windows down. International travel was completely stopped for over a year. Laboratory analyses were also affected by COVID-19 related global supply chain challenges and prolonged customs clearance procedures, resulting in late delivery of critical equipment and supplies. In addition to COVID-19, other unforeseeable challenges for the field work included the escalating conflict in Tigray, which starting in November 2020 and resulted in a declaration of a national State of Emergency in November 2021, and unusually heavy rains from August to October 2021 resulting in deteriorating road conditions and collapsing bridges. Other more typical setbacks that nonetheless required additional adaptation included reduced availability of the study population during Ramadan (May 2021), personal leave among team members (eg, multiple pregnancies, death of an infant, illness) and concerns about political insecurity surrounding general elections in June 2021. Consequent to these challenges, the originally planned follow-up period of 18 months was reduced to 12 months. Further adjustments included the timing and distribution of field team members in communities for data collection, number of samples collected per household, number of anthropometric measurements collected per infant. The protocol in this manuscript includes all changes made until the moment of writing (December 2021), with the caveat that because of increased travel times due to checkpoints and roadblocks, planned EED measurements and end-line long interviews are not yet possible and will be attempted later. Communication, both between the research team and participants, as well as among members of the research team, has been critical to the adaptive management of the project. Interaction with the study population was facilitated by a Community Advisory Board (CAB), established during formative research and made up of representatives of the participating kebeles. 58 The CAB met prior to the start of field work and facilitated the research teams' interactions with local communities. Local CAB members were informed before the team travelled to field sites and were invited into and facilitated critical research-community member communication. This included discussions between the project manager and families who had concerns about ongoing participation in the project. Realtime information from the CAGED project field teams was instrumental in adapting the project to local norms, expectations and realities, many of which shifted and were reestablished during the COVID-19 pandemic. The field teams shared formal and informal data using a variety of communication platforms. Weekly face-to-face meetings were held for all field teams to discussion and debate approaches to challenges as they occurred; these were followed same day by Zoom-based conversation with Principal Investigators to finalise changes to the weekly plan. The broader team relied heavily on WhatsApp threads for day-to-day communication, field updates or other rapid communication. Zoom-based calls were also used for trainings that targeted the field team, with content ranging from research protocol updates to refresher training on data collection methods, to the importance of COVID-19 vaccines. Data collection has only proceeded because of the research team's existing partnership, the cultural competency of local team members, and a collective openness to learning and adaptability. In near constant conversation with researchers on the ground about how stakeholders within the study areas were receiving, reacting and engaging the research programme, the team has effectively adapted the scientific protocol from its original design, without jeopardising its scientific integrity, to answer the research questions. This exercise in adaptive management was facilitated by established trust, transparent communication, local project ownership and a firm commitment to the project's success. Discussion This proposal outlines a detailed research plan for conducting a longitudinal study to understand the complex interactions between livestock and infant health. This will support implementation of the LMP to improve the nutritional status of Ethiopians, while minimising the risk of poor nutritional outcomes associated with EED and impaired linear growth. The knowledge generated in this project will benefit the local population in our study area, who will be provided with knowledge about improving domestic hygiene in relation to management of excreta from livestock and other domestic animals. It also helps the local community and national authorities to understand the determinants of stunting so that they can devise appropriate interventions, which can be evaluated in further studies. Beyond Ethiopia, findings from this study constitute a global public good and will have widespread implications, as the relationship between Campylobacter spp, livestock production, EED and infant growth is further explored. Our study has several limitations. The study has been powered for prevalence estimation and source attribution of Campylobacter infections in infants, with a minimum target sample size of 100. This sample size is small to detect associations between nutritional outcomes and putative causal factors such as Campylobacter infection. For example, the MAL-ED study enrolled 165–237 infants per study site. 12 Our sample size will allow us to detect relatively strong associations only. However, the MAL-ED study also demonstrated that the effect of the broad group of non-thermotolerant Campylobacter sp on linear growth is stronger than the effect of the well-known thermotolerant species C. jejuni . Our ability to detect several non-thermotolerant species separately by qPCR rather than at the group level by an all-inclusive immunoassay allows us to zoom in on the likely small number of species whose effect sizes may be larger than that of the whole group. The study is executed in unprecedented times of a major global pandemic and a major civil conflict in Ethiopia, leading to civic unrest, inflation and increased population mobility. This affects the ability to collect all samples and surveys as planned, resulting in an increased level of missing data. We are implementing several approaches to prevent missingness, including active communication with the CAB members, repeat visits and increasing the number of field workers to accomodate the additional workload and driving times. We will use multiple imputation methods to replace the missing data. Awareness on the association between infant's exposure to zoonotic pathogens in livestock faeces and risk of stunting is generally low, constraining initiatives to decrease exposure to these hazards. We aim to create an enabling policy and institutional environment, ensuring that smallholder livestock production and associated risk of EED receive the attention and investment this problem deserves. For example, farmers are unaware of the danger of livestock excreta or the measures they could take to reduce contamination. Health practitioners are unlikely to be equipped to diagnose and treat Campylobacter infections, especially those due to long‐term exposure to livestock. Due to lack of awareness, policy-makers and development partners may not prioritise reducing exposure to livestock faeces, in comparison to other food safety, WaSH and broader food security issues. Key strategies for this include the following: Mapping of relevant stakeholders operating at different levels (eg, academia, policy-makers, research, extension workers). Hosting of stakeholder workshops to discuss on challenges associated with livestock production and associated effects on infant health, and research project proposed solutions and findings to tackle the challenges. Feedback of emerging findings from the research project. Developing and delivery of knowledge mobilisation strategies that respond to local stakeholders' needs and actively engage them in the uptake and application of research knowledge, using existing government structures to promote safe chicken production and also transfer information and knowledge by training various actors including agricultural extensions, health workers, community health volunteers, etc. The research findings and lessons from the project will also be documented, synthesised and shared in different ways to inform policy, development practice and to be used as resource material for training farmers, extension workers and future agricultural graduates. Ethics and dissemination Ethical approval was obtained from the University of Florida Internal Review Board (IRB201903141); the Haramaya University Institutional Health Research Ethics Committee (COHMS/1010/3796/20) and the Ethiopia National Research Ethics Review Committee (SM/14.1/1059/20). Written informed consent is obtained from all participating households (husband and wife) using a form in the local language (Afan Oromo). Research findings will be disseminated to community stakeholders, including participants, through the existing CAB. Findings will be disseminated to scientific, academic, policy and development stakeholders through conferences and peer-reviewed journals and through the Feed the Future Innovation Lab for Livestock Systems. The Bill and Melinda Gates Foundation, the funder of this trial, requires an open access data policy. Therefore, all manuscripts from this funded work will be open access with the data underlying the published research results available in a public repository. The website https://www.gatesfoundation.org/how-we-work/general-information/open-access-policy provides more information on this policy. The research is part of the Feed the Future Innovation Lab for Livestock Systems. Webinars, research briefs and other targeted dissemination activities will be organised under this umbrella. The project website https://livestocklab.ifas.ufl.edu/projects/caged/ provides access to all results. A CAB including a representative of the community, religious leaders (imam), woreda and kebele administration, woreda women and children affairs, woreda bureau of health and agriculture, kebele health, and agricultural extension workers has been established to guide the research team for better understanding of local context and entry to the community and is regularly engaged in the research. Only the project manager at Haramaya University and the data manager at the University of Florida will have access to personally identifiable information in the REDCap database. Any data shared among researchers within the project will be deidentified and blinded. Materials and Data Transfer Agreements assure confidentiality of data when exchanged with international partners and others. We follow the current standard of care in Ethiopia for the management of sick children. Accordingly, treatment will not be provided for asymptomatic Campylobacter infections. All infants with acute disease (diarrhoea, fever) or severe acute malnutrition are linked with the nearby health facility. Transportation to the facility and referral arrangements are offered by the project. Supplementary Material Reviewer comments Author's manuscript Ethics statements Patient consent for publication Not applicable. Patient consent for publication Not applicable.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2063559/

Receptor palmitoylation and ubiquitination regulate anthrax toxin endocytosis

The anthrax toxin is composed of three independent polypeptide chains. Successful intoxication only occurs when heptamerization of the receptor-binding polypeptide, the protective antigen (PA), allows binding of the two enzymatic subunits before endocytosis. We show that this tailored behavior is caused by two counteracting posttranslational modifications in the cytoplasmic tail of PA receptors. The receptor is palmitoylated, and this unexpectedly prevents its association with lipid rafts and, thus, its premature ubiquitination. This second modification, which is mediated by the E3 ubiquitin ligase Cbl, only occurs in rafts and is required for rapid endocytosis of the receptor. As a consequence, cells expressing palmitoylation-defective mutant receptors are less sensitive to anthrax toxin because of a lower number of surface receptors as well as premature internalization of PA without a requirement for heptamerization. Introduction Anthrax toxin, one of the two major virulence factors produced by Bacillus anthracis , is composed of three independent polypeptide chains: the protective antigen (PA), which is involved in target cell binding; the edema factor (EF), a calmodulin-dependent adenylate cyclase; and the lethal factor (LF), a zinc-dependent metalloprotease (for reviews see Collier and Young, 2003 ; Abrami et al., 2005 ; Scobie and Young, 2005 ). Only PA is able to bind to target cells; thus, EF and LF must always act in binary combination with PA to be transported to the target cell cytosol, where they exert their activities. The two identified PA receptors, tumor endothelial marker 8 (TEM8) and capillary morphogenesis gene 2 (CMG2), are type I transmembrane proteins sharing ∼60% homology in their extracellular von Willebrand factor A domains and 68% identity in the first 145 residues of their cytoplasmic tails ( Scobie and Young, 2005 ). A relatively clear view of the mode of action of anthrax toxin has emerged over recent years (for reviews see Collier and Young, 2003 ; Abrami et al., 2005 ; Scobie and Young, 2005 ). PA is produced as an 83-kD protein that is unable to interact with EF and LF. At the target cell surface, proteolytic processing of PA83 leads to PA63, which remains receptor bound and can polymerize into a heptameric (PA 7mer ) ring called the prepore. This prepore is able to bind up to three molecules of EF and/or LF, thus leading to a large hetero-oligomeric complex containing EF–LF–PA 7mer and receptors (for reviews see Abrami et al., 2005 ; Scobie and Young, 2005 ). Once formed, this complex is rapidly internalized via a pathway that depends both on lipid rafts and clathrin ( Abrami et al., 2003 ). The complex is then delivered to early endosomes, where it associates with intraluminal vesicles ( Abrami et al., 2004 ). At the low pH of endosomes, the prepore undergoes a conformational change that leads to its membrane insertion and pore formation. Interestingly, the pH sensitivity is determined by the receptor, and a lower pH is required when PA 7mer is bound to CMG2 when compared with TEM8 ( Rainey et al., 2005 ; Wolfe et al., 2005 ). Low pH also triggers partial unfolding of EF and LF, which can translocate across the PA 7mer channel ( Krantz et al., 2005 ). Because channel formation appears to occur preferentially in the intraluminal vesicles of the multivesicular endosomes, EF and LF end up in the lumen of these vesicles ( Abrami et al., 2004 ). Final release of EF and LF from endosomes to the cytoplasm requires back fusion events between intraluminal vesicles and the limiting membrane at the level of late endosomes ( Abrami et al., 2004 ). Because EF and LF are unable to bind to cells or cross membranes on their own, binding to PA is an absolute prerequisite for internalization and intoxication. Therefore, PA receptors must remain at the cell surface until heptamerization and binding of the enzymatic subunits have taken place. In fact, PA83 is poorly internalized in comparison with PA63 (for reviews see Collier and Young, 2003 ; Abrami et al., 2005 ; Scobie and Young, 2005 ) as a result of a differential localization on the plasma membrane ( Abrami et al., 2003 ). Whereas PA83 localizes to the glycerophospholipidic region, PA63 (in particular the heptameric form) is found in specialized cholesterol-rich domains called lipid rafts. This raft association is essential for subsequent internalization of the toxin ( Abrami et al., 2003 ). Therefore, although it is receptor mediated, anthrax toxin endocytosis is actually toxin driven. To understand the molecular mechanisms that govern this well-orchestrated behavior of anthrax toxin receptors at the cell surface, we investigated the roles of posttranslational modifications of the receptor cytoplasmic tails. We focused on two modifications: S-palmitoylation and ubiquitination. S-palmitoylation is a reversible lipid modification involving the addition of a saturated 16-carbon palmitate moiety to specific cysteines via a thioester linkage. This modification is used by several hydrophilic proteins such as Ras to associate with membranes but is also found in transmembrane proteins such as the transferrin receptor ( Alvarez et al., 1990 ), LAT (linker protein for activation of T cells; Zhang et al., 1998 ), influenza HA ( Scheiffele et al., 1997 ), or the G protein of the vesicular stomatitis virus ( Mack and Kruppa, 1988 ). The exact function of this modification in membrane proteins is mostly unclear, but it might modulate the interaction of these proteins with membranes or membrane domains ( Smotrys and Linder, 2004 ) as well as with other proteins. Ubiquitination is the addition of a ubiquitin (Ub) moiety to cytoplasmic lysines by E3 Ub ligases, the third enzymes in the ubiquitination pathway ( Hicke and Dunn, 2003 ). This added Ub itself may or may not be subsequently ubiquitinated on Lys 48 or on Lys 63 leading to polyubiquitin chains ( Haglund and Dikic, 2005 ). Although polyubiquitination on Lys 48 is essentially involved in the degradation of proteins by the proteasome, monoubiquitination (the addition of a single Ub moiety to one or multiple lysines) as well as polyubiquitination on Lys 63 have been shown to be involved in endocytosis of plasma membrane proteins (G-coupled receptors, growth factors, and transporters; Hicke and Dunn, 2003 ) and sorting of internalized receptors into multivesicular bodies ( Katzmann et al., 2002 ). We report that TEM8 and CMG2 are targets for both palmitoylation and ubiquitination and that these two modifications, via counteracting effects, control receptor endocytosis. Results Characterization of TEM8/1-HA and CMG2/4-V5 TEM8 and CMG2 both exist as four isoforms, two of which, in each case, act as anthrax toxin receptors: isoforms 1 and 2 of TEM8 (TEM8/1 and TEM8/2) and isoforms 1 and 4 of CMG2 (CMG2/1 and CMG2/4; for review see Scobie and Young, 2005 ). These isoforms differ only in their cytoplasmic tails ( Fig. 1 A Figure 1. TEM8/1 and CMG2/4 are glycosylated. (A) Alignment of the cytoplasmic tails of human TEM8 and human CMG2 using the SIM software of the EXPASY server ( www.expasy.ch ). Regions of identity are shown in yellow, lysine residues in green, and cysteine residues in red. Numbering of the residues corresponds to that of TEM8/1. The asterisks label the lysine mutants in the K6R mutant described in Fig. 7. (B and C) CHO ΔATR cells untransfected and stably or transiently transfected (for 48 h; B) with a TEM8/1-HA construct were analyzed by Western blotting (40 μg of protein/lane; B) or immunofluorescence (C) using an anti-HA antibody. Bar, 10 μM. (D) CHO ΔATR cells stably expressing TEM8/1-HA were submitted or unsubmitted to surface trypsinization at 4°C and subsequently analyzed for the presence of TEM8/1-HA by Western blotting. (E) CHO ΔATR cells transiently expressing TEM8/1-HA were grown in the presence or absence of tunicamycin. Control cell extracts were subsequently left untreated or treated with N -glycosidase F or Endo H. The effects of these treatments were analyzed by SDS-PAGE followed by Western blotting against the HA tag. (B and E) Band u, unglycosilated form; p, glycosylated precursor; m, mature form. (F) Extracts of CHO ΔATR cells transiently expressing or not expressing CMG2/4-V5 were left untreated or treated with N -glycosidase F or Endo H and subsequently analyzed by SDS-PAGE followed by Western blotting against the V5 tag. ); TEM8/2 has a far shorter cytoplasmic tail than TEM8/1 and CMG2/1, and CMG2/4 differs only in the last 12–13 residues. Even between TEM8 and CMG2, the long tails have a very high degree of conservation ( Fig. 1 A ). Therefore, we have focused this study mainly on TEM8/1 but repeated key experiments with the short TEM8/2 isoform as well as with CMG2/4. TEM8/1 was tagged with an HA epitope at the COOH terminus and was either transiently transfected or stably expressed in receptor-deficient CHO cells (CHO ΔATR ; Liu and Leppla, 2003 ). Expression levels were overall higher upon transient transfection ( Fig. 1 B ) but were still readily detectable in the stably TEM8/1-HA–expressing cells ( Fig. 1, B and C ). Most of the expressed receptor was present at the cell surface as indicated by its sensitivity to cell surface trypsinization ( Fig. 1 D ). TEM8-HA always migrated as a doublet with an upper smeared band ( Fig. 1 B , m) and a lower well-defined band ( Fig. 1 B , p), the intensity of which varied greatly from experiment to experiment. We investigated whether this migration pattern was caused by glycosylation because the extracellular domain of TEM8 has three predicted N -glycosylation sites. Treatment of cells with tunicamycin, an inhibitor of N -glycosylation, led to the appearance of a third, lower mobility band (u; unglycosilated) with a concomitant decrease in the intensity of bands m and p ( Fig. 1 E ). N -glycosidase F treatment of control cell extracts led to the complete disappearance of bands m and p to the benefit of band u ( Fig. 1 E ), showing that m and p both correspond to glycosylated TEM8. Band m was endoglycosidase H (Endo H) resistant, and band p was Endo H sensitive ( Fig. 1 E ). This shows that TEM8 in band m (mature) had acquired complex Golgi-modified sugars, whereas that in band p (precursor) contained incompletely modified sugars. CMG2/4 was expressed with a COOH-terminal V5 tag ( Dowling et al., 2003 ). As for TEM8, CMG2/4-V5 migrated as a smear plus a lower molecular weight band ( Fig. 1 F ). Both bands were sensitive to N -glycosidase F treatment, and, as for TEM8, only the lower band was sensitive to Endo H ( Fig. 1 F ). Altogether, this indicates that CMG2/4 is also glycosylated, as predicted by the presence of two N -glycosylation sites in the ectodomain. PA DRM association is palmitoylation dependent We have previously shown that the anthrax toxin modifies the surface distribution of its receptor, inducing its clustering in cholesterol-rich raftlike domains. We wondered whether posttranslational modifications of the receptor tail could be involved in regulating interactions of TEM8 and CMG2 with membrane domains. TEM8/1 and /2 as well as CMG2/1 and /4 all contain cysteine residues in their cytoplasmic tails, which are potential sites for S-palmitoylation ( Linder and Deschenes, 2004 ). More specifically, all four proteins have two conserved juxtamembranous cysteines ( Fig. 1 A ; Cys-346 and Cys-347 in TEM8), a third cysteine is conserved between TEM8/1 (Cys-481) and CMG2, and a fourth unconserved cysteine is found in TEM8/1 (Cys-521) and CMG2/1 (Cys482). To test whether palmitoylation plays a role in anthrax toxin endocytosis, we analyzed whether the palmitoylation inhibitor 2-bromopalmitate ( Webb et al., 2000 ) would affect the raft association of PA63 in BHK cells, a cell line that expresses transmembrane isoforms of both TEM8 and CMG2 (unpublished data). We first verified that PA binding was not affected by the treatment ( Fig. 2 A Figure 2. Palmitoylation events are required for DRM association and internalization of PA. (A and B) Control BHK cells were pretreated or untreated with bromopalmitate and were incubated with 500 ng/ml nicked PA63 and 20 ng/ml aerolysin for 1 h at 4°C followed by 10 min at 37°C. (A) Cell extracts were submitted to SDS-PAGE followed by Western blotting to reveal PA63, aerolysin, and caveolin-1 (Cav-1). (B) Cells were solubilized in 1% Triton X-100, run on an OptiPrep gradient, and each fraction was analyzed by SDS-PAGE followed by Western blotting against PA, aerolysin, and caveolin-1. (C) BHK cells were pretreated with bromopalmitate and incubated with 500 ng/ml nicked PA63 for 1 h at 4°C followed by different times at 37°C. Cell extracts (40 μg of protein) were analyzed by SDS-PAGE and Western blotting to reveal the SDS-resistant PA 7mer pore and MEK1 (NH 2 -terminal directed). To detect the prepore (SDS-sensitive nonmembrane-inserted PA 7mer ), cell extracts were submitted to an acid pulse before SDS analysis. ). Raft association was then monitored by following the association with detergent-resistant membranes (DRMs; Brown and London, 1998 ). PA63 was associated with DRMs in control cells as previously observed ( Abrami et al., 2003 ) but shifted to the detergent-soluble fractions in drug-treated cells ( Fig. 2 B ). This was not caused by a general disruption of lipid rafts because glycosyl-phosphatidylinositol–anchored proteins, followed here using the glycosyl-phosphatidylinositol–specific bacterial toxin aerolysin ( Fivaz et al., 2002 ) as well as caveolin-1, remained primarily in the DRM fractions. Although caveolin-1 is palmitoylated on three cysteines, the modification is not required for DRM association ( Dietzen et al., 1995 ). The inhibitory effect of 2-bromopalmitate on PA63-raft association was also accompanied by the inhibition of endocytosis as reflected by a lack of SDS-resistant PA 7mer pore and the absence of LF-mediated cleavage of one of the LF substrates, the MAPK kinase MEK1 ( Fig. 2 C ; for review see Collier and Young, 2003 ). To test whether 2-bromopalmitate inhibited the heptamerization process itself, we designed an assay to detect cell surface–formed prepores. These are SDS sensitive and, therefore, are not detected by SDS-PAGE. However, they can be converted to the SDS-resistant phenotype by submitting cell extracts to low pH (pH 4.5) before SDS-PAGE analysis, as illustrated in Fig. 2 C (middle left) for control cells. The prepore was undetectable for 2-bromopalmitate–treated cells ( Fig. 2 C ). TEM8 and CMG2 are palmitoylated The aforementioned experiments show that palmitoylation events are important for anthrax toxin raft association and heptamerization. To determine whether the receptors themselves are palmitoylated, TEM8/1-HA and CMG2/4-V5 were immunoprecipitated from lysates of 3 H-palmitic acid–labeled CHO ΔATR cells transiently transfected with the respective constructs. Radiolabeled bands with motilities similar to that of TEM8/1-HA ( Fig. 3 A Figure 3. TEM8/1 and CMG2/4 anthrax toxin receptors are palmitoylated. CHO ΔATR cells transfected with TEM8/1-HA (A) or CMG2/4-V5 (B) were incubated with 3 H-palmitic acid for 2 h before immunoprecipitation using anti-tag antibodies. Immunoprecipitates were split into two, run on 4–20% gels, and analyzed either by autoradiography ( 3 H-palmitate) or Western blotting (anti-tag HA or V5). (C) Before 3 H-palmitic acid incorporation, TEM8/1-HA–expressing cells were pretreated either with cycloheximide, 2-bromopalmitate, or brefeldin A and were analyzed by autoradiography ( 3 H-palmitate) or Western blotting (anti-tag HA). (D) Autoradiograms from C were quantified by densitometry using ScanAnalysis software (Biosoft). Error bars correspond to SD ( n = 3). (E) CHO ΔATR cells transiently transfected for 48 h with TEM8/1-HA cDNA were pulsed either for 2 h with 3 H-palmitic acid or for 30 min with [ 35 S]cysteine/methionine and were chased for different times. After anti-HA immunoprecipitation, samples were analyzed by SDS-PAGE followed by autoradiography and densitometry ( 3 H) or Phosphoimager analysis ( 35 S; same curve as in Fig. 5 B ). Results were normalized to the values at time = 0. ) and CMG2/4-V5 ( Fig. 3 B ), respectively, were detected. This band was sensitive to in vitro hydroxylamine treatment ( Fig. 3 A ) or cellular treatment with 2-bromopalmitate ( Fig. 3 C ) as shown for TEM8/1-HA, indicating that palmitate addition occurred via a thioester bond. The short isoform of TEM8 containing only two cysteines was also palmitoylated (Fig. S1, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ), as was CMG2/1 (not depicted). To investigate when during the life cycle of TEM8/1 palmitoylation occurred, 3 H-palmitate labeling was performed either in the presence of cycloheximide to inhibit protein synthesis or in the presence of brefeldin A to inhibit transport from the endoplasmic reticulum to the Golgi. Neither drug significantly affected the amount of immunoprecipitated receptor ( Fig. 3 C , anti-HA Western blot). However, both drugs led to an ∼50% reduction in 3 H-palmitate incorporation, as quantified by densitometry, but not to full inhibition ( Fig. 3, C and D ), suggesting that palmitoylation occurs both in the early secretory pathway and later in the life cycle of the protein. A pulse-chase performed after 3 H-palmitate incorporation showed that the palmitate groups were lost within an hour ( Fig. 3 E ). This loss was caused by the reversible nature of the modification and not by degradation of the protein because a pulse-chase experiment using [ 35 S]cysteine/methionine showed that the half-life of TEM8/1-HA in these experiments well exceeded 5 h ( Fig. 3 E ; same 35 S curve as in Fig. 5 B ). TEM8/1 is palmitoylated on multiple cysteines To investigate which of the four TEM8/1 cytoplasmic cysteines can be palmitoylated, we first concentrated on the two juxtamembrane cysteines because TEM8/2, which contains only these two cysteines, is palmitoylated (Fig. S1). Moreover, palmitoylation sites adjacent to the transmembrane region have been previously reported for several proteins such as CD4 ( Bijlmakers and Marsh, 2003 ) or members of the SNARE family of membrane fusion proteins ( Rothman, 1994 ). Cysteines at positions 346 and 347 in TEM8/1 were changed to alanine in single (mutants AC and CA) and double (mutant AA) mutants. All three mutants were expressed to lower levels than wild type (WT; Fig. 4 A Figure 4. TEM8/1 is palmitoylated on multiple cysteines. (A–C) CHO ΔATR cells were transiently transfected for 48 h with WT (CC) or mutant TEM8/1-HA cDNA in which one or both of the first two cytoplasmic cysteines were changed to alanine (mutants AC, CA, and AA). Cells were treated (B) or untreated (A) with the proteasome inhibitor MG132. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against the HA tag. (C) CHO ΔATR cells transfected with WT or mutant TEM8/1-HA were incubated with MG132 and 3 H-palmitic acid for 2 h and submitted to immunoprecipitation against the HA tag. Samples were analyzed by SDS-PAGE followed by autoradiography and Western blotting against HA. (D–F) Similar experiments were performed on single to quadruple mutants of the four cysteine residues. (G) CHO ΔATR cells transiently transfected for 48 h with WT (CCCC) or mutant TEM8/1-HA cDNAs were incubated with 500 ng/ml PA83 for 1 h at 4°C followed by 30 min at 37°C. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against PA. ). This was not a result of lower transfection efficiencies because equivalent TEM8/1 expression levels were observed when treating cells with the proteasome inhibitor MG132 ( Fig. 4 B ). To compare cells that express similar amounts of receptor and have expression levels that allow the detection of 3 H- palmitate, incorporation was performed on MG132-treated cells. Only when the second cysteine was modified were lower levels of 3 H-palmitate incorporation observed ( Fig. 4 C ), indicating that Cys-347 is palmitoylated in the WT protein. The observation that the double mutant still incorporated significant amounts of 3 H-palmitate, however, indicated that other cysteines were modified. Thus, Cys-481 and Cys-521 in TEM8/1 were also changed to alanine in single to quadruple mutants. Once more, all mutants were expressed at lower levels than the WT receptor ( Fig. 4 D ), an effect that could similarly be overcome by treating cells with MG132 ( Fig. 4 E ). These mutants were expressed at the cell surface as indicated by their ability to bind PA ( Fig. 4 G ). The quadruple mutant AAAA and the AAAC mutant did not incorporate 3 H-palmitate ( Fig. 4 F ). However, the fact that the CCCA mutant in repeated experiments had a somewhat lower incorporation than the WT TEM8/1 suggests that Cys-521 can be palmitoylated, possibly in a subpopulation of receptors, but only when other palmitoylation sites are present. The AAC 481 A mutant was always significantly modified, as were all of the mutants with a cysteine at position 481. Thus, repeated palmitoylation experiments showed that mutations of Cys-347 and Cys-481 always led to a drastic decrease in 3 H-palmitate incorporation, whereas the mutation of Cys-521 had a milder but significant effect. It has previously been observed that the mutation of palmitoylation sites leads to the aberrant palmitoylation of remaining cysteines, especially in double-cysteine motifs ( Percherancier et al., 2001 ; Sims and Wiedmer, 2001 ; Wiedmer et al., 2003 ). Therefore, we focused further experiments on the quadruple AAAA mutant in comparison with the WT (CCCC) receptor. Palmitoylation modulates the half-life of TEM8 The lower expression levels of all palmitoylation mutants when compared with WT suggested that palmitoylation affects the half-life of the receptor. Therefore, pulse-chase experiments using [ 35 S]cysteine/methionine labeling were performed in transiently transfected CHO ΔATR cells. The initial level of synthesis was very similar for the WT and AAAA mutant receptors ( Fig. 5 A Figure 5. Palmitoylation-deficient TEM8/1 has a reduced half-life. (A) CHO ΔATR cells transiently transfected for 30 h with plasmids expressing WT or AAAA TEM8/1 were submitted to a pulse-chase analysis with [ 35 S]methionine/cysteine. Immunoprecipitated receptors were followed by autoradiography (8-d exposure). Mature TEM8 is labeled m, and the Endo H–sensitive precursor is labeled p. The band labeled with an asterisk is an unknown coimmunoprecipitated protein that is not detected in other cells, such as HeLa. (B) TEM8/1 radioactivity was quantified using a Phosphoimager (Bio-Rad Laboratories). Results correspond to the mean of two experiments and were normalized to the radioactivity at time = 0. (C–E) Cell extract for CHO ΔATR cells stably expressing WT (CCCC) or mutant AAAA TEM8/1-HA were analyzed by Western blotting (C and D) or immunofluorescence (E) against HA. (D and E) Cells were left untreated, treated with MG132, or fed with leupeptin. In E, the exposure times were identical for all images, but a 75% cut-off filter on the excitation beam was used for the MG132 condition. Untransfected CHO ΔATR cells are shown for comparison. Bar, 10 μM. ). However, 50% of the AAAA mutant was lost after ∼130 min ( Fig. 5 B ), whereas >60% of the WT receptors were still present after 5 h. The 30% loss in AAAA TEM8/1 during the first hour of chase ( Fig. 5, A and B ) suggests that palmitoylation might somewhat affect folding/trafficking through the early secretory pathway. Because this cannot account for the drastic reduction in receptor half-life, we investigated whether defective palmitoylation could cause premature targeting of TEM8/1 to lysosomes. We first generated stable cell lines expressing AAAA TEM8/1 and found, as expected, lower steady-state expression levels ( Fig. 5 C ). Cells were then fed with an inhibitor of lysosomal enzymes, leupeptin (inhibitor of serine and cysteine proteases). This treatment led to some protection of the WT receptor (at the low exposure shown in Fig. 5 D , CCCC TEM8/1 was only detected in leupeptin-treated cells). However, the effect was far more pronounced for AAAA TEM8/1 ( Fig. 5 D ). Intracellular accumulation was confirmed by immunofluorescence ( Fig. 5 E ). Although in the absence of treatment, AAAA TEM8/ 1-HA was undetectable by fluorescence microscopy, leupeptin feeding led to the appearance of punctate perinuclear structures ( Fig. 5 E ), which were presumably late endosomes/lysosomes. This was in contrast with the effect of MG132, which also led to a massive increase in staining but primarily at the plasma membrane ( Fig. 5 E ). Thus, palmitoylation of TEM8/1 appears to be crucial in preventing premature targeting to lysosomes. Palmitoylation is a negative regulator of TEM8 DRM association We next investigated whether the palmitoylation of TEM8/1 is involved in regulating its association with DRMs ( Abrami et al., 2003 ). Stably expressed WT TEM8/1 was found in detergent-sensitive fractions ( Fig. 6 A Figure 6. Palmitoylation-deficient TEM8/1 associates constitutively with DRMs. (A) DRMs were prepared from CHO ΔATR cells stably expressing CCCC or AAAA TEM8/1-HA, and the distribution of receptors was analyzed by Western blotting against HA. A higher exposure is shown for the AAAA mutant. (B) CHO ΔATR cells stably expressing AAAA TEM8/1-HA were left untreated or treated with 2-bromopalmitate. DRMs were prepared, and the distribution of mutant TEM8/1-HA and caveolin-1 was analyzed by Western blotting. Mature TEM8/1-HA is labeled m, and the Endo H–sensitive precursor is labeled p. , left) as previously described ( Abrami et al., 2003 ). In contrast, AAAA TEM8/1 was almost entirely associated with DRMs ( Fig. 6 A , right). These observations indicate that palmitoylation acts as a negative regulator of TEM8/1 DRM association. The drastic difference between the AAAA mutant and the WT receptor suggests that the bulk of the WT receptor is palmitoylated at steady state. These observations also raise the interesting possibility, which is not addressed in this study, that depalmitoylation by a protein-thioesterase activity could be a mechanism for the regulation of raft association. Regulation of protein localization by palmitoylation/depalmitoylation has been proposed for soluble proteins in particular ( Drenan et al., 2005 ; Rocks et al., 2005 ). The observation that palmitoylation-deficient TEM8/1 (AAAA) is exclusively found in DRMs ( Fig. 6 A ) is in apparent contradiction with the fact that 2-bromopalmitate inhibited DRM association of PA63 ( Fig. 2 B ). Therefore, we tested whether the drug would also affect the DRM association of AAAA TEM8/1. The mature Endo H–resistant form of the mutant receptor entirely relocalized to the detergent-soluble fractions ( Fig. 6 B ), indicating the involvement of a palmitoylation event in the association of AAAA TEM8/1 with DRMs, the substrate of which must be a protein other than the receptor itself. PA induced ubiquitination of anthrax toxin receptors We have previously shown that heptamerization of PA not only triggers the redistribution of the toxin–receptor complex to lipid rafts but also triggers its rapid endocytosis ( Abrami et al., 2003 ). We investigated whether this toxin-induced uptake could be caused by a second posttranslational modification of the receptor cytoplasmic tail that would be raft dependent. We focused our attention on ubiquitination ( Haglund and Dikic, 2005 ). CHO ΔATR cells transiently transfected with TEM8/1-HA or CMG2/4-V5 were treated with PA for various times. After immunoprecipitation of the receptors, Western blots were performed with anti-tag and anti-Ub antibodies. The addition of the toxin clearly led to the appearance of a smeared Ub-positive band both for TEM8/1 ( Fig. 7 A Figure 7. PA triggers ubiquitination of its receptor. CHO ΔATR cells were transiently transfected for 48 h with WT (A and B) or mutant forms (C and D) of TEM8/1-HA or CMG2/4-V5. 1 μg/ml PA83 was either added or not added to cells for 1 h at 4°C and shifted for different times to 37°C. After immunoprecipitation against HA (A and C) or V5 (B and D), samples were analyzed by Western blotting using anti-Ub, anti-HA, or V5 and anti-PA. In TEM8/1 K6R, lysines 352, 372, 373, 374, 412, and 414 were changed to arginine. HC, heavy chain. ) and for CMG2/4 ( Fig. 7 B ). The absence of ladder was suggestive of monoubiquitination rather than polyubiquitination with long chains, as observed for Ub 48 ubiquitination and proteasomal degradation ( Haglund and Dikic, 2005 ). Experiments performed with the shorter TEM8/2 isoform similarly led to the detection of a ubiquitinated band (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ). The ubiquitinated band detected for TEM8/2 was smaller than that detected for TEM8/1, suggesting that the receptor itself was the modified protein rather than an interacting partner. To confirm this, lysine mutagenesis was performed. Of the 16 lysines in TEM8/1 and 14 in CMG2/4, we first changed Lys-352 to arginine in TEM8 and the corresponding Lys-350 in CMG2/4 because (1) it is the only lysine common to TEM8/1 and TEM8/2; (2) it is conserved between TEM8 and CMG2; and (3) Valdez-Taubas and Pelham (2005) suggested that the palmitoylation of Tlg1 prevents access of juxtamembranous lysines by E3 ligases. Although the ubiquitinated band could still be detected after the addition of PA to K352R TEM8/1 expression cells, ubiquitination was greatly diminished ( Fig. 7 C ). A stronger effect was obtained when mutating 6 of the 16 lysines to arginine (K 6 R mutant in which all lysines labeled with an asterisk in Fig. 1 A were changed to arginine, including Lys-352). These experiments confirm that TEM8/1 is the substrate of the ubiquitination reaction and that Lys-352 is one of the modified sites but that additional lysines might be modified. The effect of mutating the first conserved juxtamembranous lysine to arginine was even more drastic in CMG2/4. No Ub-positive band could be detected in PA-treated K350R CMG2/4–transfected cells ( Fig. 7 D ). To investigate whether ubiquitination of the receptor is important for endocytosis of the anthrax toxin, we monitored the appearance of the SDS-resistant PA 7mer pore in lysine mutant-expressing cells. As shown in Fig. 7 (C and D) , the appearance of the SDS-resistant pore was either delayed or strongly diminished in the TEM8/1 and CMG2/4 lysine mutant–expressing cells, demonstrating that ubiquitination of the receptor is important for efficient endocytosis. TEM8 ubiquitination is DRM mediated Because endocytosis of TEM8 requires both raft association ( Abrami et al., 2003 ) and ubiquitination ( Fig. 7 ), we investigated whether these two events were linked. DRMs were isolated from transiently transfected CHO ΔATR cells. The bulk of TEM8/1 was found in detergent-soluble fractions ( Fig. 8 A Figure 8. Endocytosis of anthrax toxin receptor requires DRM-mediated ubiquitination and the E3 ligase Cbl. (A) CHO ΔATR cells transfected for 48 h with WT TEM8/1-HA were incubated with 1 μg/ml PA83 for 1 h at 4°C followed by 40 min at 37°C, solubilized in Triton X-100 at 4°C, and separated on an OptiPrep gradient. TEM8/1-HA was immunoprecipitated from each fraction and analyzed by SDS-PAGE and Western blotting using anti-Ub, anti-HA, and anti-PA antibodies. (B) CHO ΔATR cells transfected for 48 h with WT TEM8/1-HA were treated with βMCD to extract cholesterol, incubated with 1 μg/ml PA83 for 1 h at 4°C, and shifted for different times at 37°C. After immunoprecipitation with anti-HA beads, samples were analyzed by Western blotting using anti-Ub and anti-HA antibodies. (C) HeLa cells were transfected or untransfected with siRNAs against Cbl for 72 h and incubated with 500 ng/ml PA83 for different times at 37°C. Cell extracts were blotted for Cbl, tubulin (as an equal loading marker), and PA. (D) HeLa cells were untransfected or transfected with siRNAs against Cbl for a total of 72 h in total. 24 h later, these cells were additionally transfected with TEM8/1-HA for 48 h and incubated with 500 ng/ml PA83 for different times at 37°C. TEM8/1-HA was immunoprecipitated from each fraction and analyzed by SDS-PAGE and Western blotting using anti-Ub and anti-HA antibodies. , left), as also observed in Fig. 6 A . In marked contrast, ubiquinated TEM8/1 was detected exclusively in DRMs ( Fig. 8 A , left). When cells were treated with the toxin before Triton X-100 solubilization, PA63 was associated with DRMs (as in Fig. 2 B ) and led to the recruitment of toxin-bound TEM8/1 to this fraction ( Fig. 8 A , right). Concomitantly, the ubiquitinated form of TEM8/1 was increased in the same fraction ( Fig. 8 A ). Raft impairment by cholesterol extraction using the sequestering agent β-methylcyclodextrin (ßMCD) led to a strong inhibition of PA-induced TEM8/1 ubiquitination ( Fig. 8 B ). This observation suggests that microdomain association precedes and is required for this posttranslational modification. The E3 Ub ligase Cbl is required for anthrax toxin endocytosis Cbl is an E3 Ub ligase that can interact with lipid rafts ( Lafont and Simons, 2001 ; Haglund et al., 2004 ). To test for the involvement of Cbl in anthrax toxin endocytosis, we decided to perform RNA silencing. HeLa cells were used because of their human origin (the sequence of hamster Cbl is not available) and their high transfection efficiencies. These cells express TEM8 as indicated by the pH sensitivity of PA channel formation ( Rainey et al., 2005 ). As shown in Fig. 8 C , Cbl could be efficiently silenced by this method. The absence of Cbl did not affect binding of the toxin as indicated by the unaltered presence of PA83/PA63. However, appearance of the PA 7mer pore was drastically inhibited ( Fig. 8 C ). To investigate the effect of Cbl RNA interference on the ubiquitination of TEM8 itself, RNA interference–treated cells were transfected with TEM8/1-HA, and toxin-induced ubiquitination after immunoprecipitation of the receptors was measured. As shown Fig. 8 D , the ubiquitinated form of TEM8/1 could not longer be detected. Theses experiments show that Cbl is responsible for the ubiquitination of TEM8/1 and its subsequent internalization. Constitutive ubiquitination of palmitoylation-deficient TEM8/1 and the consequences as an anthrax toxin receptor We found that AAAA TEM8/1 is constitutively associated with DRMs and that TEM8 ubiquitination is a raft-dependent modification. Therefore, we wondered whether AAAA TEM8/1 would be constitutively ubiquitinated. As shown in Fig. 9 A Figure 9. Cells expressing the palmitoylation-deficient TEM8/1 are less sensitive to anthrax toxin. (A and B) Anti-HA immunoprecipitation experiments were performed on CHO ΔATR cells stably or transiently (48 h) expressing CCCC or AAAA mutant TEM8/1-HA. Western blotting was performed against Ub and HA. (B) Cells were left untreated or treated with βMCD before lysis and immunoprecipitation. (C) CHO ΔATR cells stably expressing WT or AAAA TEM8/1-HA were treated with 500 ng/ml PA83 and 200 ng/ml LF at 37°C for different times. 40 μg of cell extracts were analyzed by SDS-PAGE followed by Western blotting against PA, HA, and MEK1 (NH 2 -terminal–directed antibody). (D) CHO ΔATR cells transiently (48 h) expressing WT and AAAA mutant TEM8/1-HA were treated with 500 ng/ml of the furin-resistant PA SNKE mutant for 1 h at 4°C and shifted to 37°C for different times. Surface-bound toxin was shaved off with trypsin (10 min at 37°Δ). To prevent lysosomal degradation of the internalized PA, cells were treated with 10 μM nocodazole to block microtubule-dependent transport to late endosomes. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against PA and HA. , the steady-state ubiquitination level of AAAA TEM8/1 was markedly higher than that of WT CCCC TEM8/1 (especially when comparing the levels of Ub vs. HA) both in stable cell lines and upon transient transfection (note that after immunoprecipitation of TEM8/1-HA from transiently transfected cells, the levels of expressed receptors seem to be similar even though analysis of total extracts shows a lower abundance of the AAAA mutant). Ubiquitination of AAAA TEM8/1 was dependent on the integrity of lipid rafts because the removal of cholesterol using ßMCD led to a drastic reduction in the level of AAAA TEM8/1 ubiquitination. This later observation also indicates that increased ubiquitination of AAAA TEM8/1 is not a consequence of misfolding of the cytoplasmic tail as a consequence of mutagenesis, because such an event would have been insensitive to acute cholesterol extraction from the plasma membrane. Constitutive ubiquitination of the AAAA TEM8/1 and, thus, its constitutive endocytosis are likely to affect its ability to act as an anthrax toxin receptor. To address this issue directly, we monitored the cleavage kinetics of the LF target MEK1. Whereas MEK1 underwent LF-dependent cleavage in WT TEM8/1–expressing cells, the MAPK kinase remained intact in the AAAA TEM8/1–expressing cells during the time course of the experiment ( Fig. 9 C ). To test whether this lack of cleavage was only a result of the reduced number of surface-expressed receptors (levels of TEM-HA), we treated AAAA TEM8/1–expressing cells with a higher concentration of PA to reach similar amounts of bound PA as on WT TEM8/1–expressing cells (Fig. S3, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ). Interestingly, even under these conditions, MEK1 cleavage in AAAA TEM8/1–expressing cells was minimal (Fig. S3), indicating that reduced PA binding does fully account for the reduced MEK1 cleavage in these cells. Because AAAA TEM8/1 undergoes significant constitutive endocytosis, we tested whether it would mediate the internalization of PA83, which is an event that does not occur with the WT receptor and for which PA heptamerization is required ( Abrami et al., 2003 ; Liu and Leppla, 2003 ). In this study, we made use of a mutant PA (PA SNKE ; Abrami et al., 2003 ) that is modified in the furin consensus cleavage site and, thus, remains in the PA83 form. As expected ( Abrami et al., 2003 ), PA SNKE was not internalized by WT receptors and was sensitive to surface trypsinization ( Fig. 9 D ). In contrast, all cell-bound PA SNKE became trypsin resistant in AAAA TEM8/ 1–expressing cells, indicating that it had been completely endocytosed ( Fig. 9 C , right). Characterization of TEM8/1-HA and CMG2/4-V5 TEM8 and CMG2 both exist as four isoforms, two of which, in each case, act as anthrax toxin receptors: isoforms 1 and 2 of TEM8 (TEM8/1 and TEM8/2) and isoforms 1 and 4 of CMG2 (CMG2/1 and CMG2/4; for review see Scobie and Young, 2005 ). These isoforms differ only in their cytoplasmic tails ( Fig. 1 A Figure 1. TEM8/1 and CMG2/4 are glycosylated. (A) Alignment of the cytoplasmic tails of human TEM8 and human CMG2 using the SIM software of the EXPASY server ( www.expasy.ch ). Regions of identity are shown in yellow, lysine residues in green, and cysteine residues in red. Numbering of the residues corresponds to that of TEM8/1. The asterisks label the lysine mutants in the K6R mutant described in Fig. 7. (B and C) CHO ΔATR cells untransfected and stably or transiently transfected (for 48 h; B) with a TEM8/1-HA construct were analyzed by Western blotting (40 μg of protein/lane; B) or immunofluorescence (C) using an anti-HA antibody. Bar, 10 μM. (D) CHO ΔATR cells stably expressing TEM8/1-HA were submitted or unsubmitted to surface trypsinization at 4°C and subsequently analyzed for the presence of TEM8/1-HA by Western blotting. (E) CHO ΔATR cells transiently expressing TEM8/1-HA were grown in the presence or absence of tunicamycin. Control cell extracts were subsequently left untreated or treated with N -glycosidase F or Endo H. The effects of these treatments were analyzed by SDS-PAGE followed by Western blotting against the HA tag. (B and E) Band u, unglycosilated form; p, glycosylated precursor; m, mature form. (F) Extracts of CHO ΔATR cells transiently expressing or not expressing CMG2/4-V5 were left untreated or treated with N -glycosidase F or Endo H and subsequently analyzed by SDS-PAGE followed by Western blotting against the V5 tag. ); TEM8/2 has a far shorter cytoplasmic tail than TEM8/1 and CMG2/1, and CMG2/4 differs only in the last 12–13 residues. Even between TEM8 and CMG2, the long tails have a very high degree of conservation ( Fig. 1 A ). Therefore, we have focused this study mainly on TEM8/1 but repeated key experiments with the short TEM8/2 isoform as well as with CMG2/4. TEM8/1 was tagged with an HA epitope at the COOH terminus and was either transiently transfected or stably expressed in receptor-deficient CHO cells (CHO ΔATR ; Liu and Leppla, 2003 ). Expression levels were overall higher upon transient transfection ( Fig. 1 B ) but were still readily detectable in the stably TEM8/1-HA–expressing cells ( Fig. 1, B and C ). Most of the expressed receptor was present at the cell surface as indicated by its sensitivity to cell surface trypsinization ( Fig. 1 D ). TEM8-HA always migrated as a doublet with an upper smeared band ( Fig. 1 B , m) and a lower well-defined band ( Fig. 1 B , p), the intensity of which varied greatly from experiment to experiment. We investigated whether this migration pattern was caused by glycosylation because the extracellular domain of TEM8 has three predicted N -glycosylation sites. Treatment of cells with tunicamycin, an inhibitor of N -glycosylation, led to the appearance of a third, lower mobility band (u; unglycosilated) with a concomitant decrease in the intensity of bands m and p ( Fig. 1 E ). N -glycosidase F treatment of control cell extracts led to the complete disappearance of bands m and p to the benefit of band u ( Fig. 1 E ), showing that m and p both correspond to glycosylated TEM8. Band m was endoglycosidase H (Endo H) resistant, and band p was Endo H sensitive ( Fig. 1 E ). This shows that TEM8 in band m (mature) had acquired complex Golgi-modified sugars, whereas that in band p (precursor) contained incompletely modified sugars. CMG2/4 was expressed with a COOH-terminal V5 tag ( Dowling et al., 2003 ). As for TEM8, CMG2/4-V5 migrated as a smear plus a lower molecular weight band ( Fig. 1 F ). Both bands were sensitive to N -glycosidase F treatment, and, as for TEM8, only the lower band was sensitive to Endo H ( Fig. 1 F ). Altogether, this indicates that CMG2/4 is also glycosylated, as predicted by the presence of two N -glycosylation sites in the ectodomain. PA DRM association is palmitoylation dependent We have previously shown that the anthrax toxin modifies the surface distribution of its receptor, inducing its clustering in cholesterol-rich raftlike domains. We wondered whether posttranslational modifications of the receptor tail could be involved in regulating interactions of TEM8 and CMG2 with membrane domains. TEM8/1 and /2 as well as CMG2/1 and /4 all contain cysteine residues in their cytoplasmic tails, which are potential sites for S-palmitoylation ( Linder and Deschenes, 2004 ). More specifically, all four proteins have two conserved juxtamembranous cysteines ( Fig. 1 A ; Cys-346 and Cys-347 in TEM8), a third cysteine is conserved between TEM8/1 (Cys-481) and CMG2, and a fourth unconserved cysteine is found in TEM8/1 (Cys-521) and CMG2/1 (Cys482). To test whether palmitoylation plays a role in anthrax toxin endocytosis, we analyzed whether the palmitoylation inhibitor 2-bromopalmitate ( Webb et al., 2000 ) would affect the raft association of PA63 in BHK cells, a cell line that expresses transmembrane isoforms of both TEM8 and CMG2 (unpublished data). We first verified that PA binding was not affected by the treatment ( Fig. 2 A Figure 2. Palmitoylation events are required for DRM association and internalization of PA. (A and B) Control BHK cells were pretreated or untreated with bromopalmitate and were incubated with 500 ng/ml nicked PA63 and 20 ng/ml aerolysin for 1 h at 4°C followed by 10 min at 37°C. (A) Cell extracts were submitted to SDS-PAGE followed by Western blotting to reveal PA63, aerolysin, and caveolin-1 (Cav-1). (B) Cells were solubilized in 1% Triton X-100, run on an OptiPrep gradient, and each fraction was analyzed by SDS-PAGE followed by Western blotting against PA, aerolysin, and caveolin-1. (C) BHK cells were pretreated with bromopalmitate and incubated with 500 ng/ml nicked PA63 for 1 h at 4°C followed by different times at 37°C. Cell extracts (40 μg of protein) were analyzed by SDS-PAGE and Western blotting to reveal the SDS-resistant PA 7mer pore and MEK1 (NH 2 -terminal directed). To detect the prepore (SDS-sensitive nonmembrane-inserted PA 7mer ), cell extracts were submitted to an acid pulse before SDS analysis. ). Raft association was then monitored by following the association with detergent-resistant membranes (DRMs; Brown and London, 1998 ). PA63 was associated with DRMs in control cells as previously observed ( Abrami et al., 2003 ) but shifted to the detergent-soluble fractions in drug-treated cells ( Fig. 2 B ). This was not caused by a general disruption of lipid rafts because glycosyl-phosphatidylinositol–anchored proteins, followed here using the glycosyl-phosphatidylinositol–specific bacterial toxin aerolysin ( Fivaz et al., 2002 ) as well as caveolin-1, remained primarily in the DRM fractions. Although caveolin-1 is palmitoylated on three cysteines, the modification is not required for DRM association ( Dietzen et al., 1995 ). The inhibitory effect of 2-bromopalmitate on PA63-raft association was also accompanied by the inhibition of endocytosis as reflected by a lack of SDS-resistant PA 7mer pore and the absence of LF-mediated cleavage of one of the LF substrates, the MAPK kinase MEK1 ( Fig. 2 C ; for review see Collier and Young, 2003 ). To test whether 2-bromopalmitate inhibited the heptamerization process itself, we designed an assay to detect cell surface–formed prepores. These are SDS sensitive and, therefore, are not detected by SDS-PAGE. However, they can be converted to the SDS-resistant phenotype by submitting cell extracts to low pH (pH 4.5) before SDS-PAGE analysis, as illustrated in Fig. 2 C (middle left) for control cells. The prepore was undetectable for 2-bromopalmitate–treated cells ( Fig. 2 C ). TEM8 and CMG2 are palmitoylated The aforementioned experiments show that palmitoylation events are important for anthrax toxin raft association and heptamerization. To determine whether the receptors themselves are palmitoylated, TEM8/1-HA and CMG2/4-V5 were immunoprecipitated from lysates of 3 H-palmitic acid–labeled CHO ΔATR cells transiently transfected with the respective constructs. Radiolabeled bands with motilities similar to that of TEM8/1-HA ( Fig. 3 A Figure 3. TEM8/1 and CMG2/4 anthrax toxin receptors are palmitoylated. CHO ΔATR cells transfected with TEM8/1-HA (A) or CMG2/4-V5 (B) were incubated with 3 H-palmitic acid for 2 h before immunoprecipitation using anti-tag antibodies. Immunoprecipitates were split into two, run on 4–20% gels, and analyzed either by autoradiography ( 3 H-palmitate) or Western blotting (anti-tag HA or V5). (C) Before 3 H-palmitic acid incorporation, TEM8/1-HA–expressing cells were pretreated either with cycloheximide, 2-bromopalmitate, or brefeldin A and were analyzed by autoradiography ( 3 H-palmitate) or Western blotting (anti-tag HA). (D) Autoradiograms from C were quantified by densitometry using ScanAnalysis software (Biosoft). Error bars correspond to SD ( n = 3). (E) CHO ΔATR cells transiently transfected for 48 h with TEM8/1-HA cDNA were pulsed either for 2 h with 3 H-palmitic acid or for 30 min with [ 35 S]cysteine/methionine and were chased for different times. After anti-HA immunoprecipitation, samples were analyzed by SDS-PAGE followed by autoradiography and densitometry ( 3 H) or Phosphoimager analysis ( 35 S; same curve as in Fig. 5 B ). Results were normalized to the values at time = 0. ) and CMG2/4-V5 ( Fig. 3 B ), respectively, were detected. This band was sensitive to in vitro hydroxylamine treatment ( Fig. 3 A ) or cellular treatment with 2-bromopalmitate ( Fig. 3 C ) as shown for TEM8/1-HA, indicating that palmitate addition occurred via a thioester bond. The short isoform of TEM8 containing only two cysteines was also palmitoylated (Fig. S1, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ), as was CMG2/1 (not depicted). To investigate when during the life cycle of TEM8/1 palmitoylation occurred, 3 H-palmitate labeling was performed either in the presence of cycloheximide to inhibit protein synthesis or in the presence of brefeldin A to inhibit transport from the endoplasmic reticulum to the Golgi. Neither drug significantly affected the amount of immunoprecipitated receptor ( Fig. 3 C , anti-HA Western blot). However, both drugs led to an ∼50% reduction in 3 H-palmitate incorporation, as quantified by densitometry, but not to full inhibition ( Fig. 3, C and D ), suggesting that palmitoylation occurs both in the early secretory pathway and later in the life cycle of the protein. A pulse-chase performed after 3 H-palmitate incorporation showed that the palmitate groups were lost within an hour ( Fig. 3 E ). This loss was caused by the reversible nature of the modification and not by degradation of the protein because a pulse-chase experiment using [ 35 S]cysteine/methionine showed that the half-life of TEM8/1-HA in these experiments well exceeded 5 h ( Fig. 3 E ; same 35 S curve as in Fig. 5 B ). TEM8/1 is palmitoylated on multiple cysteines To investigate which of the four TEM8/1 cytoplasmic cysteines can be palmitoylated, we first concentrated on the two juxtamembrane cysteines because TEM8/2, which contains only these two cysteines, is palmitoylated (Fig. S1). Moreover, palmitoylation sites adjacent to the transmembrane region have been previously reported for several proteins such as CD4 ( Bijlmakers and Marsh, 2003 ) or members of the SNARE family of membrane fusion proteins ( Rothman, 1994 ). Cysteines at positions 346 and 347 in TEM8/1 were changed to alanine in single (mutants AC and CA) and double (mutant AA) mutants. All three mutants were expressed to lower levels than wild type (WT; Fig. 4 A Figure 4. TEM8/1 is palmitoylated on multiple cysteines. (A–C) CHO ΔATR cells were transiently transfected for 48 h with WT (CC) or mutant TEM8/1-HA cDNA in which one or both of the first two cytoplasmic cysteines were changed to alanine (mutants AC, CA, and AA). Cells were treated (B) or untreated (A) with the proteasome inhibitor MG132. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against the HA tag. (C) CHO ΔATR cells transfected with WT or mutant TEM8/1-HA were incubated with MG132 and 3 H-palmitic acid for 2 h and submitted to immunoprecipitation against the HA tag. Samples were analyzed by SDS-PAGE followed by autoradiography and Western blotting against HA. (D–F) Similar experiments were performed on single to quadruple mutants of the four cysteine residues. (G) CHO ΔATR cells transiently transfected for 48 h with WT (CCCC) or mutant TEM8/1-HA cDNAs were incubated with 500 ng/ml PA83 for 1 h at 4°C followed by 30 min at 37°C. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against PA. ). This was not a result of lower transfection efficiencies because equivalent TEM8/1 expression levels were observed when treating cells with the proteasome inhibitor MG132 ( Fig. 4 B ). To compare cells that express similar amounts of receptor and have expression levels that allow the detection of 3 H- palmitate, incorporation was performed on MG132-treated cells. Only when the second cysteine was modified were lower levels of 3 H-palmitate incorporation observed ( Fig. 4 C ), indicating that Cys-347 is palmitoylated in the WT protein. The observation that the double mutant still incorporated significant amounts of 3 H-palmitate, however, indicated that other cysteines were modified. Thus, Cys-481 and Cys-521 in TEM8/1 were also changed to alanine in single to quadruple mutants. Once more, all mutants were expressed at lower levels than the WT receptor ( Fig. 4 D ), an effect that could similarly be overcome by treating cells with MG132 ( Fig. 4 E ). These mutants were expressed at the cell surface as indicated by their ability to bind PA ( Fig. 4 G ). The quadruple mutant AAAA and the AAAC mutant did not incorporate 3 H-palmitate ( Fig. 4 F ). However, the fact that the CCCA mutant in repeated experiments had a somewhat lower incorporation than the WT TEM8/1 suggests that Cys-521 can be palmitoylated, possibly in a subpopulation of receptors, but only when other palmitoylation sites are present. The AAC 481 A mutant was always significantly modified, as were all of the mutants with a cysteine at position 481. Thus, repeated palmitoylation experiments showed that mutations of Cys-347 and Cys-481 always led to a drastic decrease in 3 H-palmitate incorporation, whereas the mutation of Cys-521 had a milder but significant effect. It has previously been observed that the mutation of palmitoylation sites leads to the aberrant palmitoylation of remaining cysteines, especially in double-cysteine motifs ( Percherancier et al., 2001 ; Sims and Wiedmer, 2001 ; Wiedmer et al., 2003 ). Therefore, we focused further experiments on the quadruple AAAA mutant in comparison with the WT (CCCC) receptor. Palmitoylation modulates the half-life of TEM8 The lower expression levels of all palmitoylation mutants when compared with WT suggested that palmitoylation affects the half-life of the receptor. Therefore, pulse-chase experiments using [ 35 S]cysteine/methionine labeling were performed in transiently transfected CHO ΔATR cells. The initial level of synthesis was very similar for the WT and AAAA mutant receptors ( Fig. 5 A Figure 5. Palmitoylation-deficient TEM8/1 has a reduced half-life. (A) CHO ΔATR cells transiently transfected for 30 h with plasmids expressing WT or AAAA TEM8/1 were submitted to a pulse-chase analysis with [ 35 S]methionine/cysteine. Immunoprecipitated receptors were followed by autoradiography (8-d exposure). Mature TEM8 is labeled m, and the Endo H–sensitive precursor is labeled p. The band labeled with an asterisk is an unknown coimmunoprecipitated protein that is not detected in other cells, such as HeLa. (B) TEM8/1 radioactivity was quantified using a Phosphoimager (Bio-Rad Laboratories). Results correspond to the mean of two experiments and were normalized to the radioactivity at time = 0. (C–E) Cell extract for CHO ΔATR cells stably expressing WT (CCCC) or mutant AAAA TEM8/1-HA were analyzed by Western blotting (C and D) or immunofluorescence (E) against HA. (D and E) Cells were left untreated, treated with MG132, or fed with leupeptin. In E, the exposure times were identical for all images, but a 75% cut-off filter on the excitation beam was used for the MG132 condition. Untransfected CHO ΔATR cells are shown for comparison. Bar, 10 μM. ). However, 50% of the AAAA mutant was lost after ∼130 min ( Fig. 5 B ), whereas >60% of the WT receptors were still present after 5 h. The 30% loss in AAAA TEM8/1 during the first hour of chase ( Fig. 5, A and B ) suggests that palmitoylation might somewhat affect folding/trafficking through the early secretory pathway. Because this cannot account for the drastic reduction in receptor half-life, we investigated whether defective palmitoylation could cause premature targeting of TEM8/1 to lysosomes. We first generated stable cell lines expressing AAAA TEM8/1 and found, as expected, lower steady-state expression levels ( Fig. 5 C ). Cells were then fed with an inhibitor of lysosomal enzymes, leupeptin (inhibitor of serine and cysteine proteases). This treatment led to some protection of the WT receptor (at the low exposure shown in Fig. 5 D , CCCC TEM8/1 was only detected in leupeptin-treated cells). However, the effect was far more pronounced for AAAA TEM8/1 ( Fig. 5 D ). Intracellular accumulation was confirmed by immunofluorescence ( Fig. 5 E ). Although in the absence of treatment, AAAA TEM8/ 1-HA was undetectable by fluorescence microscopy, leupeptin feeding led to the appearance of punctate perinuclear structures ( Fig. 5 E ), which were presumably late endosomes/lysosomes. This was in contrast with the effect of MG132, which also led to a massive increase in staining but primarily at the plasma membrane ( Fig. 5 E ). Thus, palmitoylation of TEM8/1 appears to be crucial in preventing premature targeting to lysosomes. Palmitoylation is a negative regulator of TEM8 DRM association We next investigated whether the palmitoylation of TEM8/1 is involved in regulating its association with DRMs ( Abrami et al., 2003 ). Stably expressed WT TEM8/1 was found in detergent-sensitive fractions ( Fig. 6 A Figure 6. Palmitoylation-deficient TEM8/1 associates constitutively with DRMs. (A) DRMs were prepared from CHO ΔATR cells stably expressing CCCC or AAAA TEM8/1-HA, and the distribution of receptors was analyzed by Western blotting against HA. A higher exposure is shown for the AAAA mutant. (B) CHO ΔATR cells stably expressing AAAA TEM8/1-HA were left untreated or treated with 2-bromopalmitate. DRMs were prepared, and the distribution of mutant TEM8/1-HA and caveolin-1 was analyzed by Western blotting. Mature TEM8/1-HA is labeled m, and the Endo H–sensitive precursor is labeled p. , left) as previously described ( Abrami et al., 2003 ). In contrast, AAAA TEM8/1 was almost entirely associated with DRMs ( Fig. 6 A , right). These observations indicate that palmitoylation acts as a negative regulator of TEM8/1 DRM association. The drastic difference between the AAAA mutant and the WT receptor suggests that the bulk of the WT receptor is palmitoylated at steady state. These observations also raise the interesting possibility, which is not addressed in this study, that depalmitoylation by a protein-thioesterase activity could be a mechanism for the regulation of raft association. Regulation of protein localization by palmitoylation/depalmitoylation has been proposed for soluble proteins in particular ( Drenan et al., 2005 ; Rocks et al., 2005 ). The observation that palmitoylation-deficient TEM8/1 (AAAA) is exclusively found in DRMs ( Fig. 6 A ) is in apparent contradiction with the fact that 2-bromopalmitate inhibited DRM association of PA63 ( Fig. 2 B ). Therefore, we tested whether the drug would also affect the DRM association of AAAA TEM8/1. The mature Endo H–resistant form of the mutant receptor entirely relocalized to the detergent-soluble fractions ( Fig. 6 B ), indicating the involvement of a palmitoylation event in the association of AAAA TEM8/1 with DRMs, the substrate of which must be a protein other than the receptor itself. PA induced ubiquitination of anthrax toxin receptors We have previously shown that heptamerization of PA not only triggers the redistribution of the toxin–receptor complex to lipid rafts but also triggers its rapid endocytosis ( Abrami et al., 2003 ). We investigated whether this toxin-induced uptake could be caused by a second posttranslational modification of the receptor cytoplasmic tail that would be raft dependent. We focused our attention on ubiquitination ( Haglund and Dikic, 2005 ). CHO ΔATR cells transiently transfected with TEM8/1-HA or CMG2/4-V5 were treated with PA for various times. After immunoprecipitation of the receptors, Western blots were performed with anti-tag and anti-Ub antibodies. The addition of the toxin clearly led to the appearance of a smeared Ub-positive band both for TEM8/1 ( Fig. 7 A Figure 7. PA triggers ubiquitination of its receptor. CHO ΔATR cells were transiently transfected for 48 h with WT (A and B) or mutant forms (C and D) of TEM8/1-HA or CMG2/4-V5. 1 μg/ml PA83 was either added or not added to cells for 1 h at 4°C and shifted for different times to 37°C. After immunoprecipitation against HA (A and C) or V5 (B and D), samples were analyzed by Western blotting using anti-Ub, anti-HA, or V5 and anti-PA. In TEM8/1 K6R, lysines 352, 372, 373, 374, 412, and 414 were changed to arginine. HC, heavy chain. ) and for CMG2/4 ( Fig. 7 B ). The absence of ladder was suggestive of monoubiquitination rather than polyubiquitination with long chains, as observed for Ub 48 ubiquitination and proteasomal degradation ( Haglund and Dikic, 2005 ). Experiments performed with the shorter TEM8/2 isoform similarly led to the detection of a ubiquitinated band (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ). The ubiquitinated band detected for TEM8/2 was smaller than that detected for TEM8/1, suggesting that the receptor itself was the modified protein rather than an interacting partner. To confirm this, lysine mutagenesis was performed. Of the 16 lysines in TEM8/1 and 14 in CMG2/4, we first changed Lys-352 to arginine in TEM8 and the corresponding Lys-350 in CMG2/4 because (1) it is the only lysine common to TEM8/1 and TEM8/2; (2) it is conserved between TEM8 and CMG2; and (3) Valdez-Taubas and Pelham (2005) suggested that the palmitoylation of Tlg1 prevents access of juxtamembranous lysines by E3 ligases. Although the ubiquitinated band could still be detected after the addition of PA to K352R TEM8/1 expression cells, ubiquitination was greatly diminished ( Fig. 7 C ). A stronger effect was obtained when mutating 6 of the 16 lysines to arginine (K 6 R mutant in which all lysines labeled with an asterisk in Fig. 1 A were changed to arginine, including Lys-352). These experiments confirm that TEM8/1 is the substrate of the ubiquitination reaction and that Lys-352 is one of the modified sites but that additional lysines might be modified. The effect of mutating the first conserved juxtamembranous lysine to arginine was even more drastic in CMG2/4. No Ub-positive band could be detected in PA-treated K350R CMG2/4–transfected cells ( Fig. 7 D ). To investigate whether ubiquitination of the receptor is important for endocytosis of the anthrax toxin, we monitored the appearance of the SDS-resistant PA 7mer pore in lysine mutant-expressing cells. As shown in Fig. 7 (C and D) , the appearance of the SDS-resistant pore was either delayed or strongly diminished in the TEM8/1 and CMG2/4 lysine mutant–expressing cells, demonstrating that ubiquitination of the receptor is important for efficient endocytosis. TEM8 ubiquitination is DRM mediated Because endocytosis of TEM8 requires both raft association ( Abrami et al., 2003 ) and ubiquitination ( Fig. 7 ), we investigated whether these two events were linked. DRMs were isolated from transiently transfected CHO ΔATR cells. The bulk of TEM8/1 was found in detergent-soluble fractions ( Fig. 8 A Figure 8. Endocytosis of anthrax toxin receptor requires DRM-mediated ubiquitination and the E3 ligase Cbl. (A) CHO ΔATR cells transfected for 48 h with WT TEM8/1-HA were incubated with 1 μg/ml PA83 for 1 h at 4°C followed by 40 min at 37°C, solubilized in Triton X-100 at 4°C, and separated on an OptiPrep gradient. TEM8/1-HA was immunoprecipitated from each fraction and analyzed by SDS-PAGE and Western blotting using anti-Ub, anti-HA, and anti-PA antibodies. (B) CHO ΔATR cells transfected for 48 h with WT TEM8/1-HA were treated with βMCD to extract cholesterol, incubated with 1 μg/ml PA83 for 1 h at 4°C, and shifted for different times at 37°C. After immunoprecipitation with anti-HA beads, samples were analyzed by Western blotting using anti-Ub and anti-HA antibodies. (C) HeLa cells were transfected or untransfected with siRNAs against Cbl for 72 h and incubated with 500 ng/ml PA83 for different times at 37°C. Cell extracts were blotted for Cbl, tubulin (as an equal loading marker), and PA. (D) HeLa cells were untransfected or transfected with siRNAs against Cbl for a total of 72 h in total. 24 h later, these cells were additionally transfected with TEM8/1-HA for 48 h and incubated with 500 ng/ml PA83 for different times at 37°C. TEM8/1-HA was immunoprecipitated from each fraction and analyzed by SDS-PAGE and Western blotting using anti-Ub and anti-HA antibodies. , left), as also observed in Fig. 6 A . In marked contrast, ubiquinated TEM8/1 was detected exclusively in DRMs ( Fig. 8 A , left). When cells were treated with the toxin before Triton X-100 solubilization, PA63 was associated with DRMs (as in Fig. 2 B ) and led to the recruitment of toxin-bound TEM8/1 to this fraction ( Fig. 8 A , right). Concomitantly, the ubiquitinated form of TEM8/1 was increased in the same fraction ( Fig. 8 A ). Raft impairment by cholesterol extraction using the sequestering agent β-methylcyclodextrin (ßMCD) led to a strong inhibition of PA-induced TEM8/1 ubiquitination ( Fig. 8 B ). This observation suggests that microdomain association precedes and is required for this posttranslational modification. The E3 Ub ligase Cbl is required for anthrax toxin endocytosis Cbl is an E3 Ub ligase that can interact with lipid rafts ( Lafont and Simons, 2001 ; Haglund et al., 2004 ). To test for the involvement of Cbl in anthrax toxin endocytosis, we decided to perform RNA silencing. HeLa cells were used because of their human origin (the sequence of hamster Cbl is not available) and their high transfection efficiencies. These cells express TEM8 as indicated by the pH sensitivity of PA channel formation ( Rainey et al., 2005 ). As shown in Fig. 8 C , Cbl could be efficiently silenced by this method. The absence of Cbl did not affect binding of the toxin as indicated by the unaltered presence of PA83/PA63. However, appearance of the PA 7mer pore was drastically inhibited ( Fig. 8 C ). To investigate the effect of Cbl RNA interference on the ubiquitination of TEM8 itself, RNA interference–treated cells were transfected with TEM8/1-HA, and toxin-induced ubiquitination after immunoprecipitation of the receptors was measured. As shown Fig. 8 D , the ubiquitinated form of TEM8/1 could not longer be detected. Theses experiments show that Cbl is responsible for the ubiquitination of TEM8/1 and its subsequent internalization. Constitutive ubiquitination of palmitoylation-deficient TEM8/1 and the consequences as an anthrax toxin receptor We found that AAAA TEM8/1 is constitutively associated with DRMs and that TEM8 ubiquitination is a raft-dependent modification. Therefore, we wondered whether AAAA TEM8/1 would be constitutively ubiquitinated. As shown in Fig. 9 A Figure 9. Cells expressing the palmitoylation-deficient TEM8/1 are less sensitive to anthrax toxin. (A and B) Anti-HA immunoprecipitation experiments were performed on CHO ΔATR cells stably or transiently (48 h) expressing CCCC or AAAA mutant TEM8/1-HA. Western blotting was performed against Ub and HA. (B) Cells were left untreated or treated with βMCD before lysis and immunoprecipitation. (C) CHO ΔATR cells stably expressing WT or AAAA TEM8/1-HA were treated with 500 ng/ml PA83 and 200 ng/ml LF at 37°C for different times. 40 μg of cell extracts were analyzed by SDS-PAGE followed by Western blotting against PA, HA, and MEK1 (NH 2 -terminal–directed antibody). (D) CHO ΔATR cells transiently (48 h) expressing WT and AAAA mutant TEM8/1-HA were treated with 500 ng/ml of the furin-resistant PA SNKE mutant for 1 h at 4°C and shifted to 37°C for different times. Surface-bound toxin was shaved off with trypsin (10 min at 37°Δ). To prevent lysosomal degradation of the internalized PA, cells were treated with 10 μM nocodazole to block microtubule-dependent transport to late endosomes. 40 μg of cell extracts were analyzed by SDS-PAGE and Western blotting against PA and HA. , the steady-state ubiquitination level of AAAA TEM8/1 was markedly higher than that of WT CCCC TEM8/1 (especially when comparing the levels of Ub vs. HA) both in stable cell lines and upon transient transfection (note that after immunoprecipitation of TEM8/1-HA from transiently transfected cells, the levels of expressed receptors seem to be similar even though analysis of total extracts shows a lower abundance of the AAAA mutant). Ubiquitination of AAAA TEM8/1 was dependent on the integrity of lipid rafts because the removal of cholesterol using ßMCD led to a drastic reduction in the level of AAAA TEM8/1 ubiquitination. This later observation also indicates that increased ubiquitination of AAAA TEM8/1 is not a consequence of misfolding of the cytoplasmic tail as a consequence of mutagenesis, because such an event would have been insensitive to acute cholesterol extraction from the plasma membrane. Constitutive ubiquitination of the AAAA TEM8/1 and, thus, its constitutive endocytosis are likely to affect its ability to act as an anthrax toxin receptor. To address this issue directly, we monitored the cleavage kinetics of the LF target MEK1. Whereas MEK1 underwent LF-dependent cleavage in WT TEM8/1–expressing cells, the MAPK kinase remained intact in the AAAA TEM8/1–expressing cells during the time course of the experiment ( Fig. 9 C ). To test whether this lack of cleavage was only a result of the reduced number of surface-expressed receptors (levels of TEM-HA), we treated AAAA TEM8/1–expressing cells with a higher concentration of PA to reach similar amounts of bound PA as on WT TEM8/1–expressing cells (Fig. S3, available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 ). Interestingly, even under these conditions, MEK1 cleavage in AAAA TEM8/1–expressing cells was minimal (Fig. S3), indicating that reduced PA binding does fully account for the reduced MEK1 cleavage in these cells. Because AAAA TEM8/1 undergoes significant constitutive endocytosis, we tested whether it would mediate the internalization of PA83, which is an event that does not occur with the WT receptor and for which PA heptamerization is required ( Abrami et al., 2003 ; Liu and Leppla, 2003 ). In this study, we made use of a mutant PA (PA SNKE ; Abrami et al., 2003 ) that is modified in the furin consensus cleavage site and, thus, remains in the PA83 form. As expected ( Abrami et al., 2003 ), PA SNKE was not internalized by WT receptors and was sensitive to surface trypsinization ( Fig. 9 D ). In contrast, all cell-bound PA SNKE became trypsin resistant in AAAA TEM8/ 1–expressing cells, indicating that it had been completely endocytosed ( Fig. 9 C , right). Discussion TEM8 and CMG2 are ideal anthrax toxin receptors because of their low steady-state endocytosis rate and rapid uptake upon clustering ( Abrami et al., 2003 ). These properties guarantee that PA is not internalized unless it has heptamerized and, thus, bound EF and/or LF. In this study, we have analyzed the mechanisms that govern the cell surface behavior of the receptors. Receptor palmitoylation Although the number of cytoplasmic cysteines varies from two to four in the various transmembrane TEM8 and CMG2 isoforms, this study indicates that these isoforms are all palmitoylated at least on the second juxtamembranous cysteine. In addition, long isoforms, such as those shown here for TEM8/1, can be palmitoylated on the more distal cysteines. These different palmitoylation sites could be the substrate of more than one palmitoyl transferases, 23 of which have been identified in the human genome ( Fukata et al., 2004 ; Linder and Deschenes, 2004 ). The first palmitoylation event appears to occur in the early secretory pathway, as indicated by the ∼50% inhibition of palmitate incorporation induced by cycloheximide or brefeldin A treatment and also indirectly suggested by the increased degradation of the palmitoylation-deficient mutant at early times after synthesis. Based on the study of the palmitoylation-deficient AAAA TEM8/1 mutant, palmitoylation regulates the lifetime of TEM8/1 by preventing premature internalization and targeting to lysosomes. Lysosomal targeting in the absence of proper palmitoylation has previously been observed for other transmembrane proteins such as CCR5 ( Percherancier et al., 2001 ) or the yeast SNARE Tlg1 ( Valdez-Taubas and Pelham, 2005 ). Increased turnover of membrane proteins in the absence of normal palmitoylation, however, is not systematic because the yeast SNAREs Suc1 and Syn8 are not prematurely targeted to degradation in palmitoylation-deficient cells ( Valdez-Taubas and Pelham, 2005 ). However, the most surprising finding concerning the palmitoylation-deficient mutant was its high affinity for DRMs. This appears to be in contradiction with the accepted consensus in the lipid raft field that palmitoylation allows association with cholesterol-rich domains. Careful analysis of the data illustrates that the effect of palmitoylation on a transmembrane protein cannot be predicted because a variety of situations have been described. LAT is palmitoylated, and this modification is essential for DRM association ( Zhang et al., 1998 ); caveolin-1 is triple palmitoylated, but this modification is not necessary for DRM association ( Dietzen et al., 1995 ); and, finally, the transferrin receptor ( Alvarez et al., 1990 ) and the G protein of the vesicular stomatitis virus ( Mack and Kruppa, 1988 ), which are two well-characterized nonraft proteins, are both palmitoylated. We now illustrate a fourth scenario in which palmitoylation actually prevents DRM association. Interestingly, the DRM association of TEM8 was still sensitive to 2-bromopalmitate, indicating the requirement for other palmitoylated proteins. Based on this observation, we would like to propose that palmitoylation of TEM8 regulates its interaction with a partner protein, the palmitoylation of which is required for the toxin-induced raft association of WT TEM8. This interaction would be essential for the toxin's action based on the inhibitory effect of 2-bromopalmitate on PA surface heptamerization, which is a prerequisite for endocytosis via the WT receptor ( Fig. 10 Figure 10. Schematic representation of the toxin-induced behavior of TEM8/CMG2 at the cell surface. Cell surface anthrax toxin receptor is palmitoylated and localizes to the glycerophospholipidic region of the membrane, where PA83 binds to it. Furin cleavage and heptamerization of PA trigger raft association of the toxin–receptor complex ( Abrami et al., 2003 ) in a manner that probably involves a conformational change of the clustered receptors and the association with partner proteins, one of which is predicted to be palmitoylated. Within rafts, the receptors become accessible to its E3 Ub ligase, possibly Cbl that modifies the cytoplasmic tails, thus allowing the interaction with proteins of the endocytic machinery harboring Ub-interacting domains. The clathrin-dependent endocytic machinery is thus recruited, and the toxin–receptor complex is internalized. ). Toxin-induced receptor ubiquitination The cytoplasmic tails of all isoforms of TEM8 and CMG2 contain lysine residues in numbers that vary from 3 in TEM8/2 to 14 in CMG2/4 and 16 in TEM8/1, all of which are potential ubiquitination sites. We found that anthrax PA triggers the ubiquitination of each of these receptors. Indeed, Lys-352 in TEM8/1 and the corresponding Lys-350 in CMG2/4 acquire this modification, and, thus, mutation of these lysines leads to the inhibition of PA endocytosis. Ubiquitination TEM8 is mediated by the E3 Ub ligase Cbl as shown by RNA silencing on HeLa cells. The modification is raft dependent because it occurred preferentially in DRMs and was inhibited by cholesterol extraction from the plasma membrane. Moreover, DRM association appears to be sufficient to induce ubiquitination because the AAAA TEM8/1 palmitoylation-deficient mutant is constitutively DRM associated as well as ubiquitinated. The findings described in this study—i.e., that both palmitoylation and ubiquitination of TEM8 are important for anthrax toxin action—appear to be in disagreement with previous work showing that anthrax PA could be endocytosed in cells expressing a variety of COOH-terminal TEM8 truncation mutants ( Liu and Leppla, 2003 ). In this latter study, all transmembrane-truncated TEM8 mutants, with the exception of the mutant lacking the entire cytoplasmic tail, retained palmitoylation and ubiquitination sites because they were all longer than TEM8 isoform 2. The completely tailless mutant was still able to internalize PA ( Liu and Leppla, 2003 ). However, the conditions used were very different (higher PA concentrations and longer incubation times) than those in this study, and kinetics were not measured. Concluding remarks We have identified two posttranslational modifications of the anthrax toxin receptors, palmitoylation and ubiquitination, that have opposite effects on the uptake of receptors and play a crucial role in the mode of toxin action. Palmitoylation serves to spatially segregate the receptors away from their Ub ligase, which, in turn, controls their endocytosis and turnover. In terms of anthrax toxin's mode of action, an intricate sequence of events can be envisioned ( Fig. 10 ). PA binds to its receptors, and subsequent furin processing and heptamerization likely trigger a conformational change in the receptor, leading to its association with partner proteins and/or depalmitoylation. Altogether, this induces the redistribution of the toxin–receptor complex to lipid rafts, where an encounter with the E3 Ub ligase Cbl occurs. Ubiquitination of TEM8/CMG2 allows an interaction with proteins of the endocytic machinery harboring Ub-interacting domains ( Hicke and Dunn, 2003 ). One of these is likely to be Eps15, which is involved in the formation of clathrin-coated pits, contains three Ub-interacting motifs (for review see Polo et al., 2003 ), and is required for anthrax toxin endocytosis ( Abrami et al., 2003 ). Recruitment of the endocytosis machinery leads to rapid uptake and transport to early endosomes, where anthrax toxin receptor ubiquitination might play a second role in sorting into multivesicular bodies ( Katzmann et al., 2002 ). This study also illustrates a novel role for the palmitoylation of membrane proteins. It has recently been proposed that palmitoylation of the SNARE protein snf1 prevents recognition by the E3 ligases of critical lysines adjacent to the transmembrane domain by affecting protein conformation ( Valdez-Taubas and Pelham, 2005 ). The absence of palmitoylation might also facilitate access of E3 ligase to lysines on TEM8/CMG2. However, more importantly, palmitoylation serves to spatially segregate these membrane proteins away from their Ub ligase, thus controlling their endocytosis. Receptor palmitoylation Although the number of cytoplasmic cysteines varies from two to four in the various transmembrane TEM8 and CMG2 isoforms, this study indicates that these isoforms are all palmitoylated at least on the second juxtamembranous cysteine. In addition, long isoforms, such as those shown here for TEM8/1, can be palmitoylated on the more distal cysteines. These different palmitoylation sites could be the substrate of more than one palmitoyl transferases, 23 of which have been identified in the human genome ( Fukata et al., 2004 ; Linder and Deschenes, 2004 ). The first palmitoylation event appears to occur in the early secretory pathway, as indicated by the ∼50% inhibition of palmitate incorporation induced by cycloheximide or brefeldin A treatment and also indirectly suggested by the increased degradation of the palmitoylation-deficient mutant at early times after synthesis. Based on the study of the palmitoylation-deficient AAAA TEM8/1 mutant, palmitoylation regulates the lifetime of TEM8/1 by preventing premature internalization and targeting to lysosomes. Lysosomal targeting in the absence of proper palmitoylation has previously been observed for other transmembrane proteins such as CCR5 ( Percherancier et al., 2001 ) or the yeast SNARE Tlg1 ( Valdez-Taubas and Pelham, 2005 ). Increased turnover of membrane proteins in the absence of normal palmitoylation, however, is not systematic because the yeast SNAREs Suc1 and Syn8 are not prematurely targeted to degradation in palmitoylation-deficient cells ( Valdez-Taubas and Pelham, 2005 ). However, the most surprising finding concerning the palmitoylation-deficient mutant was its high affinity for DRMs. This appears to be in contradiction with the accepted consensus in the lipid raft field that palmitoylation allows association with cholesterol-rich domains. Careful analysis of the data illustrates that the effect of palmitoylation on a transmembrane protein cannot be predicted because a variety of situations have been described. LAT is palmitoylated, and this modification is essential for DRM association ( Zhang et al., 1998 ); caveolin-1 is triple palmitoylated, but this modification is not necessary for DRM association ( Dietzen et al., 1995 ); and, finally, the transferrin receptor ( Alvarez et al., 1990 ) and the G protein of the vesicular stomatitis virus ( Mack and Kruppa, 1988 ), which are two well-characterized nonraft proteins, are both palmitoylated. We now illustrate a fourth scenario in which palmitoylation actually prevents DRM association. Interestingly, the DRM association of TEM8 was still sensitive to 2-bromopalmitate, indicating the requirement for other palmitoylated proteins. Based on this observation, we would like to propose that palmitoylation of TEM8 regulates its interaction with a partner protein, the palmitoylation of which is required for the toxin-induced raft association of WT TEM8. This interaction would be essential for the toxin's action based on the inhibitory effect of 2-bromopalmitate on PA surface heptamerization, which is a prerequisite for endocytosis via the WT receptor ( Fig. 10 Figure 10. Schematic representation of the toxin-induced behavior of TEM8/CMG2 at the cell surface. Cell surface anthrax toxin receptor is palmitoylated and localizes to the glycerophospholipidic region of the membrane, where PA83 binds to it. Furin cleavage and heptamerization of PA trigger raft association of the toxin–receptor complex ( Abrami et al., 2003 ) in a manner that probably involves a conformational change of the clustered receptors and the association with partner proteins, one of which is predicted to be palmitoylated. Within rafts, the receptors become accessible to its E3 Ub ligase, possibly Cbl that modifies the cytoplasmic tails, thus allowing the interaction with proteins of the endocytic machinery harboring Ub-interacting domains. The clathrin-dependent endocytic machinery is thus recruited, and the toxin–receptor complex is internalized. ). Toxin-induced receptor ubiquitination The cytoplasmic tails of all isoforms of TEM8 and CMG2 contain lysine residues in numbers that vary from 3 in TEM8/2 to 14 in CMG2/4 and 16 in TEM8/1, all of which are potential ubiquitination sites. We found that anthrax PA triggers the ubiquitination of each of these receptors. Indeed, Lys-352 in TEM8/1 and the corresponding Lys-350 in CMG2/4 acquire this modification, and, thus, mutation of these lysines leads to the inhibition of PA endocytosis. Ubiquitination TEM8 is mediated by the E3 Ub ligase Cbl as shown by RNA silencing on HeLa cells. The modification is raft dependent because it occurred preferentially in DRMs and was inhibited by cholesterol extraction from the plasma membrane. Moreover, DRM association appears to be sufficient to induce ubiquitination because the AAAA TEM8/1 palmitoylation-deficient mutant is constitutively DRM associated as well as ubiquitinated. The findings described in this study—i.e., that both palmitoylation and ubiquitination of TEM8 are important for anthrax toxin action—appear to be in disagreement with previous work showing that anthrax PA could be endocytosed in cells expressing a variety of COOH-terminal TEM8 truncation mutants ( Liu and Leppla, 2003 ). In this latter study, all transmembrane-truncated TEM8 mutants, with the exception of the mutant lacking the entire cytoplasmic tail, retained palmitoylation and ubiquitination sites because they were all longer than TEM8 isoform 2. The completely tailless mutant was still able to internalize PA ( Liu and Leppla, 2003 ). However, the conditions used were very different (higher PA concentrations and longer incubation times) than those in this study, and kinetics were not measured. Concluding remarks We have identified two posttranslational modifications of the anthrax toxin receptors, palmitoylation and ubiquitination, that have opposite effects on the uptake of receptors and play a crucial role in the mode of toxin action. Palmitoylation serves to spatially segregate the receptors away from their Ub ligase, which, in turn, controls their endocytosis and turnover. In terms of anthrax toxin's mode of action, an intricate sequence of events can be envisioned ( Fig. 10 ). PA binds to its receptors, and subsequent furin processing and heptamerization likely trigger a conformational change in the receptor, leading to its association with partner proteins and/or depalmitoylation. Altogether, this induces the redistribution of the toxin–receptor complex to lipid rafts, where an encounter with the E3 Ub ligase Cbl occurs. Ubiquitination of TEM8/CMG2 allows an interaction with proteins of the endocytic machinery harboring Ub-interacting domains ( Hicke and Dunn, 2003 ). One of these is likely to be Eps15, which is involved in the formation of clathrin-coated pits, contains three Ub-interacting motifs (for review see Polo et al., 2003 ), and is required for anthrax toxin endocytosis ( Abrami et al., 2003 ). Recruitment of the endocytosis machinery leads to rapid uptake and transport to early endosomes, where anthrax toxin receptor ubiquitination might play a second role in sorting into multivesicular bodies ( Katzmann et al., 2002 ). This study also illustrates a novel role for the palmitoylation of membrane proteins. It has recently been proposed that palmitoylation of the SNARE protein snf1 prevents recognition by the E3 ligases of critical lysines adjacent to the transmembrane domain by affecting protein conformation ( Valdez-Taubas and Pelham, 2005 ). The absence of palmitoylation might also facilitate access of E3 ligase to lysines on TEM8/CMG2. However, more importantly, palmitoylation serves to spatially segregate these membrane proteins away from their Ub ligase, thus controlling their endocytosis. Materials and methods Proteins and antibodies Anthrax toxin subunits ( Leppla, 1988 ; Gordon et al., 1995 ) and aerolysin ( Buckley, 1990 ) were purified as described previously. PA63 was generated by trypsin cleavage of PA83 ( Abrami et al., 2003 ). Antibodies against PA ( Liu and Leppla, 2003 ) and aerolysin ( Fivaz et al., 2002 ) were polyclonals developed in our laboratories. Proteins and antibodies were obtained from the following companies: Anti NH 2 -terminal MEK1 antibodies from Upstate Biotechnology; anti–caveolin-1 from Transduction Laboratories; anti-Ub (sc-8017) as well as antibodies and siRNA against human Cbl from Santa Cruz Biotechnology, Inc.; anti-HA coupled to beads, labeled, or unlabeled with HRP and anti-GFP from Roche; anti-V5 labeled or unlabeled with HRP from Invitrogen; HRP secondary antibodies from Pierce Chemical Co.; and FITC-conjugated secondary antibodies from Invitrogen. Cells, plasmids, and transfections BHK, HeLa, and anthrax toxin receptor–deficient CHO (here designated as CHO ΔATR ) cells were grown as described previously ( Abrami et al., 2003 , 2004 ; Liu and Leppla, 2003 ). The human CMG2 (isoform 4) gene tagged with a V5 epitope and cloned in the pcDNA3.1/ V5-HIS-TOPO expression vector was provided by J. Martignetti (Mount Sinai School of Medicine, New York, NY; Dowling et al., 2003 ). TEM8 isoforms 1 and 2 tagged with GFP and cloned in the pHS001-EGFP expression vector was provided by J. Young (Salk Institute, San Diego, CA; Rainey et al., 2005 ). Isoform 1 of human TEM8 gene tagged with an HA epitope was in the pIREShyg2 vector ( Liu and Leppla, 2003 ). Cysteine to alanine mutant constructs were generated using the QuikChange Site-Directed Mutagenesis Kit (Stratagene) and were transfected into CHO ΔATR (1 μg cDNA/9.6-cm 2 plate) using Fugene (Roche). Stable lines were selected by two rounds of hygromycin resistance. Colonies were isolated by limited dilution. To silence Cbl, HeLa cells were transfected with 200 pmol/9.2-cm2 dish of siRNA using OligofectAMINE (Invitrogen) transfection reagent. Toxin treatment and analysis Confluent cells were incubated in incubation medium (IM; Glasgow minimal essential medium buffered with 10 mM Hepes, pH 7.4) at 4°C for 1 h with various combinations of proaerolysin, PA, and LF, washed, and placed at 37°C in IM for different times. Cells were lysed by incubation for 30 min at 4°C with radioimmunoprecipitation buffer (1% NP-40, 50 mM Tris-HCl, pH 7.4, 0.25% sodium deoxycholate, 150 mm NaCl, 1 mM EDTA, and a cocktail of protease inhibitors; Roche). Protein concentrations of extracts were determined with bicinchoninic acid (Pierce Chemical Co.). To convert surface PA 7mer to an SDS-resistant form, cell extracts were incubated at room temperature for 10 min with 145 mM NaCl and 20 mM MES-Tris, pH 4.5. SDS-PAGE was performed using 4–20% gels (NOVEX) under nonreducing conditions. Gels were transferred onto nitrocellulose as described previously ( Abrami et al., 2004 ). Drug and enzymatic treatments To inhibit N -glycosylation, cells were treated with 10 μg/ml tunicamycin (Sigma-Aldrich) during the last 16 h of growth. For N -deglycosylation, cell extracts were boiled for 5 min with 1% SDS and 1% β-mercaptoethanol, diluted fivefold in 40 mM phosphate buffer, pH 7.0, containing 10 mM EDTA, 1% Triton X-100, 2.5 mM PMSF, and 1% β-mercaptoethanol, and were incubated for 16 h at 37°C with 10 U/ml N -glycosidase F. Endo H treatment was performed according to the manufacturer's instructions (New England Biolabs, Inc.). Palmitoylation was inhibited by pretreating cells with 100 μM 2-bromopalmitate (Sigma-Aldrich) for 1 h at 37°C . Chemical removal of S-palmitoylation was performed by treating cell extracts for 1 h at room temperature with 1 M hydroxylamine hydrochloride, pH 7.2. Protein synthesis was blocked by a 30-min treatment with 10 μg/ml cycloheximide at 37°C. Endoplasmic reticulum–to-Golgi transport was blocked by 20 μg/ml brefeldin A pretreatment for 30 min at 37°C. The proteasome inhibitor MG132 (Sigma-Aldrich) was used at 10 μM during the 16 h in culture medium. To block lysosomal enzymes, cells were fed for 16 h with 250 μg/ml leupeptin. To extract cholesterol, cells were treated with 10 mM βMCD (Sigma-Aldrich) in IM for 30 min at 37°C, leading to a 59.3 ± 3.8% decrease in total cholesterol as measured by thin layer chromatography ( Abrami and van der Goot, 1999 ). Biochemical methods DRMs were prepared using OptiPrep gradients as described previously ( Abrami et al., 2003 ). Six fractions were collected from the top, and the total protein content of each fraction was precipitated with trichloroacetic acid ( Abrami et al., 2003 ). For immunoprecipitations of TEM8 or CMG2, cells were lysed for 30 min at 4°C in immunoprecipitation buffer (0.5% NP-40, 500 mM Tris-HCl, pH 7.4, 20 mM EDTA, 10 mM NaF, 2 mM benzamidine, 1 mM N -ethyl-maleimide, and a cocktail of protease inhibitors), centrifuged for 3 min at 2,000 g , and supernatants were incubated for 2 h at 4°C with either HA-coupled agarose beads (Roche) or protein G–coupled beads (GE Healthcare) with 2 μg monoclonal antibody against V5. After washing of the beads, samples were boiled for 5 min under reducing conditions. Radiolabeling experiments To follow palmitoylation, TEM8/CMG2-expressing cells were incubated for 2 h at 37°C in IM with 200 μCI /ml 3 H-palmitic acid (9,10- 3 H(N); American Radiolabeled Chemicals, Inc), were washed, and submitted to immunoprecipitation. Beads were incubated for 30 min at 60°C in nonreducing sample buffer before SDS-PAGE. After fixation (25% isopropanol, 65% H 2 O, and 10% acetic acid), gels were incubated for 30 min in enhancer Amplify NAMP100 (GE Healthcare), dried, and exposed to a Hyperfilm Multipurpose (GE Healthcare). For metabolic labeling, CHO cells were transiently transfected for 30 h with TEM8/1-HA cDNAs, washed with methionine/cysteine-free medium, incubated for a 30-min pulse at 37°C with 50 μCi/ml [ 35 S]methionine/cysteine (Hartman Analytics), washed, and further incubated for different times at 37°C in complete medium with a 10-fold excess of nonradioactive methionine and cysteine. Receptors were immunoprecipitated and analyzed by SDS-PAGE. Immunofluorescence microscopy CHO cells were fixed with 3% formaldehyde, permeabilized with 0.1% Triton X-100, and labeled with anti-HA monoclonal antibodies followed by fluorescein isothiocyanate–conjugated goat anti–mouse IgG. Images were acquired using a 100× lens on a microscope (Axiophot; Carl Zeiss MicroImaging, Inc.) equipped with a cooled camera (model Orca; Hamamatsu) using the Openlab acquisition software (Improvision). Online supplemental material Three supplemental figures are provided. Fig. S1 shows that isoform 2 of TEM8, which has a very short cytoplasmic tail ( Fig. 1 A ), is palmitolyted. Fig. S2 shows that this isoform is also ubiquitinated. Finally, Fig. S3 shows than when toxin concentrations are adapted so that CCCC TEM8/1– and AAAA TEM8/1–expressing cells bind similar amounts of PA, MEK1 cleavage is observed in the former cells within an hour but not in the latter cells. Online supplemental material is available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 . Proteins and antibodies Anthrax toxin subunits ( Leppla, 1988 ; Gordon et al., 1995 ) and aerolysin ( Buckley, 1990 ) were purified as described previously. PA63 was generated by trypsin cleavage of PA83 ( Abrami et al., 2003 ). Antibodies against PA ( Liu and Leppla, 2003 ) and aerolysin ( Fivaz et al., 2002 ) were polyclonals developed in our laboratories. Proteins and antibodies were obtained from the following companies: Anti NH 2 -terminal MEK1 antibodies from Upstate Biotechnology; anti–caveolin-1 from Transduction Laboratories; anti-Ub (sc-8017) as well as antibodies and siRNA against human Cbl from Santa Cruz Biotechnology, Inc.; anti-HA coupled to beads, labeled, or unlabeled with HRP and anti-GFP from Roche; anti-V5 labeled or unlabeled with HRP from Invitrogen; HRP secondary antibodies from Pierce Chemical Co.; and FITC-conjugated secondary antibodies from Invitrogen. Cells, plasmids, and transfections BHK, HeLa, and anthrax toxin receptor–deficient CHO (here designated as CHO ΔATR ) cells were grown as described previously ( Abrami et al., 2003 , 2004 ; Liu and Leppla, 2003 ). The human CMG2 (isoform 4) gene tagged with a V5 epitope and cloned in the pcDNA3.1/ V5-HIS-TOPO expression vector was provided by J. Martignetti (Mount Sinai School of Medicine, New York, NY; Dowling et al., 2003 ). TEM8 isoforms 1 and 2 tagged with GFP and cloned in the pHS001-EGFP expression vector was provided by J. Young (Salk Institute, San Diego, CA; Rainey et al., 2005 ). Isoform 1 of human TEM8 gene tagged with an HA epitope was in the pIREShyg2 vector ( Liu and Leppla, 2003 ). Cysteine to alanine mutant constructs were generated using the QuikChange Site-Directed Mutagenesis Kit (Stratagene) and were transfected into CHO ΔATR (1 μg cDNA/9.6-cm 2 plate) using Fugene (Roche). Stable lines were selected by two rounds of hygromycin resistance. Colonies were isolated by limited dilution. To silence Cbl, HeLa cells were transfected with 200 pmol/9.2-cm2 dish of siRNA using OligofectAMINE (Invitrogen) transfection reagent. Toxin treatment and analysis Confluent cells were incubated in incubation medium (IM; Glasgow minimal essential medium buffered with 10 mM Hepes, pH 7.4) at 4°C for 1 h with various combinations of proaerolysin, PA, and LF, washed, and placed at 37°C in IM for different times. Cells were lysed by incubation for 30 min at 4°C with radioimmunoprecipitation buffer (1% NP-40, 50 mM Tris-HCl, pH 7.4, 0.25% sodium deoxycholate, 150 mm NaCl, 1 mM EDTA, and a cocktail of protease inhibitors; Roche). Protein concentrations of extracts were determined with bicinchoninic acid (Pierce Chemical Co.). To convert surface PA 7mer to an SDS-resistant form, cell extracts were incubated at room temperature for 10 min with 145 mM NaCl and 20 mM MES-Tris, pH 4.5. SDS-PAGE was performed using 4–20% gels (NOVEX) under nonreducing conditions. Gels were transferred onto nitrocellulose as described previously ( Abrami et al., 2004 ). Drug and enzymatic treatments To inhibit N -glycosylation, cells were treated with 10 μg/ml tunicamycin (Sigma-Aldrich) during the last 16 h of growth. For N -deglycosylation, cell extracts were boiled for 5 min with 1% SDS and 1% β-mercaptoethanol, diluted fivefold in 40 mM phosphate buffer, pH 7.0, containing 10 mM EDTA, 1% Triton X-100, 2.5 mM PMSF, and 1% β-mercaptoethanol, and were incubated for 16 h at 37°C with 10 U/ml N -glycosidase F. Endo H treatment was performed according to the manufacturer's instructions (New England Biolabs, Inc.). Palmitoylation was inhibited by pretreating cells with 100 μM 2-bromopalmitate (Sigma-Aldrich) for 1 h at 37°C . Chemical removal of S-palmitoylation was performed by treating cell extracts for 1 h at room temperature with 1 M hydroxylamine hydrochloride, pH 7.2. Protein synthesis was blocked by a 30-min treatment with 10 μg/ml cycloheximide at 37°C. Endoplasmic reticulum–to-Golgi transport was blocked by 20 μg/ml brefeldin A pretreatment for 30 min at 37°C. The proteasome inhibitor MG132 (Sigma-Aldrich) was used at 10 μM during the 16 h in culture medium. To block lysosomal enzymes, cells were fed for 16 h with 250 μg/ml leupeptin. To extract cholesterol, cells were treated with 10 mM βMCD (Sigma-Aldrich) in IM for 30 min at 37°C, leading to a 59.3 ± 3.8% decrease in total cholesterol as measured by thin layer chromatography ( Abrami and van der Goot, 1999 ). Biochemical methods DRMs were prepared using OptiPrep gradients as described previously ( Abrami et al., 2003 ). Six fractions were collected from the top, and the total protein content of each fraction was precipitated with trichloroacetic acid ( Abrami et al., 2003 ). For immunoprecipitations of TEM8 or CMG2, cells were lysed for 30 min at 4°C in immunoprecipitation buffer (0.5% NP-40, 500 mM Tris-HCl, pH 7.4, 20 mM EDTA, 10 mM NaF, 2 mM benzamidine, 1 mM N -ethyl-maleimide, and a cocktail of protease inhibitors), centrifuged for 3 min at 2,000 g , and supernatants were incubated for 2 h at 4°C with either HA-coupled agarose beads (Roche) or protein G–coupled beads (GE Healthcare) with 2 μg monoclonal antibody against V5. After washing of the beads, samples were boiled for 5 min under reducing conditions. Radiolabeling experiments To follow palmitoylation, TEM8/CMG2-expressing cells were incubated for 2 h at 37°C in IM with 200 μCI /ml 3 H-palmitic acid (9,10- 3 H(N); American Radiolabeled Chemicals, Inc), were washed, and submitted to immunoprecipitation. Beads were incubated for 30 min at 60°C in nonreducing sample buffer before SDS-PAGE. After fixation (25% isopropanol, 65% H 2 O, and 10% acetic acid), gels were incubated for 30 min in enhancer Amplify NAMP100 (GE Healthcare), dried, and exposed to a Hyperfilm Multipurpose (GE Healthcare). For metabolic labeling, CHO cells were transiently transfected for 30 h with TEM8/1-HA cDNAs, washed with methionine/cysteine-free medium, incubated for a 30-min pulse at 37°C with 50 μCi/ml [ 35 S]methionine/cysteine (Hartman Analytics), washed, and further incubated for different times at 37°C in complete medium with a 10-fold excess of nonradioactive methionine and cysteine. Receptors were immunoprecipitated and analyzed by SDS-PAGE. Immunofluorescence microscopy CHO cells were fixed with 3% formaldehyde, permeabilized with 0.1% Triton X-100, and labeled with anti-HA monoclonal antibodies followed by fluorescein isothiocyanate–conjugated goat anti–mouse IgG. Images were acquired using a 100× lens on a microscope (Axiophot; Carl Zeiss MicroImaging, Inc.) equipped with a cooled camera (model Orca; Hamamatsu) using the Openlab acquisition software (Improvision). Online supplemental material Three supplemental figures are provided. Fig. S1 shows that isoform 2 of TEM8, which has a very short cytoplasmic tail ( Fig. 1 A ), is palmitolyted. Fig. S2 shows that this isoform is also ubiquitinated. Finally, Fig. S3 shows than when toxin concentrations are adapted so that CCCC TEM8/1– and AAAA TEM8/1–expressing cells bind similar amounts of PA, MEK1 cleavage is observed in the former cells within an hour but not in the latter cells. Online supplemental material is available at http://www.jcb.org/cgi/content/full/jcb.200507067/DC1 . Supplementary Material Supplemental Material Index

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3587502/

Surface Plasmon Resonance Biosensor for Detection of Bacillus anthracis , the Causative Agent of Anthrax from Soil Samples Targeting Protective Antigen

Bacillus anthracis , the causative agent of anthrax is one of the most important biological warfare agents. In this study, surface plasmon resonance (SPR) technology was used for indirect detection of B. anthracis by detecting protective antigen (PA), a common toxin produced by all live B. anthracis bacteria. For development of biosensor, a monoclonal antibody raised against B. anthracis PA was immobilized on carboxymethyldextran modified gold chip and its interaction with PA was characterized in situ by SPR and electrochemical impedance spectroscopy. By using kinetic evaluation software, K D (equilibrium constant) and B max (maximum binding capacity of analyte) were found to be 20 fM and 18.74, respectively. The change in Gibb's free energy (∆G = −78.04 kJ/mol) confirmed the spontaneous interaction between antigen and antibody. The assay could detect 12 fM purified PA. When anthrax spores spiked soil samples were enriched, PA produced in the sample containing even a single spore of B. anthracis could be detected by SPR. PA being produced only by the vegetative cells of B. anthracis , confirms indirectly the presence of B. anthracis in the samples. The proposed method can be a very useful tool for screening and confirmation of anthrax suspected environmental samples during a bio-warfare like situation.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9356517/

Medical cannabis and cannabidiol: A new harvest for Malawi

In February 2020 parliament passed the Cannabis Regulation Bill (2020) which regulates the cultivation and production of industrial hemp and medical cannabis. The country will only fully benefit from this development if the medical and scientific community can take the lead in enabling the country to exploit the plant's potential to help address some of our economic and public health challenges. This special communication provides some basic information on cannabis and discusses its history and medical uses. Cannabidiol (CBD) has emerged as one of the most important cannabis-derived phytochemicals and has formed the basis for the growth of the medical cannabis industry. The scientific data on the mechanisms of the effects of CBD on the human neuroendocrine-immune network is reviewed and the first effective cannabis-based FDA-approved treatment for epilepsy discussed. Some clinical research that is being done on the antipsychotic and neuroprotective properties of CBD is also reviewed. A case is made for the potential of CBD as a neuroprotective adjunctive therapy for the prevention of neuropsychological sequelae associated with complicated malaria. The safety profile of CBD is reviewed and finally, the potential importance of the re-medicalization of cannabis-based therapies for the broader field of phytomedicine is pointed out. Background On the 28th February 2020, the Malawi National Assembly passed the Cannabis Regulation Bill. The bill regulates the cultivation, propagation, production, processing, storage, exportation, importation, distribution and use of medical cannabis and industrial hemp. It is the culmination of over 20 years of lobbying efforts by Malawian advocates which resulted in adoption of a motion in the National Assembly in June 2016 1 to explore cannabis as an alternative crop to tobacco as well as the successful completion of industrial hemp cultivation trials at Chitedze Agricultural Research Station and other sites under the Ministry of Agriculture and private companies. Within the region, similar legislation has passed in Uganda (2015), Lesotho, Zimbabwe, South Africa (2018) and Zambia (2019). Globally, there has been a resurgence of innovation, and commercialization related to cannabis and in recent years, a significant amount of scientific research and progress in medical uses of cannabis. Medical publications about cannabis in Malawi have been limited to issues relating to marijuana use and its causes, consequences and associations. For health professionals, scientists and researchers in Malawi, it is important to be aware of the growing body of scientific data that has accrued around cannabis to be able to advance the discourse. This special communication is aimed at providing some basic information and highlighting some of the potential medical benefits of cannabis. Special emphasis is placed on cannabidiol (CBD) which was historically the lesser known of the major cannabis derived compounds since it was thought to be non-psychoactive. In the last decade however, CBD has undergone an increase in popularity and significance with respect to medical uses of cannabis. Marijuana, industrial hemp and their current legal status Cannabis is a genus of flowering plants mainly comprising two species, sativa and indica. Cannabis sativa is the oldest recorded cultivated crop 2 and is grown for its fibre, seeds and flowers. The fibre produced by cannabis is referred to as hemp fibre and is used to produce many products such as rope, clothing, fabric, paper and building materials. The seeds of the cannabis plant are used as food for both humans and livestock and are pressed to produce oil. Hempseed oil is an important nutritional supplement and ingredient in soaps and cosmetics and many other products. The flowers of cannabis plants produce a thick and sticky resin. This resin is known to contain the psychoactive ingredient of marijuana, Δ 9 -tetrahydrocannabinol (THC). Cultivars of Cannabis grown as marijuana can contain anywhere from 2% to over 20% THC 3 . Industrial hemp refers to cannabis that is bred specifically for the industrial uses of its derived products other than THC. Such varieties have been bred to have a very high fibre (30–40%) and seed content. In contrast to marijuana, these varieties have very low THC content and are consequently non-psychoactive. In Europe where there are established hemp industries the European Union (EU) standards for industrial hemp state that the THC content cannot exceed 0.2% 4 while a 1% THC content is acceptable in more tropical areas. More recently, the term medical cannabis has become popular. Medical cannabis includes all varieties of cannabis which are grown specifically for medicinal purposes and may include varieties with varying levels of THC and other cannabinoids such as CBD. The legality of different types of medical cannabis is determined by the legislative framework prevailing in the country and may vary from low THC varieties to marijuana varieties that can be used for both medical and recreational purposes. In Malawi, prior the new act of 2020, legislation related to cannabis was covered the Noxious Weeds Act of 1936 and the Dangerous Drugs Act of 1956 together with their Regulations. Under these laws, cultivation of cannabis was made illegal. However provisions were made for the use of cannabis for industrial, food and medicinal purposes. Regulations enabled certain persons to be authorized to acquire, possess and supply cannabis under license for research and medical purposes. However, these laws did not distinguish between different varieties of cannabis and lacked a workable framework for effective regulation. The Cannabis Regulation Bill of 2020 clearly distinguishes between marijuana (which remains illegal), industrial hemp and medical cannabis. It establishes a Cannabis Regulation Authority responsible for licensing and regulating the legal forms of cannabis and sets out framework for requirements and enforcement of all matters related to cannabis. The functional details of all of these aspects will be contained in regulations which are currently being drafted. Historical medical uses of cannabis Cannabis has been used medicinally for millennia. It has been used by healers in ancient China and India, by ancient Greek and Roman doctors, by Arab doctors during the Middle Ages and by British Victorian and continental European physicians 5 . In Africa the oldest recorded uses of medicinal cannabis date back to ancient Egypt where it was used in suppositories for relieving the pain of haemorrhoids 6 and as a treatment for sore eyes 7 . During colonial times, European travellers documenting their travels in Africa found well-established traditions of cannabis use entrenched in many African cultures with religious and medical uses prominent 8 , 9 . In Malawi, there is much anthropological evidence that cannabis (known locally as chamba) has been a part of the culture since ancient times. This is seen in traditions such as curing the bud by tightly wrapping it in banana leaf (the so-called Malawi Cob) and the use of the term "chamba cha makolo" to denote an ancient tradition 10 . Chamba was traditionally used in treatment of anthrax, dysentery, fevers, malaria and treatment of snake bites 11 . European medicine was slow in recognizing the medicinal value of cannabis with the first major description of its actions only happening in 1839 12 . Since then it has reportedly been used for: controlling nausea and vomiting, treating cholera and rabies, diminishing muscle spasms, stimulating appetite, lowering intraocular pressure in glaucoma, relieving phantom limb pain and other pain syndromes, alleviating menstrual cramps, promoting uterine contractions in labour, treating addiction and withdrawal symptoms, preventing seizures and reducing anxiety 13 . Cannabis was very widely used by American doctors during the early twentieth century as well as by English doctors until the 1970's. The passage of the Marijuana Tax Act (1937) in the US and later on, the UN Convention on Narcotics made cannabis illegal in most places by the middle of the 20th century which meant that cannabis based remedies could not be effectively studied using modern methodologies such as randomised clinical studies. In 1971 an overhaul of the British Pharmaceutical Codex resulted in the removal of a variety of non-evidence-based remedies including cannabis. In the US where cannabis therapy had been widely used in the treatment of epilepsy, its use diminished with the introduction of phenobarbital (1912) and phenytoin (1937) 14 . Currently there is a resurgence in the development of cannabis based medicines driven by research on biomedical and molecular properties and activities of cannabis. Medical cannabis is currently either being used or being investigated in the treat a long list of conditions including: epilepsy, Huntington disease, Parkinson's disease, pain relief, Alzheimer's disease, multiple sclerosis pain, inflammatory bowel disease, sleep disorders, arthritis, glaucoma, eating disorders, asthma, anxiety, stress, addiction, PTSD, cancer and protection of the brain after stroke 13 . Cannabinoids and their mechanisms of action Of the more than 560 distinct compounds that have been isolated from cannabis species, the most important and abundant group are known as cannabinoids 15 . The best studied of these cannabinoids are THC and CBD. The psychoactive effects of cannabis are mediated by THC. Because of this, it was assumed that many of the therapeutic properties of cannabis were also mediated by THC and this resulted in CBD being largely ignored by many earlier researchers. This has changed dramatically over the years with the perception of CBD moving from an inactive cannabinoid to a drug with a wide spectrum of action throughout the neuroendocrine-immune network. A PUBMED search for cannabidiol generated 50 reports for the period 1999–2002, 225 reports for 2003–2007 and 1,205 reports for 2008 to 2014 16 . Cannabinoids produce many of their effects through the endocannabinoid system via cannabinoid (CB) receptors. This system influences synaptic communication and modulates eating, anxiety, learning and memory, reward processing and growth and development 17 . The endogenous ligands of the CB receptors are collectively referred to as endocannabinoids (EC) and are known to modulate neuronal excitability 18 . Recent work has indicated that EC and EC-like compounds are present in human milk and play a role in establishing suckling responses in infants 19 . CB1 receptors are found primarily in the brain but also in several peripheral tissues. CB2 receptors are found mainly in immune and hematopoietic cells but can be upregulated in other tissues 20 . THC exerts its action by binding to CB 1 R and CB 2 R. In contrast, CBD is able to exert some of its most important neural effects through some non-CB-R mechanisms 21 . CBD has a very wide range of targets which are reviewed by Watt and Karl 22 . CBD has low affinity for both CB 1 R and CB 2 R 23 , but has very strong neuromodulatory effects via the serotonergic receptor 5-hydroxytryptophan 1A (5HT-1A) 21 and enhances levels of EC 24 , 25 . The discovery of the mechanisms of action of have led to further studies which illustrated the anxiolytic, immunomodulatory, anti-emetic, anti-psychotic, anti-oxidative and neuroprotective properties of CBD. The greatest potential value of CBD is in neuropsychology and neuropsychiatry. Even at low doses CBD is known to have beneficial physiological effects which promote and maintain health (anti-oxidative, anti-inflammatory, neuroprotection). The pharmacology and potential therapeutic role in epilepsy and other neuropsychiatric disorders is reviewed by Devinsky et al. (2014) 26 . The use of cannabinoid based therapies for the treatment of spasticity, pain and anorexia demonstrated to clinicians and pharmaceutical companies that it is possible to develop and commercialize cannabinoids for human disease. Anti-epileptic properties In cases where traditional antiepileptic drugs had failed or had intolerable side effects many patients had turned to medical marijuana. Accounts of dramatic improvements with cannabis based products with high CBD:THC (e.g, >20:1) ratios sparked serious interest in rigorous scientific study of CBD as an anti-seizure medication 27 . The most significant advances have been made in the development of CBD as a treatment for children with Dravet syndrome (DS), an early onset "treatment resistant" severe epilepsy. Supported by good results in a series of pre-clinical studies 28 , the safety and effectiveness of a standardized oral solution of a 98% oil-based CBD extract called Epidiolex was successfully tested in an open label trial involving 214 children and young adults 29 . This culminated in a randomized, double-blind, placebo-controlled trial of CBD as a treatment for DS in 2017 30 . CBD was effective in reducing the frequency of convulsive seizures with 5% of patients becoming seizure free. Epidiolex is now considered to be at the forefront of cannabis based therapies and is Food and Drug Administration (FDA) approved for treatment of DS. There are plans to conduct similar trials for other forms of epilepsy since the neuroprotective activity of CBD is likely to have a similar effect 26 . Antipsychotic properties Cannabidiol is currently being used for the treatment of a variety of conditions including psychiatric disorders such as schizophrenia and dementia 31 . Systematic reviews of the antipsychotic properties of CBD in humans and the effect of CBD on cognitive function are provided by Iseger & Bossong, (2015) 27 and Osborne et al., (2017) 32 respectively and Watt & Karl, (2017) 22 have conducted a mini review of the in vivo evidence for therapeutic properties of CBD for Alzheimer's disease. Cannabidiol has the ability to counteract psychotic symptoms and cognitive impairment associated with acute THC administration and may lower the risk of developing cannabis -related psychosis. Five small scale clinical studies with CBD treatment of patients with psychotic symptoms all showed improvement of symptomology or psychometric test performance with no side effects. Several pre-clinical mouse models demonstrate improvements in recognition, social recognition and spatial memory in Alzheimer's disease but there are no human clinical trials to date 32 . Neuroprotection from trauma and infectious diseases Trauma and infections of the CNS result in neurologic dysfunction. Two pre-clinical studies in murine ischemia models showed CBD is strongly neuroprotecive and can attenuate and prevent learning and memory deficits induced by brain hypoixia 32 – 34 . Three pre-clinical studies in murine models of hepatic encephalopathy (HE) from liver failure showed CBD improved spatial learning and memory and working memory deficits 35 – 37 . In Malawi, cerebral malaria (CM) is the most common complication in Plasmodium infection. A prospective study of survivors of CM in Blantyre, The Blantyre Malaria Project Epilepsy Study (BMPES) 38 found that almost a third of retinopathy-positive CM survivors developed epilepsy or other neurological sequelae such as disruptive behavioural disorders and neurodisabilities. Neuroprotective clinical trials aimed at maintaining normothermia and seizure control are warranted. Notably, previous trials that have attempted to achieve seizure control using phenobarbital in paediatric CM have been associated with increased mortality likely due to respiratory suppression as a side effect of the drug 39 . There has been no human study to evaluate CBD in CM. However, Campos et al., (2015) 40 evaluated the effects of CBD in a murine model of CM. Mice were challenged with Plasmodium berghei ANKA and treated with CBD. Mice were also treated with Artesunate. Cytokines and neurotrophins in the prefrontal cortex and hippocampus were also measured. Challenged mice displayed memory deficits and increases in anxiety-like behaviours at the peak of disease and after the clearance of the parasitemia. Both of these effects were prevented by CBD treatment. Treatment with CBD resulted in an increase in Brain Derived Neurotrophic Factor (BDNF) expression in the hippocampus which is associated with the improvement of cognitive performance 41 , 42 . The treatment also resulted in decreased levels of proinflammatory cytokines in the hippocampus which are related with improved memory and less behaviour impairment 43 , 44 This mechanism of protection involving regulation of neurogenesis and neuroinflammatory markers can be extended to murine hepatic encephalopathy model, where CBD chronic treatment restored BDNF levels and decreased the mRNA expression of the type-1 TNF-α receptors in the brain 36 . An intravital microscopy study showed that CBD reduced vascular changes, CNS leukocyte migration and TNF-α expression induced by the administration of LPS in rodents 45 . Taken together, the evidence suggests that CBD may be a promising candidate as an adjunctive therapy in combination with various antimicrobial agents to prevent brain damage and neurological outcomes of various infections in humans Safety Anecdotal evidence has suggested that CBD has a favorable side effect profile, which may improve compliance and adherence to treatment. From the numerous clinical studies that have since been performed on CBD, its safety, even at high doses, has been proven in the context of various medical conditions. The 39th Meeting of the World Health Organization's Expert Committee on Drug Dependence in Geneva held from 6–10 November 2017 found that there is no evidence of recreational use of CBD or any public health related problems associated with the use of pure CBD 46 . Bergamaschi et al., (2011) 31 reviewed 132 original studies on CBD's safety and side effects. All studies consistently find CBD to be safe, that catalepsy is not induced and physiological parameters such as heart rate, blood pressure and body temperature are unaltered. Psychological and psychomotor functions are not adversely affected and neither are gastrointestinal transit and food intake. There is no toxicity and chronic use and high doses of up to 1,500 mg per day are well tolerated. Another review of clinical data and relevant animal studies related to the safety and side effects of CBD 47 noted that because of its importance as adjunct therapy more emphasis needs to be placed on clinical research which looks at CBD's interaction with hepatic (drug)-metabolizing enzymes, such as those belonging to the cytochrome P450 family, drug-transporters and interactions with other drugs. CBD is metabolized, amongst others, via the CYP3A4 enzyme which is also involved in the metabolism of estimated 60% of clinically prescribed drugs 31 . Various drugs such as ketoconazol, itraconazol, ritonavir and clarithromycin inhibit this enzyme. When co-administered with CBD this leads to slower CBD degradation and can consequently lead to higher CBD doses that are longer pharmaceutically active. In contrast, phenobarbital, rifampicin, carbamazepine and phenytoin induce CYP3A4 causing reduced CBD bioavailability. These interactions can have both positive and negative effects. For example, in a clinical study of children treated for epilepsy with clobazam and CBD the CBD interaction with isozymes CYP3A4 and CYP2C19 caused increased clobazam bioavailability, making it possible to reduce the dose of the anti-epileptic drug, which in turn reduced its side effects 48 . Some of the parameters summarized by Bergamaschi et al., (2011), which were observed in animal experiments, are yet to be studied in humans 49 . Given that the endocannabinoid system also plays an important role in endocrine regulation, further research of CBD's off-target effects in this area is needed. Conclusions and future directions Cannabidiol is an effective and safe therapeutic agent with the potential to help in addressing many of Malawi's public health challenges. It is a proven and approved anti-seizure therapy. The efficacy of CBD in models of neuronal injury, neurodegeneration and neuroinflammation suggest that it could also be effective in a wide range of conditions as a preventive or therapeutic agent. This is especially true for the cerebral malaria where pre-clinical data show CBD could be an effective adjunctive therapy for prevention of neurological sequelae. Taken together with its good safety profile, a case can be made for the consideration of clinical trials of CBD based therapies in Malawi. The commendable record of clinical research in malaria in Malawi and the commercial availability of high quality CBD formulations globally positions Malawi well to advance in this area. Malawi is home to one of the most globally sought after Cannabis sativa strains but is yet to benefit in any meaningful way from this unique aspect of our natural heritage. An immediate priority has to be the development and production of CBD formulations from local cannabis varieties. The significant steps taken thus far in legislative reform, advocacy and awareness need to be built upon. Genetic studies and breeding would allow Malawi to develop high grade cannabis varieties with unique characteristics for medicinal uses. Registration of plant varieties and geographic indicators for strains from Malawi is key. In recent years there has been increased focus on the importance of plant based therapies such as herbal and traditional medicines in public health. The scientific and medical community has responded proactively by focussing and mobilising research capacity and investment in the field. The Africa Centre of Excellence in Public Health and Herbal Medicine (ACEPHEM) is an example. The explosion of interest and investment in cannabis based phytochemicals could be a windfall for phytopharmaceuticals in general. Opportunities exist for cross pollination between the approaches and tools that have been developed for cannabis based medicines and those in the broader herbal medicine arena. Importantly, information gathered from the development of herbal medicines could give indications on the potential medicinal value of some of the other phytocannabinoids that are only now beginning to be investigated.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3911078/

Improved Short-Sequence-Repeat Genotyping of Mycobacterium avium subsp. paratuberculosis by Using Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry

Accurate sequence analysis of mononucleotide repeat regions is difficult, complicating the use of short sequence repeats (SSRs) as a tool for bacterial strain discrimination. Although multiple SSR loci in the genome of Mycobacterium avium subsp. paratuberculosis allow genotyping of M. avium subsp. paratuberculosis isolates with high discriminatory power, further characterization of the most discriminatory loci is limited due to inherent difficulties in sequencing mononucleotide repeats. Here, a method was evaluated using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) as an alternative to Sanger sequencing to further differentiate the dominant mycobacterial interspersed repetitive-unit (MIRU)–variable-number tandem-repeat (VNTR) M. avium subsp. paratuberculosis type ( n = 37) in Canadian dairy herds by targeting a highly discriminatory mononucleotide SSR locus. First, PCR-amplified DNA was digested with two restriction enzymes to yield a sufficiently small fragment containing the SSR locus. Second, MALDI-TOF MS was performed to identify the mass, and thus repeat length, of the target. Sufficiently intense, discriminating spectra were obtained to determine repeat lengths up to 15, an improvement over the limit of 11 using traditional sequencing techniques. Comparison to synthetic oligonucleotides and Sanger sequencing results confirmed a valid and reproducible assay that increased discrimination of the dominant M. avium subsp. paratuberculosis MIRU-VNTR type. Thus, MALDI-TOF MS was a reliable, fast, and automatable technique to accurately resolve M. avium subsp. paratuberculosis genotypes based on SSRs.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2730335/

Anthrax of the Gastrointestinal Tract

When swallowed, anthrax spores may cause lesions from the oral cavity to the cecum. Gastrointestinal anthrax is greatly underreported in rural disease-endemic areas of the world. The apparent paucity of this form of anthrax reflects the lack of facilities able to make the diagnosis in these areas. The spectrum of disease, ranging from subclinical infection to death, has not been fully recognized. In some community-based studies, cases of gastrointestinal anthrax outnumbered those of cutaneous anthrax. The oropharyngeal variant, in particular, is unfamiliar to most physicians. The clinical features of oropharyngeal anthrax include fever and toxemia, inflammatory lesion(s) in the oral cavity or oropharynx, enlargement of cervical lymph nodes associated with edema of the soft tissue of the cervical area, and a high case-fatality rate. Awareness of gastrointestinal anthrax in a differential diagnosis remains important in anthrax-endemic areas but now also in settings of possible bioterrorism.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9324797/

Signatures of selection in core and accessory genomes indicate different ecological drivers of diversification among Bacillus cereus clades

Abstract Bacterial clades are often ecologically distinct, despite extensive horizontal gene transfer (HGT). How selection works on different parts of bacterial pan‐genomes to drive and maintain the emergence of clades is unclear. Focusing on the three largest clades in the diverse and well‐studied Bacillus cereus sensu lato group, we identified clade‐specific core genes (present in all clade members) and then used clade‐specific allelic diversity to identify genes under purifying and diversifying selection. Clade‐specific accessory genes (present in a subset of strains within a clade) were characterized as being under selection using presence/absence in specific clades. Gene ontology analyses of genes under selection revealed that different gene functions were enriched in different clades. Furthermore, some gene functions were enriched only amongst clade‐specific core or accessory genomes. Genes under purifying selection were often clade‐specific, while genes under diversifying selection showed signs of frequent HGT. These patterns are consistent with different selection pressures acting on both the core and the accessory genomes of different clades and can lead to ecological divergence in both cases. Examining variation in allelic diversity allows us to uncover genes under clade‐specific selection, allowing ready identification of strains and their ecological niche. 1 INTRODUCTION Bacterial strains often appear grouped together in distinct phylogenetic clusters, or "clades," despite frequent homogenizing horizontal gene transfer (HGT; Buckee et al., 2008 ; Fraser et al., 2007 ; Schloss & Handelsman, 2004 ). Although uncovered by methods that are blind to ecology (Carroll et al., 2020 ; Guinebretière et al., 2008 ; Priest et al., 2004 ), these clades are often ecologically distinct from each other, both in phenotype and in genome content (Cohan, 2016 ; Hanage et al., 2006 ). How distinct bacterial phylogenetic clades appear is not fully understood (Doolittle & Papke, 2006 ). A key question in the debate is whether ecological differentiation is determined primarily by selection on the core genome (genes shared by all strains within a clade) or the accessory genome (genes shared only by a subset of those strains; Maistrenko et al., 2020 ; McInerney et al., 2017 ; Tettelin et al., 2008 ). Selection in bacteria can be divided into three main categories: purifying selection, which removes deleterious alleles from a population; diversifying selection, which increases allelic diversity when rare alleles confer an advantage (McNally et al., 2019 ; Molina & Van Nimwegen, 2008 ); and directional selection, where alleles of genes are replaced by fitter variants (Cohan, 2016 ). Because recognizing directional selection requires data from a large number of isolates over a substantial time period (Buckee et al., 2008 ; Chen et al., 2006 ; Lefébure & Stanhope, 2007 ), we will focus on purifying and diversifying selection in this study. Purifying selection is prevalent amongst microbes (McNally et al., 2019 ) and common amongst core genes, either because they are integral to cellular processes or vital for survival in a given habitat (Cohan, 2016 , 2017 ). Purifying selection may maintain cohesion within a clade by purging diversity, isolating clades from each other and maintaining their distinctiveness (Cohan, 2011 , 2016 ). Diversifying selection also plays a key role in microbial evolution by maintaining multiple allelic variants within a population, a pattern which is common in genes linked to host colonization, phage resistance, and responses to vaccines and antibiotics (Harrow et al., 2021 ; McNally et al., 2019 ). Quantifying the relative impact of selection on bacterial divergence is challenging, as this is dependent upon grouping multiple strains together as a single species, which has proven more difficult for bacteria than for plants and animals (Robinson et al., 2017 ). In addition to difficulties in recognizing directional selection, identifying which regions in the core genome are under selection is challenging due to inconsistent selection across sites within a gene and over time, the ubiquity of negative selection across genomes and mutation rate heterogeneity (Chen et al., 2006 ; Fay & Wu, 2003 ; Zhang et al., 2005 ). In contrast, selection on accessory genes is more simply inferred using presence/absence (Méric, Mageiros, Pascoe, et al., 2018 ; Vasquez‐Rifo et al., 2019 ). In this study, we aimed to identify regions under strong selection while accounting for other factors that influence the rate of molecular evolution (Méric, Mageiros, Pascoe, et al., 2018 ). Genes with higher or lower allelic diversity compared to the genomic average are probably under diversifying and purifying selection respectively, so we combined our expectations of allelic diversity with a method used to infer differences in allelic diversity between genes (Cohan, 2016 ; Méric, Mageiros, Pensar, et al., 2018 ; Shea et al., 2011 ). We applied this method to closely related bacterial clades with distinct ecological niches which we hypothesize have undergone divergent selection pressures. This methodology allows us to uncover the effect of selection within bacterial core genomes and compare selective pressures acting on different bacterial clades. The Bacillus cereus ( Bc ) group contains a number of species with clinical or industrial importance: Bacillus anthracis ( Ba ), the causative agent of anthrax (Turnbull, 2002 ); Bacillus cereus sensu stricto , a causative agent of food‐poisoning (Messelhäußer & Ehling‐Schulz, 2018 ); Bacillus thuringiensis ( Bt ), a group of specialized invertebrate pathogens widely exploited as biopesticides (Bravo et al., 2007 ); and Bacillus mycoides ( Bm ), a psychrotolerant species which incorporates the former Bacillus weihenstephanensis (Lechner et al., 1998 ; Liu et al., 2018 ). The Bc group has been well studied, and there is increasing evidence that different clades have distinct ecological niches (Manktelow et al., 2021 ; Zheng et al., 2017 ), making this group ideal for exploring the importance of niche‐specific selection in driving clade divergence. For instance, carriage of enterotoxins and insecticidal toxin genes is known to vary strongly between clades (Cardazzo et al., 2008 ; Méric, Mageiros, Pascoe, et al., 2018 ); clade also correlates with habitat, thermal niche and cytotoxicity (Guinebretière et al., 2008 , 2010 ; Raymond et al., 2010 ). Thermal niches and clade also predict relative fitness at different temperatures and fitness in a model insect host (Manktelow et al., 2021 ) and are linked to differences in biogeographical distribution (Drewnowska et al., 2020 ). The phylogenetic structure of the group is well established and recoverable when alignments of multiple housekeeping genes (multilocus sequence typing [MLST]) or of the entire core genome are used to create phylogenies (Méric, Mageiros, Pascoe, et al., 2018 ; Priest et al., 2004 ). However, the taxonomy of the Bc group is much disputed (Carroll et al., 2020 ; Helgason et al., 2000 ; Liu et al., 2015 ). Different authors subdivide the group based on different levels of genetic distinctiveness; consequently, the number of informally recognized clades ranges between five and seven (Guinebretière et al., 2008 ; Méric, Mageiros, Pascoe, et al., 2018 ). In this study we will use the five‐clade structure initially recovered by MLST (Raymond et al., 2010 ), as these groups are clearly separated by large phylogenetic distances. This is important to our methodology as selection leading to clade divergence should have occurred far in the past. The gene‐by‐gene approach used here—which relies on known loci across multiple genomes—means that our methods cannot identify recent directional selection that has occurred only within a subset of a given clade or new imports that do not belong to a recognized locus (Sheppard et al., 2012 ) and so has not been designed to identify "ecotypes" with recent evolutionary origins (Cohan, 2016 ). We will focus on clades 1, 2 and 3 in the Bc group, originally named the " anthracis ," " kurstaki " and " weihenstephanensis " clades respectively (Priest et al., 2004 ). While Bacillus cereus sensu stricto strains are found in all three clades (Patiño‐Navarrete & Sanchis, 2017 ), Clade 1 contains all Ba isolates, Clade 2 contains the majority of insecticidal Bt isolates while Clade 3 corresponds to the psychrotolerant Bacillus mycoides species (Liu et al., 2018 ). Bacteria in these clades are readily isolated from both clinical and natural environments and are well represented in genomic databases. Here, we hypothesized that the three Bc clades are ecologically distinct due to selection on their core genomes. We predicted that different genes would be found within the clade‐specific core genomes of each clade, and that these genes would have different levels of allelic diversity in each clade, due to differences in selection pressure. We also hypothesized that ecological selection acts on the accessory genome and that HGT would be more frequent amongst diversifying genes, promoting diversification between clades. A large collection of Bc isolate genomes were used to reconstruct the five‐clade phylogeny identified in previous studies (Méric, Mageiros, Pascoe, et al., 2018 ; Priest et al., 2004 ). Based on comparisons of gene‐level allelic diversity to the Bc strict core genome average (Chattopadhyay et al., 2009 ; Méric, Mageiros, Pensar, et al., 2018 ), we identified genes core to each clade under selection, while presence/absence was used to identify accessory genes under selection (Méric, Mageiros, Pascoe, et al., 2018 ; Vasquez‐Rifo et al., 2019 ). These genes were subjected to Gene Ontology (GO) analyses to determine functional enrichment, while consistency indices (CInds) were used to estimate rates of HGT (Méric, Mageiros, Pensar, et al., 2018 ). 2 MATERIALS AND METHODS 2.1 Isolate selection Bacillus cereus ( Bc ) sequence assemblies were gathered from the Multispecies BIGSdb database (Jolley & Maiden, 2010 ; https://sheppardlab.com/resources/ ). The isolates belonged to a recognized Bc sl species (Bazinet, 2017 ), were assembled from fewer than 3,000 contigs and had genome sizes in line with previous estimates for the group (Chun et al., 2012 ; Li et al., 2015 ; Méric, Mageiros, Pascoe, et al., 2018 ; Yi et al., 2016 ). In total, 352 isolate genomes met the selection criteria; of these, 24 isolates could not be assigned to clades with certainty and were removed from the analysis, leaving 328 isolate genomes (Table S1 ). 2.2 Creation of a reference pan‐genome The assemblies were aligned using the mafft algorithm (Katoh & Standley, 2013 ) and a gene‐by‐gene approach. Assembly was conducted in the BIGSdb database (Sheppard et al., 2012 ). Contiguous sequences for each isolate were exported and entered into the Pan‐genome Iterative Refinement And Threshold Evaluation ( pirate ) toolbox (Bayliss et al., 2019 ). In the pirate toolbox, genome sequences are passed through multiple cluster thresholds to account for different selection strengths between isolates, avoiding over‐clustering and over‐splitting of groups (Bayliss et al., 2019 ). Sequences are filtered from input files and cd‐hit used to create sequence clusters. Markov Cluster (MCL) processes are repeated by pirate at default amino acid identity thresholds; the initial clustering at the lowest threshold identified "gene families" and continued until the highest user‐specified threshold. Unique MCL clusters at the highest threshold (95% amino acid identity) were classified as "unique alleles" (Bayliss et al., 2019 ). Paralogues were identified and loci were classified, then gene families with multiple loci were checked for over‐clustering. Genes were annotated using prokka (Seemann, 2014 ). pirate produced a gene presence/absence matrix, with each gene possessing its own identifier (Méric et al., 2014 ). The strict core genome for the entire data set was identified in excel by ordering genes based on the percentage of isolates within the group containing this gene. Genes were considered "strict core" if present in all isolates. 2.3 Phylogenetic analysis A maximum‐likelihood phylogeny was produced using 1,004 "strict core" gene sequences. These strict core genes were present in all isolates used in this study. The concatenated sequences were aligned using mafft (Katoh & Standley, 2013 ). A maximum‐likelihood phylogeny (Gadagkar et al., 2005 ; Saitou & Imanishi, 1989 ) was produced using iq‐tree (Minh et al., 2020 ) with ModelFinder (Kalyaanamoorthy et al., 2017 ); the substitution model selected was GTR+F+R10. Inclusion of isolates assigned to clades in a previous study helped with clade recovery (Méric, Mageiros, Pascoe, et al., 2018 ). The tree was visualized using the r package ggtree (Yu et al., 2017 ). 2.4 Identifying core and accessory genes under selection within clades To derive clade‐specific core and accessory genomes, strains within clades 1–3 were extracted in R by using ggtree (Yu et al., 2017 ). From these, we reconstructed clade‐specific core genomes consisting of genes present in ≥95% of the isolates within each clade. This led to a reduced chance of rejecting "clade‐defining" genes that have been lost in very derived isolates. Based on previous observations of allelic diversity and selection (Cohan, 2016 ; Dugatkin et al., 2005 ; Shea et al., 2011 ), genes of low allelic diversity were considered to be under purifying selection (i.e., selection leading to a reduced number of different alleles) while genes of high allelic diversity were considered to be under diversifying selection (i.e., selection leading to a greater number of alleles). All alleles of each gene in the strict core and clade‐specific core genomes were found through comparison of the isolates to a representative FASTA sequence in the Multispecies BIGSdb Genome Comparator under default parameters. Incomplete loci were ignored for pairwise comparison and paralogues were excluded entirely (Jolley & Maiden, 2010 ). We produced alignments using the mafft algorithm (Katoh & Standley, 2013 ). Diversity per locus was calculated for each gene by dividing the number of distinct alleles by the number of isolates containing that gene (Méric, Mageiros, Pascoe, et al., 2018 ; Méric, Mageiros, Pensar, et al., 2018 ). To distinguish selected regions from neutral ones while accounting for other factors influencing molecular evolution, the allelic diversity of each clade‐specific core gene was compared to the overall within‐clade diversity of the strict core genome (Fay & Wu, 2003 ; Méric, Mageiros, Pensar, et al., 2018 ). Those genes that lay outside two standard deviations of the core genome average (i.e., ~5% of the genes) were considered to have significantly low or high diversity and were therefore considered to be under selection (Cohan, 2016 ; Dugatkin et al., 2005 ; Shea et al., 2011 ). Clade‐specific accessory genes were defined as genes present in under 95% of a clade; gene presence/absence was used to identify accessory genes under selection as in previous studies (Méric, Mageiros, Pascoe, et al., 2018 ; Vasquez‐Rifo et al., 2019 ). 2.5 Gene Ontology analysis To determine whether selected clade‐specific core and accessory genes were enriched for certain functions, each gene was assigned an identification number from the Universal Protein Resource Knowledge Base (UniProtKB; Boutet et al., 2007 ), based on pirate 's prediction of their gene name and function (Bayliss et al., 2019 ). Bacillus subtilis identification codes were used because the list of B. subtilis UniProtKB codes is more comprehensive than for the list for Bc , and gene names and functions are equivalent between the species. UniProtKB codes were also assigned for the strict core genes. Where a gene coded for a hypothetical function or had no suitable orthologue amongst B. subtilis , the gene was excluded from the analysis. Codes for each set of genes were entered into the Gene Enrichment Analysis tool on the Gene Ontology website, which uses the panther classification system (Mi et al., 2013 ). Over‐ and under‐representation of biological processes compared to the strict core genome was calculated using binomial testing (Rupert Jr, 2012 ) with replacement, approximating the hypergeometric distribution due to sample size (Rivals et al., 2007 ). A Bonferroni correction was used to account for multiple testing (Weisstein, 2004 ). 2.6 Inference of HGT using consistency indices To examine the impact of HGT on clade formation, consistency indices (CInds) were used to estimate the level of HGT amongst genes under selection (Méric, Mageiros, Pensar, et al., 2018 ). CInds were created to detect homoplasy by comparing the fit of genetic alignment data to a phylogenetic tree. An alignment of allelic sequences from the same gene is compared to a reliable phylogeny produced using multiple conserved genes (Saitou & Imanishi, 1989 ) to produce a consistency index; lower indices indicate a greater degree of homoplasy. Homoplasy can be caused by independent mutation but is commonly assumed to be caused mainly by homologous recombination (Sanderson & Donoghue, 1989 ; Schliep, 2011 ), meaning that CInds can be used to infer levels of HGT within a group of bacterial strains. Only genes that were present in all strains in the phylogeny and were considered under either purifying or diversifying selection in at least one clade were included in the consistency index analysis. Consistency indices were calculated for each gene using the r package phangorn (Schliep, 2011 ) and the maximum‐likelihood group phylogeny was used for comparisons. The process was repeated for all genes in the strict core genome ( n = 1,004). The average CInd of each gene set was compared using a Wilcoxon–Mann–Whitney test. The frequency distribution of CInds for both gene sets was also examined. Both analyses have previously been conducted to test for significant differences in CInds between sets of genes (Méric, Mageiros, Pensar, et al., 2018 ). 2.1 Isolate selection Bacillus cereus ( Bc ) sequence assemblies were gathered from the Multispecies BIGSdb database (Jolley & Maiden, 2010 ; https://sheppardlab.com/resources/ ). The isolates belonged to a recognized Bc sl species (Bazinet, 2017 ), were assembled from fewer than 3,000 contigs and had genome sizes in line with previous estimates for the group (Chun et al., 2012 ; Li et al., 2015 ; Méric, Mageiros, Pascoe, et al., 2018 ; Yi et al., 2016 ). In total, 352 isolate genomes met the selection criteria; of these, 24 isolates could not be assigned to clades with certainty and were removed from the analysis, leaving 328 isolate genomes (Table S1 ). 2.2 Creation of a reference pan‐genome The assemblies were aligned using the mafft algorithm (Katoh & Standley, 2013 ) and a gene‐by‐gene approach. Assembly was conducted in the BIGSdb database (Sheppard et al., 2012 ). Contiguous sequences for each isolate were exported and entered into the Pan‐genome Iterative Refinement And Threshold Evaluation ( pirate ) toolbox (Bayliss et al., 2019 ). In the pirate toolbox, genome sequences are passed through multiple cluster thresholds to account for different selection strengths between isolates, avoiding over‐clustering and over‐splitting of groups (Bayliss et al., 2019 ). Sequences are filtered from input files and cd‐hit used to create sequence clusters. Markov Cluster (MCL) processes are repeated by pirate at default amino acid identity thresholds; the initial clustering at the lowest threshold identified "gene families" and continued until the highest user‐specified threshold. Unique MCL clusters at the highest threshold (95% amino acid identity) were classified as "unique alleles" (Bayliss et al., 2019 ). Paralogues were identified and loci were classified, then gene families with multiple loci were checked for over‐clustering. Genes were annotated using prokka (Seemann, 2014 ). pirate produced a gene presence/absence matrix, with each gene possessing its own identifier (Méric et al., 2014 ). The strict core genome for the entire data set was identified in excel by ordering genes based on the percentage of isolates within the group containing this gene. Genes were considered "strict core" if present in all isolates. 2.3 Phylogenetic analysis A maximum‐likelihood phylogeny was produced using 1,004 "strict core" gene sequences. These strict core genes were present in all isolates used in this study. The concatenated sequences were aligned using mafft (Katoh & Standley, 2013 ). A maximum‐likelihood phylogeny (Gadagkar et al., 2005 ; Saitou & Imanishi, 1989 ) was produced using iq‐tree (Minh et al., 2020 ) with ModelFinder (Kalyaanamoorthy et al., 2017 ); the substitution model selected was GTR+F+R10. Inclusion of isolates assigned to clades in a previous study helped with clade recovery (Méric, Mageiros, Pascoe, et al., 2018 ). The tree was visualized using the r package ggtree (Yu et al., 2017 ). 2.4 Identifying core and accessory genes under selection within clades To derive clade‐specific core and accessory genomes, strains within clades 1–3 were extracted in R by using ggtree (Yu et al., 2017 ). From these, we reconstructed clade‐specific core genomes consisting of genes present in ≥95% of the isolates within each clade. This led to a reduced chance of rejecting "clade‐defining" genes that have been lost in very derived isolates. Based on previous observations of allelic diversity and selection (Cohan, 2016 ; Dugatkin et al., 2005 ; Shea et al., 2011 ), genes of low allelic diversity were considered to be under purifying selection (i.e., selection leading to a reduced number of different alleles) while genes of high allelic diversity were considered to be under diversifying selection (i.e., selection leading to a greater number of alleles). All alleles of each gene in the strict core and clade‐specific core genomes were found through comparison of the isolates to a representative FASTA sequence in the Multispecies BIGSdb Genome Comparator under default parameters. Incomplete loci were ignored for pairwise comparison and paralogues were excluded entirely (Jolley & Maiden, 2010 ). We produced alignments using the mafft algorithm (Katoh & Standley, 2013 ). Diversity per locus was calculated for each gene by dividing the number of distinct alleles by the number of isolates containing that gene (Méric, Mageiros, Pascoe, et al., 2018 ; Méric, Mageiros, Pensar, et al., 2018 ). To distinguish selected regions from neutral ones while accounting for other factors influencing molecular evolution, the allelic diversity of each clade‐specific core gene was compared to the overall within‐clade diversity of the strict core genome (Fay & Wu, 2003 ; Méric, Mageiros, Pensar, et al., 2018 ). Those genes that lay outside two standard deviations of the core genome average (i.e., ~5% of the genes) were considered to have significantly low or high diversity and were therefore considered to be under selection (Cohan, 2016 ; Dugatkin et al., 2005 ; Shea et al., 2011 ). Clade‐specific accessory genes were defined as genes present in under 95% of a clade; gene presence/absence was used to identify accessory genes under selection as in previous studies (Méric, Mageiros, Pascoe, et al., 2018 ; Vasquez‐Rifo et al., 2019 ). 2.5 Gene Ontology analysis To determine whether selected clade‐specific core and accessory genes were enriched for certain functions, each gene was assigned an identification number from the Universal Protein Resource Knowledge Base (UniProtKB; Boutet et al., 2007 ), based on pirate 's prediction of their gene name and function (Bayliss et al., 2019 ). Bacillus subtilis identification codes were used because the list of B. subtilis UniProtKB codes is more comprehensive than for the list for Bc , and gene names and functions are equivalent between the species. UniProtKB codes were also assigned for the strict core genes. Where a gene coded for a hypothetical function or had no suitable orthologue amongst B. subtilis , the gene was excluded from the analysis. Codes for each set of genes were entered into the Gene Enrichment Analysis tool on the Gene Ontology website, which uses the panther classification system (Mi et al., 2013 ). Over‐ and under‐representation of biological processes compared to the strict core genome was calculated using binomial testing (Rupert Jr, 2012 ) with replacement, approximating the hypergeometric distribution due to sample size (Rivals et al., 2007 ). A Bonferroni correction was used to account for multiple testing (Weisstein, 2004 ). 2.6 Inference of HGT using consistency indices To examine the impact of HGT on clade formation, consistency indices (CInds) were used to estimate the level of HGT amongst genes under selection (Méric, Mageiros, Pensar, et al., 2018 ). CInds were created to detect homoplasy by comparing the fit of genetic alignment data to a phylogenetic tree. An alignment of allelic sequences from the same gene is compared to a reliable phylogeny produced using multiple conserved genes (Saitou & Imanishi, 1989 ) to produce a consistency index; lower indices indicate a greater degree of homoplasy. Homoplasy can be caused by independent mutation but is commonly assumed to be caused mainly by homologous recombination (Sanderson & Donoghue, 1989 ; Schliep, 2011 ), meaning that CInds can be used to infer levels of HGT within a group of bacterial strains. Only genes that were present in all strains in the phylogeny and were considered under either purifying or diversifying selection in at least one clade were included in the consistency index analysis. Consistency indices were calculated for each gene using the r package phangorn (Schliep, 2011 ) and the maximum‐likelihood group phylogeny was used for comparisons. The process was repeated for all genes in the strict core genome ( n = 1,004). The average CInd of each gene set was compared using a Wilcoxon–Mann–Whitney test. The frequency distribution of CInds for both gene sets was also examined. Both analyses have previously been conducted to test for significant differences in CInds between sets of genes (Méric, Mageiros, Pensar, et al., 2018 ). 3 RESULTS 3.1 The Bacillus cereus group phylogeny has a distinct clade structure The strict core genome phylogeny divided Bc isolates into genetically distinct clades. In total, 328 genomes from the Multispecies BIGSdb database (Jolley & Maiden, 2010 ) met criteria for the study, with an average size of ~5.6 ± 0.3 Mb (Table S1 ) and an average contig number of 285. Variation in assembly sequence size and contig number was consistent with other published estimates of Bacillus group genome sizes (Chun et al., 2012 ; Li et al., 2015 ; Méric, Mageiros, Pascoe, et al., 2018 ; Takeno et al., 2012 ; Yi et al., 2016 ). The group pan‐genome produced by pirate contained 36,687 genes, consisting of 1,004 strict core genes excluding homologues and 35,679 accessory genes. A maximum‐likelihood tree was produced using the concatenated strict core genome sequences and was consistent with the five‐clade phylogeny proposed by previous studies (Méric, Mageiros, Pascoe, et al., 2018 ; Sorokin et al., 2006 ; Figure 1a ). The three largest clades, clades 1–3, contained 94, 95 and 78 isolates respectively (Figure 1a ). FIGURE 1 Identifying core and accessory genes under selection within the Bacillus cereus sensu lato ( Bc sl ) phylogeny. (a) Maximum‐likelihood phylogeny of the Bc sl group strains used in this study. The concatenated core genome sequences were aligned using mafft and fed into iq‐tree . Clade identity was determined through reference to type strains and consultation of existing clade metadata. (b) The process by which genes of low and high diversity were identified. The graph shows the frequency distribution of allelic diversity values across genes within a clade's flexible core genome. The solid line shows the mean diversity of the strict core genome within the clade and the dashed lines show the second standard deviation interval of the strict core genome 3.2 Functional enrichment is dependent on clade and whether the genes are core or accessory Analysis of clade‐specific core genes under selection suggests different selective pressures acting on each Bc clade. Allelic diversity was calculated for each gene that was present in all strains within a specific clade—the clade‐specific core genes—and compared to the strict core genome average to identify genes under purifying or diversifying selection (Figure 1b ). Out of 4,383 clade‐specific core genes across three clades, 261 had allelic diversity significantly lower than the within‐clade strict core genome average (two standard deviations below the mean), while 161 had significantly higher allelic diversity than the within‐clade strict core genome average (two standard deviations above the mean; Table S2 ). Despite some genes appearing in multiple clade‐specific core genomes, most genes were conserved or diverse only within one clade (Figure 2 ). Genes found to be conserved or diverse in previous studies were also found to be conserved or diverse respectively in this study. These included the cspA gene, coding for a highly conserved cold‐shock protein used to classify the psychrotolerant Bm (Lechner et al., 1998 ), and the hag gene which encodes a diverse bacterial flagella protein (Xu & Côté, 2006 ). Genes linked to functions such as protein export were conserved in all clades (Bost & Belin, 1997 ; Fröderberg et al., 2004 ; Table S2 ) and, as expected, Clade 3 contained many highly conserved cold‐shock proteins (Ermolenko & Makhatadze, 2002 ). Genes under diversifying selection in all clades included genes coding for flagellin (Xu & Côté, 2006 ) and the bacteriophage membrane receptor yueB (São‐José et al., 2004 ). A notable gene under diversifying selection in Clade 2 was emrB , a multidrug export protein (Lomovskaya & Lewis, 1992 ). FIGURE 2 Venn diagram showing patterns of purifying and diversifying selection in the clade‐specific core genomes of the three largest Bacillus cereus sensu lato ( Bc sl ) clades. Numbers indicate the total number of genes that are experiencing selection (either purifying or diversifying), and location of numbers indicates whether the genes are experiencing selection in one clade or in multiple clades. (a) Genes under purifying selection ( n = 261). (b) Genes under diversifying selection ( n = 161) Clade‐specific accessory genes under selection were identified through presence/absence to a specific clade. In total, 5,239, 7,559 and 5,605 genes were found only in Clade 1, Clade 2 and Clade 3 respectively and present in less than 95% of the clade. Accessory genes under positive selection in each clade showed functions that are distinct to each clade. Of these, several are worthy of note; the Clade 1‐specific accessory genome included the gene InlA , which codes for internalin‐A and allows the invasion of mammal cells (Dhar et al., 2000 ), the Clade 2‐specific accessory genome included Cry toxins—key Bt insecticidal toxins—such as cry2Ab (Zheng et al., 2017 ), and the Clade 3‐specific accessory genome contained the gene binA , which produces a homologue to an insecticidal binary toxin component (Palma et al., 2014 ; Table S2 ). 3.3 GO analyses suggest clade‐specific selection acting on the core and accessory genomes of each Bc clade Binomial testing was used to measure the functional enrichment of biological processes (Ashburner et al., 2000 ; Gene Ontology Consortium, 2019 ) within clade‐specific core and accessory genomes (Mi et al., 2013 ) by comparison to the strict core genome. This methodology allowed ecological characterization of the clades and avoided a priori assumptions of relevance. Additionally, it avoids characterizing a clade by the possession of any one gene, as has often been the case in the Bc sl group (Bravo et al., 2007 ; Lechner et al., 1998 ). There was significant functional enrichment of biological processes amongst conserved and diverse clade‐specific core genes of all clades; conserved clade‐specific genes were often linked to translation (Figure 3a ). However, some enrichment was clade‐specific: Clade 3 contained a greater number of conserved genes linked to negative regulation of transcription and fewer conserved genes linked to biosynthesis and stimulus response than would be expected based on the strict core genome (Figure 3a ). The same was found to be the case for diverse clade‐specific genes; genes with uncharacterized functions were more common than expected within Clade 1 and less common than expected in Clades 2 and 3, but only Clade 2 showed unique functional enrichment, with more genes linked to antibiotic and antimicrobial resistance than expected. Functional enrichment of biological processes was robust when the criteria for considering genes under selection within a clade were relaxed to include ~10% of the clade‐specific core genomes as opposed to ~5% as described above. FIGURE 3 Significant enrichment of biological processes within clade‐specific conserved core genes and clade‐specific accessory genes across the major Bacillus cereus sensu lato ( Bc sl ) clades. Enrichment values were calculated using the Gene Ontology Enrichment analysis software available online, using a binomial test with Bonferroni correction. Only significant enrichments are shown Like the clade‐specific core genomes, there were disparities in functional enrichment between the clade‐specific accessory genomes. While some processes—such as antibiotic biosynthesis—were enriched in all clade‐specific accessory genomes, there were differences between the clades regarding the enrichment of other biological processes (Figure 3b ). Interestingly, biological processes enriched within clade‐specific accessory genomes were not the same as those enriched within that clade's specific core genome. For instance, Clade 3 accessory genes were more likely to be linked to motility and secondary metabolism, while its clade‐specific core genome was not. Clade 3 was also not significantly enriched for accessory genes linked to negative regulation of transcription, while its core genome was (Figure 3b ). 3.4 Genes under diversifying selection undergo more frequent HGT Two sets of clade‐specific core genes were suitable for CInd analysis; 42 genes with low allelic diversity and 24 genes with high allelic diversity were present in all 328 strains and therefore their gene phylogeny could be compared to the strict core genome phylogeny to check for inconsistencies that suggest HGT. CInds were calculated for clade‐specific core genes of high and low diversity, as well as for all genes in the strict core genome. The CInds of each gene set suggest that genes under diversifying selection undergo frequent HGT, while HGT is uncommon amongst conserved genes (Figure 4a,b ); the mean CInd of conserved clade‐specific core genes (0.46 ± 0.02) was significantly higher than the mean of the strict core genome (0.34 ± 0.003; Wilcoxon–Mann–Whitney test; U = 33722, p = 4.435e −11 ). In contrast, the mean CInd of diverse clade‐specific core genes (0.28 ± 0.018) was significantly lower than for the strict core genome (Wilcoxon–Mann–Whitney test; U = 7893, p =.00385; Figure 5 ). FIGURE 4 Consistency between phylogenies based on a single gene and a core genome. (a) Consistency between the conserved cold shock protein ( CspA ) gene phylogeny and the strict core genome ( n = 1,004) of the Bacillus cereus sensu lato ( Bc sl ) group. A given isolate in each tree is connected to the same isolate in the other tree by a line. (b) Consistency between the diverse flagellin ( hag ) gene phylogeny and the strict core genome ( n = 1,004) of the Bc sl group. A given isolate in each tree is connected to the same isolate in the other tree by a line FIGURE 5 Consistency index distribution amongst genes under purifying selection ( n = 42) and diversifying selection ( n = 24). Consistency indices were calculated for each gene using the phangorn package in r and a maximum‐likelihood phylogeny was created using alignments of the concatenated core genome. The average consistency index of each gene set was compared to that of the strict core genome ( n = 1,004). Normalized numbers of genes represent the number of genes with a given consistency index while controlling for the size of the data set 3.1 The Bacillus cereus group phylogeny has a distinct clade structure The strict core genome phylogeny divided Bc isolates into genetically distinct clades. In total, 328 genomes from the Multispecies BIGSdb database (Jolley & Maiden, 2010 ) met criteria for the study, with an average size of ~5.6 ± 0.3 Mb (Table S1 ) and an average contig number of 285. Variation in assembly sequence size and contig number was consistent with other published estimates of Bacillus group genome sizes (Chun et al., 2012 ; Li et al., 2015 ; Méric, Mageiros, Pascoe, et al., 2018 ; Takeno et al., 2012 ; Yi et al., 2016 ). The group pan‐genome produced by pirate contained 36,687 genes, consisting of 1,004 strict core genes excluding homologues and 35,679 accessory genes. A maximum‐likelihood tree was produced using the concatenated strict core genome sequences and was consistent with the five‐clade phylogeny proposed by previous studies (Méric, Mageiros, Pascoe, et al., 2018 ; Sorokin et al., 2006 ; Figure 1a ). The three largest clades, clades 1–3, contained 94, 95 and 78 isolates respectively (Figure 1a ). FIGURE 1 Identifying core and accessory genes under selection within the Bacillus cereus sensu lato ( Bc sl ) phylogeny. (a) Maximum‐likelihood phylogeny of the Bc sl group strains used in this study. The concatenated core genome sequences were aligned using mafft and fed into iq‐tree . Clade identity was determined through reference to type strains and consultation of existing clade metadata. (b) The process by which genes of low and high diversity were identified. The graph shows the frequency distribution of allelic diversity values across genes within a clade's flexible core genome. The solid line shows the mean diversity of the strict core genome within the clade and the dashed lines show the second standard deviation interval of the strict core genome 3.2 Functional enrichment is dependent on clade and whether the genes are core or accessory Analysis of clade‐specific core genes under selection suggests different selective pressures acting on each Bc clade. Allelic diversity was calculated for each gene that was present in all strains within a specific clade—the clade‐specific core genes—and compared to the strict core genome average to identify genes under purifying or diversifying selection (Figure 1b ). Out of 4,383 clade‐specific core genes across three clades, 261 had allelic diversity significantly lower than the within‐clade strict core genome average (two standard deviations below the mean), while 161 had significantly higher allelic diversity than the within‐clade strict core genome average (two standard deviations above the mean; Table S2 ). Despite some genes appearing in multiple clade‐specific core genomes, most genes were conserved or diverse only within one clade (Figure 2 ). Genes found to be conserved or diverse in previous studies were also found to be conserved or diverse respectively in this study. These included the cspA gene, coding for a highly conserved cold‐shock protein used to classify the psychrotolerant Bm (Lechner et al., 1998 ), and the hag gene which encodes a diverse bacterial flagella protein (Xu & Côté, 2006 ). Genes linked to functions such as protein export were conserved in all clades (Bost & Belin, 1997 ; Fröderberg et al., 2004 ; Table S2 ) and, as expected, Clade 3 contained many highly conserved cold‐shock proteins (Ermolenko & Makhatadze, 2002 ). Genes under diversifying selection in all clades included genes coding for flagellin (Xu & Côté, 2006 ) and the bacteriophage membrane receptor yueB (São‐José et al., 2004 ). A notable gene under diversifying selection in Clade 2 was emrB , a multidrug export protein (Lomovskaya & Lewis, 1992 ). FIGURE 2 Venn diagram showing patterns of purifying and diversifying selection in the clade‐specific core genomes of the three largest Bacillus cereus sensu lato ( Bc sl ) clades. Numbers indicate the total number of genes that are experiencing selection (either purifying or diversifying), and location of numbers indicates whether the genes are experiencing selection in one clade or in multiple clades. (a) Genes under purifying selection ( n = 261). (b) Genes under diversifying selection ( n = 161) Clade‐specific accessory genes under selection were identified through presence/absence to a specific clade. In total, 5,239, 7,559 and 5,605 genes were found only in Clade 1, Clade 2 and Clade 3 respectively and present in less than 95% of the clade. Accessory genes under positive selection in each clade showed functions that are distinct to each clade. Of these, several are worthy of note; the Clade 1‐specific accessory genome included the gene InlA , which codes for internalin‐A and allows the invasion of mammal cells (Dhar et al., 2000 ), the Clade 2‐specific accessory genome included Cry toxins—key Bt insecticidal toxins—such as cry2Ab (Zheng et al., 2017 ), and the Clade 3‐specific accessory genome contained the gene binA , which produces a homologue to an insecticidal binary toxin component (Palma et al., 2014 ; Table S2 ). 3.3 GO analyses suggest clade‐specific selection acting on the core and accessory genomes of each Bc clade Binomial testing was used to measure the functional enrichment of biological processes (Ashburner et al., 2000 ; Gene Ontology Consortium, 2019 ) within clade‐specific core and accessory genomes (Mi et al., 2013 ) by comparison to the strict core genome. This methodology allowed ecological characterization of the clades and avoided a priori assumptions of relevance. Additionally, it avoids characterizing a clade by the possession of any one gene, as has often been the case in the Bc sl group (Bravo et al., 2007 ; Lechner et al., 1998 ). There was significant functional enrichment of biological processes amongst conserved and diverse clade‐specific core genes of all clades; conserved clade‐specific genes were often linked to translation (Figure 3a ). However, some enrichment was clade‐specific: Clade 3 contained a greater number of conserved genes linked to negative regulation of transcription and fewer conserved genes linked to biosynthesis and stimulus response than would be expected based on the strict core genome (Figure 3a ). The same was found to be the case for diverse clade‐specific genes; genes with uncharacterized functions were more common than expected within Clade 1 and less common than expected in Clades 2 and 3, but only Clade 2 showed unique functional enrichment, with more genes linked to antibiotic and antimicrobial resistance than expected. Functional enrichment of biological processes was robust when the criteria for considering genes under selection within a clade were relaxed to include ~10% of the clade‐specific core genomes as opposed to ~5% as described above. FIGURE 3 Significant enrichment of biological processes within clade‐specific conserved core genes and clade‐specific accessory genes across the major Bacillus cereus sensu lato ( Bc sl ) clades. Enrichment values were calculated using the Gene Ontology Enrichment analysis software available online, using a binomial test with Bonferroni correction. Only significant enrichments are shown Like the clade‐specific core genomes, there were disparities in functional enrichment between the clade‐specific accessory genomes. While some processes—such as antibiotic biosynthesis—were enriched in all clade‐specific accessory genomes, there were differences between the clades regarding the enrichment of other biological processes (Figure 3b ). Interestingly, biological processes enriched within clade‐specific accessory genomes were not the same as those enriched within that clade's specific core genome. For instance, Clade 3 accessory genes were more likely to be linked to motility and secondary metabolism, while its clade‐specific core genome was not. Clade 3 was also not significantly enriched for accessory genes linked to negative regulation of transcription, while its core genome was (Figure 3b ). 3.4 Genes under diversifying selection undergo more frequent HGT Two sets of clade‐specific core genes were suitable for CInd analysis; 42 genes with low allelic diversity and 24 genes with high allelic diversity were present in all 328 strains and therefore their gene phylogeny could be compared to the strict core genome phylogeny to check for inconsistencies that suggest HGT. CInds were calculated for clade‐specific core genes of high and low diversity, as well as for all genes in the strict core genome. The CInds of each gene set suggest that genes under diversifying selection undergo frequent HGT, while HGT is uncommon amongst conserved genes (Figure 4a,b ); the mean CInd of conserved clade‐specific core genes (0.46 ± 0.02) was significantly higher than the mean of the strict core genome (0.34 ± 0.003; Wilcoxon–Mann–Whitney test; U = 33722, p = 4.435e −11 ). In contrast, the mean CInd of diverse clade‐specific core genes (0.28 ± 0.018) was significantly lower than for the strict core genome (Wilcoxon–Mann–Whitney test; U = 7893, p =.00385; Figure 5 ). FIGURE 4 Consistency between phylogenies based on a single gene and a core genome. (a) Consistency between the conserved cold shock protein ( CspA ) gene phylogeny and the strict core genome ( n = 1,004) of the Bacillus cereus sensu lato ( Bc sl ) group. A given isolate in each tree is connected to the same isolate in the other tree by a line. (b) Consistency between the diverse flagellin ( hag ) gene phylogeny and the strict core genome ( n = 1,004) of the Bc sl group. A given isolate in each tree is connected to the same isolate in the other tree by a line FIGURE 5 Consistency index distribution amongst genes under purifying selection ( n = 42) and diversifying selection ( n = 24). Consistency indices were calculated for each gene using the phangorn package in r and a maximum‐likelihood phylogeny was created using alignments of the concatenated core genome. The average consistency index of each gene set was compared to that of the strict core genome ( n = 1,004). Normalized numbers of genes represent the number of genes with a given consistency index while controlling for the size of the data set 4 DISCUSSION This study aimed to explore ecological differentiation between closely related bacterial clades and the role of selection in driving and maintaining this distinctiveness. To accomplish this, we tested bacterial genomes from an economically important and well‐studied model group for signatures of selection. The Bc group contains many different strains, all thought to be well‐adapted to exploit protein‐rich food such as cadavers (Manktelow et al., 2021 ; Rasigade et al., 2018 ). Despite high levels of genetic similarity, the clade structure of the group is distinct and robust to multiple phylogenetic methods. Clades have been associated with differences in fitness and virulence gene complement, as well as with distinct biogeographical and thermal niches (Cardazzo et al., 2008 ; Drewnowska et al., 2020 ; Guinebretière et al., 2008 , 2010 ; Manktelow et al., 2021 ; Méric, Mageiros, Pascoe, et al., 2018 ; Zheng et al., 2017 ); here, we show that clade‐specific core and accessory genomes bear signatures consistent with niche‐specific selection. We identified genes under putative purifying and diversifying selection within clade‐specific core genomes by comparison to diversity in the strict core genome. As mentioned, identifying genes undergoing selection presents computational and data sampling challenges (Buckee et al., 2008 ; Zhang et al., 2005 ); additionally, selection must be distinguished from other factors affecting allelic diversity (Chen et al., 2006 ; Fay & Wu, 2003 ; Zhang et al., 2005 ). This was achieved by using allelic diversity and comparison to the average genomic diversity to identify outliers under strong selection (Méric, Mageiros, Pensar, et al., 2018 ). Genes with very low or very high allelic diversity compared to the average are likely to be under strong purifying or diversifying selection (Cohan, 2016 ; Dugatkin et al., 2005 ; Shea et al., 2011 ). Amongst gene sets with non‐normally distributed allelic diversity values, using percentile values to encapsulate the most extreme 5% of the data would be suitable; however, due to a normal distribution of the data in this study, mean and SD filtering of allelic diversity provided a way to quickly identify genes under strong selection. It should be noted that low allelic diversity may occur due to purifying selection or due to directional selection combined with HGT (i.e., gene‐specific sweeps; Cohan, 2016 ); this may explain the low numbers of conserved genes within Clade 2. However, because the majority of conserved genes also showed low levels of HGT (Figure 5 ), we feel confident that the majority of conserved genes are the result of purifying selection; an in‐depth examination could identify genes from among these sets that are more likely to have undergone gene‐specific sweeps. Analysis of clade‐specific conserved core genes suggested that core genes under purifying selection differed significantly between clades and supported previous hypotheses about the ecological distinctiveness of major Bc clades (Figure 3a ). For example, consider our analysis of Clade 3, now recognized as Bm (Carroll et al., 2020 ). Here, the analysis of clade‐specific core genes identified the cold‐shock protein gene cspA , a unique sequence signature of which was used to originally classify the psychrotolerant Bm species (Lechner et al., 1998 ). Furthermore, Clade 3 possessed many conserved genes linked to ribosome assembly and negative regulation of transcription, and few linked to metabolism, biosynthetic processes and external stimuli responses (Ermolenko & Makhatadze, 2002 ). These features are characteristic of adaptation to low temperatures, where metabolic functions are downregulated in response to cold (Barria et al., 2013 ; López‐Maury et al., 2008 ; Tribelli & López, 2018 ). This supports other studies indicating that strains within Clade 3 are psychrotolerant specialists (Lechner et al., 1998 ; Liu et al., 2018 ; Manktelow et al., 2021 ) and demonstrates how the methodology used here can identify important genes with specific variants within ecologically distinct groups. Different patterns of enrichment amongst conserved clade‐specific core genes also suggest that the clades are ecologically distinct, and purifying selection may maintain new species by purging novel variation caused by mutation and HGT (Cohan, 2016 , 2017 ). We found evidence that diversifying selection within clade‐specific core genomes acts on different genes depending on the clade. While the hag flagellin gene was extremely diverse across all three clades, only genes of high allelic diversity within Clade 2 were enriched for functions linked to flagellum‐dependent motility. Flagellin is a common receptor for bacteriophages, and because variations in flagellin structure may prevent phage infection, this is a trait likely to be under diversifying selection (Nobrega et al., 2018 ). Clade 2 also has the largest proportion of isolates encoding insecticidal toxins and carries a greater number of insecticidal toxins than other clades (Méric, Mageiros, Pascoe, et al., 2018 ; Zheng et al., 2017 ). This supports the hypothesis that this clade is dominated by specialist insect pathogens (Raymond & Bonsall, 2013 ; Raymond & Federici, 2017 ; Raymond et al., 2010 ) and provides further evidence for the ecological distinctiveness of the clades. Flagellar motility may also be important during the early stages of insect infection (Mazzantini et al., 2016 ); Bt mutants with reduced flagellar motility have reduced virulence when infecting larvae (Zhang et al., 1993 ). Diverse Bt genes were also more likely to be linked to antimicrobial resistance (Table S2 ). Antimicrobial resistance mechanisms are common in Bc strains (Abriouel et al., 2011 ; Bernhard et al., 1978 ) and are often under diversifying selection, which can result in the emergence and maintenance of allelic diversity for that trait (Levin, 1988 ; McNally et al., 2019 ). Diversifying selection on antibiotic resistance may be prevalent amongst Clade 2 strains because competition to enter insect cadavers first is intense (Garbutt et al., 2011 ; Van Leeuwen et al., 2015 ). Therefore, overcoming host defences and securing the first infection of a host may provide an advantage in pathogenic bacteria that is not seen in necrotrophic bacteria. One of the aims of this study was to assess the importance of selection in maintaining bacterial species. Alternative drift‐based models of bacterial speciation assume that genetic differences between taxa are self‐reinforcing (Fraser et al., 2007 ). HGT can erode differences between neutrally diverging lineages and greater genetic distance leads to reduced HGT via a range of mechanisms (Fraser et al., 2007 ) There is evidence for these kinds of forces operating in the Bc sl group; for instance, HGT predominantly occurs within clades (Didelot et al., 2009 ). Nevertheless, one notable result of this study was the variation in inferred levels of HGT between loci under different forms of selection. Here, we used CInds to infer the prevalence of HGT. High CInds amongst conserved genes—such as the cspA gene—indicate low levels of HGT (Méric, Mageiros, Pensar, et al., 2018 ); in contrast, low CInds in diverse genes such as the hag gene imply high levels of HGT (Figures 4 and 5 ). At a fundamental level, all chromosomal genes undergo HGT at similar rates (Gogarten et al., 2002 ). However, the subsequent fate of horizontally transferred alleles differs depending on gene and gene function; this may be due to variation in selection strength and type between genes (Kivisaar, 2019 ; Nakamura et al., 2004 ). Our results indicate that the effects of HGT are strongly modulated by selection in the Bc sl group. When novel allelic diversity is favoured under diversifying selection, HGT can supply that diversity. However, purifying selection can also purge clade‐specific allelic variants that incur strong selective disadvantages in the "wrong" genetic background (Vos et al., 2015 ). Moderate levels of HGT therefore do not impede speciation, as seen in other species (Melendrez et al., 2016 ). Background levels of HGT are important, but selection can clearly act to promote clade identity and genetic coherence in the face of HGT. While unlikely to be an issue in Bc due to intermediate levels of homologous recombination (Patiño‐Navarrete & Sanchis, 2017 ), CInds are probably most effective at identifying patterns of HGT when levels are low or intermediate; at high rates of HGT genes may be spread sufficiently widely so that genes received via HGT cannot be distinguished from genes received via linear descent (Andam & Gogarten, 2011 ; Sanderson & Donoghue, 1989 ). Spotting inconsistencies may also be difficult in conserved genes due to the small number of differences between genes. However, given levels of HGT are roughly intermediate for all gene sets (~0.5) and that conserved genes with small differences are sufficiently different to be used for reconstructing phylogenies (Saitou & Imanishi, 1989 ), these would seem to be minor concerns. The role of accessory genomes in ecological specialization is widely accepted (Brockhurst et al., 2019 ; Cobo‐Simón & Tamames, 2017 ); Bt , which carries key virulence factors primarily on large plasmids, is a well‐known example (Zheng et al., 2017 ). As with the core genome analysis, accessory genes unique to each clade were significantly enriched for specific biological processes. Furthermore, the processes enriched within a clade‐specific core genome often differed from the processes enriched within the specific accessory genome of the same clade. For instance, the Clade 3 accessory genome was enriched for genes linked to motility and secondary metabolic processes, while its core genome was not. The utility of presence/absence for identifying accessory genes under selection is still debated, as strains accumulate a mix of deleterious, beneficial and neutral genes and the frequency of beneficial accessory genes is unclear (Vos & Eyre‐Walker, 2017 ). Despite this, presence/absence of specific accessory genes has been found to be biologically meaningful in other studies (Cohen et al., 2013 ; Méric, Mageiros, Pascoe, et al., 2018 ; Vasquez‐Rifo et al., 2019 ). With this considered, our results would suggest that both the core and accessory genome determine a strain's ecology. While these results indicate the importance of chromosomal core and accessory genes to strain ecology, they should be taken with caution for three reasons. First, enrichment within the accessory genome may not be representative of all strains within a clade; the majority of genes in bacterial pan‐genomes are either common ("core" or nearly core) or extremely rare (accessory; Haegeman & Weitz, 2012 ). Because the Bc sl clades consist of isolates assigned to different species or ecotypes—for instance, both Clades 1 and 2 contain strains identified as Bt (Méric, Mageiros, Pascoe, et al., 2018 )—the enrichment of certain biological processes within a clade's accessory genome may be due to high numbers of rare genes that are possessed by a minority of the clade in question. Second, the different functional enrichment in clade‐specific core and accessory genomes may reflect differences in selection over time as opposed to differences in function; within one species of bacterium, accessory gene content change occurs at faster rates but is retained less readily than amino acid substitution in the core genome (Wielgoss et al., 2016 ), implying that accessory genomes reflect current selection and core genomes reflect past selection. Third, this study did not attempt to incorporate plasmid sequences into the analysis. While it was not possible to differentiate between chromosomal and plasmid DNA in all isolates, we did not explicitly analyse plasmid sequences in this study. While some plasmids are stably associated with Bc lineages and therefore considered part of the "core genome" (Méric, Mageiros, Pascoe, et al., 2018 ; Zheng et al., 2017 ), many plasmids are highly mobile and carry genes encoding several key virulence traits (Patiño‐Navarrete & Sanchis, 2017 ; Schnepf et al., 1998 ). While analysis of the selection pressures that formed the Bc clades will benefit by excluding plasmid sequences (by reducing the confounding effect that highly mobile plasmids may have on analysis), future researchers may wish to incorporate these important parts of the Bc sl pan‐genome. Therefore, future iterations of this methodology may benefit from two modifications: splitting analysis of the accessory genome into genes of intermediate and low frequency within a clade (Inglin et al., 2018 ) and the incorporation of plasmid sequence data. It is interesting that Clade 1 does not appear to possess any significant clade‐specific enrichments, aside from deficiencies in certain biosynthetic processes (Figure 3b ) and the possession of the internalin‐A protein gene inlA (allowing for epithelial cell invasion) in its clade‐specific core genome (Dhar et al., 2000 ). We hypothesized that the anthracis clade would consist of necromenic (cadaver‐associated) bacteria that may specialize on vertebrates (Manktelow et al., 2021 ) although Ba itself is a clonal expansion and represents only a small part of the diversity in this group. Clade 1 includes at least six currently recognized species, though one proposed revision suggests lumping all these groups into a single taxon based on a 92.5% average nucleotide identify (ANI; Carroll et al., 2020 ). Regardless of current taxonomic disputes, the clade splits into two groups separated by a 94% ANI. These two branches of Clade 1 were previously described as PanC Groups II and III, corresponding to Bacillus paranthracis and allies and Bacillus albus / wiedmannii and allies respectively (Guinebretière et al., 2008 , 2010 ). There is evidence for differences in phenotype and biogeography between these groups (Drewnowska et al., 2020 ; Guinebretière et al., 2008 , 2010 ). "Lumping" these groups into a single clade may be obscuring ecological distinctiveness in Clade 1. While useful for identifying the selection pressures that formed the Bc clades and that are currently creating diversity within each clade, our results should not be taken to mean that the clades are ecologically monolithic. Repeating this analysis using the seven‐clade phylogeny of Guinebretière et al. ( 2008 ) and with greater representation in these subgroups may reveal ecological distinctions that were not seen in this study. This possibility suggests how this selection‐informed analysis may be used for refining taxonomic decision‐making. Methods based on raw genetic differences, such as ANI, appear highly objective; however, decisions still need to be made on how to apply rules and what level of differentiation is appropriate for describing species in a particular group (Carroll et al., 2020 ; Vos, 2011 ). There are advantages in describing species as units with real ecological and phenotypic distinctiveness; if groups recognized by ANI‐based decisions also show coherent patterns of selection, it provides another means of assessing whether a species definition is of practical value. Another pragmatic application of genome‐wide analysis of conserved genes is its value in identifying key ecological traits and single loci that can be used for species‐level identification; one example from this study is the wealth of psychrotolerance traits found in Clade 3, exemplified by the conserved cold‐shock gene cspA . In conclusion, this study showed that functional enrichment in both core and accessory genes is heavily dependent on clade in the Bc bacterial group. Key ecological traits associated with Bacillus species—such as antimicrobial and insecticidal activity in Bt strains and psychrotolerance in Bm strains—were among those enriched in specific clades, supporting the hypothesis that clades within the group formed due to different selection pressures and have distinct ecologies. The core and accessory genomes of each clade appear to experience selection on different traits, highlighting the importance of considering both when determining clade ecology. High levels of HGT amongst diversifying core genes suggest that HGT plays a key role in promoting diversification within the Bc sl group. Lastly, this analysis identified genes, such as the cspA gene in Clade 3, that can be used to identify strains to the clade level and to infer their ecological niche, allowing easier determination of strains' potential to harm humans and to act as biopesticides, with the commensurate benefits to agricultural and medical practices. AUTHOR CONTRIBUTIONS H.W. designed the study, performed research, analysed data and was the primary writer for the manuscript. S.K.S. and B.R. helped design the study, and S.K.S. contributed the use of pirate and the BIGSdb database. S.K.S., B.R. and M.V. all contributed to the writing of the manuscript. CONFLICT OF INTEREST The authors declare no competing interests. OPEN RESEARCH BADGES This article has earned an Open Data Badge for making publicly available the digitally‐shareable data necessary to reproduce the reported results. The data is available at https://doi.org/10.24378/exe.3992 , https://hdl.handle.net/10871/129565 . OPEN RESEARCH BADGES This article has earned an Open Data Badge for making publicly available the digitally‐shareable data necessary to reproduce the reported results. The data is available at https://doi.org/10.24378/exe.3992 , https://hdl.handle.net/10871/129565 . Supporting information Table S1‐S3 Click here for additional data file. DATA AVAILABILITY STATEMENT Genetic data can be accessed from public databases by referring to the strain accession numbers in Table S1 . Sample metadata are available from the Multispecies BIGSdb (Jolley & Maiden, 2010 ; https://sheppardlab.com/resources/ ) and are available in Table S1 . Metadata include Multispecies BIGSdb ID, the clade the strain was assigned to in this study, isolate identifier, aliases, pathotype, species source, lineage, serovar, clinical isolate, sequence length (bp) and accession number. The pirate Pipeline is available through GitHub ( https://github.com/SionBayliss/PIRATE ) . Details of the clade‐specific core genes that showed extremely high and low allelic diversity can be found in Table S2 a,b. Details of clade‐specific accessory genes can also be found in Table S2 c. UniprotKB codes are available for each gene from the UniProt Knowledgebase (UniProtKB; https://www.uniprot.org/ ) and are listed next to their respective gene in Table S2 . Metadata include pirate ID number, the clades in which a gene was conserved/diverse/accessory, consensus gene name, consensus gene product and UniProtKB code. Raw output from the pirate pipeline (both the excel summary and the identified "gene family".FASTA files), the maximum‐likelihood tree file and iq‐tree command lines, output from the Gene Ontology analysis tool, and the raw output from analysis of consistency indices will be made available publicly through Open Research Exeter (ORE; https://ore.exeter.ac.uk/repository/handle/10036/10890 ) upon acceptance and publication. The iq‐tree and R scripts used to generate relevant output (maximum‐likelihood phylogeny Figure 1a , Gene Ontology graph Figure 3 and consistency index graph Figure 5 ) will also be stored here.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1430378/

Characterization of Dominant-Negative Forms of Anthrax Protective Antigen

Certain mutations within the protective antigen (PA) moiety of anthrax toxin endow the protein with a dominant-negative (DN) phenotype, converting it into a potent antitoxin. Proteolytically activated PA oligomerizes to form ring-shaped heptameric complexes that insert into the membrane of an acidic intracellular compartment and promote translocation of bound edema factor and/or lethal factor to the cytosol. DN forms of PA co-oligomerize with the wild-type protein and block the translocation process. We prepared and characterized 4 DN forms: a single, a double, a triple, and a quadruple mutant. The mutants were made by site-directed mutation of the cloned form of PA in Escherichia coli and tested by various assays conducted on CHO cells or in solution. All 4 mutant PAs were competent for heptamerization and ligand binding but were defective in the pH-dependent functions: pore formation, ability to convert to the SDS-resistant heptamer, and ability to translocate bound ligand. The single mutant (F427K) showed less attenuation than the others in the pH-dependent functions and lower DN activity in a CHO cell assay. The quadruple (K397D + D425K + F427A + 2β2-2β3) deletion showed the most potent DN activity at low concentrations but also gave indications of low stability in a urea-mediated unfolding assay. The double mutant (K397D + D425K) and the triple (K397D + D425K + F427A) showed strong DN activity and slight reduction in stability relative to the wild-type protein. The properties of the double and the triple mutants make these forms worthy of testing in vivo as a new type of antitoxic agent for treatment of anthrax.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7261280/

Hepatitis B virus X protein promotes liver cell pyroptosis under oxidative stress through NLRP3 inflammasome activation

Objective Hepatitis B virus X protein (HBx) is a pivotal factor for HBV-induced hepatitis. Herein, we sought to investigate HBx-mediated NLR pyrin domain containing 3 (NLRP3) inflammasome activation and pyroptosis under oxidative stress. Methods The effect of HBx on the NLRP3 inflammasome was analyzed by enzyme-linked immunosorbent assays, quantitative reverse transcription-polymerase chain reaction, western blotting, and immunofluorescence in hepatic HL7702 cells. Pyroptosis was evaluated by western blotting, lactate dehydrogenase release, propidium iodide staining, and transmission electron microscopy. NLRP3 expression in the inflammasome from liver tissues was assessed by immunohistochemistry. Results In hydrogen peroxide (H 2 O 2 )-stimulated HL7702 cells, HBx triggered the release of pro-inflammatory mediators apoptosis-associated speck-like protein containing a CARD (ASC), interleukin (IL)-1β, IL-18, and high-mobility group box 1 (HMGB1); activated NLRP3; and initiated pro-inflammatory cell death (pyroptosis). HBx localized to the mitochondria, where it induced mitochondrial damage and production of mitochondrial reactive oxygen species (mitoROS). Treatment of HL7702 cells with a mitoROS scavenger attenuated HBx-induced NLRP3 activation and pyroptosis. Expression levels of NLRP3, ASC, and IL-1β in liver tissues from patients were positively correlated with HBV DNA concentration. Conclusions The NLRP3 inflammasome was activated by elevated mitoROS levels and mediated HBx-induced liver inflammation and hepatocellular pyroptosis under H 2 O 2 -stress conditions. Electronic supplementary material The online version of this article (10.1007/s00011-020-01351-z) contains supplementary material, which is available to authorized users. Objective Hepatitis B virus X protein (HBx) is a pivotal factor for HBV-induced hepatitis. Herein, we sought to investigate HBx-mediated NLR pyrin domain containing 3 (NLRP3) inflammasome activation and pyroptosis under oxidative stress. Methods The effect of HBx on the NLRP3 inflammasome was analyzed by enzyme-linked immunosorbent assays, quantitative reverse transcription-polymerase chain reaction, western blotting, and immunofluorescence in hepatic HL7702 cells. Pyroptosis was evaluated by western blotting, lactate dehydrogenase release, propidium iodide staining, and transmission electron microscopy. NLRP3 expression in the inflammasome from liver tissues was assessed by immunohistochemistry. Results In hydrogen peroxide (H 2 O 2 )-stimulated HL7702 cells, HBx triggered the release of pro-inflammatory mediators apoptosis-associated speck-like protein containing a CARD (ASC), interleukin (IL)-1β, IL-18, and high-mobility group box 1 (HMGB1); activated NLRP3; and initiated pro-inflammatory cell death (pyroptosis). HBx localized to the mitochondria, where it induced mitochondrial damage and production of mitochondrial reactive oxygen species (mitoROS). Treatment of HL7702 cells with a mitoROS scavenger attenuated HBx-induced NLRP3 activation and pyroptosis. Expression levels of NLRP3, ASC, and IL-1β in liver tissues from patients were positively correlated with HBV DNA concentration. Conclusions The NLRP3 inflammasome was activated by elevated mitoROS levels and mediated HBx-induced liver inflammation and hepatocellular pyroptosis under H 2 O 2 -stress conditions. Electronic supplementary material The online version of this article (10.1007/s00011-020-01351-z) contains supplementary material, which is available to authorized users. Introduction Hepatitis B virus (HBV) is an oncogenic virus presently responsible for approximately 350 million cases of chronic infections worldwide [ 1 ]. HBV's persistence is reportedly associated with an increased risk of cirrhosis and hepatocellular carcinoma (HCC) [ 2 ]. Specifically, the hepatitis viral X protein (HBx), encoded by the HBV X gene, is implicated in HBV-related hepatitis, cirrhosis, and the initiation of HCC [ 3 , 4 ]. As a multifunctional oncoprotein, HBx localizes in the cytoplasm, nucleus, and mitochondria, where it affects signal transduction, transcription, and mitochondrial function [ 5 , 6 ]. NLR pyrin domain containing 3 (NLRP3) is a cytoplasmic pattern recognition receptor that is widely distributed in hepatic parenchymal cells and non-substantial cells [ 7 – 9 ]. The NLRP3 inflammasome, which consists of NLRP3, inflammasome adaptor protein apoptosis-associated speck-like protein containing CARD (ASC), and pro-caspase-1, requires two signals to be activated. The initiation signal is mediated by nuclear factor (NF)-κB, which upregulates expression of the inflammasome-related proteins; while, the second signal is mediated by endogenous or exogenous hazard signals [ 10 – 12 ]. The activation of NLRP3 promotes the production of active caspase-1, which contains two heterodimers of p20 and p10. This activation then induces the maturation and secretion of inflammatory cytokines, namely interleukin (IL)-1β, IL-18, and high-mobility group box 1 protein (HMGB1), as well as the induction of inflammatory necrosis (pyroptosis) [ 13 – 16 ]. Increasing evidence indicates that the inflammasome is involved in various liver diseases, including liver injury, hepatitis, liver fibrosis, and cirrhosis; however, whether the NLRP3 inflammasome participates in HBx-induced hepatitis remains unclear. The mitochondrial ROS (mitoROS) model is a widely accepted mechanistic explanation for NLRP3 activation [ 11 , 17 ]. Physiological levels of ROS maintain normal cell signaling and homeostasis; however, abnormally high levels of ROS activate several signaling molecules, including NF-κB, mitogen-activated protein kinases (MAPKs), protein kinase B (Akt), and signal transducer and activator of transcription 3 (STAT3), resulting in cellular inflammation and apoptosis [ 18 ]. Given that the mitochondrial oxidative respiratory chain serves as the primary source of intracellular ROS and that the liver is a mitochondria-rich organ, it is plausible that the mitoROS model may contribute significantly to the development and progression of liver diseases. Further, our previous studies showed that HBx interacts with cytochrome c oxidase subunit 3 (COXIII), a protein related to mitochondrial respiratory chains, and causes an increase in mitoROS levels, resulting in decreased membrane potential, ATP synthesis disorder, and cytosolic calcium overload, ultimately causing mitochondrial dysfunction [ 19 , 20 ]. In the current work, we sought to investigate whether HBx promoted mitoROS-mediated liver inflammatory injury via activation of the NLRP3 inflammasome. We also examined the role of HBx in hepatocyte pyroptosis under oxidative stress. Materials and methods Patient tissue and serum samples Archived paraffin-embedded HCC tissues and matched non-tumor tissues collected from 51 patients from 2014 to 2017 at Union Hospital of Fujian Medical University, China, were randomly selected. Written informed consent was obtained before surgical resection. Additionally, 84 serum samples, including 23 HBV-negative and 61 HBV-positive samples from patients collected between 2017 and 2018, were evaluated (tissue and serum samples were from different subjects). The inclusion criterion was patients who were first diagnosed with HBV infection and did not receive antiviral therapy. The exclusion criterion were patients with other hepatitis virus infections, nonviral hepatitis (alcoholic or non-alcoholic hepatitis, drug-induced hepatitis, etc.), and autoimmune hepatitis. All clinical samples were collected according to protocols approved by the Medical Faculty of Fujian Medical University Ethics Committee (Approval number 2019Y001). Cell culture and plasmids Normal human liver HL7702 cells (Shanghai Cell Biology Institute of Chinese Academy of Science, Shanghai, China) were cultured in Roswell Park Memorial Institute (RPMI) medium supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, USA). For induction of oxidative stress, cells were treated with 100-μM hydrogen peroxide (H 2 O 2 ) for 12 h after 36-h plasmid transfection. The other groups that were transfected with plasmids for 48 h, however, did not undergo induced oxidative stress. The pHBx plasmid expressing HBx and pcDNA3.1 (pNC) was maintained in our laboratory. The pGEM-4Z (pGEM) plasmid was purchased from Promega (#P2161; Madison, WI, USA). The pHBV plasmid expressing full-length wild-type HBV genomic complementary DNA (cDNA) and pHBV-HBx-expressing cDNA of an HBx-deficient HBV mutant were gifted from Professor M.J. Bouchard (Drexel University, Philadelphia, PA, USA) [ 21 , 22 ]. Plasmid transfections were performed using Lipofectamine 3000 (#L3000-008; Invitrogen, Carlsbad, CA, USA). N -acetyl- l -cysteine (NAC; #A7250; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in distilled deionized water (ddH 2 O) and added to cells at a final concentration of 5 nM prior to H 2 O 2 treatment for 60 min. Mito-TEMPO (#SML0737; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in ddH 2 O and added to cells at a final concentration of 50 μM before H 2 O 2 treatment for 60 min. Cell viability Cell viability was determined using a Cell Counting Kit-8 (CCK-8; #CK04-11; Dojindo, Kumamoto, Japan) as described by the manufacturer. Briefly, HL7702 cells were plated into 96-well plates and grown to 70–80% confluence. The cells were then treated with different H 2 O 2 concentrations (0, 25, 50, 75, 100, 125, 150, 175, and 200 μM). After 12 h of H 2 O 2 treatment, 10-µl CCK-8 reagent was added to each well and incubated for 1 h at 37 °C. The absorbance of each well was read at 450 nm using a spectrophotometric reader. Enzyme-linked immunosorbent assay (ELISA) The levels of ASC, IL-1β, IL-18, and HMGB1 in cellular medium were detected by double-antibody sandwich ELISA using Quantikine ELISA kits (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions. The optical density (OD) at 450 nm for each well was determined using a plate reader (Bio-Rad, Hercules, CA, USA). Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) Total RNA was extracted using TRIzol Reagent (#15596026; Invitrogen, Carlsbad, CA, USA). Following reverse transcription, the quantitative PCR reactions were performed using SYBR Green PCR Master Mix (#17747200; Roche, Basel, Switzerland) according to the manufacturer's protocol. The primer sequences are provided in Table 1 . The data from five independent experiments were analyzed using the 2 −∆∆ C t method values. Table 1 Primers used for RT-PCR Gene Primer sequences NLRP3 Forward 5′-GGTGGAGTGTCGGAGAAG-3′ Reverse 5′-CTGTCATTGTCCTGGTGTCT-3′ ASC Forward 5′-GCTGCTGGATGCTCTGTA-3′ Reverse 5′-AGGCTGGTGTGAAACTGAA-3′ Caspase-1 Forward 5′-GAGCAGCCAGATGGTAGAG-3′ Reverse 5′-CCCACAGACATTCATACAGTTTC-3′ IL-1β Forward 5′-TCACCTCTCCTACTCACT-3′ Reverse 5′-CGGTTGCTCATCAGAATG-3′ IL-18 Forward 5′-GACCTTCCAGATCGCTTCCTC-3′ Reverse 5′-GATGCAATTGTCTTCTACTGGTTC-3′ HMGB1 Forward 5′-TCAAAGGAGAACATCCTGGCCTGT-3′ Reverse 5′-CTGCTTGTCATCTGCAGCAGTGTT-3′ GSDMD Forward 5′-AGACCATCTCCAAGGAACTG-3′ Reverse 5′-GGACAACACCAGGCACTC-3′ GAPDH Forward 5′-GAAGGTGAAGGTCGGAGTC-3′ Reverse 5′-GAAGATGGTGATGGGATTTC-3′ NLRP3 NLR pyrin domain containing 3, ASC apoptosis-associated speck-like protein containing a caspase recruitment domain (CARD), HMGB1 , high-mobility group box 1, GSDMD gasdermin D, GAPDH glyceraldehyde-3-phosphate dehydrogenase Western blotting Protein expression was evaluated by western blotting, as described previously [ 23 ]. Antibodies specific for NF-κB p65 (#8242; Cell Signaling Technology, Danvers, MA, USA), NLRP3 (#sc-66846; Santa Cruz Biotechnology, Santa Cruz, CA, USA), pro-caspase-1 (#IMG-5028; Novus Biologicals, Littleton, CO, USA), caspase-1 (#sc-56036; Santa Cruz Biotechnology, Santa Cruz, CA, USA), IL-1β (#ab9722; Abcam, Cambridge, MA, USA), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (#2118; Cell Signaling Technology, Danvers, MA, USA), ASC (#ab180799; Abcam, Cambridge, MA, USA), gasdermin D (GSDMD) (#ab210070; Abcam, Cambridge, MA, USA), HBx (#ab39716; Abcam, Cambridge, MA, USA), hepatitis B virus core protein (HBc) (#MAB16989; Merck Millipore, Billerica, MA, USA), and cytochrome c oxidase subunit 4 (COXIV) (#GB11250; Wuhan Goodbio Technology, Wuhan, China) were used. Immunofluorescence HL7702 cells were plated into 24-well plates and grown to 50% confluence. Following plasmid transfection and H 2 O 2 induction, immunofluorescence was performed. Briefly, cells were fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.25% Triton X-100 for 15 min. Cells were then blocked with 10% donkey serum for 60 min, incubated with primary antibodies overnight at 4 °C, and immunostained with secondary antibodies for 60 min at 23–25 °C. Primary antibodies specific for ASC (#ab180799; Abcam, Cambridge, MA, USA) and HBx (#ab235; Abcam, Cambridge, MA, USA) were used. The secondary donkey anti-mouse IgG (H + L), Alexa Fluor 488 (#A-21202; Invitrogen, Carlsbad, CA, USA), donkey anti-rabbit IgG (H + L), and Alexa Fluor 555 (#A-31572; Invitrogen, Carlsbad, CA, USA) antibodies were used. Finally, cells were observed using fluorescence confocal microscopy (Zeiss FM780, Jena, Germany). Pictures were taken under 63 × oil microscope and excitation light wavelengths of 405 nm, 488 nm, and 594 nm were selected for layer scanning with a scanning interval of 0.35 um/layer. The fluorescence excitation displayed blue, green, and red, respectively. Mitochondria isolation Mitochondrial proteins were isolated according to the introduction of cellular mitochondria isolation kit (#89874; Thermo Fisher Scientific, Waltham, MA, USA). Briefly, cells with 100-μM H 2 O 2 treating for 12 h after 36-h plasmid transfection were firstly harvested with trypsin and homogenized with Dounce Tissue Grinder on ice. Then, the post-nuclear fractions were obtained by centrifuging at 700× g for 10 min at 4 °C. To obtain a more purified fraction of mitochondria, the post-nuclear fractions were centrifuged at 3000× g for 15 min. Finally, the mitochondria pellet was dissociated with lysis buffer (#C3601-4; Beyotime Biotechnology, Shanghai, China) and then subjected to western blotting analysis. Lactate dehydrogenase (LDH) release assay HL7702 cells were cultured in 96-well plates at a density of 5000 cells/well. After transfection with plasmids and H 2 O 2 induction, the culture medium was collected. The release of lactate dehydrogenase (LDH) was determined using an LDH Cytotoxicity Detection Kit (#04744926001; Roche, Basel, Switzerland) according to the manufacturer's protocol. The OD at 490 nm for each sample was determined using a microplate reader (Bio-Rad, Hercules, CA, USA). Propidium iodide (PI) staining Membrane pore formation was detected by PI and Hoechst 33342 staining (#C1056; Beyotime Biotechnology, Shanghai, China). HL7702 cells were seeded into 6-well plates, transfected with plasmids, and exposed to H 2 O 2 . After treatment, the cells were washed with phosphate-buffered saline (PBS) and stained with Hoechst 33342 (5 µl) and PI (5 µl) for 20 min at 37 °C. The cells were washed with PBS three times and observed using a fluorescence microscope (TE2000-U, NIKON, Tokyo, Japan). The percentage of PI-positive cells was determined in three random microscopic fields. Transmission electron microscopy (TEM) Cell membrane integrity and mitochondrial morphological changes were determined using TEM. Cells were collected and fixed overnight at 4 °C in 0.1-M sodium cacodylate solution (pH 7.3) containing 2% fresh glutaraldehyde. The fixed cells were treated with 1% osmium tetroxide in 0.1-M cacodylate solution at 4 °C for 1 h and then stained with 1% uranyl acetate, dehydrated with ethanol, and embedded in epoxy resin. Finally, 90-nm ultrathin sections were cut and post-stained with uranyl acetate and bismuth subnitrate. The samples were visualized by TEM (TEC-NAI 1200, FEI Italia, Milan, Italy). Measurement of mitochondrial ROS levels Mitochondrial levels of ROS were determined using MitoSOX™ Red (#M36008; Molecular Probes, Eugene, OR, USA) according to the manufacture's protocol. Cells in 6-well plates were transfected and treated with or without 100 μM H 2 O 2 . After treatment, the culture medium was removed, and the trypsinized cells were treated with 5 µM MitoSOX™ Red reagent at 37 °C for 10 min. Cells were then washed with warm PBS thrice. Fluorescence intensity was determined using BD Accuri™ C6 Flow Cytometer (BD Biosciences, San Jose CA, USA). The FlowJo software v7.6 was used to measure the mean fluorescence intensity. Immunohistochemistry Immunohistochemistry was performed using Polink-2 plus Polymer HRP Detection System (#PV9001 and PV9003; Zhongshan Golden Bridge Biotechnology, Beijing, China) according to the manufacture's protocol. Primary antibodies specific for NLRP3 (#ab4207; Abcam, Cambridge, MA, USA), ASC (#ab180799; Abcam, Cambridge, MA, USA), and IL-1β (#ab9722; Abcam, Cambridge, MA, USA) were incubated with sections overnight at 4 °C. Representative images of 10 random fields per section were obtained microscopically (DMil, Leica, Wetzlar, Germany). The mean integrated optical density (IOD) was determined using Image-Pro Plus software v6.0 (Media Cybernetics, Rockville, MD, USA) to evaluate the expression levels of proteins. Statistical analysis Data were assessed to evaluate the normal distribution and homogeneity of variance by the Shapiro–Wilk normality test and Levene's test, respectively. Comparisons between two groups following a normal distribution and equal variance were performed using a two-tailed Student's t test; whereas, one-way analysis of variance (ANOVA) with Tukey post-test was used for multiple comparisons. For a non-normal distribution or uneven variance, the Wilcoxon rank-sum and Kruskal–Wallis test with Dunnett's post hoc test were performed to compare two groups or multiple groups. Data are shown as mean values ± standard deviation (SD). P  10 5 IU/ml) had the highest cytoplasmic staining of NLRP3 (Fig. 6 a–d). Correlation analysis further demonstrated that the expression level of these proteins positively correlated with HBV DNA titers (Fig. 6 e–g). Fig. 6 Association between the expression levels of NLRP3 inflammasome components and hepatitis B virus (HBV) DNA load. Analysis by immunohistochemistry of expression levels of NLRP3, ASC, and IL-1β in tumor-adjacent tissues of HCC patients with or without HBV infection. Representative images are shown, and the mean integrated optical density (IOD) of protein expression was statistically analyzed, n = 51 ( a – d ). Pearson correlation analysis of NLRP3, ASC, and IL-1β expression and HBV DNA copy number in a cohort of 51 patients, n = 51 ( e – g ). Serum ASC in patients with varied HBV DNA load was detected by ELISA, n = 84 ( h ). The height of the histogram represents the DNA copy number counts. Data are shown as mean ± SD. * P  10 5 IU/ml) had the highest cytoplasmic staining of NLRP3 (Fig. 6 a–d). Correlation analysis further demonstrated that the expression level of these proteins positively correlated with HBV DNA titers (Fig. 6 e–g). Fig. 6 Association between the expression levels of NLRP3 inflammasome components and hepatitis B virus (HBV) DNA load. Analysis by immunohistochemistry of expression levels of NLRP3, ASC, and IL-1β in tumor-adjacent tissues of HCC patients with or without HBV infection. Representative images are shown, and the mean integrated optical density (IOD) of protein expression was statistically analyzed, n = 51 ( a – d ). Pearson correlation analysis of NLRP3, ASC, and IL-1β expression and HBV DNA copy number in a cohort of 51 patients, n = 51 ( e – g ). Serum ASC in patients with varied HBV DNA load was detected by ELISA, n = 84 ( h ). The height of the histogram represents the DNA copy number counts. Data are shown as mean ± SD. * P < 0.05 and ** P < 0.01 Next, given that no previous research has detected the expression of serum ASC in HBV patients, we evaluated ASC concentrations in 84 serum samples from patients with or without HBV infection. In accordance with the result of ASC content in the cell culture medium (Fig. 2 a–b) and cytoplasmic staining of ASC in tissues (Fig. 5 a), HBV-positive patients exhibited higher ASC concentrations in their sera compared to those of HBV-negative patients (Fig. 6 h). These results strongly suggest that the NLRP3 inflammasome pathway is involved in HBV-related hepatitis. Discussion Inflammation is the hallmark of acute and chronic liver diseases. Persistent chronic inflammation can lead to liver fibrosis, cirrhosis, and even liver cancer [ 28 ]. Among HBV proteins, HBx is a leading mediator of hepatic inflammation [ 29 , 30 ]. Considering that early HBV infection is primarily characterized by chronic liver inflammation, rather than employ a model of liver cancer development, we instead chose to use the standard liver cell line HL7702 to explore activation of intrahepatic cytokine networks induced by HBx. In the current study, we found that under oxidative stress, HBx enhanced NLRP3 inflammasome-mediated inflammation and pyroptosis through upregulating mitoROS production. Interestingly, in a setting of HBV replication without H 2 O 2 stress, HBx played a similar role in mediating the NLRP3-associated inflammatory process. To the best of our knowledge, this study provides the first evidence that HBx-mediated activation of the NLRP3 inflammasome and pyroptosis required conditions, such as oxidative stress or HBV replication. Oxidative stress is a prominent indicator of HBV infection. It is characterized by increased oxidative products, reduced antioxidant enzymes, and elevated ROS levels, including those of H 2 O 2 [ 31 – 33 ]. Similarly, we found that HBx enhanced the levels of malondialdehyde, a product of fatty acid oxidation, and reduced the levels of the antioxidant superoxide dismutase in mice (data not shown). However, our previous studies demonstrated that HBx alone is not sufficient to induce oxidative stress [ 20 ]. Therefore, in this study, we used H 2 O 2 to induce oxidative stress in cells prior to evaluating the effect of HBx on the NLRP3 inflammasome, to simulate the intrahepatic oxidative stress environment. We found that HBx-expressing cells promoted the secretion of ASC, IL-1β, IL-18, and HMGB1 under H 2 O 2 stress. Also, HBV-expressing cells without H 2 O 2 intervention also promoted the release of inflammatory cytokines. Hence, other HBV proteins may induce oxidative stress after which HBx may promote NLRP3 activation. The observed ROS levels between the pGEM and pHBV-HBx groups (Fig. 5 d) confirmed this hypothesis. We also observed opposing roles for HBx on apoptosis. Similarly, Huang et al. found that HBx suppresses apoptosis of hepatoma cells during starvation by enhancing mitophagy; however, it exhibits the opposite role under well-fed conditions [ 23 ]. Taken together, these results suggest that HBx may exert specific and unique roles in different environments, while also stimulating the release of pro-inflammatory cytokines only in the presence of other specific factors, such as oxidative stress, nutrient deprivation, or the expression of other HBV proteins. Mitochondrial injury often leads to the generation of ROS and subsequently activates the ROS signaling pathway. Rahmani et al. showed that HBx interacts with the mitochondrial human voltage-dependent anion-selective channel protein 3 (HVDAC3) and alters mitochondrial membrane potential [ 34 ]. Consistent with this, we found that HBx was present in the mitochondrial fraction isolated from HBx-expressing cells. TEM results demonstrated morphological changes consistent with mitochondrial damage. Moreover, exposure to H 2 O 2 caused an increase in mitoROS levels in HBx-expressing cells compared to that in the control groups, indicating that HBx facilitated the oxidative stress-related damage. To date, the mechanism of HBx-mediated liver cell death has primarily focused on non-inflammatory cell death, such as autophagy and apoptosis [ 23 , 35 ]. Our current study shows, for the first time, a new form of cell death, HBx-induced pyroptosis, in normal hepatocytes exposed to H 2 O 2 . Pyroptosis refers to inflammatory cell death and may play a more significant role in HBV infection than other types of cell death. It is generally accepted that HBx can induce the production of pro-inflammatory cytokines in various liver cells. However, the underlying mechanistic details associated with HBx-induced inflammation remain unclear. Ample evidence suggests that NLRP3 inflammation occurs in hepatocytes, sinusoidal endothelial cells, and non-substantial cells, including Kupffer cells and hepatic stellate cells [ 8 , 9 , 36 ]. In comparison with that in other tissues, the expression level of caspase-1 in the liver is higher, and Kupffer cells can produce a large amount of IL-1β by activating NLRP3 [ 37 ]. Based on these data, pyroptosis may be a pivotal mechanism involved in HBx-mediated inflammation, at least in HL7702 liver cells. Contrary to the findings of Yu et al., which state that HBV suppresses lipopolysaccharide (LPS)-mediated NLRP3 activation [ 38 ], we found LPS-induced NLRP3 activation, primarily in Kupffer cells. One reason for this discrepancy may be differences in study aims and design. For instance, they primarily focused on the hepatitis B e antigen (HBeAg), which inhibits NLRP3 activation. Moreover, although they conducted experiments using clinical samples, the sample size was relatively small, and they did not consider the effect of HBV DNA titer on NLRP3. We demonstrated that the inflammasome expression signal was enhanced in the liver tissue of patients with increased HBV DNA copy number. Recently, some studies have showed that ASC can be released into the extracellular space from cells via inflammasome activation and accumulate in tissue. Then, ASC is phagocytosed by the surrounding immune cells, which promotes the production of pro-inflammatory cytokines and spread of inflammation [ 39 , 40 ]. In addition, Franklin et al. demonstrated that subcutaneous and intraperitoneal injections of ASC cause acute inflammation in mice [ 41 ]. As ASC is a newly identified inflammatory protein that can mediate the spread of inflammation, we tested the serum levels of ASC in patients with various HBV DNA loads. Interestingly, higher serum ASC levels were observed in HBV-positive patients, similar to that observed in HBx-expressing cells. Collectively, these results suggest that HBx mediates the spread of inflammation via the promotion of ASC secretion. Further studies are needed to explore the mechanism of inflammation mediated by ASC. Another study demonstrated that the NLRP3 inflammasome expression signal in normal, hepatitis-related, and cirrhotic tissue shows a continuously increasing trend; however, it becomes significantly reduced in hepatoma tissues [ 42 ]. Similarly, our data showed that inflammasome expression levels in liver cancer tissues from HBV-infected patients were significantly down-regulated compared with that in matched healthy para-cancer tissues (data not shown). Moreover, the mRNA and protein expression levels of the inflammasome in HepG2.2.15 cells, a human hepatoma cell line that stably expresses HBV, were decreased compared with those of control HepG2 cells (data not shown). The discrepant roles of HBV on the NLRP3 inflammasome may be dependent on the different stages of a natural HBV infection. It is possible that during the early stage of HBV infection, NLRP3 may be activated, resulting in inflammation and cell pyroptosis. Alternatively, long-lasting chronic inflammation may cause the development of liver cancer. When hepatocytes become cancerous, HBV may reduce the expression of the inflammasome and the secretion of pro-inflammatory mediators, thereby evading recognition and elimination by immune cells. Numerous studies have shown that the NLRP3 inflammasome can resist the formation of colon cancer [ 43 ], further supporting this speculation. However, the specific mechanism requires further investigation. Additionally, many studies showed that HMGB1 is an upstream activator of the inflammasomes [ 44 , 45 ]. However, some studies have suggested that HMGB1 is a downstream protein of inflammasomes. Lamkanfi et al. showed that mice with knockout caspase 1 down-regulates LPS-induced HMGB1 production compared to wild-type mice. Also, the LPS-induced release of HMGB1 from macrophages depends on the activation of the NLRP3 inflammasome [ 46 ]. Hou et al. reported that pyroptosis enhances the secretion of HMGB1 in macrophages [ 47 ]. Chen et al. demonstrated that an adipokine (visfatin) can activate the NLRP3 inflammasome in vascular endothelial cells and promote the production of HMGB1, which leads to damage to the intercellular connections in the vascular endothelium [ 48 ]. Similarly, we observed that HBx increases the secretion of HMGB1 in hepatocytes under H 2 O 2 -stimulation, whereas NAC reduces the secretion of HMGB1 via inhibition of the activation of the NLRP3 inflammasome. The inconsistencies in the above results may be related to the complexity of cell signaling pathway regulation and the different types of cells and tissues. While important new findings are provided, limitations in the current work have been noted. For instance, only a single human normal hepatic cell line was used in this study. Evaluation of other normal hepatic cell lines is, therefore, required. Also, the current study did not include any animal studies, which may provide relevant and essential insights. In summary, under conditions of oxidative stress, HBx activated NLRP3 in normal hepatocytes and promoted pyroptosis via the mitochondrial ROS pathway, ultimately causing the release of ASC, IL-1β, IL-18, and HMGB1. Our study may provide novel insights into the mechanisms involved in HBV-induced hepatitis. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 (TIFF 31952 kb) Supplementary material 2 (DOC 25 kb) Below is the link to the electronic supplementary material. Supplementary material 1 (TIFF 31952 kb) Supplementary material 2 (DOC 25 kb)

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8537463/

Lemierre’s Syndrome: Case Presentation of a Rare and Possibly Life-Threatening Condition

Lemierre's syndrome is, presently, a very rare condition, but a life-threatening one. The syndrome was first described in 1936 by Andre Lemierre and comprises an oropharyngeal infection (most commonly associated with anaerobic bacteria Fusobacterium necrophorum ), internal jugular vein thrombophlebitis and, possibly, secondary septic metastasis (common sites are lungs or brain). We describe such a rare case diagnosed at our Infectious Diseases Department in September 2019. 1. Introduction In 1936, Andre Lemierre, a French physician, described 20 cases of anaerobic septicemia (18 deaths) [ 1 ]. The syndrome (named after him) comprises an initial site of infection in the oropharyngeal region and, subsequently, internal jugular vein thrombophlebitis; through the bloodstream, germs may be transported to other locations and generate septic metastases (mainly to the brain and lungs, but also to the kidneys or joints). Most cases involve young healthy patients. However, variants have been described that are atypical in the age of onset or in the location of the infection or septic metastases [ 2 ]. Fusobacterium necrophorum is the most isolated germ responsible for this syndrome [ 3 ]; however, other bacteria may be encountered. 2. Case Report PCM, 18 years old, female, from the Gorj County, was admitted on 21 September 2019 to the Infectious Diseases Department with high fever (39–40 °C), chills, intense sore throat and headache. The onset of the disease was two weeks earlier with fever and sore throat. She had consulted the general practitioner who established the diagnosis of exudative tonsillitis and recommended treatment with 625 mg amoxicillin/clavulanate three times per day and 200 mg ibuprofen two times per day. The patient followed the treatment for 12 days, but the condition worsened, the fever gradually went up and a bulge appeared on the left side of the neck. Her past medical history was inconspicuous; the patient denied consumption of tobacco, alcohol, recreational drugs or oral birth control drugs. She had received all the mandatory immunization (according to the Romanian national program). There was no contact with other sick persons with similar symptoms. When she was admitted to the hospital, her general condition was consistent with moderate sepsis; she had fever (38.9 °C), elevated pulse rate (137 bpm), normal blood pressure (116/74 mm Hg) and normal oxygen saturation (97%). A significant lymph nodes enlargement (about 6 cm in diameter) was noted on the left side of the neck, in front and behind the sternocleidomastoid muscle, and left supraclavicular adenopathy; there was limited movement of the enlarged lymph nodes and they were hard on palpation. Both tonsils showed white exudate on their surface. There was no abnormal lung or heart sounds, no liver or spleen enlargement and no neurological abnormalities. The urine output was normal. The initial laboratory tests are shown in Table 1 . The chest X-ray showed consolidation on the lower right lobe. An ultrasound of the left side of the neck highlighted the lymph nodes enlargement, intense vascularization and a deep edema. A computed tomography (CT) of the neck (see Figure 1 and Figure 2 ) and the upper thorax showed segmental partial thrombosis of the left internal jugular vein. The Ear-Nose-Throat (ENT) examination established the diagnosis of cryptic tonsillitis and cervical adenitis. The throat swab culture on blood agar media and blood culture using BACTEC vials for aerobic germs were negative. An enzyme-linked immunosorbent assay (ELISA) test for HIV was also negative. The diagnosis of Lemierre's syndrome was established and the initial prognosis was reserved (the patient met the old criteria for sepsis); she started antimicrobial treatment with penicillin G, 4 international megaunits (MUI)/day, intravenous (i.v.), and metronidazole, 1.5 g/day, per os (p.o.). She also received enoxaparin, 0.6 mL twice a day, s.c., and diclofenac, 100 mg/day, p.o. The fever gradually decreased, the patient returned to normal temperature after 6 days of treatment, and the general condition improved; however, the lymph nodes remained enlarged throughout the period of hospital admission. The follow-up laboratory tests are shown in Table 2 . The patient was discharged after 16 days; she received a prescription for metronidazole, 1.5 g/day, p.o., for another 14 days and for rivaroxaban, 20 mg/day, p.o., for 30 days; she continued with this regimen at home as recommended. A month later, she was seen in our department: she was in good condition, there were no signs of infection, but the ultrasound still showed signs of thrombosis of the jugular vein. She refused the surgical intervention to remove the thrombus. 3. Discussion Today, Lemierre's syndrome is a rare condition (1 case per million annually) [ 2 , 4 ], probably due to the extensive use of antimicrobials. This was the first case diagnosed in our clinic since its opening in 1984 and probably one of the first descriptions of this condition in a Romanian patient. A search on PubMed and Medscape returned only one result of a fellow Romanian diagnosed and treated in Greece [ 4 ]. From the clinical point of view, our case is a classic one, but we were not able to detect the causative germ, probably because only aerobic culture media were used and prior antimicrobial therapy was administered. Unfortunately, anaerobic culture media were not available at the hospital at that time. However, given the fact that the starting point was exudative pharyngitis, we suspected bacterial involvement, which is considered sufficient according to the contemporary diagnostic criteria [ 5 ]. A number of alternative diagnoses were taken into account prior to the laboratory examinations. First, we considered streptococcal tonsillitis complicated with a left peritonsillar abscess, but there was no trismus or tonsillar protrusion. Infectious mononucleosis might lead to exudative pharyngitis, but the lymph nodes should be bilaterally enlarged and much smaller; also, a rash should be noted after the treatment with aminopenicillin. Given the aspect of the lymph nodes, there was another idea of leukemia or malignant lymphoma. There are many cases of tuberculosis in Romania, but oral involvement, usually due to Mycobacterium bovis , is extremely rare, and the girl denied consumption of raw milk. Diphtheria has not been observed at our clinic for the last 40 years and the patient was immunized against it. We also dismissed the possibility of oropharyngeal anthrax or tonsillar cancer due to human papillomavirus (HPV). After the diagnosis of Lemierre's syndrome, the initial prognosis was reserved due to the mortality rate which even today remains significant (4–25% [ 4 , 5 , 6 , 7 , 8 , 9 ]). A recent meta-analysis found a mortality rate of 4% [ 10 ]. Our case also met the 2001 criteria of sepsis [ 11 , 12 , 13 ]. The patient was initially treated with amoxicillin/clavulanate and there are three probable causes why there was no good result: either non-susceptible bacteria were involved or the dose was insufficient or the antimicrobial was not able to properly penetrate the tonsillar crypts. Based on the available literature data, we empirically chose the antimicrobial combination of penicillin and metronidazole, with good clinical and biological results. The duration of the treatment varies from 2 to 6 weeks according to the medical literature [ 2 , 4 , 5 , 6 , 7 ]; in our case, it was about 6 weeks (including pre- and post-hospitalization). Another problem was whether or not to use anticoagulation. We found arguments for and against it in some case reports or small case series [ 2 , 3 , 4 , 5 , 7 , 8 , 9 , 10 , 14 ]. A recent meta-analysis from 2020 that collected the data of 712 patients diagnosed with Lemierre's syndrome found that in 5.2% of the cases, the patients experienced new venous thromboembolism, in 11.7% of the cases—new peripheral septic lesions; however, the rates of both events were lower if anticoagulation was used; the number of bleeding events was found to be low [ 10 , 12 , 15 ]. Our choice was to use enoxaparin concomitant with antimicrobials (which should kill any bacteria potentially released due to the lysis of the thrombus). Because the thrombus persisted and the girl refused surgical intervention, she continued therapy with rivaroxaban, which is more convenient to use at home, having an oral route of administration. Our report includes information on the course of blood tests during the clinical course of Lemierre's syndrome that have not been described by most previous reports. Similarly to other forms of non-septic venous thrombophlebitis, D-dimers were considerably elevated upon diagnosis; while this may have been due to sepsis, this finding suggests that the clinical value of this parameter in supporting diagnosis should be explored by future research. In addition, we analyzed laboratory parameters over the course of hospitalization, showing the timing of their progressive improvement in the case in which both antimicrobial and anticoagulant treatments were initiated. They can be used as reference for comparison with cases in which only antimicrobial treatment was used. 4. Conclusions Lemierre's syndrome is a very rare disease with reserved prognosis and a significant mortality rate which requires personalized therapy with an antimicrobial combination in association with anticoagulants.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7541672/

Anthrax Protective Antigen Retargeted with Single-Chain Variable Fragments Delivers Enzymes to Pancreatic Cancer Cells

The nontoxic, anthrax protective antigen/lethal factor N-terminal domain (PA/LF N ) complex is an effective platform for translocating proteins into the cytosol of cells. Mutant PA (mPA) was recently fused to epidermal growth factor (EGF) to retarget delivery of LF N to cells bearing EGF receptors (EGFR), but the requirement for a known cognate ligand limits the applicability of this approach. Here, we render practical protective antigen retargeting to a variety of receptors with mPA single-chain variable fragment (scFv) fusion constructs. Our design enables the targeting of two pancreatic cancer-relevant receptors, EGFR and carcinoembryonic antigen. We demonstrate that fusion to scFvs does not disturb the basic functions of mPA. Moreover, mPA–scFv fusions enable cell-specific delivery of diphtheria toxin catalytic domain and Ras/Rap1-specific endopeptidase to pancreatic cancer cells. Importantly, mPA–scFv fusion-based treatments display potent cell-specific toxicity in vitro, opening fundamentally new routes toward engineered immunotoxins and providing a potential solution to the challenge of targeted protein delivery to the cytosol of cancer cells.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5310721/

Protein S-palmitoylation in cellular differentiation

Reversible protein S-palmitoylation confers spatiotemporal control of protein function by modulating protein stability, trafficking and activity, as well as protein–protein and membrane–protein associations. Enabled by technological advances, global studies revealed S-palmitoylation to be an important and pervasive posttranslational modification in eukaryotes with the potential to coordinate diverse biological processes as cells transition from one state to another. Here, we review the strategies and tools to analyze in vivo protein palmitoylation and interrogate the functions of the enzymes that put on and take off palmitate from proteins. We also highlight palmitoyl proteins and palmitoylation-related enzymes that are associated with cellular differentiation and/or tissue development in yeasts, protozoa, mammals, plants and other model eukaryotes. Introduction Protein S-acylation or S-palmitoylation involves the posttranslational addition of fatty acyl chains, typically a palmitate (C16:0), to cysteine residues of proteins via thioester linkages ( Figure 1A ). In contrast with other forms of protein lipidation, S-palmitoylation is uniquely reversible due to the high-energy thioester bond formed between the acyl group and the cysteine side chain, potentially allowing for rapid spatiotemporal control of protein function akin to protein phosphorylation. S-Palmitoylation predominantly serves to target proteins to specific membrane compartments and/or microdomains [ 1 – 4 ]. While typically not the primary membrane association signal, S-palmitoylation often acts in concert with other lipid modifications, such as N-myristoylation and prenylation, to determine the intracellular distribution of soluble proteins [ 2 – 4 ]. By influencing protein localization and trafficking, palmitoylation/depalmitoylation can have critical effects on protein function as epitomized by compartmentalized Ras signaling, where spatial segregation of distinct signal transduction modules diversifies the signaling outputs by a single protein [ 5 , 6 ]. Palmitoylation may also regulate protein activity in diverse ways, such as inducing protein conformational changes, modulating protein stability and protein–protein interactions and interacting with other posttranslational modifications [ 7 – 12 ]. These functional consequences of protein palmitoylation are not mutually exclusive, and a single palmitoylation event can simultaneously modulate multiple aspects of protein function [ 13 ]. Figure 1. Protein S-palmitoylation and analytical strategies. ( A ) Dynamic S-palmitoylation is mediated by DHHC-containing PATs (DHHC-PATs) and acyl protein thioesterases that put on and take off palmitate from cysteine residues of proteins, respectively. ( B ) Metabolic and bioorthogonal labeling strategy using fatty acid chemical reporters as well as ( C ) the selective chemical labeling strategy for analysis of protein S-palmitoylation. In both strategies, the incorporation of detection or enrichment probes in the final step allows for direct detection or enrichment and mass spectrometry identification of tagged proteins. S-Palmitoylation of intracellular proteins is mainly mediated by an evolutionarily conserved family of palmitoyl acyltransferases (PATs), which are identified by the characteristic catalytic Asp-His-His-Cys (DHHC) motif embedded within a cysteine-rich domain. First discovered in budding yeast Saccharomyces cerevisiae [ 14 , 15 ], PAT orthologs are found in all eukaryotes, ranging from 5 PATs in fission yeast Schizosaccharomyces pombe to ∼23 in mammals and 24 in Arabidopsis . While there is overlapping substrate specificity between multiple PATs [ 16 , 17 ], it is clear that there are additional specificity determinants in the variable domains of the PATs [ 18 ] as well as regulatory mechanisms that determine the context-dependent protein substrate pool and function of individual enzymes. Regulation of specific PATs at the transcriptional [ 19 ], translational [ 20 ] and posttranslational [ 21 ] levels has been reported with significant changes in the palmitoylation state of the corresponding protein substrates. Additionally, PAT activity can be determined by its subcellular trafficking [ 22 ], oligomerization state [ 23 ] or additional protein subunits [ 14 , 24 , 25 ]. Notably, these are isolated examples in specific host systems looking at a few protein substrates. In general, the in vivo substrate selectivity and regulation of PATs as well as the physiological significance of PAT-mediated protein palmitoylation remain to be fully characterized. Until recently, depalmitoylation of cytosolic cysteine residues has been attributed to only two related acyl protein thioesterases APT1/LYPLA1 and APT2/LYPLA2 [ 26 ]. Since APT1 and APT2 are both palmitoylated, autoregulatory mechanisms are proposed where the depalmitoylating enzymes control their physical access to protein substrates [ 27 , 28 ]. Studies focusing on serine hydrolases led to the independent discovery of the ABHD17 proteins that depalmitoylate N-Ras and PSD-95 in mammalian cells [ 29 , 30 ]. ABHD12 and ABHD13 also exhibited depalmitoylating activities, albeit weaker than ABHD17, when tested against PSD-95 as the substrate [ 29 ]. Interestingly, an ABDH13 ortholog in Toxoplasma , TgPpt1, was demonstrated to depalmitoylate proteins and to have important functions in regulating host cell invasion by the parasite [ 31 ]. These findings suggest that the family of depalmitoylating enzymes may be larger and more diverse than previously appreciated. Like the PATs, the substrate selectivity and biological importance of depalmitoylating enzymes remain to be determined, but it is clear that cellular context matter and factors, such as cell type and physiology, can affect depalmitoylating activity toward specific protein substrates [ 29 , 30 ]. This review focuses on protein S-palmitoylation in the context of cellular differentiation, which is the process by which cells becomes more specialized to perform specific functions. We will introduce the main chemical strategies and various chemical tools to analyze in vivo protein palmitoylation, and to interrogate the activity of palmitoylating and depalmitoylating enzymes in different cellular states. We will then highlight palmitoyl proteins and palmitoylation-related enzymes that are involved with cellular and tissue development in yeasts, protozoa, mammals, plants and other model eukaryotes. Other forms of protein acylation such as N-myristoylation, N-palmitoylation and O-palmitoleoylation of secreted proteins (e.g. Wnt and Hedgehog) and the family of membrane-bound O -acyltransferases that mediate these modifications will not be included in this review [ 32 , 33 ]. Chemical strategies and tools to study protein S-palmitoylation Advances in our understanding of protein S-palmitoylation over the last decade can be largely attributed to the development of chemical tools that enable rapid quantitative analyses of palmitoylated proteins. These chemical tools can be broadly categorized into two main strategies. In the first strategy ( Figure 1B ), cells are metabolically labeled with fatty acid chemical reporters, which are site-specifically installed onto cysteines of target proteins via native thioester linkages by the endogenous enzymatic machinery. Visualization or enrichment probes are subsequently introduced using bioorthogonal labeling reactions to allow detection or identification of modified proteins [ 34 , 35 ]. Because these bioorthogonal fatty acid chemical reporters can be competed away by endogenous substrates, they have been employed in pulse-chase experiments to monitor palmitate turnover kinetics on proteins [ 36 , 37 ]. Fatty acid chemical reporters with cross-linking functionalities have also been developed to interrogate protein–protein interactions that depend on protein palmitoylation [ 38 ]. The second strategy involves the selective chemical modification of thioester-linked cysteines in S-palmitoylated proteins ( Figure 1C ). After initial capping of free thiols on proteins, selective cleavage of thioester linkages liberates thiols that can be selectively labeled to facilitate detection and/or enrichment by biochemical methods or mass spectrometry [ 39 , 40 ]. Methods, such as acyl-biotin exchange (ABE) and acyl-resin-assisted capture, employ this strategy. Notably, the selective hydrolysis and removal of thioester-linked acyl groups facilitate the identification of palmitoylation sites since direct mass spectrometry detection of acylated peptides can be challenging [ 41 ]. In a recent variation of the latter strategy, acyl-PEG switch/exchange uses PEGylated thiol-reactive reagents to induce electrophoretic mobility shift of modified proteins to monitor the relative abundance of palmitoylated versus nonpalmitoylated forms of target proteins and the number of palmitoylation sites [ 29 , 42 , 43 ]. Both strategies have been successfully employed to detect and identify S-palmitoylated proteins in global and focused studies of various cell types. The ability to visualize and profile dynamically S-palmitoylated proteins on a global scale has greatly expanded the scope of the modification by revealing new palmitoylated proteins and regulatory roles of protein palmitoylation in eukaryotic physiology and disease. Since the first global profiling of palmitoylated proteins in yeast by Roth et al. [ 16 ], dozens of palmitoyl proteomes consisting of ∼10% of the proteomes in yeasts, protozoans [ 44 ], plants [ 45 , 46 ] and mammalian systems have been reported [ 47 – 49 ]. Despite the lack of a 'consensus sequence' surrounding palmitoylation sites, the ever increasing number of palmitoyl proteomes led to the development and refinement of in silico predictive programs [ 50 – 53 ], which are robust in predicting modification sites for proteins with canonical palmitoylation motifs (e.g. dual acylation, cysteines near to prenylation motifs or transmembrane domains) and for those with validated palmitoylated homologs or orthologs. Our predictive ability for proteins with atypical modification sites will likely continue to improve as the list and diversity of experimentally validated proteins and palmitoylation sites expand. Improving the quality, analysis and curation of palmitoyl proteomes will be important as the field moves toward comparative proteomics to identify palmitoylation events that are critical for various biological phenomena and disease states. Notably, the overlap between palmitoyl proteomes obtained using the chemical reporter strategy and those obtained using the selective chemical modification strategy is limited, reflecting the different sources of false positives from each technique ( Table 1 ) [ 54 ]. With metabolic labeling, metabolism of the fatty acid chemical reporters can lead to enrichment of proteins with other forms of protein lipidation, although alk-16 (also known as 16-ODYA) has been shown to be preferentially incorporated into palmitoylated proteins [ 34 , 48 ]. The biotin switch strategy will enrich for proteins with any thioester-linked modifications, which also include SUMOylation and ubiquitination. Marrying these independent complementary approaches minimizes these pitfalls for more accurate global analyses of protein S-palmitoylation in cells [ 44 ]. Since bona fide palmitoylated proteins are more likely to be identified using multiple methods, Blanc et al. [ 54 ] combined the data from existing databases, global and focused palmitoylation studies to generate a high confidence SwissPalm database for improved prediction of palmitoylated proteins and modification sites. Comparison of SwissPalm with other databases uncovered potentially extensive cross-talk between palmitoylation and various posttranslational modifications [ 54 ]. Table 1 Chemical tools to study protein S-palmitoylation Fatty acid probes Principle of method Detection Identification of modified proteins Variations/other applications Metabolic labeling Radiolabelled (e.g. 3 H, 13 C, 125 I) fatty acids Autoradiographic detection Sensitive and quantitative X Indirect identification (e.g. immunoprecipitation and overexpression) Used in pulse-chase experiments to determine turnover kinetics X Cumbersome and requires long exposure periods or additional safety protocols X False positives as a result of fatty acid metabolism, including crosstalk between different forms of protein fatty acylation X Requires prior knowledge of candidate proteins Bioorthogonal fatty acid chemical reporters (e.g. azide-, alkyne-functionalized) Introduction of custom detection and/or enrichment probes post-metabolic labeling via bioorthogonal reactions Rapid, sensitive and quantitative non-radioactive detection Selective enrichment of palmitoylproteomes Used in pulse-chase experiments to determine turnover kinetics Multiplex detection is possible with orthogonal reporters and detection probes X False positives as a result of fatty acid metabolism, including crosstalk between different forms of protein fatty acylation X Challenging site ID due to difficult MS detection of acylated peptides Fatty acid chemical reporters with crosslinking functionalities enable the interrogation of palmitoylation-specific protein-protein interactions X False positives as a result of fatty acid metabolism, including crosstalk between different forms of protein fatty acylation Method Principle of method Detection Identification of modified proteins Variations/other applications Selective chemical labeling of palmitoylated cysteines Acyl-biotin exchange (ABE) or Acyl-resin-assisted capture (acyl-Rac) After initial capping of free thiols, selective cleavage of thioester bonds liberates free thiols for reaction with thiol-reactive reagents that enable detection and enrichment Rapid, sensitive and quantitative non-radioactive detection Selective enrichment of palmitoylproteomes The population of modified proteins can be determined using thiol-reactive reagent that alters protein electophoretic mobility (acyl-PEG switch/exchange) Identification of modification sites Acyl-PEG-exchange (APE) or Acyl-PEG-switch X False positives with other thioester-linked PTMs and also with incomplete thiol capping X False positives with other thioester-linked PTMs and also with incomplete thiol capping X symbol indicates the limitations of each strategy. PTMs, posttranslational modifications. In addition to biochemical analyses, the ability to rapidly and selectively perturb the modification in cells will be valuable toward understanding the biological significance of palmitoylation in different physiological contexts. Chemical inhibitors provide a rapid and convenient method to perturb protein palmitoylation, especially in organisms and cells that are challenging to manipulate genetically. Readers are referred to a review that includes a dedicated section on the different classes of small molecule inhibitors targeting palmitoylating and depalmitoylating enzymes [ 55 ]. 'Clickable' forms of some inhibitors have also been used to discover and validate enzymes that modulate protein palmitoylation in cells [ 30 , 31 , 36 ]. It is important to note that although 2-bromopalmitate has been and remains one of the most widely used inhibitors of protein palmitoylation, it also inhibits fatty lipid metabolism and should not be used in isolation to prove or interrogate the function of protein S-palmitoylation. Undoubtedly, the development of potent and highly selective chemical inhibitors for PATs and depalmitoylating thioesterases will be pivotal toward dissecting their functional contributions and uncovering insights into the regulation of dynamic protein palmitoylation. S-Palmitoylation in yeast cellular differentiation The fission yeast S. pombe is an emerging model organism for palmitoylation studies due to its genetic tractability and relatively simple palmitoylation machinery compared with other eukaryotes. Studies have associated palmitoylation of several important signal transduction proteins with S. pombe sexual differentiation, which involves mating between haploid cells of opposite mating types, entry and progression through meiosis to yield haploid spores. Using the fission yeast system, where a single Ras ortholog is involved in two distinct signaling pathways with quantifiable phenotypes, the Chang laboratory showed that Ras1 signal transduction is spatially compartmentalized with cellular morphogenesis regulated by nonpalmitoylated Ras1 at endomembranes and mating requiring palmitoylated Ras1 at the plasma membrane [ 5 ]. Similar to the S. cerevisiase PAT ERF2 that preferentially modifies heterolipidated GTPases [ 16 ], the S. pombe Erf2 ortholog is the primary PAT that palmitoylates Ras1 and two other small GTPases — Rho2 and Rho3 [ 19 , 56 ]. High expression levels of SpErf2 and its noncatalytic protein cofactor SpErf4 during meiosis are required for Rho3 palmitoylation, and dysregulation of Rho3 palmitoylation triggers aberrant meiotic divisions in sensitized cells [ 19 ]. Rho2 palmitoylation is needed for morphogenesis and cell wall integrity of vegetative cells via the Pmk1 pathway that is antagonistically regulated by Rho3 [ 56 ]. Future work will reveal the regulatory roles and mechanisms of SpErf2-mediated palmitoylation in coordinating the signaling outputs of these small GTPases as cells exit vegetative growth and undergo sexual differentiation. Cryptococcus neoformans is a facultative intracellular fungal pathogen that is able to survive and proliferate in the harsh environment of macrophage phagolysosomes. Recent investigations uncover a major role for PAT-mediated protein palmitoylation in the virulent potential of C. neoformans . The PFA4 gene encoding a DHHC-PAT was identified from screening genes that influenced host–pathogen interactions, and loss of PFA4 function has dramatic effects on morphology, stress tolerance and virulence potential of C. neoformans [ 57 ]. Comparative palmitoyl proteome profiling identified CnPFA4 protein substrates that are involved in cell wall synthesis, membrane transport, signal transduction and membrane trafficking, which is consistent with the pleiotropic defects observed for pfa4Δ cells [ 57 ]. CnPFA4 was also identified as the major PAT responsible for Ras1 palmitoylation, which is required for Ras1 localization at the plasma membrane and pathogenesis in a cryptococcosis murine model [ 58 , 59 ]. Ras1 palmitoylation is not required for C. neoformans sexual differentiation [ 58 ], suggesting palmitoylation-dependent compartmentalization of Ras1 signaling. Protein palmitoylation can affect the function of nuclear proteins. In S. cerevisiae , PFA4-mediated palmitoylation of the telomere-binding protein RIF1 altered heterochromatin dynamics and transcriptional silencing [ 60 ]. It is unclear if RIF1 palmitoylation is regulated, but with 25% of the putative palmitoylated proteins in mouse and human cells being nuclear proteins [ 54 ], this finding raises the interesting possibility that palmitoylation of nuclear proteins may be important in direct modulation of global gene expression during cellular transitions. S-Palmitoylation in the virulence and transmission of protozoan parasites Apicomplexan parasites are a large group of obligate intracellular protozoan parasites. Most members of this group have complex asexual and sexual reproduction cycles within multiple hosts, with survival requiring rapid adjustment to distinct environments and precise spatiotemporal coordination of key cellular processes for host cell invasion, replication and egress. Recent work demonstrated the pervasive roles of protein S-palmitoylation in the developmental life cycles of Toxoplasma gondii and Plasmodium falciparum and, by extension, their pathogenesis. Treatment with 2-bromopalmitate yielded pleiotropic developmental defects and reduced the invasive capacity of T. gondii and P. falciparum [ 44 , 61 ]. Interestingly, chemical inhibition of depalmitoylation activity by palmitoyl protein thioesterase 1 TgPPT1 significantly enhanced host cell invasion by T. gondii [ 31 ]. Palmitoyl proteome analyses in T. gondii and P. falciparum reveal palmitoyl proteins that are needed for the function of specialized invasion organelles, providing further insights into the apparently essential roles of dynamic protein palmitoylation on host cell invasion by these parasites [ 44 , 62 ]. Many components of the invasion motor glideosome complex, including the well-known GAP45, are palmitoylated, disruption of which is associated with motility and invasion defects [ 44 , 62 , 63 ]. Palmitoylation of TgAMA1 and Pf/TgARO is required for proper apical localization of specialized invasion-associated secretory organelles called rhoptries [ 62 , 64 , 65 ]. Consistent with the importance of protein palmitoylation in rhoptry function, disruption of rhoptry-localized TgDHHC7 responsible for TgARO palmitoylation blocks host invasion [ 64 , 66 ]. Global studies of PATs in Toxoplasma and Plasmodium support functional specialization by the enzymes across different life cycle stages of the parasites, with PATs' function determined by differential subcellular localization, expression patterns and posttranslational modifications [ 66 – 68 ]. These studies also implicate protein palmitoylation in other developmental stages of apicomplexan parasites. For example, PfDHHC2 is essential for the progression through both asexual and sexual stages in the mammalian and mosquito hosts, respectively [ 69 ]. In Plasmodium berghei , PbDHHC10 expression is translationally repressed in gametocyetes and briefly translated during ookinete formation to mediate palmitoylation events that are required for crystalloid formation and parasite transmission from the mammalian host to the mosquito vector [ 70 ]. PbHHC3 and PbDHHC9 appear to have functionally overlapping roles in mediating parasite sexual differentiation in the insect host [ 71 , 72 ]. Protein palmitoylation also plays a key role in the life cycle of a different protozoan parasite Giardia lamblia , which is one of the major global causes of diarrheal disease. Encystation of the parasite to form cysts is critical for the survival of the parasite outside the host and its transmission. In addition to a changing palmitoyl proteome during encystation, Merino et al. [ 73 ] showed that genetic and chemical perturbation of protein palmitoylation by the different PATs in G. lamblia can negatively affect parasite differentiation into cysts. Identifying the palmitoyl proteins and the respective PATs participating in this key cellular transition will open up opportunities for therapeutic interventions. S-Palmitoylation in mammalian tissue development Meta-analysis of mammalian palmitoyl proteomes reveals dominant roles for protein palmitoylation in neural development and function, with a striking proportion of synaptic genes encoding for palmitoyl proteins [ 74 ], including neurotransmitter receptors, transporters, adhesion molecules, scaffolding proteins and vesicular trafficking proteins [ 75 ]. Reduced ZDHHC2 expression was observed in degenerating dopaminergic neurons and in patients with incipient Parkinson's disease [ 76 ]. Additionally, mutations in PATs, including ZDHHC17/ZDHHC13 (HIP14/HIP14L) and ZDHHC8, are associated with neurological disorders such as Huntington's disease and schizophrenia, respectively [ 18 , 75 ]. The spine density deficits in a schizophrenia mouse model are palmitoylation-dependent and can be rescued in vivo by overexpressing ZDHHC8 or one of its substrates — the constitutively active brain-specific splice isoform of Cdc42-palm [ 47 , 77 – 79 ]. By regulating its subcellular localization and RhoGDI (Rho guanine nucleotide dissociation inhibitor) binding, palmitoylation of Cdc42-palm is required for normal dendritic spine development during synaptogenesis [ 47 , 80 ]. Palmitoylation of other proteins, including paralemmin [ 81 ] and actin regulator LIMK1 [ 82 ], have also been shown to affect dendritic spine maturation that is critical for neuronal plasticity. The modulation of global protein palmitoylation in response to synaptic activity and cellular signals further supports the regulatory roles of protein palmitoylation in neuronal plasticity and development [ 47 ]. Palmitate turnover of the most abundant neural scaffolding protein PSD-95 is accelerated upon glutamate receptor activation, and this down-regulates receptor signaling activity [ 83 ]. On the other hand, suppression of neural activity triggers the ZDHHC2 translocation to postsynaptic membrane, leading to increased PSD-95 palmitoylation and synaptic accumulation [ 22 ]. Synaptic activity increases ZDHHC5-mediated palmitoylation of intracellular cadherin-binding protein δ-catenin and stabilizes synaptic cadherin adhesion complexes that are critical for synaptic plasticity [ 84 ]. In cultured neuronal stem cells, induction of neural differentiation led to the rapid degradation of ZDHHC5 and reduced flotillin palmitoylation [ 21 ], implicating protein palmitoylation in neural stem cell differentiation. Palmitoylation has been shown to stabilize EID1, an inhibitor of the CREB-binding protein/p300 epigenetic regulator for neural stem cell differentiation [ 85 ]. The pivotal role for PATs and their substrates in the development of other tissues that are not part of the nervous system is indicated by the associations between dysregulated protein palmitoylation with cancers, which reflect defective control over cellular proliferation and differentiation [ 74 ]. The involvement of palmitoyl proteins, associated enzymes in cellular transformation and tumorigenesis are covered in a detailed review by Yeste-Velasco et al. [ 55 ]. Here, we focus on the role of protein palmitoylation in cellular and tissue development, particularly in modulating growth factor and hormone signaling, which mediate intercellular communication and regulate gene expression programs that drive cellular differentiation in various developmental stages. Palmitoylation of the epidermal growth factor receptor (EGFR) by ZDHHC20 limits EGFR signaling by inhibiting autophosphorylation and recruitment of the downstream adapter protein Grb2 and increasing receptor turnover [ 86 ]. For estrogen receptors (ERs), progesterone receptors and the androgen receptors, palmitoylation by ZDHHC7 and ZDHHC21 is crucial for their plasma membrane localization and function in mediating rapid tissue-specific responses to steroid hormones [ 87 – 89 ]. Upon estradiol binding, ERα is depalmitoylated and dissociates from caveolin-1, after which it is available to downstream signaling targets in the RK/MAPK and PI3K/AKT pathways [ 89 ]. Mice that are deficient in ZDHHC21 activity showed defects in maintaining skin homeostasis and hair follicle differentiation that results in hair loss [ 90 ], though it remains to be determined if these phenotypes can be attributed to defective steroid hormone signaling. Other protein substrates of ZDHHC7 or ZDHHC21 include Scribble [ 91 ] as well as the death receptor Fas and kinase Lck, highlighting the additional roles of these two PATs in modulating cellular polarity and proliferation [ 92 , 93 ]. ZDHHC13 knockout mice with pleiotropic developmental abnormalities of the hair, skin and bone hint at yet to be identified protein palmitoylation events that affect normal mammalian cellular and tissue development [ 94 ]. S-Palmitoylation in plants Compared with yeast and mammalian systems, the study of protein S-palmitoylation in plants is in the early stages. There are no reports of plant palmitoyl protein thioesterases to date. Of the 24 Arabidopsis thaliana DHHC-PATs [ 95 ], only a few have been characterized in any detail in terms of their biological roles. AtPAT10 is involved in vacuolar and tonoplast function, while AtPAT24/TIP1 has been implicated in developmental processes including pollen tube and root hair growth, shoot branching and cell polarity [ 96 , 97 ]. Recently, AtPAT13 and AtPAT14 have been shown to be involved in leaf senescence [ 98 , 99 ]. Identifying their protein substrates will be integral toward understanding how PATs and protein palmitoylation coordinate cellular differentiation and development in plants, but may require the use of independent methods to confidently establish PAT–substrate relationships [ 45 ]. Among the ∼500 candidate S-palmitoylated proteins in plants identified using the biotin switch strategy, protein kinases are overrepresented [ 45 , 46 ], suggesting that S-palmitoylation may have a major role in modulating phosphorylation signaling cascades in plants. These include RLK (receptor-like kinase) superfamily members that contain conserved cysteines adjacent to predicted transmembrane domains or N-myristoylated sites [ 45 ]. S-Palmitoylation of the LIP1 and LIP2 receptor-like cytoplasmic kinases is needed for directing pollen tube growth [ 100 ]. Non-RLKs such as the calcium-dependent protein kinase OSCPK2 in rice require both N-myristoylation and S-palmitoylation for proper subcellular localization [ 101 ]. Calcineurin B-like proteins, CBL1 and CBL2, which recruit Ser/Thr protein kinases during Ca 2+ signaling, require S-palmitoylation for targeting to the right membrane compartments in the cell [ 102 , 103 ]. Palmitoylation-deficient CBL2 fails to localize to the tonoplast and led to seed germination defects [ 102 ]. Besides individual proteins, palmitoylation can also direct the cellular distribution of large protein complexes. Kumar et al. [ 104 ] demonstrated that all catalytic subunits within the cellulose synthase complex are S-acylated and that the modification is required for correct localization of the complex to the plasma membrane and normal cellulose synthesis. S-Palmitoylation in other model organisms In zebrafish, strong phenotypes observed with various PAT deficiencies support the importance of protein palmitoylation in cellular differentiation and animal development. ZDHHC13 modulates bone morphogenetic protein signaling for lineage specification during embryogenesis [ 105 ]. A preliminary study using 2-bromopalmitate further implicates PATs and protein palmitoylation in mediating the transition from maternal to zygotic transcriptional programs after embryo fertilization [ 106 ]. Knockdown of DHHC15b and DHHC16 negatively affects forebrain development via dysregulation of neuronal differentiation and neural stem cell proliferation, respectively [ 107 , 108 ]. The forebrain developmental defect observed in animals deficient in ZDHHC15b activity is further associated with poor learning ability [ 108 ]. In Caenorhabditis elegans , lysosome-related fibrous body-membrane organelles (FB-MOs) are important for asymmetric cytoplasmic partitioning. Worms with mutations in spe-10 , which encodes a DHHC-containing protein localized to FB-MOs, showed defective spermatogenesis and are sterile, suggesting a role for SPE-10-mediated protein palmitoylation in establishing cellular polarity during cellular differentiation [ 109 ]. A systematic study of the other 14 DHHC-containing proteins in C. elegans using single and double RNAi knockdowns, however, did not yield any obvious phenotypes [ 110 ]. A similar functional analysis remains to be performed for Drosophila , in which tissue- and sex-specific expression of specific DHHC-containing genes have been observed [ 111 ]. Perspectives This is an exciting time to be studying protein S-palmitoylation. Posttranslational modifications increase the proteome complexity, and pervasive reversible modifications like protein palmitoylation have the potential to orchestrate diverse biological processes involved as cells transition from one state to another. The tools are in place to monitor in vivo palmitoylation stoichiometry and dynamics in various eukaryotic systems. Increasingly, selective small molecule inhibitors complement genetic approaches to profile and rapidly interrogate the functional contributions of palmitoylating and depalmitoylating enzymes. Ongoing efforts to minimize false positives, validate and curate palmitoyl proteomes will continue to enhance in silico predictive programs and establish the framework for comparative proteomics studies. We are now poised to identify and dissect critical palmitoylation events that regulate and/or are regulated during cellular differentiation and navigate the complex regulatory networks governing eukaryotic physiology and disease. Competing Interests The Authors declare that there are no competing interests associated with the manuscript.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7356292/

Rapid Microscopic Detection of Bacillus anthracis by Fluorescent Receptor Binding Proteins of Bacteriophages

Bacillus anthracis , the etiological agent of anthrax disease, is typically diagnosed by immunological and molecular methods such as polymerase chain reaction (PCR). Alternatively, mass spectrometry techniques may aid in confirming the presence of the pathogen or its toxins. However, because of the close genetic relationship between B. anthracis and other members of the Bacillus cereus sensu lato group (such as Bacillus cereus or Bacillus thuringiensis ) mis- or questionable identification occurs frequently. Also, bacteriophages such as phage gamma (which is highly specific for B. anthracis ) have been in use for anthrax diagnostics for many decades. Here we employed host cell-specific receptor binding proteins (RBP) of (pro)-phages, also known as tail or head fibers, to develop a microscopy-based approach for the facile, rapid and unambiguous detection of B. anthracis cells. For this, the genes of (putative) RBP from Bacillus phages gamma, Wip1, AP50c and from lambdoid prophage 03 located on the chromosome of B. anthracis were selected. Respective phage genes were heterologously expressed in Escherichia coli and purified as fusions with fluorescent proteins. B. anthracis cells incubated with either of the reporter fusion proteins were successfully surface-labeled. Binding specificity was confirmed as RBP fusion proteins did not bind to most isolates of a panel of other B. cereus s.l. species or to more distantly related bacteria. Remarkably, RBP fusions detected encapsulated B. anthracis cells, thus RBP were able to penetrate the poly-γ- d -glutamate capsule of B. anthracis . From these results we anticipate this RBP-reporter assay may be useful for rapid confirmative identification of B. anthracis . 1. Introduction Bacillus anthracis causing the zoonotic infectious disease anthrax in mammals and humans phylogenetically belongs to the Bacillus cereus sensu lato group of very closely related Firmicutes bacteria. The group comprises several familiar species, including Bacillus cereus sensu stricto , Bacillus thuringiensis , Bacillus weihenstephanensis , Bacillus mycoides and a variety of lesser characterized members [ 1 ]. Classical, culture-based techniques and classification upon phenotypic traits such as susceptibility against penicillin or lack of hemolysis are ambiguous and often fail to reliably differentiate B. anthracis from its close relatives. When comparing the 16S rRNA gene sequences, a very high degree of agreement can be observed among these species [ 2 ], thus far essentially disfavoring assays for species identification targeting these genetic elements. Similar challenges arise when using techniques such as multi locus sequence typing on members of the B. cereus s.l . group. In fact, most species of this group should be regarded as a single species [ 1 ]. However, some species carry characteristic virulence plasmids on which the genetic information for certain toxins is encoded. These include megaplasmid pCER270 for production of cereulide toxin in a clade of B. cereus sensu stricto strains [ 3 ] or plasmid pXO1 encoding a three-partite AB toxin from B. anthracis better known as lethal and edema toxin, respectively [ 4 ]. These phenotypic characteristics facilitate clinical differentiation, but do not always constitute reliable criteria for rapid identification of individual species. For example, virulence plasmids typical for B. anthracis (pXO1 and pXO2) can also be found in certain B. cereus isolates [ 1 ]. The crucial need for species identification without necessitating live bacteria is typically met by applying molecular methods such as polymerase chain reaction (PCR). For the identification of the tier 1 agent B. anthracis , chromosomal markers such as PL3 [ 5 ], dhp61 (BA5345) [ 6 ] or a nonsense mutation within the plcR -gene are frequently interrogated for [ 7 ]. Plasmids pXO1 and pXO2 are identified by virulence factor genes pagA , lef , cya , capB or capC , respectively [ 5 , 8 , 9 ]. In addition, immunological tests have been established which, due to their sensitivities to specific proteins, can not only detect antibodies after infection but also the pathogen's antigens in the host blood such as the poly-γ- d -glutamic acids forming the bacterial capsule [ 10 ] or the toxin-subunit protective antigen (PA) during acute infection [ 11 ]. However, the challenge of species-specificity remains. Finally, a newer approach, matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS), has proven successful because it facilitates rapid identification of difficult-to-identify pathogens such as B. anthracis [ 12 ]. In contrast to these assays which strongly rely on financial investments in equipment and consumables, the application of the classical bacteriophage (phage) plaque assay is both resource saving and easy to perform. As phages are viruses that only infect target bacteria, some phages have a very narrow host range accepting just a single species or even only a few strains within a species [ 13 , 14 ]. A number of virulent bacteriophages have been described in the literature that infect and multiply in B. anthracis . The most B. anthracis -specific phages can be assigned to the families Siphorviridae [ 15 , 16 ] and Tectiviridae [ 17 , 18 ] and always feature double-stranded DNA (dsDNA) as their genetic material. Brown et al. (1955) discovered the γ (gamma) phage which is a Siphovirus [ 16 ]. Phage γ has been introduced as a standard for plaque assay identification of B. anthracis [ 19 , 20 , 21 ], even though newer work has found a number of additional non- B.-anthracis strains susceptible to the phage [ 15 ]. Another B. anthracis specific phage named Wip1 (worm intestinal phage 1) is from the Tectiviridae family [ 18 ]. This phage was first isolated from the earthworm Eisenia fetida [ 18 ]. Schuch et al. (2010) compared Wip1 and γ phages for their host specificities towards B. anthracis and B. cereus s. s. strains. Remarkably, phage Wip1 achieved higher specificity than the γ phage [ 18 , 22 ]. Another Tectivirus phage that is very specific for B. anthracis is called AP50c [ 17 ]. This lytic phage was derived from temperate parental phage AP50t isolated from soil [ 23 ] and is genomically very similar to phage Wip1 but not identical [ 18 ]. Genome sequencing has revealed that the genome of B. anthracis contains four (inactive) prophages which have been named LambdaBa01-04 [ 24 ]. The presence of these prophages in a genome is also very specific for B. anthracis . This is especially true for LambdaBa01, 03 and 04 which were only found in strains of this species but not in close relatives of the B. cereus s.l. group [ 24 ]. The particular host specificity of phages is usually determined by receptor binding proteins (RBP) which enable the phage to recognize and bind to cell wall structures of the host bacterium [ 13 , 25 ]. In the above-mentioned specific "anthrax" phages, these receptor binding proteins (RBP) comprise the so-called tail (Siphorviridae) or head (Tectiviridae) fibers [ 25 ]. The RBP of phages Wip1 and γ were already provisionally characterized by in silico analysis and subsequent experimentation [ 18 , 26 ] but not yet the RBP of phage AP50c or of prophage LambdaBa03. The structural make-up of the typically homotrimeric RBP is similar in many phages [ 27 , 28 ]. RBP feature two critical domains: at the N -terminus, the RBP is anchored to the phage (head or tail) while the recognition and binding domain is located at the C -terminus of the protein. This binding domain can either confer narrow or broad specificity. The corresponding surface structures of the bacteria (i.e., the receptors), which are responsible for recognition and adsorption of the phage or its RBP can be quite different, including such diverse entities as polysaccharides, teichoic acids, structural or capsule proteins [ 27 , 28 ]. Davison et al. (2005) showed that for the binding of the γ phage, the receptor protein GamR of B. anthracis is essential [ 29 ]. The RBP of γ phage was identified as the product of the gp14 gene on the phage genome [ 26 ]. For phage Wip1 the receptor of B. anthracis has not yet been unambiguously identified but it has been proposed from earlier work that the surface layer protein Sap (surface array protein) is involved in binding by the RBP either directly or indirectly [ 18 ]. The CsaB protein, a cell-surface anchoring protein, was found to be required for phage AP50c adsorption [ 30 ]. Because Sap is anchored by CsaB, Sap is the likely receptor for the B. anthracis specific phage AP50c [ 31 ], yet no indication of the RBP involved was given. From these previous works we further characterized Bacillus (pro)-phage RBP and developed tools to be used in routine DNA-independent, fluorescence microscopic rapid identification of the highly pathogenic bacterium B. anthracis . 2. Materials and Methods 2.1. Bacterial Culture and Inactivation Unless specified differently, B. anthracis strains and other Bacilli were cultivated at 37 °C on tryptic soy agar plates (TSA, Merck KGaA, Darmstadt, Germany) or in 250 mL baffled flasks containing 50 mL tryptic soy broth (TSB, Merck KGaA) with shaking at 110 rpm. All risk group 3 (RG-3) B. anthracis strains were grown in the biosafety level 3 (BSL-3) laboratory at the Bundeswehr Institute of Microbiology (IMB) and then chemically inactivated before further use [ 32 ]. Inactivation of RG 2 strains for subsequent RBP reporter tests was carried out by pelleting 1 mL of a bacterial culture at 5000× g for 3 min and resuspending the cell pellet in aqueous peracetic acid solution (2% Terralin PAA, Schülke & Mayr GmbH, Norderstedt, Germany) or 4% paraformaldehyde (Merck KGaA) and incubating at room temperature for 30 or 60 min, respectively. For heat inactivation another sample was resuspended in PBS and incubated at 98 °C for 30 min (with heated lid cover). After inactivation, all samples were washed twice with PBS. For cultivation of encapsulated B. anthracis cells, a fresh colony from a TSA plate was used to inoculate 5 mL of Luria Bertani (LB) broth (Merck KGaA) containing 0.8% NaHCO 3 in cell culture flaks (Nunc EasYFlask 25 cm 2 ; ThermoFisher Scientific, Darmstadt, Germany) followed by incubation with 10% CO 2 atmosphere at 37 °C for 4 h or overnight. Escherichia coli cultures were grown in LB broth or on Luria Bertani (LB)-agar (Merck KGaA) with 20 µg/mL of gentamycin and 100 µg/mL carbenicillin (Carl Roth, Karlsruhe, Germany) where required. 2.2. Spore Preparation Sporulation and subsequent spore purification of B. anthracis and other Bacilli was done as previously described [ 33 ] with slight modifications. A colony of a fresh overnight culture of B. anthracis (or other Bacilli) was used to inoculate 50 mL sporulation medium containing 0.8% nutrient broth (Merck KGaA) amended with 0.05 mM MnCl 2 , 0.7 mM CaCl 2 and 1.0 mM MgCl 2 [ 34 ] in 500 mL baffled flasks. After incubation at 37 °C and 110 rpm shaking for 72 h, Tween 80 was added to a final concentration of 3% and the culture incubated for another 24 h. The culture was transferred to a 50 mL centrifugation tube and harvested by centrifugation at 2000× g and 20 °C for 10 min. The supernatant was discarded and the pellet washed twice with 25 mL 3% Tween 80 and further incubated on a rotary shaker at 150 rpm for 24 h. Purity of spore suspensions was checked by phase contrast microscopy and spores were harvested by centrifugation (2000× g and 20 °C for 10 min) when purity was above 95% (fewer than 5% vegetative cells present). If purity was less than 95%, spores were washed again with 25 mL 3% Tween 80 and incubated for another 24 h until purity was sufficient. Finally, the spore pellet was resuspended in 3 mL ice-cold ultrapure H 2 O and stored at 4 °C until further use. Spore preparations reached concentrations of up to 10 9 spores per mL. 2.3. Isolation of DNA DNA from Bacilli and bacteriophages was isolated using MasterPure™ Gram Positive DNA Purification kit (Lucigen, Middleton, WI, USA) and DNA concentrations quantified using the Qubit dsDNA HS Assay Kit (ThermoFisher Scientific) according to the manufacturers' protocols. DNA preparations were stored at −20 °C until further use. 2.4. Cloning of RBP-Fusion Constructs For construction of genetic RBP-mCherry -fusions, first the mCherry open reading frame from plasmid mCherry-pBAD (mCherry-pBAD was a gift from Davidson, Shaner and Tsien, University of California at San Diego, La Jolla, CA, USA; Addgene plasmid #54630 (company, Watertown, MA, USA; http://n2t.net/addgene:54630 ; RRID:Addgene_54630; [ 35 ]) was PCR amplified creating overhangs containing restriction sites for Esp3I for cloning into pASG-IBA105 expression plasmid downstream of the twin-strep-tag epitope sequence. Primer overhangs also introduced recognition sites for restriction enzymes SalI, EcoRI and BsrGI as well as XhoI, PstI and BsiWI upstream and downstream of the mCherry gene, respectively, for subsequent insertion of RBP genes. Primer sequences are listed in Table 1 . One-step Esp3I digestion and ligation was carried out using StarGate Direct transfer cloning System (IBA GmbH, Göttingen, Germany) as described in the manufacturer's protocol and plasmids were transformed into NEB Turbo cells (New England Biolabs GmbH, Frankfurt am Main, Germany). Clones were confirmed by Sanger sequencing (Eurofins Genomics Germany GmbH, Ebersberg, Germany) of their recombinant pASG-mCherry plasmids using primers flanking the insert. Next, for generation of mCherry-RBP fusions, respective RBP genes were PCR amplified from purified DNA generating XhoI and BsiWI overhangs and were digested with XhoI and BsiWI alongside pASG-mCherry. After ligation of the fragments, constructs were transformed and plasmid sequences of clones checked as described above. 2.5. Expression and Purification of Strep-Tagged mCherry-RBP Fusion Reporters The pASG-mCherry::RBP plasmids were transformed into E. coli ArcticExpress cells (Agilent Technologies Inc., Waldbronn, Germany). Single colonies harboring recombinant plasmids with fusion constructs were used to inoculate 5 mL of LB medium with gentamycin and carbenicillin in a 50 mL centrifugation tube. After overnight incubation at 37 °C with shaking at 150 rpm, 400 µL of the culture were added to a 1000 mL baffled flask containing 200 mL prewarmed LB medium and incubated at 30 °C with shaking at 110 rpm until the optical density (OD 600 ) of the culture reached 0.6–0.8. Temperature was decreased to 12 °C and gene expression derepressed with a final concentration of 0.2 ng/mL anhydrotetracycline (AHT; IBA GmbH, Göttingen, Germany) for 24 h. The culture was harvested by centrifugation and the cell pellet resuspended in 50 mL lysis buffer containing 100 mM Tris-HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl, 40 µg/mL lysozyme and 1% Halt Protease-Inhibitor Cocktail, EDTA-free (ThermoFisher Scientific). Mechanical lysis was carried out using a French press system (Emulsiflex-C3; Avestin Europe GmbH, Mannheim, Germany) and lysate was then centrifuged at 10,000× g , at 4 °C for 10 min and filtered through a 0.45 µm pore size syringe filter. For subsequent affinity chromatography using the Äkta pure system (GE Healthcare Life Science, Munich, Germany), the cleared lysate was loaded onto 1 mL Streptactin XT columns (IBA GmbH, Göttingen, Germany), washed with 20 mL buffer W (100 mM Tris-HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl) and the protein was eluted with buffer BXT (buffer W containing 50 mM biotin). After dialysis against a 1000-fold volume of HEPES buffer (50 mM HEPES, 50 mM NaCl, 5 mM EDTA, pH 7.5) using SnakeSkin 10K MWCO dialysis membrane (ThermoFisher Scientific), protein concentrations were measured with Pierce BCA Protein Assay Kit (ThermoFisher Scientific). Next, Amicon Ultra 15 Centrifugal Filters 10K MWCO (Merck KGaA) were used to adjust protein concentrations to 1 mg/mL and protein aliquots were directly stored at −80 °C until further use or first amended with 50% glycerol (final concentration) as cryo-protectant and kept at −20 °C for testing in RBP-fusion reporter assays. 2.6. SDS-PAGE and Western Blot Protein samples were mixed with 10× NuPAGE Sample Reducing Agent (ThermoFisher Scientific) and 4× NuPAGE LDS Sample Buffer (ThermoFisher Scientific), denatured at 95 °C for 5 min and applied to a polyacrylamide gel (Novex NuPAGE 4–12% Bis-Tris protein-gel, 1.0 mm, 10-well; ThermoFisher Scientific) using a mini gel tank (ThermoFisher Scientific). SDS-PAGE was performed at 200 V for 60 min with MOPS running buffer (ThermoFisher Scientific). Then the proteins were transferred onto a 0.45 µm pore size nitrocellulose membrane (ThermoFisher Scientific) at 30 V for 75 min via semidry blotting (Novex Semi-Dry Blotter, ThermoFisher Scientific) in NuPAGE transfer buffer. Pierce Reversible Protein Stain Kit (ThermoFisher Scientific) was used to stain whole blotted protein before detection of Strep-tagged proteins, which was carried out using Strep-MAB-Classic (HRP conjugate, IBA GmbH) based chemiluminescence detection and Clarity Western ECL substrate (Bio-Rad Laboratories, Munich, Germany) according to the manufacturers' protocols. A ChemiDoc MP imaging system (Bio-Rad Laboratories) and image Lab 5.2 software (Bio-Rad Laboratories) were used for documentation. 2.7. RBP Testing for Binding to Host Cells An overnight culture of B. anthracis or other Bacilli was used to inoculate 50 mL of fresh TSB in a 250 mL baffled shaking flask to an optical density (OD 600 ) of 0.05 and the culture was incubated at 37 °C and 110 rpm. For growth phase experiments starting from spores, 10 7 spores were used to inoculate 50 mL of brain heart infusion (BHI, Merck KGaA) broth containing 10% fetal bovine serum (Merck KGaA). At different time points, OD 600 was measured and a sample taken equivalent to 100 µL of an OD 600 of 1 (for non-germinated spores as inoculum a volume of 1 mL of the inoculated culture was used as first sample at T 0 ). Treatment was the same for inactivated or encapsulated B. anthracis cells. Samples were pelleted by centrifugation at 5000× g for 2 min in 1.5 mL centrifugation tubes, resuspended in 100 µL of Ringer-HEPES buffer (50 mM HEPES, 1.5 mM CaCl 2 , 1.5 mM KCl, 100 mM NaCl, 0.6 mM NaHCO 3 , pH 7.4) and mixed with 5 µg of purified RBP fusions. After 5 min incubation at room temperature, cells were washed once with Ringer-HEPES (5000× g for 2 min) and 3 µL were transferred into a well of a chamber slide with lid (µ-slide 8 Well, Ibidi GmbH, Martinsried, Germany). When encapsulated cells were tested, samples were mixed with an equal volume of ink prior to transfer to the chamber slide. For proper microscopic analysis of cells, suspensions were covered with a 1 mm thick agar-agar pad serving as a coverslip (1% molten agar-agar solidified between two microscopy slides). Samples were analyzed for cells emitting mCherry signal (extinction: 587 nm, emission: 610 nm) from bound RBP fusions using Axio Observer Z1 700 Confocal Laser Scanning Microscope (Carl Zeiss, Oberkochen, Germany). 2.1. Bacterial Culture and Inactivation Unless specified differently, B. anthracis strains and other Bacilli were cultivated at 37 °C on tryptic soy agar plates (TSA, Merck KGaA, Darmstadt, Germany) or in 250 mL baffled flasks containing 50 mL tryptic soy broth (TSB, Merck KGaA) with shaking at 110 rpm. All risk group 3 (RG-3) B. anthracis strains were grown in the biosafety level 3 (BSL-3) laboratory at the Bundeswehr Institute of Microbiology (IMB) and then chemically inactivated before further use [ 32 ]. Inactivation of RG 2 strains for subsequent RBP reporter tests was carried out by pelleting 1 mL of a bacterial culture at 5000× g for 3 min and resuspending the cell pellet in aqueous peracetic acid solution (2% Terralin PAA, Schülke & Mayr GmbH, Norderstedt, Germany) or 4% paraformaldehyde (Merck KGaA) and incubating at room temperature for 30 or 60 min, respectively. For heat inactivation another sample was resuspended in PBS and incubated at 98 °C for 30 min (with heated lid cover). After inactivation, all samples were washed twice with PBS. For cultivation of encapsulated B. anthracis cells, a fresh colony from a TSA plate was used to inoculate 5 mL of Luria Bertani (LB) broth (Merck KGaA) containing 0.8% NaHCO 3 in cell culture flaks (Nunc EasYFlask 25 cm 2 ; ThermoFisher Scientific, Darmstadt, Germany) followed by incubation with 10% CO 2 atmosphere at 37 °C for 4 h or overnight. Escherichia coli cultures were grown in LB broth or on Luria Bertani (LB)-agar (Merck KGaA) with 20 µg/mL of gentamycin and 100 µg/mL carbenicillin (Carl Roth, Karlsruhe, Germany) where required. 2.2. Spore Preparation Sporulation and subsequent spore purification of B. anthracis and other Bacilli was done as previously described [ 33 ] with slight modifications. A colony of a fresh overnight culture of B. anthracis (or other Bacilli) was used to inoculate 50 mL sporulation medium containing 0.8% nutrient broth (Merck KGaA) amended with 0.05 mM MnCl 2 , 0.7 mM CaCl 2 and 1.0 mM MgCl 2 [ 34 ] in 500 mL baffled flasks. After incubation at 37 °C and 110 rpm shaking for 72 h, Tween 80 was added to a final concentration of 3% and the culture incubated for another 24 h. The culture was transferred to a 50 mL centrifugation tube and harvested by centrifugation at 2000× g and 20 °C for 10 min. The supernatant was discarded and the pellet washed twice with 25 mL 3% Tween 80 and further incubated on a rotary shaker at 150 rpm for 24 h. Purity of spore suspensions was checked by phase contrast microscopy and spores were harvested by centrifugation (2000× g and 20 °C for 10 min) when purity was above 95% (fewer than 5% vegetative cells present). If purity was less than 95%, spores were washed again with 25 mL 3% Tween 80 and incubated for another 24 h until purity was sufficient. Finally, the spore pellet was resuspended in 3 mL ice-cold ultrapure H 2 O and stored at 4 °C until further use. Spore preparations reached concentrations of up to 10 9 spores per mL. 2.3. Isolation of DNA DNA from Bacilli and bacteriophages was isolated using MasterPure™ Gram Positive DNA Purification kit (Lucigen, Middleton, WI, USA) and DNA concentrations quantified using the Qubit dsDNA HS Assay Kit (ThermoFisher Scientific) according to the manufacturers' protocols. DNA preparations were stored at −20 °C until further use. 2.4. Cloning of RBP-Fusion Constructs For construction of genetic RBP-mCherry -fusions, first the mCherry open reading frame from plasmid mCherry-pBAD (mCherry-pBAD was a gift from Davidson, Shaner and Tsien, University of California at San Diego, La Jolla, CA, USA; Addgene plasmid #54630 (company, Watertown, MA, USA; http://n2t.net/addgene:54630 ; RRID:Addgene_54630; [ 35 ]) was PCR amplified creating overhangs containing restriction sites for Esp3I for cloning into pASG-IBA105 expression plasmid downstream of the twin-strep-tag epitope sequence. Primer overhangs also introduced recognition sites for restriction enzymes SalI, EcoRI and BsrGI as well as XhoI, PstI and BsiWI upstream and downstream of the mCherry gene, respectively, for subsequent insertion of RBP genes. Primer sequences are listed in Table 1 . One-step Esp3I digestion and ligation was carried out using StarGate Direct transfer cloning System (IBA GmbH, Göttingen, Germany) as described in the manufacturer's protocol and plasmids were transformed into NEB Turbo cells (New England Biolabs GmbH, Frankfurt am Main, Germany). Clones were confirmed by Sanger sequencing (Eurofins Genomics Germany GmbH, Ebersberg, Germany) of their recombinant pASG-mCherry plasmids using primers flanking the insert. Next, for generation of mCherry-RBP fusions, respective RBP genes were PCR amplified from purified DNA generating XhoI and BsiWI overhangs and were digested with XhoI and BsiWI alongside pASG-mCherry. After ligation of the fragments, constructs were transformed and plasmid sequences of clones checked as described above. 2.5. Expression and Purification of Strep-Tagged mCherry-RBP Fusion Reporters The pASG-mCherry::RBP plasmids were transformed into E. coli ArcticExpress cells (Agilent Technologies Inc., Waldbronn, Germany). Single colonies harboring recombinant plasmids with fusion constructs were used to inoculate 5 mL of LB medium with gentamycin and carbenicillin in a 50 mL centrifugation tube. After overnight incubation at 37 °C with shaking at 150 rpm, 400 µL of the culture were added to a 1000 mL baffled flask containing 200 mL prewarmed LB medium and incubated at 30 °C with shaking at 110 rpm until the optical density (OD 600 ) of the culture reached 0.6–0.8. Temperature was decreased to 12 °C and gene expression derepressed with a final concentration of 0.2 ng/mL anhydrotetracycline (AHT; IBA GmbH, Göttingen, Germany) for 24 h. The culture was harvested by centrifugation and the cell pellet resuspended in 50 mL lysis buffer containing 100 mM Tris-HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl, 40 µg/mL lysozyme and 1% Halt Protease-Inhibitor Cocktail, EDTA-free (ThermoFisher Scientific). Mechanical lysis was carried out using a French press system (Emulsiflex-C3; Avestin Europe GmbH, Mannheim, Germany) and lysate was then centrifuged at 10,000× g , at 4 °C for 10 min and filtered through a 0.45 µm pore size syringe filter. For subsequent affinity chromatography using the Äkta pure system (GE Healthcare Life Science, Munich, Germany), the cleared lysate was loaded onto 1 mL Streptactin XT columns (IBA GmbH, Göttingen, Germany), washed with 20 mL buffer W (100 mM Tris-HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl) and the protein was eluted with buffer BXT (buffer W containing 50 mM biotin). After dialysis against a 1000-fold volume of HEPES buffer (50 mM HEPES, 50 mM NaCl, 5 mM EDTA, pH 7.5) using SnakeSkin 10K MWCO dialysis membrane (ThermoFisher Scientific), protein concentrations were measured with Pierce BCA Protein Assay Kit (ThermoFisher Scientific). Next, Amicon Ultra 15 Centrifugal Filters 10K MWCO (Merck KGaA) were used to adjust protein concentrations to 1 mg/mL and protein aliquots were directly stored at −80 °C until further use or first amended with 50% glycerol (final concentration) as cryo-protectant and kept at −20 °C for testing in RBP-fusion reporter assays. 2.6. SDS-PAGE and Western Blot Protein samples were mixed with 10× NuPAGE Sample Reducing Agent (ThermoFisher Scientific) and 4× NuPAGE LDS Sample Buffer (ThermoFisher Scientific), denatured at 95 °C for 5 min and applied to a polyacrylamide gel (Novex NuPAGE 4–12% Bis-Tris protein-gel, 1.0 mm, 10-well; ThermoFisher Scientific) using a mini gel tank (ThermoFisher Scientific). SDS-PAGE was performed at 200 V for 60 min with MOPS running buffer (ThermoFisher Scientific). Then the proteins were transferred onto a 0.45 µm pore size nitrocellulose membrane (ThermoFisher Scientific) at 30 V for 75 min via semidry blotting (Novex Semi-Dry Blotter, ThermoFisher Scientific) in NuPAGE transfer buffer. Pierce Reversible Protein Stain Kit (ThermoFisher Scientific) was used to stain whole blotted protein before detection of Strep-tagged proteins, which was carried out using Strep-MAB-Classic (HRP conjugate, IBA GmbH) based chemiluminescence detection and Clarity Western ECL substrate (Bio-Rad Laboratories, Munich, Germany) according to the manufacturers' protocols. A ChemiDoc MP imaging system (Bio-Rad Laboratories) and image Lab 5.2 software (Bio-Rad Laboratories) were used for documentation. 2.7. RBP Testing for Binding to Host Cells An overnight culture of B. anthracis or other Bacilli was used to inoculate 50 mL of fresh TSB in a 250 mL baffled shaking flask to an optical density (OD 600 ) of 0.05 and the culture was incubated at 37 °C and 110 rpm. For growth phase experiments starting from spores, 10 7 spores were used to inoculate 50 mL of brain heart infusion (BHI, Merck KGaA) broth containing 10% fetal bovine serum (Merck KGaA). At different time points, OD 600 was measured and a sample taken equivalent to 100 µL of an OD 600 of 1 (for non-germinated spores as inoculum a volume of 1 mL of the inoculated culture was used as first sample at T 0 ). Treatment was the same for inactivated or encapsulated B. anthracis cells. Samples were pelleted by centrifugation at 5000× g for 2 min in 1.5 mL centrifugation tubes, resuspended in 100 µL of Ringer-HEPES buffer (50 mM HEPES, 1.5 mM CaCl 2 , 1.5 mM KCl, 100 mM NaCl, 0.6 mM NaHCO 3 , pH 7.4) and mixed with 5 µg of purified RBP fusions. After 5 min incubation at room temperature, cells were washed once with Ringer-HEPES (5000× g for 2 min) and 3 µL were transferred into a well of a chamber slide with lid (µ-slide 8 Well, Ibidi GmbH, Martinsried, Germany). When encapsulated cells were tested, samples were mixed with an equal volume of ink prior to transfer to the chamber slide. For proper microscopic analysis of cells, suspensions were covered with a 1 mm thick agar-agar pad serving as a coverslip (1% molten agar-agar solidified between two microscopy slides). Samples were analyzed for cells emitting mCherry signal (extinction: 587 nm, emission: 610 nm) from bound RBP fusions using Axio Observer Z1 700 Confocal Laser Scanning Microscope (Carl Zeiss, Oberkochen, Germany). 3. Results 3.1. Cloning of Phage RBP Genes and Production of Recombinant RBP-Fusion Reporters in E. coli When initiating this work we performed sequence similarity database searches in order to identify relatives of RBP from phages γ (protein Gp14) and Wip1 (P23) [ 36 ] and did protein sequence alignments in order to identify possible RBP of phage AP50c ( Figure S1 ). We recognized that a hypothetical protein, BA4079, very similar to RBPγ was encoded on the B. anthracis chromosome located within previously identified prophage LambdaBa03 [ 24 ]. Protein alignment of BA4079 with RBPγ (Gp14) revealed amino acid (aa) identities of 83.0% (similarity 89.0%) across the entire length (500 aa) of the alignment ( Figure S1 ). These values increased to 95.2% and 98.4% when only the C -terminal section of the proteins, comprising a continuous stretch without gaps of 374 aa were reanalyzed. The second information gained from this database search relevant for the study at hand, was that there is not any (hypothetical) protein encoded on the genome of phage AP50c that has significant similarity to the one identified from phage Wip1. However, we recognized a corresponding (hypothetical) polypeptide to Wip1 P24 in phage AP50c. P24 from phage Wip1 was found to be required for RBP Wip (P23) activity [ 18 ]. The respective gene encoding the hypothetical AP50c protein P29 is located directly downstream of a gene for yet another hypothetical protein, P28, without any relatives in the database. When P28 (phage AP50c) was aligned with P23 (phage Wip1) the identity score was low, only 32 out of 151 aa (21.2%) with a similarity of 36.4% but featuring 52 gap positions ( Figure S1 ). Remarkably, the first seven aa residues of both polypeptides (MGLKKPS) were a perfect match. Thus, by genomic position and the short identical stretch to P23, we selected putative protein P28 to be further studied as RBP candidate of phage AP50c. As a basis of a versatile expression plasmid for production of fluorescent reporter fusions, a plasmid chassis was constructed. For this, the PCR-amplified gene of mCherry was cloned in E. coli using expression vector pASG-IBA105, which contains a twin-strep-tag –encoding sequence ( tst ), resulting in pASG-mCherry. The previously identified RBP genes from phages Gamma and Wip1, as well as putative RBP genes from phage AP50c and prophage λ03 were PCR-amplified from genomic DNA and inserted in-frame downstream to the fluorescent protein gene in vector pASG-mCherry, to yield a set of plasmids of the following composition; pASG:: tst :: mCherry :: RBP γ/Wip/AP50/λ03 . In case of constructs harboring RBP from phages Wip1 and AP50c the gene downstream of the RBP gene on the phage genome was cloned as a transcriptional fusion to the RBP gene. This is because in a previous study the necessity of this adventitious protein for RBP function has been demonstrated [ 18 ]. Thus, if not stated otherwise for RBP Wip/AP50 the term RBP comprises two polypeptides in our study (P23+P24 for phage Wip1 and P28+P29 for phage AP50c). Nevertheless, we also included production of P23 or P28, respectively, alone in plasmids pASG:: tst :: mCherry :: P23 Wip /P28 AP50 . Also, we included 5′-truncated versions of the RBP λ03 gene. Aiming at improving solubility of the corresponding protein, coding regions of the following peptides were cloned as well: RBP λ03Δ1-120 , RBP λ03Δ1-139 , RBP λ03Δ1-316 , RBP λ03Δ1-342 . All fusion proteins were produced in E. coli ArcticExpress, as other E. coli expression strains tested were found to produce insoluble proteins mostly incorporated into inclusion bodies. Sizes of fusion proteins purified by affinity chromatography were confirmed by Western blotting as shown in Figure 1 and Figure S2 . Protein yield of RBPγ was low, most of the protein was found as insoluble inclusion bodies. Proteins could also be obtained for RBP λ03Δ 1-139 , RBP λ03Δ1-316 , and RBP λ03Δ1-342 ( Figure S2 ) as well as for P23 and P28 alone. Truncated RBP λ03Δ1-120 was soluble and gave higher protein yields than full-length RBP λ03 ( Figure 1 ). A minor degradation signal was detected for this RBP λ03Δ1-120 by protein staining ( Figure 1 a). This byproduct lacked a TST epitope because it was not visible after TST detection ( Figure 1 b). A faint smaller-sized degradation product of RBP λ03Δ1-120 featuring the TST was observed by Western blotting ( Figure 1 b). More prominently degraded TST-labeled RBP reporters were detected for RBP λ03 and RBPγ ( Figure 1 b). 3.2. (Pro)-Phage RBP Bind to Bacillus anthracis Cells The RBP fusion proteins were next tested for their abilities to bind to B. anthracis cells, especially with regard to putative RBP AP50 P28(+P29), as the RBP of phage AP50c has not been identified thus far. A second emphasis was on putative RBP λ03 and its truncated derivatives. For testing of RBP binding, 2–3 h old vegetative cultures of B. anthracis CDC 1014 were used. The RBP λ03 reporter showed binding to B. anthracis as microscopically detectable fluorescence and cell surfaces were visibly labeled ( Figure 2 ). Of the deliberately truncated RBP λ03 only RBP λ03Δ1-120 was able to bind to cells ( Figure 2 ). Binding to cell surfaces of this truncated RBP λ03Δ1-120 was stronger than that of the full-length parent RBP λ03 ( Figure 2 ). Related RBPγ reporter also yielded signals, however, similar to full length RBP λ03 most of the protein was found in insoluble inclusion bodies. Thus, further testing of low-yield RBPγ and RBP λ03 as well as of non-binding derivatives RBP λ03Δ1-139 , RBP λ03Δ1-316 and RBP λ03Δ1-342 was abandoned in favor of the other RBP reporters including RBP λ03Δ1-120 . Fluorescent labeling of B. anthracis cells was also achieved for RBP AP50 P28(+P29) ( Figure 2 ), which supported our initial hypothesis that P28(+P29) is the actual RBP of phage AP50c. When AP50c P28 was tested by itself (and also when Wip1 P23 was tested by itself) only a very weak binding signal was observed ( Figure S3 ) and thus, P28 and P23 alone were also abandoned for further testing. These results suggest P29 playing a pivotal role for proper function of P28 as RBP. Also, Figure 2 supports our hypothesis that locus BA4079 of prophage Lambda03 encodes for a RBP (RBP λ03 ) and that its truncated derivative RBP λ03; Δ1-120 functions as a B. anthracis reporter. 3.3. Binding of (Pro)-Phage RBP to B. anthracis Cells Is Growth Phase Dependent During the initial RBP reporter binding experiments we observed that RBP fusion proteins exhibited variations in their binding patterns. We reasoned this is most likely dependent on the host's growth phase since we did not use synchronized B. anthracis cultures in initial binding experiments. In particular, RBP λ03Δ1-120 fusion proteins showed declining binding signals on cell surfaces of B. anthracis CDC 1014 or Sterne the longer cultures grew ( Figure 2 and Figure 3 ). To investigate this observation in more detail, growth experiments were carried out for B. anthracis CDC 1014 or Sterne cultures starting from spores. Since spores need time to germinate, differences in RBP binding patterns should occur as a function of cultivation time. RBP binding was monitored from culture samples taken at intervals of typically 30 min during a period of 0 to 8 h including a final 24 h sample. From the micrographs depicted in Figure 3 (time point T 0 min), it can be seen that none of the RBP fusion reporters of (pro)-phages AP50c, LambdaBa03 or Wip1 showed any detectable fluorescence signals when tested on non-germinated spores. It is thus likely that RBP do not bind to spores under the conditions tested. This finding was corroborated by incubating these RBP reporters with spores of B. anthracis Sterne and B. cereus strains 10987 and 4342 as well as B. thuringiensis 10792. Conversely, all RBP reporters produced significant fluorescent signals on cell surfaces of germinated B. anthracis spores at the latest after 120 min, with the RBP λ03; Δ1-120 fusion being the only one that already showed binding after 90 min. The RBP λ03Δ1-120 reporter reached maximum binding signal after 120 min, whereupon the signal remained strong, decreasing after 180 min and was no longer detectable after 240 min. However, this complete lack of binding in later growth phases did not occur in each growth experiment conducted. If, for example, vegetative cells of an overnight culture were inoculated instead of spores, the signal was also retained in later phases and even after 24 h, which was certainly due to the unsynchronized cell division. For the RBP AP50 reporter, the first fluorescence signal on germinating spores was detected after 120 min, which continuously intensified and reached its peak after about 180 to 240 min, whereupon it remained constant for several hours and only became slightly weaker between 8 and 24 h. A similar fluorescence signal pattern was observed when RBP Wip was tested. The strongest binding signal was scored 180 to 240 min after germination was initiated. In the further course the signal became significantly weaker between 8 and 24 h ( Figure 3 ) and featured incompletely distributed, patchy fluorescence signals on the cell surfaces (e.g., RBP Wip at 480 min; also compare Figure 2 ). This "tiger stripes" pattern also appeared yet more weakly on ageing cells labeled with RBP AP50 or RBP λ03Δ1-120 , respectively. In order to show the temporal RBP binding pattern on B. anthracis cells in a semi-quantitative manner, we next correlated RBP reporter signal strength with B. anthracis growth phases during growth experiments (growth curves). Analysis showed that all three RBP reporters feature maximum binding to host cells via fluorescence during logarithmic growth phase of B. anthracis cultures ( Figure 4 ). The earliest response exhibited the RBP λ03Δ1-120 reporter from early to mid-logarithmic growth phase ( Figure 4 ), the latest, RBP Wip , peaking near the end of logarithmic growth ( Figure 4 ). In contrast, the RBP AP50 reporter yielded measurable signals from the mid-logarithmic growth phase, climaxing at late logarithmic-phase to early stationary phase yet remained clearly detectable until the end of the experiments ( Figure 4 ). Thus, it appears that the RBP AP50 reporter was the most versatile for this RBP recognized cells in the widest range of growth phases, except spores and freshly germinated spores (<2 h) (compare Figure 3 , e.g., RBP Wip at 480 min). Next, we compared these results with that of not-synchronized cultures featuring cells of different growth phases. In contrast to that of synchronized cultures, the results here were quite erratic, as would be expected. Some patterns, however, emerged. Binding of the RBP Wip reporter was maximum at the start of the cultures and after 3 to 5 h. RBP λ03Δ1-120 recognized cells best between 1 and 2.5 h. Binding of RBP AP50 was most constant; weaker signals were obtained only around 3 h, 7 h and after 24 h. In contrast to synchronized cultures, weak fluorescence signals could be obtained at any time using any of the three RBP reporters on non-synchronized cultures. 3.4. Inactivated Cells of B. anthracis Can Be Labeled With (Pro)-Phage RBP Reporters Often times it is not possible to work with live cultures of B. anthracis e.g., in field laboratory settings lacking required equipment or in the absence of a fluorescence microscope in BSL-3 facilities. Also, mindful of laboratory work safety, we were curious whether it was possible to use the RBP reporters on inactivated B. anthracis cells. To test this, we evaluated different in-house validated B. anthracis inactivation regimens for suitability of subsequent RBP reporter binding on inactivated cells of B. anthracis strains Sterne or CDC1014. Cultures were inactivated either by heat, formaldehyde or peracetic acid. Cells inactivated by heat yielded strong fluorescence signals upon binding of the RBP AP50 and RBP λ03Δ1-120 reporters similar to fluorescence levels of non-inactivated cells. Conversely, heat-inactivated cells were only poorly labeled by the RBP Wip reporter ( Figure 5 ). Similarly, formaldehyde-inactivation did not prevent the binding of the RBP AP50 and RBP λ03Δ1-120 reporters but the RBP Wip reporter did not bind. In contrast, inactivation with peracetic acid yielded fluorescence signals for all three RBP reporters upon binding to inactivated cells, however, of lower intensities than the controls ( Figure 5 ). Nevertheless, this line of experiments made possible the further use of inactivated B. anthracis cells and of inactivated cells of other pathogenic Bacilli. Thus, we then tested binding of the RBP reporters on inactivated cells of fully virulent B. anthracis isolates of risk group 3 (RG 3) from our collection. These strains were of diverse phylogenetic composition from all three major branches A, B and C [ 37 ]. RBP reporter binding was done on overnight cultures and on fresh, 4 h old cultures inoculated thereof. Cultures of RG 2 strains were inoculated under the same conditions as controls as some of these strains have been used for the experiments described above. All RG 3 strains were successfully labeled by the three RBP reporters yet with different signal strengths ( Table 2 ). Most strains yielded strong fluorescent signals for any of the three reporters, yet cells of the C-branch isolate A1074 were labeled less efficiently. Also, cells of B-branch strains seemed to be accessible to the three RBP reporters, though binding of RBP λ03Δ1-120 was more efficient than binding of RBP AP50 or RBP Wip . A similar pattern was observed for A-branch strain Vollum ( Table 2 ). 3.5. Encapsulated Cells of Bacillus anthracis Can Be Labeled with (Pro)-Phage RBP Reporters When grown in host mammals, B. anthracis produces a poly- d -glutamyl capsule for averting the host's immune response. We tested thus to which extend this capsule would hinder penetration and binding of RBP reporters to B. anthracis cells. For this, pXO2 (capsule) positive B. anthracis strain Ames and the other six RG-3 strains from Table 2 were grown under capsule inducing conditions, after 4 h of growth in fresh inducing culture, cells were inactivated using peracetic acid and negative-stained with ink. The three RBP reporters were added and samples subjected to fluorescence microscopy. Figure 6 shows that the capsule did not prevent cell labeling by the three RBP reporters. All samples gave strong fluorescence signals, clearly showing binding of the RBP reporters amidst the capsule and the bacterial cell as exemplified by the Ames strain ( Figure 6 ). Even the largest of the three RBP reporters, RBP λ03Δ1-120 , was able to label encapsulated cells. Care had to be taken when adjusting the ink concentration, otherwise capsule visualization by negative staining with black ink eclipsed fluorescence signals noticeably. Notwithstanding this caveat, this line of experiments clearly demonstrated that all three RBP reporters, RBP AP50 , RBP λ03Δ1-120 and RBP Wip were capable of labeling encapsulated cells of B. anthracis . 3.6. Binding of (Pro)-Phage RBP Is Specific to B. anthracis Cells Next, we determined RBP reporter binding to a panel of Bacillus strains closely related to B. anthracis . Of 56 non- anthracis Bacillus ssp. tested, most (42%) did not bind any of the three RBP reporters at all and a small number (12%), only very weakly ( Table 3 ; Figure 7 ). Three strains ( B. cereus 3094, B. paranthracis 2002 and B. weihenstephanensis B0293) were significantly labeled by the RBP λ03Δ1-120 reporter, yet clearly yielding a weaker signal than B. anthracis host cells, even distinct from signals of cells of rare B. anthracis C-branch strain A1074 ( Figure 7 ). Cells of a single one of these strains ( B. cereus 3094) was also markedly labeled by RBP AP50 and RBP Wip reporters. Again higher fluorescence signals upon RBP reporter binding were observed when B. anthracis cells (even the few colored cells of C-branch strain A1074 were more uniformly labeled) were used as positive control hosts ( Table 3 ; Figure 7 ). Thus, from analysis of Table 3 and mindful of the results depicted in Figure 7 we suggest specificities of the RBP reporters may be as high as 98% (1 false positive out of 56 non- anthracis Bacilli) for RBP AP50 ( B. cereus 3093) and RBP Wip ( B. cereus 2700) or 95% (three false positives out of 56 non- anthracis Bacilli) for RBP λ03Δ1-120 ( B. cereus 3094, B. weihenstephanensis B0293 and B. paranthracis 2002). 3.1. Cloning of Phage RBP Genes and Production of Recombinant RBP-Fusion Reporters in E. coli When initiating this work we performed sequence similarity database searches in order to identify relatives of RBP from phages γ (protein Gp14) and Wip1 (P23) [ 36 ] and did protein sequence alignments in order to identify possible RBP of phage AP50c ( Figure S1 ). We recognized that a hypothetical protein, BA4079, very similar to RBPγ was encoded on the B. anthracis chromosome located within previously identified prophage LambdaBa03 [ 24 ]. Protein alignment of BA4079 with RBPγ (Gp14) revealed amino acid (aa) identities of 83.0% (similarity 89.0%) across the entire length (500 aa) of the alignment ( Figure S1 ). These values increased to 95.2% and 98.4% when only the C -terminal section of the proteins, comprising a continuous stretch without gaps of 374 aa were reanalyzed. The second information gained from this database search relevant for the study at hand, was that there is not any (hypothetical) protein encoded on the genome of phage AP50c that has significant similarity to the one identified from phage Wip1. However, we recognized a corresponding (hypothetical) polypeptide to Wip1 P24 in phage AP50c. P24 from phage Wip1 was found to be required for RBP Wip (P23) activity [ 18 ]. The respective gene encoding the hypothetical AP50c protein P29 is located directly downstream of a gene for yet another hypothetical protein, P28, without any relatives in the database. When P28 (phage AP50c) was aligned with P23 (phage Wip1) the identity score was low, only 32 out of 151 aa (21.2%) with a similarity of 36.4% but featuring 52 gap positions ( Figure S1 ). Remarkably, the first seven aa residues of both polypeptides (MGLKKPS) were a perfect match. Thus, by genomic position and the short identical stretch to P23, we selected putative protein P28 to be further studied as RBP candidate of phage AP50c. As a basis of a versatile expression plasmid for production of fluorescent reporter fusions, a plasmid chassis was constructed. For this, the PCR-amplified gene of mCherry was cloned in E. coli using expression vector pASG-IBA105, which contains a twin-strep-tag –encoding sequence ( tst ), resulting in pASG-mCherry. The previously identified RBP genes from phages Gamma and Wip1, as well as putative RBP genes from phage AP50c and prophage λ03 were PCR-amplified from genomic DNA and inserted in-frame downstream to the fluorescent protein gene in vector pASG-mCherry, to yield a set of plasmids of the following composition; pASG:: tst :: mCherry :: RBP γ/Wip/AP50/λ03 . In case of constructs harboring RBP from phages Wip1 and AP50c the gene downstream of the RBP gene on the phage genome was cloned as a transcriptional fusion to the RBP gene. This is because in a previous study the necessity of this adventitious protein for RBP function has been demonstrated [ 18 ]. Thus, if not stated otherwise for RBP Wip/AP50 the term RBP comprises two polypeptides in our study (P23+P24 for phage Wip1 and P28+P29 for phage AP50c). Nevertheless, we also included production of P23 or P28, respectively, alone in plasmids pASG:: tst :: mCherry :: P23 Wip /P28 AP50 . Also, we included 5′-truncated versions of the RBP λ03 gene. Aiming at improving solubility of the corresponding protein, coding regions of the following peptides were cloned as well: RBP λ03Δ1-120 , RBP λ03Δ1-139 , RBP λ03Δ1-316 , RBP λ03Δ1-342 . All fusion proteins were produced in E. coli ArcticExpress, as other E. coli expression strains tested were found to produce insoluble proteins mostly incorporated into inclusion bodies. Sizes of fusion proteins purified by affinity chromatography were confirmed by Western blotting as shown in Figure 1 and Figure S2 . Protein yield of RBPγ was low, most of the protein was found as insoluble inclusion bodies. Proteins could also be obtained for RBP λ03Δ 1-139 , RBP λ03Δ1-316 , and RBP λ03Δ1-342 ( Figure S2 ) as well as for P23 and P28 alone. Truncated RBP λ03Δ1-120 was soluble and gave higher protein yields than full-length RBP λ03 ( Figure 1 ). A minor degradation signal was detected for this RBP λ03Δ1-120 by protein staining ( Figure 1 a). This byproduct lacked a TST epitope because it was not visible after TST detection ( Figure 1 b). A faint smaller-sized degradation product of RBP λ03Δ1-120 featuring the TST was observed by Western blotting ( Figure 1 b). More prominently degraded TST-labeled RBP reporters were detected for RBP λ03 and RBPγ ( Figure 1 b). 3.2. (Pro)-Phage RBP Bind to Bacillus anthracis Cells The RBP fusion proteins were next tested for their abilities to bind to B. anthracis cells, especially with regard to putative RBP AP50 P28(+P29), as the RBP of phage AP50c has not been identified thus far. A second emphasis was on putative RBP λ03 and its truncated derivatives. For testing of RBP binding, 2–3 h old vegetative cultures of B. anthracis CDC 1014 were used. The RBP λ03 reporter showed binding to B. anthracis as microscopically detectable fluorescence and cell surfaces were visibly labeled ( Figure 2 ). Of the deliberately truncated RBP λ03 only RBP λ03Δ1-120 was able to bind to cells ( Figure 2 ). Binding to cell surfaces of this truncated RBP λ03Δ1-120 was stronger than that of the full-length parent RBP λ03 ( Figure 2 ). Related RBPγ reporter also yielded signals, however, similar to full length RBP λ03 most of the protein was found in insoluble inclusion bodies. Thus, further testing of low-yield RBPγ and RBP λ03 as well as of non-binding derivatives RBP λ03Δ1-139 , RBP λ03Δ1-316 and RBP λ03Δ1-342 was abandoned in favor of the other RBP reporters including RBP λ03Δ1-120 . Fluorescent labeling of B. anthracis cells was also achieved for RBP AP50 P28(+P29) ( Figure 2 ), which supported our initial hypothesis that P28(+P29) is the actual RBP of phage AP50c. When AP50c P28 was tested by itself (and also when Wip1 P23 was tested by itself) only a very weak binding signal was observed ( Figure S3 ) and thus, P28 and P23 alone were also abandoned for further testing. These results suggest P29 playing a pivotal role for proper function of P28 as RBP. Also, Figure 2 supports our hypothesis that locus BA4079 of prophage Lambda03 encodes for a RBP (RBP λ03 ) and that its truncated derivative RBP λ03; Δ1-120 functions as a B. anthracis reporter. 3.3. Binding of (Pro)-Phage RBP to B. anthracis Cells Is Growth Phase Dependent During the initial RBP reporter binding experiments we observed that RBP fusion proteins exhibited variations in their binding patterns. We reasoned this is most likely dependent on the host's growth phase since we did not use synchronized B. anthracis cultures in initial binding experiments. In particular, RBP λ03Δ1-120 fusion proteins showed declining binding signals on cell surfaces of B. anthracis CDC 1014 or Sterne the longer cultures grew ( Figure 2 and Figure 3 ). To investigate this observation in more detail, growth experiments were carried out for B. anthracis CDC 1014 or Sterne cultures starting from spores. Since spores need time to germinate, differences in RBP binding patterns should occur as a function of cultivation time. RBP binding was monitored from culture samples taken at intervals of typically 30 min during a period of 0 to 8 h including a final 24 h sample. From the micrographs depicted in Figure 3 (time point T 0 min), it can be seen that none of the RBP fusion reporters of (pro)-phages AP50c, LambdaBa03 or Wip1 showed any detectable fluorescence signals when tested on non-germinated spores. It is thus likely that RBP do not bind to spores under the conditions tested. This finding was corroborated by incubating these RBP reporters with spores of B. anthracis Sterne and B. cereus strains 10987 and 4342 as well as B. thuringiensis 10792. Conversely, all RBP reporters produced significant fluorescent signals on cell surfaces of germinated B. anthracis spores at the latest after 120 min, with the RBP λ03; Δ1-120 fusion being the only one that already showed binding after 90 min. The RBP λ03Δ1-120 reporter reached maximum binding signal after 120 min, whereupon the signal remained strong, decreasing after 180 min and was no longer detectable after 240 min. However, this complete lack of binding in later growth phases did not occur in each growth experiment conducted. If, for example, vegetative cells of an overnight culture were inoculated instead of spores, the signal was also retained in later phases and even after 24 h, which was certainly due to the unsynchronized cell division. For the RBP AP50 reporter, the first fluorescence signal on germinating spores was detected after 120 min, which continuously intensified and reached its peak after about 180 to 240 min, whereupon it remained constant for several hours and only became slightly weaker between 8 and 24 h. A similar fluorescence signal pattern was observed when RBP Wip was tested. The strongest binding signal was scored 180 to 240 min after germination was initiated. In the further course the signal became significantly weaker between 8 and 24 h ( Figure 3 ) and featured incompletely distributed, patchy fluorescence signals on the cell surfaces (e.g., RBP Wip at 480 min; also compare Figure 2 ). This "tiger stripes" pattern also appeared yet more weakly on ageing cells labeled with RBP AP50 or RBP λ03Δ1-120 , respectively. In order to show the temporal RBP binding pattern on B. anthracis cells in a semi-quantitative manner, we next correlated RBP reporter signal strength with B. anthracis growth phases during growth experiments (growth curves). Analysis showed that all three RBP reporters feature maximum binding to host cells via fluorescence during logarithmic growth phase of B. anthracis cultures ( Figure 4 ). The earliest response exhibited the RBP λ03Δ1-120 reporter from early to mid-logarithmic growth phase ( Figure 4 ), the latest, RBP Wip , peaking near the end of logarithmic growth ( Figure 4 ). In contrast, the RBP AP50 reporter yielded measurable signals from the mid-logarithmic growth phase, climaxing at late logarithmic-phase to early stationary phase yet remained clearly detectable until the end of the experiments ( Figure 4 ). Thus, it appears that the RBP AP50 reporter was the most versatile for this RBP recognized cells in the widest range of growth phases, except spores and freshly germinated spores (<2 h) (compare Figure 3 , e.g., RBP Wip at 480 min). Next, we compared these results with that of not-synchronized cultures featuring cells of different growth phases. In contrast to that of synchronized cultures, the results here were quite erratic, as would be expected. Some patterns, however, emerged. Binding of the RBP Wip reporter was maximum at the start of the cultures and after 3 to 5 h. RBP λ03Δ1-120 recognized cells best between 1 and 2.5 h. Binding of RBP AP50 was most constant; weaker signals were obtained only around 3 h, 7 h and after 24 h. In contrast to synchronized cultures, weak fluorescence signals could be obtained at any time using any of the three RBP reporters on non-synchronized cultures. 3.4. Inactivated Cells of B. anthracis Can Be Labeled With (Pro)-Phage RBP Reporters Often times it is not possible to work with live cultures of B. anthracis e.g., in field laboratory settings lacking required equipment or in the absence of a fluorescence microscope in BSL-3 facilities. Also, mindful of laboratory work safety, we were curious whether it was possible to use the RBP reporters on inactivated B. anthracis cells. To test this, we evaluated different in-house validated B. anthracis inactivation regimens for suitability of subsequent RBP reporter binding on inactivated cells of B. anthracis strains Sterne or CDC1014. Cultures were inactivated either by heat, formaldehyde or peracetic acid. Cells inactivated by heat yielded strong fluorescence signals upon binding of the RBP AP50 and RBP λ03Δ1-120 reporters similar to fluorescence levels of non-inactivated cells. Conversely, heat-inactivated cells were only poorly labeled by the RBP Wip reporter ( Figure 5 ). Similarly, formaldehyde-inactivation did not prevent the binding of the RBP AP50 and RBP λ03Δ1-120 reporters but the RBP Wip reporter did not bind. In contrast, inactivation with peracetic acid yielded fluorescence signals for all three RBP reporters upon binding to inactivated cells, however, of lower intensities than the controls ( Figure 5 ). Nevertheless, this line of experiments made possible the further use of inactivated B. anthracis cells and of inactivated cells of other pathogenic Bacilli. Thus, we then tested binding of the RBP reporters on inactivated cells of fully virulent B. anthracis isolates of risk group 3 (RG 3) from our collection. These strains were of diverse phylogenetic composition from all three major branches A, B and C [ 37 ]. RBP reporter binding was done on overnight cultures and on fresh, 4 h old cultures inoculated thereof. Cultures of RG 2 strains were inoculated under the same conditions as controls as some of these strains have been used for the experiments described above. All RG 3 strains were successfully labeled by the three RBP reporters yet with different signal strengths ( Table 2 ). Most strains yielded strong fluorescent signals for any of the three reporters, yet cells of the C-branch isolate A1074 were labeled less efficiently. Also, cells of B-branch strains seemed to be accessible to the three RBP reporters, though binding of RBP λ03Δ1-120 was more efficient than binding of RBP AP50 or RBP Wip . A similar pattern was observed for A-branch strain Vollum ( Table 2 ). 3.5. Encapsulated Cells of Bacillus anthracis Can Be Labeled with (Pro)-Phage RBP Reporters When grown in host mammals, B. anthracis produces a poly- d -glutamyl capsule for averting the host's immune response. We tested thus to which extend this capsule would hinder penetration and binding of RBP reporters to B. anthracis cells. For this, pXO2 (capsule) positive B. anthracis strain Ames and the other six RG-3 strains from Table 2 were grown under capsule inducing conditions, after 4 h of growth in fresh inducing culture, cells were inactivated using peracetic acid and negative-stained with ink. The three RBP reporters were added and samples subjected to fluorescence microscopy. Figure 6 shows that the capsule did not prevent cell labeling by the three RBP reporters. All samples gave strong fluorescence signals, clearly showing binding of the RBP reporters amidst the capsule and the bacterial cell as exemplified by the Ames strain ( Figure 6 ). Even the largest of the three RBP reporters, RBP λ03Δ1-120 , was able to label encapsulated cells. Care had to be taken when adjusting the ink concentration, otherwise capsule visualization by negative staining with black ink eclipsed fluorescence signals noticeably. Notwithstanding this caveat, this line of experiments clearly demonstrated that all three RBP reporters, RBP AP50 , RBP λ03Δ1-120 and RBP Wip were capable of labeling encapsulated cells of B. anthracis . 3.6. Binding of (Pro)-Phage RBP Is Specific to B. anthracis Cells Next, we determined RBP reporter binding to a panel of Bacillus strains closely related to B. anthracis . Of 56 non- anthracis Bacillus ssp. tested, most (42%) did not bind any of the three RBP reporters at all and a small number (12%), only very weakly ( Table 3 ; Figure 7 ). Three strains ( B. cereus 3094, B. paranthracis 2002 and B. weihenstephanensis B0293) were significantly labeled by the RBP λ03Δ1-120 reporter, yet clearly yielding a weaker signal than B. anthracis host cells, even distinct from signals of cells of rare B. anthracis C-branch strain A1074 ( Figure 7 ). Cells of a single one of these strains ( B. cereus 3094) was also markedly labeled by RBP AP50 and RBP Wip reporters. Again higher fluorescence signals upon RBP reporter binding were observed when B. anthracis cells (even the few colored cells of C-branch strain A1074 were more uniformly labeled) were used as positive control hosts ( Table 3 ; Figure 7 ). Thus, from analysis of Table 3 and mindful of the results depicted in Figure 7 we suggest specificities of the RBP reporters may be as high as 98% (1 false positive out of 56 non- anthracis Bacilli) for RBP AP50 ( B. cereus 3093) and RBP Wip ( B. cereus 2700) or 95% (three false positives out of 56 non- anthracis Bacilli) for RBP λ03Δ1-120 ( B. cereus 3094, B. weihenstephanensis B0293 and B. paranthracis 2002). 4. Discussion The confirmatory specific identification of B. anthracis is often achieved by means of antigen–antibody interaction, be it in the form of enzyme-linked immunosorbent assays (ELISA) [ 38 ], lateral flow assays [ 39 , 40 ] or by the use of fluorescently labeled antibodies in microscopy [ 41 ] (further alternative detection techniques for B. anthracis are reviewed in a contribution to the special issue "An Update on Anthrax" of Microorganisms [ 42 ]). The direct fluorescent-antibody (DFA) assay [ 41 ] may be seen as being related to the study at hand insofar as both methods take advantage of fluorescence reporters for detecting presumptive B. anthracis cultures and thus helping confirming the identity of this notorious biothreat agent. In contrast to RBP fused to fluorescent protein for visual labeling of target cells, recognition via DFA was achieved using two fluorescent dye-labeled monoclonal antibodies, one specifically directed to the galactose/ N -acetylglucosamine polysaccharide cell wall antigen, the other one recognizing the capsule antigen. Only when these two antibody reporters were combined, the DFA assay reached high specificity. Of 230 B. anthracis isolates tested 227 were positive (99% specificity) and 56 of 56 non- anthracis Bacillus strains were found to be negative. This DFA assay is fast, taking less than 1 h for completion [ 41 ], thus requiring only moderately more time than the RBP reporter assay described here (see graphical abstract for a visualization of the timescale of the assay). The use of RBP (and other phage-derived proteins) for detection and identification of bacteria is not a new methodology; its utility has been extensively reviewed [ 13 , 43 , 44 ]. Depending on the specificities of utilized phage proteins, RBP can even be used for the identification and separation of different isolates of the same species according to different O-antigens on the surface of Listeria monocytogenes cells [ 45 ]. The situation might be viewed as similar to the situation of B. anthracis within the B. cereus s.l. group of bacteria. Taking into account the very close genetic relationship of B. anthracis to B. cereus , B. anthracis may also be considered a subspecies within the B. cereus s.l. group [ 46 ]. Thus, our RBP reporters would detect a subspecies as well, in this case B. anthracis . In contrast to the canonical B. anthracis typing phage γ [ 21 ], phage Wip1 showed no binding to cells of an untypical strain of B. cereus (strain ATCC 4342) [ 18 , 22 ]. This result correlates with our findings regarding lack of recognition of RBP Wip to cells of this B. cereus isolate. Kan et al. (2013) in their infection and adsorption tests showed affinity of the Wip1 phage for the B. cereus CDC 2000032805 strain [ 47 ], which is also a host for the γ phage [ 18 ]. Since this B. cereus CDC 2000032805 strain was not available to us, RBP reporter binding to this strain could not be tested and thus we were not able to assess if RBP Wip (or RBP AP50 ) would merely show marginal binding as did a small group of the other Bacilli tested ( Table 3 ) or whether this untypical host would be recognized as efficiently as B. anthracis . However, recognition of phage Wip1 RBP proteins P23(+P24) to cells of strain CDC2000032805 was shown previously as a patchy fluorescence pattern [ 18 ]. Thus, strain CDC2000032805 must be added to list of B. cereus s.l . strains able to be partly recognized by RBP Wip . Phage AP50c infected 111 of 115 tested B. anthracis strains (~97%) except for e.g., a Sap mutant of Sterne strain and none of 100 B. cereus sensu lato strains [ 17 ]. Remarkably, in the same study two out of three different Vollum derivatives also were insensitive to phage AP50c. With today's knowledge on the receptor of phage AP50 [ 31 ], these are derivatives that have likely lost their S-layers. Later, two additional B. cereus strains were found to be sensitive to phage AP50c, B. cereus RS438 (CDC2000032805) and B. cereus RS756 (better known as E33L ZK; Zebra Killer [ 48 ]), with efficiencies of plating reduced by about one third compared to a B. anthracis Sterne derivative host. Two additional strains ( B. thuringiensis serovar pulsiensis BGSC 4CC1 and B. thuringiensis serovar monterrey BGSC 4AJ1) allowed adsorption of the phage but not propagation. In contrast, B. cereus ATCC 4342 sensitive to phage γ, was insensitive to phage AP50c [ 31 ]. In the same study, Sap was identified as likely receptor of phage AP50 [ 31 ]. This finding was supported by a parallel study in which by analysis of mutants that failed to support AP50 propagation, the CsaB protein was found to be required for phage AP50 host recognition and adsorption. CsaB is required for cell-surface anchoring of the S-layer and thus including Sap [ 30 ]. Our and earlier observations [ 18 ] of "tiger stripes" binding patterns of RBP reporters ( Figure 3 ) that depend on host cell growth phases support earlier notions [ 18 ]. In their thorough characterization of the phage Wip1 genome and several protein functions, the authors suspected a temporal change in S-layer proteins Sap to EA1 (extractable antigen 1 encoded by the eag gene) when cultures of B. anthracis exit from logarithmic into stationary phase [ 49 , 50 ] and that this was the underlying reason for diminished binding of RBP P23 in their study [ 18 ]. Likely, the Sap protein may also be the reason why B. cereus strains 2700, 3093 and 3094 cross-reacted weakly with our AP50 and Wip reporters ( Table 3 ). Some B. cereus strains possess Sap proteins that have a high similarity to Sap of B. anthracis [ 31 ]. Likewise, binding of RBP λ03Δ1-120 reporter to several B. cereus strains ( Table 3 ) may be based on similarities of their GamR receptor (i.e., the receptor of phage γ and likely also the receptor for RBP λ03 ) with that of B. anthracis [ 29 ]. Because RBP reporter binding is dependent on the presence of the cognate receptors (Sap or GamR, respectively) on the bacterial host cell surface, recognition is not dependent on the presence or absence of B. anthracis virulence plasmids (pXO1 and pXO2). Thus, B. cereus strains harboring such plasmids [ 1 ] cannot per se be expected to be labeled by the RBP reporters. Conversely, the rather weak binding of RBP reporters to cells of rare B. anthracis C-branch strain A1074 may be linked to the overall poor growth of this strain in our hands on both solidified and liquid media. Interestingly, phage Bam35, a Tectivirus of B. thuringiensis genomically closely related to phages Wip1 and AP50 does not encode for proteins related to P23 or P28, respectively [ 51 ]. Attempts to identify the RBP of this Bam35 phage have thus far been unsuccessful. The protein encoded by a gene occupying the same location on the phage Bam35 genome (P29) as P23/P24 (of phage Wip1) [ 18 ] or P28/P29 (of phage AP50) did not bind to host cells, though this protein is very likely positioned on the surface of Bam35 virons [ 52 ]. Instead this phage seems to utilize different means of host cell recognition. Experiments with peptidoglycan isolated from Bacilli and E. coli demonstrated that N -acetyl-muramic acid in the bacterial cell wall is required for binding of phage Bam35 [ 52 ]. Thus, even in genomically closely related Bacillus Tectivirus phages it is not always that straightforward to identify (i) the phage RBP and (ii) the cellular receptor recognized by this RBP. Though we abandoned early experiments with phage γ RBP (Gp14) reporters because of protein yield and solubility issues, this protein nevertheless bound to cells of B. cereus ATCC 4342 ( Figure S4 ). This result agrees with earlier results on a different host cell wall binding protein, the endolysin PlyG of phage γ [ 53 ]. This PlyG is produced from phage-infected cells right before new virions are to be released from the dying host cell. PlyG depolymerizes the peptidoglycan from within after PlyG is transported across the cytoplasmic membrane. However, PlyG can also act from without, digesting the B. anthracis cell wall when added to growing cells [ 53 ]. The authors found PlyG to be highly specific for B. anthracis cells; only two non-B. anthracis Bacilli were targeted: strain B. cereus RSVF1 (identical to strain 4342 [ 53 ]) and B. cereus ATCC 10987. These two were among the isolates our RBP λ03Δ1-120 reporter recognized weakly ( Table 3 ; Figure 7 ). Though endolysins are typically investigated for as alternative antimicrobial compounds [ 54 ], the terminal cell wall binding domain (CBD) of PlyG was later used for B. anthracis detection as a bioprobe [ 55 ]. The bioprobe assay was tested for specificity on two B. anthracis and 17 Bacillus spp. strains, however, atypical B. cereus strains such as strain 4342 were not included. Notably, and in concurrence with our results using RBP reporters for detection of B. anthracis , PlyG-CDB was able to detect encapsulated cells, however, spores were not detected unless germination was induced first. While this PlyG-CBD detection seemed to be specific, the detection assay took a couple of hours to complete [ 55 ]. This PlyG-based detection assay was later further developed by shortening the PlyG-CDB down to 20, 15 or 10 aa residues and by including attached fluorescent Qdots for microscopic analysis. Remarkably, even the shortest derivative was able to bind to B. anthracis Sterne cells but not to cells of three other B. cereus s.l. strains tested [ 56 ]. However, similar to our RBP λ03 reporter, cells of B. cereus strain 4342 were also labeled. Further, while our new RBP reporter assays take about 10 min to perform from harvesting cell cultures to fluorescence microscopy (see graphical abstract for details), PlyG-Qdots detection took at least 3 h because two 90 min incubation steps are required [ 56 ]. In order to accelerate phage-based detection and identification of B. anthracis such methods have also seen significant improvements. One is combining phage-amplification in its natural host coupled with phage nucleic acid amplification by PCR [ 57 ]. This assay can be expected to be as specific as the classical plaque assay for the oligonucleotide primers used were tested for specificity against a DNA negative panel. A short phage propagation period is followed by signal (DNA) amplification by real time PCR. This approach shortened the total assay time to about 5 h (including 4 h of growth of host and phage) [ 57 ]. The host-mediated phage amplification/PCR amplification hybrid identification approach has recently seen methodological improvements. In an improved technique named phage-mediated molecular detection (PMMD) a short incubation period of bacterial host culture ( Staphylococcus aureus or B. anthracis ) with a species-specific phage is followed by RNA extraction and reverse transcription PCR (RT-PCR) on specific phage transcripts [ 58 ]. The authors thus took advantage of the high number (relative to DNA) of phage RNA molecules produced per infected cell, which can be expected to far exceed the number of nascent phage DNA genomes. Indeed, the concentration of phage RNA after host infection was sufficient for the generation of strong signals. In this assay B. anthracis was grown prior to RT-PCR for 3 h without phage followed by an infection phase of about 13 min and RNA-preparation. A further advantage of this technique is that it can be coupled to antibiotic susceptibility testing [ 58 ]. Nevertheless, in contrast to the new RBP reporter assay, B. anthracis detection by PMMD requires growth of live bacteria not always possible, especially in field settings [ 59 ], whereas RBP reporters introduced here, may be also used on inactivated cells if required. 5. Conclusions In this work, we developed RBP proteins of several (pro)-phages of B. anthracis into microscopy-based detection tools. In doing so we identified two new RBP from phage AP50c and chromosomally integrated prophage LambdaBa03. Detection can be achieved within about 10 min when live cells of B. anthracis are used, yet the assay also works very well on inactivated and on encapsulated cells. The assay is very specific, especially in the case of the RBP reporters constructed from RBP of phages AP50c and Wip1, while RBP λ03Δ1-120 exhibited a slightly broader host range basically following the specificities of their parental phages. Of note, however, our RBP reporter assay is a qualitative rather than a quantitative detection method requiring a fluorescence microscope. Because of its rapidity and specificity we envision this RBP reporter assay to be able to supplant the original phage based plaque-assay for confirmative pathogen identification in laboratories with access to fluorescence microscopy.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8980399/

Seroprevalence of sheeppox and goatpox virus in Asia and African continent: A systematic review and meta-analysis (Scientometrics)

Background and Aim: Two endemic capripox infectious diseases, sheeppox (SP) and goatpox (GP) are common in Asia, Africa, and the Middle East. Sheep and goats, in general, are considered current assets of small and marginal farmers and have significant economic value in terms of meat, wool, and skin/hide production. Sheep and goat populations in India total 148.88 million and 74.26 million, respectively. Capripox caused US$ 2.3 million (Indian Rupee [INR] 105 million) in economic damages in Maharashtra (India) alone, and it took over 6 years for a flock to recover from the outbreak. The projected yearly loss at the national level is US$ 27.47 million (INR 1250 million). As a result, Capripox diseases put small and marginal farmers under much financial strain. The present study estimates the seroprevalence of SP and GP diseases in the Asian and African continents using systematic review and meta-analysis. The results of the study will help researchers and policymakers to understand the spatial and temporal distribution of the disease and its burden. In addition, the results are also helpful to design and implement location-specific prevention and eradication measures against these diseases. Materials and Methods: Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines of Cochran collaborations were used for systematic review and subsequently meta-analysis were used. The literature was collected from various databases. Initial search string resulted in more than nine thousand articles for the period 2000 to 2020 using the different combinations of keywords and Boolean operators (or not) asterisk* and quotation marks. Out of 9398 papers, 80 studies were chosen for complete test reviews and quality bias evaluation using the inclusion and exclusion criteria. Finally, 21 articles were used for the meta-analysis. The statistical study employed fixed effects and random effects models using R. Results: Seroprevalence of SP and GP was calculated using studies with a cumulative sample size of 4352, out of which sheep and goats' samples together contribute 48%, followed by sheep (32%) and goat (21%). The result of the meta-regression revealed that detection techniques had a significant impact on the overall effect size at 5% level (Qm=14.12). Subgroup analysis of polymerase chain reaction (PCR) test with samples was further grouped into two categories based on the median, and it revealed that 62% of samples used PCR as a detecting test followed by group-II. Conclusion: From the study, it is concluded that SP and GP diseases are highly prevalent; hence, effective vaccines, proper education to farmers through extension activity, and transboundary disease movement restriction are necessary for the control and eradication of the disease. Background and Aim: Two endemic capripox infectious diseases, sheeppox (SP) and goatpox (GP) are common in Asia, Africa, and the Middle East. Sheep and goats, in general, are considered current assets of small and marginal farmers and have significant economic value in terms of meat, wool, and skin/hide production. Sheep and goat populations in India total 148.88 million and 74.26 million, respectively. Capripox caused US$ 2.3 million (Indian Rupee [INR] 105 million) in economic damages in Maharashtra (India) alone, and it took over 6 years for a flock to recover from the outbreak. The projected yearly loss at the national level is US$ 27.47 million (INR 1250 million). As a result, Capripox diseases put small and marginal farmers under much financial strain. The present study estimates the seroprevalence of SP and GP diseases in the Asian and African continents using systematic review and meta-analysis. The results of the study will help researchers and policymakers to understand the spatial and temporal distribution of the disease and its burden. In addition, the results are also helpful to design and implement location-specific prevention and eradication measures against these diseases. Materials and Methods: Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines of Cochran collaborations were used for systematic review and subsequently meta-analysis were used. The literature was collected from various databases. Initial search string resulted in more than nine thousand articles for the period 2000 to 2020 using the different combinations of keywords and Boolean operators (or not) asterisk* and quotation marks. Out of 9398 papers, 80 studies were chosen for complete test reviews and quality bias evaluation using the inclusion and exclusion criteria. Finally, 21 articles were used for the meta-analysis. The statistical study employed fixed effects and random effects models using R. Results: Seroprevalence of SP and GP was calculated using studies with a cumulative sample size of 4352, out of which sheep and goats' samples together contribute 48%, followed by sheep (32%) and goat (21%). The result of the meta-regression revealed that detection techniques had a significant impact on the overall effect size at 5% level (Qm=14.12). Subgroup analysis of polymerase chain reaction (PCR) test with samples was further grouped into two categories based on the median, and it revealed that 62% of samples used PCR as a detecting test followed by group-II. Conclusion: From the study, it is concluded that SP and GP diseases are highly prevalent; hence, effective vaccines, proper education to farmers through extension activity, and transboundary disease movement restriction are necessary for the control and eradication of the disease. Introduction Capripoxvirus (Poxviride) is a notifiable disease [ 1 ] due to its impact on small ruminant productivity. The family Poxviridae includes sheeppox virus (SPPV), goatpox virus (GTPV), and Lumpy skin diseases virus, and all these viruses are closely related serologically and have indistinguishable share over 96% sequence homology among themselves [ 2 ]. SPPV and GTPV are -borne diseases that primarily affect sheep and goats, respectively [ 3 ]. In general, both SPPV and GTPV are host-specific, but some strains are ineffective to heterologous hosts also. However, recently molecular basis of host specificity study shows some genes are associated with conferring host preference [ 4 , 5 ]. The virus spreads through aerosols generated from infected animals or through direct abraded skin/mucosal contact or indirectly through mechanical transmission by vectors [ 1 , 6 ]. There is no evidence that human infection with capripoxvirus and capripoxvirus is not pathogenic to humans [ 2 , 7 ]. The clinical signs in SPPV and GTPV are characterized by mild to severe fever, erythematous macules, vesicles, papules, pustules, and scabs on the skin. The lesions may also develop on the mucous membrane and internal organs, causing respiratory signs, diarrhea, depression, emaciation, and abortion [ 2 , 5 ]. The mortality rate is highly variable (ranging from 5 to 10% in local breeds in endemic regions, to 100% of exotic sheep and goat species in endemic regions [ 8 ], SPPV and GTPV outbreaks have become a major concern in Northern and Central Africa, the Middle East, and most of the Asian continent [ 2 , 3 ]. In India, goatpox (GP) outbreak was first reported in the year 1936, and sheeppox (SP) was first reported in Bombay (1931-1932) and Mysore [ 9 ]. Since then, frequent outbreaks have been reported from several states, causing significant economic losses [ 8 , 10 - 12 ]. A recent outbreak was reported in Himalayan gorals during April-May 2018 in the Tawang district of Arunachal Pradesh [ 13 ]. In the case of the African continent, Ethiopia is believed to have the largest livestock population with more than 49 million sheep and goat population [ 14 ]; a total of 663 SP and GP disease outbreaks were reported in all major parts of Eastern Amhara Region, Ethiopia between 2013 and 2019. From these outbreaks, 57,638 sheep and goats contracted the disease. Out of the 57,638 sick sheep and goats, 6401 animals died [ 15 ]. A study was conducted on herds and flocks affected between August 2017 and January 2018 in Bauchi, Nigeria, revealed that incidence risk and fatality rate were 53 and 34% in sheep; 50 and 33% in goats, respectively Limon et al . [ 16 ]. Seroprevalence of SP and GP in three states of Northern Nigeria (Bauchi, Kaduna, and Plateau) was 2% at a 95% confidence interval (CI) [ 17 ]. Recently outbreaks have been reported in Kazakhstan, Mongolia, Azerbaijan, Turkey, Greece, and Bulgaria of the Asian continent [ 18 ]. Enzootic capripox is responsible for direct and indirect economic loses in sheep and goats and it causes productivity losses in endemic areas, including reduced milk yield, reduced weight gain, increased abortion rates, damage to wool and hide, and increased susceptibility to pneumonia and flystrike along with mortality [ 19 ]. In addition to the above, it also reduced the free trade of animals and animal products from endemic areas to other places [ 12 ]. The economic losses due to capripox in Maharashtra (India) state alone was US$ 2.3million (Indian Rupee [INR] 105 million) and it took nearly 6 years for a flock to recover from capripox outbreak [ 20 ]; estimated annual loss at a national level is US$ 27.47 million (INR 1250 million) [ 20 ]. India has 148.88 million and 74.26 million populations of sheep and goat, respectively [ 20 ]. In the case of Ethiopia, nearly 5-7 million sheep and goat die annually and the estimated economic loss to meet industry is US$ 400 million annually [ 21 ]. Hence, considering the sheep and GP's economic importance is crucial for estimating the disease burden in the region. Keeping the above factors in mind, the present epidemiological study estimates the prevalence of SP and GP diseases in Asian and African continents based on systematic review and meta-analysis. The result could be an input for the researcher and policymakers about the disease (SGP) burden, thereby supporting the process of identification of priorities in veterinary healthcare, prevention, and policy. Materials and Methods Ethical approval Ethical approval was not necessary as the study materials were collected through the public literature database. Study period and location The present study included the literature which were collected from various databases for the period 2000 to 2020 in Asia and Africa. Literature search strategy Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines of Cochran collaborations were used for systematic review and subsequently meta-analysis [ 22 - 24 ]. A literature survey was performed systematically to collect relevant literature on the prevalence of sheep (SPPV) pox and GTPV in the Asian and African continent. The information was collected from the various databases, including PubMed, Google Scholars, Science Direct, Springer's, Biomed Central, Consortium of e-Resources in Agriculture, research proceedings/compendium of conferences, seminars, symposia, Krishikosh, and other published sources. Initial search string resulted in more than nine thousand articles for the period 2000 to 2020 using the different combinations of keywords [SP, GP, Capripoxvirus, prevalence, seroprevalence, ovine, caprine, small ruminant, epidemiology, and domestic ruminant] along with Boolean operators [or and not] asterisk* and quotation marks [" "] were used to Peer-reviewed publication with the English language were retained. Zotero 5.0 developed by George Mason University, Virginia, USA and Rayyan developed by Qatar Computing Research Institute, Qatar (the Systematic Reviews web app) were used for systematic reviews and compilations [ 25 ]. Study selection and data extraction The schematic representation of the systematic review on seroprevalence of SP and GP in Asia and the African continent is depicted in Figure-1 . Out of the 9398 studies compiled during the literature search; During the preliminary screening phase, studies were excluded based on irrelevance, duplicates, lack of temporal and spatial information, etc. Accordingly, 80 studies were selected for full-text reviews and subjected to quality of bias assessment. Finally, 21 articles were used for the meta-analysis [ 5 , 21 , 26 - 44 ]. The determinants such as author, publication year, research conducted year, region, species (sheep and goat), number of the sample tested, number of positive samples, and tests used for the analysis were extracted from the selected articles. Figure-1 Schematic diagram of selection of articles used for the systematic review of this study. Quality assessment of studies Quality assessment of the studies was done by two investigators independently. Investigator used the seven questions with 5 points Likert scale to judge the quality of each research paper. The maximum score of five indicates a very likely and 1 very unlikely article. The scores of the investigators were further used to calculate the coefficient of the validity with Aiken value [ 44 - 46 ]. V=∑S/[[n*[c-1]]] Where, Aiken V=Validity index S=Scores assigned by each rater minus the lowest score in the used S=r-lo r=rater category selection score lo=lowest scores in the scoring category c=Maximum score in the grading scale n=number of rater V value ranges from 0 to 1, where 1 indicate rater gives 100% consent to included studies concerning structured question included. Statistical modeling and analysis Meta-analysis Meta-analysis is a statistical research process used to assimilate various studies to calculate an overall summary estimate of the study using R open-source scripting software written by R core team (Comprehensive R Archive Network) version 3.2.5 and the R package used was "meta" as reported earlier [ 47 ]. The graphical representation of effect size was done through forest plot or CI plot. In a meta-analysis, predominantly fixed effect and Random effect models are used based on the variation in the studies and inconsistency (I 2 ) values. The random effect model will be used when the heterogeneity among the studies is found statistically significant in combination with inconsistency indices. Quantifying heterogeneity and Inconsistency The degree of heterogeneity in a meta-analysis decides the effort in reaching general interpretations. This degree might be estimated by assessing the variance between the different studies [ 48 , 49 ]. Indices H and I 2 values are usually calculated to summarize the impact of heterogeneity among included studies [ 50 ]. Inconsistency (I 2 ) measures the degree of inconsistency ranging from 0 to 100%. Where I 2 is preferable to test for heterogeneity in judging consistency of evidence and selection of either fixed-effect or random-effect model. If I 2 50% least-moderate heterogeneity, and I 2 >95% indicates high heterogeneity [ 25 , 46 ] in the analysis. Testing of heterogeneity It is important to consider the inconsistency among the studies to calculate heterogeneity. If CI for the results of individual studies (generally depicted graphically using horizontal lines) have poor overlap, this generally indicates the presence of statistical heterogeneity [ 46 ]. This can be calculated using Cochran's Q statistic, Tau square, H value, and p values obtained, and results are given in the last line of the forest plot [ 25 , 51 ]. The calculated Chi-squared (χ 2 , or Chi-square) test is included in the forest plots in Cochrane reviews [ 52 , 53 ] helps to assess whether observed differences in results are compatible with chance alone or not and p0.05) nullifying the effect of publication bias in the study. Figure-3 Publication bias among studies is shown in funnel plot showing asymmetry and heterogeneity. Ethical approval Ethical approval was not necessary as the study materials were collected through the public literature database. Study period and location The present study included the literature which were collected from various databases for the period 2000 to 2020 in Asia and Africa. Literature search strategy Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines of Cochran collaborations were used for systematic review and subsequently meta-analysis [ 22 - 24 ]. A literature survey was performed systematically to collect relevant literature on the prevalence of sheep (SPPV) pox and GTPV in the Asian and African continent. The information was collected from the various databases, including PubMed, Google Scholars, Science Direct, Springer's, Biomed Central, Consortium of e-Resources in Agriculture, research proceedings/compendium of conferences, seminars, symposia, Krishikosh, and other published sources. Initial search string resulted in more than nine thousand articles for the period 2000 to 2020 using the different combinations of keywords [SP, GP, Capripoxvirus, prevalence, seroprevalence, ovine, caprine, small ruminant, epidemiology, and domestic ruminant] along with Boolean operators [or and not] asterisk* and quotation marks [" "] were used to Peer-reviewed publication with the English language were retained. Zotero 5.0 developed by George Mason University, Virginia, USA and Rayyan developed by Qatar Computing Research Institute, Qatar (the Systematic Reviews web app) were used for systematic reviews and compilations [ 25 ]. Study selection and data extraction The schematic representation of the systematic review on seroprevalence of SP and GP in Asia and the African continent is depicted in Figure-1 . Out of the 9398 studies compiled during the literature search; During the preliminary screening phase, studies were excluded based on irrelevance, duplicates, lack of temporal and spatial information, etc. Accordingly, 80 studies were selected for full-text reviews and subjected to quality of bias assessment. Finally, 21 articles were used for the meta-analysis [ 5 , 21 , 26 - 44 ]. The determinants such as author, publication year, research conducted year, region, species (sheep and goat), number of the sample tested, number of positive samples, and tests used for the analysis were extracted from the selected articles. Figure-1 Schematic diagram of selection of articles used for the systematic review of this study. Quality assessment of studies Quality assessment of the studies was done by two investigators independently. Investigator used the seven questions with 5 points Likert scale to judge the quality of each research paper. The maximum score of five indicates a very likely and 1 very unlikely article. The scores of the investigators were further used to calculate the coefficient of the validity with Aiken value [ 44 - 46 ]. V=∑S/[[n*[c-1]]] Where, Aiken V=Validity index S=Scores assigned by each rater minus the lowest score in the used S=r-lo r=rater category selection score lo=lowest scores in the scoring category c=Maximum score in the grading scale n=number of rater V value ranges from 0 to 1, where 1 indicate rater gives 100% consent to included studies concerning structured question included. Statistical modeling and analysis Meta-analysis Meta-analysis is a statistical research process used to assimilate various studies to calculate an overall summary estimate of the study using R open-source scripting software written by R core team (Comprehensive R Archive Network) version 3.2.5 and the R package used was "meta" as reported earlier [ 47 ]. The graphical representation of effect size was done through forest plot or CI plot. In a meta-analysis, predominantly fixed effect and Random effect models are used based on the variation in the studies and inconsistency (I 2 ) values. The random effect model will be used when the heterogeneity among the studies is found statistically significant in combination with inconsistency indices. Quantifying heterogeneity and Inconsistency The degree of heterogeneity in a meta-analysis decides the effort in reaching general interpretations. This degree might be estimated by assessing the variance between the different studies [ 48 , 49 ]. Indices H and I 2 values are usually calculated to summarize the impact of heterogeneity among included studies [ 50 ]. Inconsistency (I 2 ) measures the degree of inconsistency ranging from 0 to 100%. Where I 2 is preferable to test for heterogeneity in judging consistency of evidence and selection of either fixed-effect or random-effect model. If I 2 50% least-moderate heterogeneity, and I 2 >95% indicates high heterogeneity [ 25 , 46 ] in the analysis. Testing of heterogeneity It is important to consider the inconsistency among the studies to calculate heterogeneity. If CI for the results of individual studies (generally depicted graphically using horizontal lines) have poor overlap, this generally indicates the presence of statistical heterogeneity [ 46 ]. This can be calculated using Cochran's Q statistic, Tau square, H value, and p values obtained, and results are given in the last line of the forest plot [ 25 , 51 ]. The calculated Chi-squared (χ 2 , or Chi-square) test is included in the forest plots in Cochrane reviews [ 52 , 53 ] helps to assess whether observed differences in results are compatible with chance alone or not and p0.05) nullifying the effect of publication bias in the study. Figure-3 Publication bias among studies is shown in funnel plot showing asymmetry and heterogeneity. Meta-analysis Meta-analysis is a statistical research process used to assimilate various studies to calculate an overall summary estimate of the study using R open-source scripting software written by R core team (Comprehensive R Archive Network) version 3.2.5 and the R package used was "meta" as reported earlier [ 47 ]. The graphical representation of effect size was done through forest plot or CI plot. In a meta-analysis, predominantly fixed effect and Random effect models are used based on the variation in the studies and inconsistency (I 2 ) values. The random effect model will be used when the heterogeneity among the studies is found statistically significant in combination with inconsistency indices. Quantifying heterogeneity and Inconsistency The degree of heterogeneity in a meta-analysis decides the effort in reaching general interpretations. This degree might be estimated by assessing the variance between the different studies [ 48 , 49 ]. Indices H and I 2 values are usually calculated to summarize the impact of heterogeneity among included studies [ 50 ]. Inconsistency (I 2 ) measures the degree of inconsistency ranging from 0 to 100%. Where I 2 is preferable to test for heterogeneity in judging consistency of evidence and selection of either fixed-effect or random-effect model. If I 2 50% least-moderate heterogeneity, and I 2 >95% indicates high heterogeneity [ 25 , 46 ] in the analysis. Testing of heterogeneity It is important to consider the inconsistency among the studies to calculate heterogeneity. If CI for the results of individual studies (generally depicted graphically using horizontal lines) have poor overlap, this generally indicates the presence of statistical heterogeneity [ 46 ]. This can be calculated using Cochran's Q statistic, Tau square, H value, and p values obtained, and results are given in the last line of the forest plot [ 25 , 51 ]. The calculated Chi-squared (χ 2 , or Chi-square) test is included in the forest plots in Cochrane reviews [ 52 , 53 ] helps to assess whether observed differences in results are compatible with chance alone or not and p0.05) nullifying the effect of publication bias in the study. Figure-3 Publication bias among studies is shown in funnel plot showing asymmetry and heterogeneity. Results Bias assessment The seroprevalence of SP and GP was calculated using a total sample size of 4352, out of which sheep and goats' samples together contribute 2084, followed by sheep (1367) and goat (901). The quality of the studies was assessed and the same is presented in Table-1 , which indicates the score given by the two independent authors with respect to seven items using the Likert scale. Based on the ratings calculated, the Aiken V value for all the studies is more than 0.7, and it indicates that the study quality is acceptable. Table 1 Studies included in the meta-analysis with their quality assessment scores. Study Region Country Study's target population is a representative of the national population with respect to relevant variables? How were the samples selected, randomly or census undertaken? Was the probability of bias was minimal among the studies? Was the data collected directly from the subjects? Was an acceptable case definition used in the study? Was the measured parameter valid and reliable? Mode of sample collection was same for all the studies? Summary on the overall risk of study bias Aiken V Index Average score * Average score * Average score * Average score * Average score * Average score * Average score * Average score * Boshra et al . [ 5 ] Asia Saudi Arabia 4.00 4.50 4.50 4.50 4.50 5.00 4.50 4.50 0.88 Fentie et al . [ 21 ] Africa Ethiopia 4.50 5.00 5.00 5.00 5.00 4.50 5.00 4.86 0.96 Hopker et al . [ 27 ] Asia India 4.50 5.00 5.00 5.00 5.00 4.50 5.00 4.86 0.96 Hota et al . [ 26 ] Asia India 5.00 5.00 5.00 5.00 5.00 4.50 5.00 4.93 0.98 Enan et al . [ 43 ] Africa Sudan 5.00 4.50 4.50 4.50 4.50 5.00 5.00 4.71 0.93 Kali et al . [ 34 ] Africa Algeria 5.00 5.00 5.00 5.00 5.00 5.00 4.00 4.86 0.96 Kardjadj et al . [ 37 ] Africa Algeria 4.50 5.00 4.50 5.00 5.00 5.00 4.50 4.79 0.95 Mansour et al . [ 36 ] Africa Sudan 5.00 4.50 4.50 4.50 4.50 5.00 4.50 4.64 0.91 Masoud et al . [ 42 ] Asia Pakistan 5.00 4.50 4.50 5.00 4.00 4.00 4.50 4.50 0.88 Pham et al . [ 40 ] Asia North Vietnam 4.50 4.00 4.50 5.00 4.00 4.50 4.50 4.43 0.86 Santhamani et al . [ 41 ] Asia India 4.00 4.00 5.00 5.00 5.00 4.50 4.00 4.50 0.88 Soundararajan et al . [ 39 ] Asia India 4.00 4.00 5.00 4.50 5.00 4.50 5.00 4.57 0.89 Ramachandran et al . [ 30 ] Asia India 4.00 5.00 5.00 5.00 5.00 5.00 5.00 4.86 0.96 Kadam [ 33 ] Asia India 4.00 5.00 4.50 4.50 5.00 5.00 5.00 4.71 0.93 Pothiappan et al . [ 38 ] Asia India 4.50 4.50 5.00 5.00 4.50 5.00 5.00 4.79 0.95 Gnanavel [ 31 ] Asia India 5.00 4.50 5.00 4.50 4.50 4.50 5.00 4.71 0.93 Chetan [ 29 ] Asia India 5.00 5.00 4.50 5.00 4.50 5.00 4.50 4.79 0.95 Ijaz et al . [ 32 ] Asia Pakistan 4.50 5.00 5.00 5.00 5.00 5.00 4.50 4.86 0.96 Bolajoka et al . [ 28 ] Africa north central Nigeria 4.50 5.00 4.50 4.50 5.00 4.50 5.00 4.71 0.93 Malmarugan et al . [ 35 ] Asia India 5.00 4.50 4.50 4.50 5.00 5.00 5.00 4.79 0.95 * Average score of two independent authors and Aiken value of 21 articles included in the meta-analysis Meta-regression to identify the factors affecting the heterogeneity Univariate meta-regression was used to identify potential covariates that likely affect the magnitude and direction of the overall estimate of heterogeneity. The result of the meta-regression ( Table-2 ) revealed that detection techniques had a significant impact on the overall heterogeneity at 5% level (Qm=14.12) variables such as region, species, sample size, and year were not statistically significant. The estimated results revealed that the subgroup analysis and sensitivity analysis are required for further fine-tuning of prevalence rates of sheep and GTPV. Table 2 The univariate meta-regression analysis of sheep pox and goat pox virus. Predictors Estimate SE z-value τ 2 I² (%) H² R² (%) Qm p-value Region 0.40 0.14 2.82 0.11 99.15 117.91 6.76 2.67 0.101 Test 0.44 0.17 2.49 0.09 98.91 91.57 25.02 13.27 0.038 * Species 0.67 0.12 5.50 0.11 99.06 106.68 8.14 3.96 0.137 Sample Size 0.00 0.00 -1.66 0.11 99.14 116.76 7.76 2.77 0.095 Year -0.01 0.02 -0.66 0.12 99.28 138.47 0.00 0.43 0.510 * Indicate the 5% level of significance Sub-group and sensitivity analysis Subgroup analysis was performed for the covariates like PCR test with a level of sample size, other tests with sample size and region as they could affect the heterogeneity ( Table-3 ). Subgroup analysis of PCR test with levels of samples size was further grouped into two categories based on the median and it revealed that 62% of the sample size used PCR as detecting test in the Group-I category with 95% CI (0.16; 0.296), I 2 =95% and τ 2 =0.030, followed by Group-II 45% prevalence (95% CI: 0.29; 0.66). In the case of the region, studies showed that the prevalence (Figures- 2 and 4 ) of SP and GP for the study period in Asia was 44% (95% CI, 0.26; 0.63 with I 2 =98% and τ 2 =0.134), followed by Africa with 16% prevalence. Table 3 Prevalence of sheep pox and goat pox were stratified according to diagnostic tests with samples for sub-group analysis. Particulars Prevalence % (95% CI) I 2 (%) τ 2 Model PCR with samples Group I (More than median) 0.62 (0.16:0.29) 95 0.030 Random Effect Model Group II (Less than median) 0.47 (0.29:0.66) 88 0.047 Random Effect Model Other tests with samples Group I (More than median) 0.22 (0.16:0.29) 93 0.007 Random Effect Model Group II (Less than median) 0.09 (0.02:0.20) 92 0.033 Random Effect Model Overall effect 0.31 (0.20:0.43) 97 0.081 Random Effect Model Region with samples Asia 0.44 (0.26:0.63) 98 0.134 Random Effect Model Africa 0.16 (0.05:0.30) 94 0.040 Random Effect Model Other tests include ELISA, Hemagglutination, Radial hemolysis, VNT, clinical examination. PCR=Polymerase chain reaction Figure-4 Forest plot for prevalence of sheeppox and goatpox in Asia and African continent based on test with samples. Bias assessment The seroprevalence of SP and GP was calculated using a total sample size of 4352, out of which sheep and goats' samples together contribute 2084, followed by sheep (1367) and goat (901). The quality of the studies was assessed and the same is presented in Table-1 , which indicates the score given by the two independent authors with respect to seven items using the Likert scale. Based on the ratings calculated, the Aiken V value for all the studies is more than 0.7, and it indicates that the study quality is acceptable. Table 1 Studies included in the meta-analysis with their quality assessment scores. Study Region Country Study's target population is a representative of the national population with respect to relevant variables? How were the samples selected, randomly or census undertaken? Was the probability of bias was minimal among the studies? Was the data collected directly from the subjects? Was an acceptable case definition used in the study? Was the measured parameter valid and reliable? Mode of sample collection was same for all the studies? Summary on the overall risk of study bias Aiken V Index Average score * Average score * Average score * Average score * Average score * Average score * Average score * Average score * Boshra et al . [ 5 ] Asia Saudi Arabia 4.00 4.50 4.50 4.50 4.50 5.00 4.50 4.50 0.88 Fentie et al . [ 21 ] Africa Ethiopia 4.50 5.00 5.00 5.00 5.00 4.50 5.00 4.86 0.96 Hopker et al . [ 27 ] Asia India 4.50 5.00 5.00 5.00 5.00 4.50 5.00 4.86 0.96 Hota et al . [ 26 ] Asia India 5.00 5.00 5.00 5.00 5.00 4.50 5.00 4.93 0.98 Enan et al . [ 43 ] Africa Sudan 5.00 4.50 4.50 4.50 4.50 5.00 5.00 4.71 0.93 Kali et al . [ 34 ] Africa Algeria 5.00 5.00 5.00 5.00 5.00 5.00 4.00 4.86 0.96 Kardjadj et al . [ 37 ] Africa Algeria 4.50 5.00 4.50 5.00 5.00 5.00 4.50 4.79 0.95 Mansour et al . [ 36 ] Africa Sudan 5.00 4.50 4.50 4.50 4.50 5.00 4.50 4.64 0.91 Masoud et al . [ 42 ] Asia Pakistan 5.00 4.50 4.50 5.00 4.00 4.00 4.50 4.50 0.88 Pham et al . [ 40 ] Asia North Vietnam 4.50 4.00 4.50 5.00 4.00 4.50 4.50 4.43 0.86 Santhamani et al . [ 41 ] Asia India 4.00 4.00 5.00 5.00 5.00 4.50 4.00 4.50 0.88 Soundararajan et al . [ 39 ] Asia India 4.00 4.00 5.00 4.50 5.00 4.50 5.00 4.57 0.89 Ramachandran et al . [ 30 ] Asia India 4.00 5.00 5.00 5.00 5.00 5.00 5.00 4.86 0.96 Kadam [ 33 ] Asia India 4.00 5.00 4.50 4.50 5.00 5.00 5.00 4.71 0.93 Pothiappan et al . [ 38 ] Asia India 4.50 4.50 5.00 5.00 4.50 5.00 5.00 4.79 0.95 Gnanavel [ 31 ] Asia India 5.00 4.50 5.00 4.50 4.50 4.50 5.00 4.71 0.93 Chetan [ 29 ] Asia India 5.00 5.00 4.50 5.00 4.50 5.00 4.50 4.79 0.95 Ijaz et al . [ 32 ] Asia Pakistan 4.50 5.00 5.00 5.00 5.00 5.00 4.50 4.86 0.96 Bolajoka et al . [ 28 ] Africa north central Nigeria 4.50 5.00 4.50 4.50 5.00 4.50 5.00 4.71 0.93 Malmarugan et al . [ 35 ] Asia India 5.00 4.50 4.50 4.50 5.00 5.00 5.00 4.79 0.95 * Average score of two independent authors and Aiken value of 21 articles included in the meta-analysis Meta-regression to identify the factors affecting the heterogeneity Univariate meta-regression was used to identify potential covariates that likely affect the magnitude and direction of the overall estimate of heterogeneity. The result of the meta-regression ( Table-2 ) revealed that detection techniques had a significant impact on the overall heterogeneity at 5% level (Qm=14.12) variables such as region, species, sample size, and year were not statistically significant. The estimated results revealed that the subgroup analysis and sensitivity analysis are required for further fine-tuning of prevalence rates of sheep and GTPV. Table 2 The univariate meta-regression analysis of sheep pox and goat pox virus. Predictors Estimate SE z-value τ 2 I² (%) H² R² (%) Qm p-value Region 0.40 0.14 2.82 0.11 99.15 117.91 6.76 2.67 0.101 Test 0.44 0.17 2.49 0.09 98.91 91.57 25.02 13.27 0.038 * Species 0.67 0.12 5.50 0.11 99.06 106.68 8.14 3.96 0.137 Sample Size 0.00 0.00 -1.66 0.11 99.14 116.76 7.76 2.77 0.095 Year -0.01 0.02 -0.66 0.12 99.28 138.47 0.00 0.43 0.510 * Indicate the 5% level of significance Sub-group and sensitivity analysis Subgroup analysis was performed for the covariates like PCR test with a level of sample size, other tests with sample size and region as they could affect the heterogeneity ( Table-3 ). Subgroup analysis of PCR test with levels of samples size was further grouped into two categories based on the median and it revealed that 62% of the sample size used PCR as detecting test in the Group-I category with 95% CI (0.16; 0.296), I 2 =95% and τ 2 =0.030, followed by Group-II 45% prevalence (95% CI: 0.29; 0.66). In the case of the region, studies showed that the prevalence (Figures- 2 and 4 ) of SP and GP for the study period in Asia was 44% (95% CI, 0.26; 0.63 with I 2 =98% and τ 2 =0.134), followed by Africa with 16% prevalence. Table 3 Prevalence of sheep pox and goat pox were stratified according to diagnostic tests with samples for sub-group analysis. Particulars Prevalence % (95% CI) I 2 (%) τ 2 Model PCR with samples Group I (More than median) 0.62 (0.16:0.29) 95 0.030 Random Effect Model Group II (Less than median) 0.47 (0.29:0.66) 88 0.047 Random Effect Model Other tests with samples Group I (More than median) 0.22 (0.16:0.29) 93 0.007 Random Effect Model Group II (Less than median) 0.09 (0.02:0.20) 92 0.033 Random Effect Model Overall effect 0.31 (0.20:0.43) 97 0.081 Random Effect Model Region with samples Asia 0.44 (0.26:0.63) 98 0.134 Random Effect Model Africa 0.16 (0.05:0.30) 94 0.040 Random Effect Model Other tests include ELISA, Hemagglutination, Radial hemolysis, VNT, clinical examination. PCR=Polymerase chain reaction Figure-4 Forest plot for prevalence of sheeppox and goatpox in Asia and African continent based on test with samples. Discussion Contagious viral infections of SP and GP diseases have been reported in different parts of the world, including the Indian subcontinent and Africa [ 2 , 55 ]. It is more common in the arid and semi-arid zone of East Africa and the Horn of Africa, and also it is endemic in Iran, Iraq, Turkey, Egypt, Sudan, Syria, Southeast Russia, Mongolia, India, Pakistan, Afghanistan, Nepal, Vietnam, Chinese Taipei, and China [ 5 , 55 , 56 ] This is in accordance with our study of seroprevalence of SP and GP, which were 44 and 16% in Asia and the African continent, respectively ( Figure-5 ). Figure-5 Forest plot for the prevalence of sheeppox and goatpox in Asia and African continent based on region. Commonly, tests such as immunofluorescence, immunoprecipitation, virus neutralization, PCR, and ELISA are employed to detect viral agent/antibodies in disease-endemic areas [ 57 ]. However, these tests are insufficient for the detection, sequencing, and differentiation of SP and GTPV. Hence, it is vital to develop and validate the latest PCR assays [ 58 - 60 ] real-time PCR, the loop-mediated isothermal amplification of DNA, nanotechnology-based fluoroimmuno assays [ 40 ], ELISAs based on inactivated SPPV antigen [ 61 ], recombinant protein and monoclonal antibody-based ELISAs to enable rapid capripoxvirus diagnosis and surveillance in Asian, the African continent. The present study also revealed that PCR tests have significance with the level of sample size compared to other confirmatory tests. Hence, the present study results are assumed to be significant for the following context; Asia and Africa together contribute more than 75% of the world's sheep and goat population [ 62 ]. Asia alone contributes 43.6 and 55.4% of the world's sheep and goat population, respectively, which is followed by Africa with 30.0 and 38.7% of sheep and goat population, respectively [ 63 ]. The statistics showed the importance of sheep and goats with respect to nutritional (meat and milk) and economic security (hides and wool) to the Asian and African countries. SP and GP are highly infectious viral diseases that cause substantial economic losses by reducing productivity and increased susceptibility to other diseases and trade deficit by reducing free trade of animal and animal products from endemic areas to other places in another way [ 21 , 64 , 65 ]. The major reasons for the prevalence of SP and GP diseases include low production and coverage of the vaccines, poor quality management, transboundary movement of infected animals and animal products, and grazing in a common pasture with poor quality water [ 66 , 67 ]. Conclusion Infectious diseases are among the major factors which limit the production and productivity of small ruminants; SP and GP are prioritized in the list. SP and GP are often associated with high mortality and morbidity and significant economic loss to the producer. Hence, seroprevalence studies are essential to understand the spatial and temporal distribution of the diseases and tests that are very suitable for identifying, controlling, and eradicating the SP and GP diseases. From the systematic review and meta-analysis, it is concluded that regular vaccination with attenuated vaccines, education of farmers through extension activities, and effective transboundary movement restrictions leads to control and eradication of the diseases will be a reality in not in the Asians and African continent but all over the world. To the best of our knowledge, this study is the first to estimate the pooled prevalence of SP and GP in Asia and the African continent using systematic review and meta-analysis. Authors' Contributions KPS: Guided and supervised every step of the work. APB: Collected the data, conducted the analysis, and wrote the manuscript. CS: Analyzed the data. RRA: Drafted the manuscript. ES: Analyzed the data and drafted and edited the manuscript. VS: Extracted the data and edited the manuscript. SPK: Extraction and documentation of data. BRS: Reviewed and edited the manuscript. SSP: Extracted the data and drafted the manuscript. All authors read and approved the final manuscript. Competing Interests The authors declare that they have no competing interests. Publisher's Note Veterinary World remains neutral with regard to jurisdictional claims in published map and institutional affiliation.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9279926/

Laboratory Screening of Control Agents Against Isolated Fungal Pathogens Causing Postharvest Diseases of Pitaya in Guizhou, China

Pitaya, or dragon fruit, is a typical tropical fruit with an appealing taste and diverse health benefits to humans. The plantation of pitaya in Guizhou province in China has greatly boosted the income of local farmers and alleviated poverty. However, the frequent occurrence of postharvest diseases has brought large economic loss. To find a solution, we set out to identify the postharvest disease-causing agents of Guizhou pitaya. Several fungi were isolated from diseased pitaya and identified as species based on the ITS1 sequence similarity. Of them, Penicillium spinulosum , Phoma herbarum , Nemania bipapillata , and Aspergillus oryzae were, for the first time, found to cause dragon fruit disease. In consideration of their prevalence in postharvest fruit diseases, Alternaria alternata H8 and Fusarium proliferatum H4 were chosen as representative pathogens for the drug susceptibility test. Among the tested drugs and plant extracts, 430 g/L tebuconazole and 45% prochloraz were found to be the most potent fungicides against H8 and H4, respectively. The research provides insights into the mechanism and control of postharvest diseases of dragon fruits in Guizhou, China, and thus could be of economic and social significance to local farmers and the government. Introduction Dragon fruit or pitaya ( Hylocereus species ) belongs to the family Cataceae, which is a typical tropical fruit. Since its introduction to Guizhou province in 2001 ( Wang et al., 2018 ), the fruit has been shown to be quite adapted to the local climate and ecology. So far, three varieties, purple dragon, crystal red dragon, and pink dragon, have been introduced and cultivated in Guizhou, and the dragon fruit-planting area in Guizhou province has increased to be the third in China ( Wang et al., 2018 ). With its renowned health benefits to consumers and appealing taste, the locally produced dragon fruit is widely accepted in the domestic market. Dragon fruit cultivation in Guizhou province has greatly boosted the local economy and been lifting local farmers out of poverty. With the planting history and area growing, the incidences of dragon fruit diseases are increasingly more frequent, especially the postharvest diseases, which lead to the decline of the yield and quality of marketable dragon fruits, affecting the economic benefits ultimately. Due to the interruption of nutrient supply, the vitality of the postharvest dragon fruits was weakened, the disease resistance was reduced, and they could be infected easily by pathogenic microorganisms during storage and transportation, resulting in illness. At present, there are kinds of diseases caused by microorganisms such as anthracnose, soft rot, canker, black spot, and wilts in dragon fruit after harvest ( Balendres and Bengoa, 2019 ). Among them, fungal diseases are more common and serious. Dragon fruit anthracnose was usually caused by Colletotrichum gloeosporioides ( Masyahit et al., 2009 ) and C. truncatum ( Guo et al., 2014 ). The pathogens that cause black spot disease of dragon fruit were reported to be Alternaria alternata ( Castro et al., 2017 ) and Bipolaris cactivora ( Tarnowski et al., 2010 ; Ben-Ze Ev et al., 2011 ). Dragon fruit soft rot is found to be caused by Neoscytalidium dimidiatum ( Pan et al., 2021 ) and Gilbertella persicaria ( Guo et al., 2012 ). However, there are few reports focusing on causative agents and control methods for postharvest diseases of dragon fruits in the Guizhou area. In order to investigate the microbial species causing postharvest diseases on dragon fruits in Guizhou, samples of diseased dragon fruits were collected from the Luodian County of Guizhou province in China. Microbes were isolated from diseased fruit tissue and reinoculated on healthy fruits to confirm pathogenicity according to Koch's rule. Pathogenic microorganisms were identified by rDNA-ITS sequence similarity analysis. Finally, a variety of prevention and control reagents were screened for inhibition efficacy against selected pathogens by an indoor experiment. This study provides a helpful understanding of the mechanism of postharvest diseases and control measures for dragon fruits in Guizhou province and the neighboring area. Materials and Methods Sample Collection and Microbial Isolation Samples of diseased dragon fruits were collected in Luodian County in Guizhou Province. Pathogenic microorganisms were isolated by conventional tissue isolation methods in the laboratory. The potato dextrose agar (PDA) plates were used for separation and purification. The isolated strains were stored at −20°C in 40% glycerol. Pathogenicity Test Healthy dragon fruits were selected and soaked in 75% alcohol for 1 min, followed by repeated washes in sterile water, and surface-dried. Mycelia "cakes" (blank agar "cakes" as control) were chopped aseptically from the culture plate and placed onto the surface (non-injury inoculation) and the stabbing wound (stab inoculation) of dragon fruits, which was then allowed to incubate at 28°C, and the status of infection was observed along during incubation. Morphological Characterization of Pathogens The pathogenic microorganisms were cultured on PDA and incubated at 28°C. The colony morphology was checked regularly, and the mycelia were observed under a Model EX30 inverted microscope (Ningbo Shunyu Tech. Co. Ltd., Zhejiang, China). DNA Extraction and Molecular Identification The fungus was cultured in potato dextrose broth (PDB) at 28°C for 3 days, and the mycelia were collected for genomic DNA (gDNA) extraction. Fungal gDNA was extracted according to the users' instruction of the fungal genomic DNA rapid extraction kit (Sangon). The rDNA-ITS1 fragment of the gDNA was PCR-amplified using universal ITS1 primers and subjected to nucleotide sequencing. All the obtained sequences were searched for similarity against the NCBI nucleotide collection (nr) database with default parameters ( https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome ). The phylogenetic tree was constructed using the neighbor-joining method to determine the taxonomic status. Preparation of Tested Fungicides and Plant Extracts All tested fungicides are commercially available. The plant extracts were prepared as such: the plant was air-dried and milled to powder. For the procedure, 10 g of the powder was extracted with 95% ethanol and heated for 4 h with refluxing. The extract was filtered and evaporated under reduced pressure. The residue was re-dissolved with hot water, cooled, and adjusted to a concentration of 500 mg/mL as stock solution. Indoor Screening for Control Agents Tested fungicide or plant extracts (Tween 20 as control) with appropriate quantity was added to melted PDA and cooled to make plates. The plates were then inoculated by placing an inoculum (4 mm disc from cultures of A. alternata and F. proliferatum ) at the center and incubated at the ambient temperature of 28°C. Each test was performed in triplicate. After 6 days of incubation, the colony diameter was recorded, and the inhibition rate was calculated using the following formula: Inhibition rate = d i a m e t e r o f c o n t r o l − d i a m e t e r o f t r e a t m e n t d i a m e t e r o f c o n t r o l − 4 m m × 100 % . Furthermore, the efficacy for the tested control agent was expressed as the half-maximal effective concentration (EC 50 , the concentration at which the tested fungicide reduced mycelial growth by 50%) determined by regression of the inhibition rate against the log10 values of the fungicide concentrations. Sample Collection and Microbial Isolation Samples of diseased dragon fruits were collected in Luodian County in Guizhou Province. Pathogenic microorganisms were isolated by conventional tissue isolation methods in the laboratory. The potato dextrose agar (PDA) plates were used for separation and purification. The isolated strains were stored at −20°C in 40% glycerol. Pathogenicity Test Healthy dragon fruits were selected and soaked in 75% alcohol for 1 min, followed by repeated washes in sterile water, and surface-dried. Mycelia "cakes" (blank agar "cakes" as control) were chopped aseptically from the culture plate and placed onto the surface (non-injury inoculation) and the stabbing wound (stab inoculation) of dragon fruits, which was then allowed to incubate at 28°C, and the status of infection was observed along during incubation. Morphological Characterization of Pathogens The pathogenic microorganisms were cultured on PDA and incubated at 28°C. The colony morphology was checked regularly, and the mycelia were observed under a Model EX30 inverted microscope (Ningbo Shunyu Tech. Co. Ltd., Zhejiang, China). DNA Extraction and Molecular Identification The fungus was cultured in potato dextrose broth (PDB) at 28°C for 3 days, and the mycelia were collected for genomic DNA (gDNA) extraction. Fungal gDNA was extracted according to the users' instruction of the fungal genomic DNA rapid extraction kit (Sangon). The rDNA-ITS1 fragment of the gDNA was PCR-amplified using universal ITS1 primers and subjected to nucleotide sequencing. All the obtained sequences were searched for similarity against the NCBI nucleotide collection (nr) database with default parameters ( https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome ). The phylogenetic tree was constructed using the neighbor-joining method to determine the taxonomic status. Preparation of Tested Fungicides and Plant Extracts All tested fungicides are commercially available. The plant extracts were prepared as such: the plant was air-dried and milled to powder. For the procedure, 10 g of the powder was extracted with 95% ethanol and heated for 4 h with refluxing. The extract was filtered and evaporated under reduced pressure. The residue was re-dissolved with hot water, cooled, and adjusted to a concentration of 500 mg/mL as stock solution. Indoor Screening for Control Agents Tested fungicide or plant extracts (Tween 20 as control) with appropriate quantity was added to melted PDA and cooled to make plates. The plates were then inoculated by placing an inoculum (4 mm disc from cultures of A. alternata and F. proliferatum ) at the center and incubated at the ambient temperature of 28°C. Each test was performed in triplicate. After 6 days of incubation, the colony diameter was recorded, and the inhibition rate was calculated using the following formula: Inhibition rate = d i a m e t e r o f c o n t r o l − d i a m e t e r o f t r e a t m e n t d i a m e t e r o f c o n t r o l − 4 m m × 100 % . Furthermore, the efficacy for the tested control agent was expressed as the half-maximal effective concentration (EC 50 , the concentration at which the tested fungicide reduced mycelial growth by 50%) determined by regression of the inhibition rate against the log10 values of the fungicide concentrations. Results Isolation and Purification of Pathogenic Fungi A number of fungal isolates were separated from diseased dragon fruit tissue, and seven fungal strains were preliminarily established based on colony and mycelium morphology and labeled as H1, H2, H4, H6, H7, H8, and H9. Colony and conidia morphology are shown in Figure 1 . The appearance of the H1 colony on PDA is brown with white slowly growing edges. The spores are elliptical to round with varying sizes but connected in tandem to each other. The H2 colony is whitish gray and grows fast. Its spores are rod-shaped under the microscope and distributed around the mycelia. The H4 strain colony exhibits a white appearance and grows fast with elliptical spores arranged in clusters around the top of the spore stalk. The colony of strain H6 is yellowish-brown with dense hyphae, and the spores appear round when observed under a microscope. The H7 colony is white in appearance and grows slowly on PDA. The spores are irregularly rod-shaped. The H8 colony is white with dense mycelia. The spores are oval under the microscope and densely clustered around spore stalks. The H9 colony is dark green and grows fast; the spores are spherical and connected in tandem to each other to form long chains. FIGURE 1 Morphological characteristics of pathogens and pathogenicity confirmation. (A) Colony morphology on the PDA plate. (B) Microscopic image of conidia taken with ×400 magnification, (C) non-injury inoculation, and (D) injury inoculation; fruits on the right were inoculated with mycelia "cakes," and the ones on the left are blank controls. Pathogenicity of Isolated Microorganisms The isolated strains were re-inoculated onto fruits in two ways, a non-injured inoculation and a stab-injured inoculation. As shown in Figure 1 , all the seven strains can infect dragon fruits with or without an injury. Despite the presence of a stabbing wound, no observable infection occurred in control experiments where blank agar was used for 'inoculation'. Taxonomic Identification The ITS1 segments were PCR-amplified from corresponding genomic DNA extracted from the seven pathogenic fungi. Their nucleotide sequences were used as a query for blast searches, and top hits are listed in Table 1 . According to the search results, the pathogenic strains were roughly assigned as Phoma herbarum H1, Colletotrichum nymphaeae H2, Alternaria alternata H4, Aspergillus oryzae H6, Penicillium spinulosum H7, Fusarium proliferatum H8, and Nemania bipapillata H9. Phylogenetic relationships based on the ITS1 sequence similarity between the strains and selected top hits are illustrated in Figure 2 . TABLE 1 NCBI blast results of samples. Strains Taxonomy Sequence similarity (%) Top hit H1 Phoma herbarum 99 MT367635.1 H2 Colletotrichum nymphaeae 100 MW217266.1 H4 Alternaria alternata 100 MN944587.1 H6 Aspergillus oryzae 99 MH345908.1 H7 Penicillium spinulosum 100 JQ639057.1 H8 Fusarium proliferatum 99 MG543763.1 H9 Nemania bipapillata 99 JQ341104.1 FIGURE 2 Molecular phylogenetic tree based on the rDNA-ITS sequence similarity. Sequence alignment and tree building were performed by MEGA5.0 using the neighbor-joining method, and phylogeny was tested by 500 bootstrap replications. The numbers on branches were calculated as bootstrap values. Accession numbers of ITS1 sequences from isolated strains were, namely, ON514545.1 (H1), ON514546.1 (H2), ON514547.1 (H4), ON514548.1 (H6), ON514549.1 (H7), ON514550.1 (H8), and ON514551.1 (H9). Preliminary Screening of Agro-Agents for Fungicides Against Representative Pathogenic Fungi In consideration of the prevalence of A. alternata and F. proliferatum in fruit disease, the two strains H4 and H8 were chosen as representative pathogenic microorganisms of dragon fruit and tested their susceptibility to a series of agricultural agents. As shown in Table 2 , a total of 20 agro-agents were screened for indoor toxicity at concentrations of 50 μg/ml and 10 μg/ml on A. alternata H4 and F. proliferatum H8. For A. alternata H4, at 50 μg/ml concentration, 10% difenoconazole, 430 g/L tebuconazole, and 3% zhongshengmycin showed the highest inhibition rate, while at 10 μ g/ml concentration, 10% difenoconazole, 430 g/L tebuconazole, 50% iprodione were the three most potent candidates. For F. proliferatum H8, 3% benziothiazolinone, 430 g/L tebuconazole, and 45% prochloraz exhibited the highest toxicity at a concentration of 50 μg/ml, while at the 10 μg/ml level, 10% difenoconazole, 50% iprodione, and 45% prochloraz were the top three candidates. Notably, three microbial preparations were included in the screening; the insecticidal Beauveria bassiana showed the highest inhibition rate at 50 μg/ml concentration, and the fungicidal Bacillus cereus was the most potent at 10 μg/ml. TABLE 2 Inhibitory effects of 20 agro-agents at the concentrations of 50 µg/ml and 10 µg/ml on A. alternata H4 and F. proliferatum H8. Agro-agents Inhibition rate (%) 50 µg/ml 10 µg/ml A. alternata H4 F. proliferatum H8 A. alternata H4 F. proliferatum H8 80% mancozeb 47.39 ± 1.72 33.52 ± 2.34 6.43 ± 1.93 17.96 ± 1.51 10% difenoconazole 92.46 ± 2.85 87.92 ± 3.60 80.76 ± 2.93 82.93 ± 3.37 2% wuyiencin 52.14 ± 2.40 56.36 ± 2.58 46.71 ± 2.91 38.08 ± 2.93 3% benziothiazolinone 73.68 ± 2.78 84.05 ± 3.56 54.44 ± 2.80 46.54 ± 1.38 50% chloroisobromine cyanuric acid 10.61 ± 2.33 12.49 ± 2.56 4.96 ± 2.80 6.01 ± 1.69 50% iprodione 70.25 ± 1.63 74.32 ± 2.79 56.37 ± 3.67 67.54 ± 2.79 50% benzylpenicillin 39.78 ± 1.61 56.12 ± 2.78 5.26 ± 2.70 16.51 ± 1.75 80% ethylicin 76.09 ± 2.85 82.54 ± 2.24 54.11 ± 2.00 61.54 ± 2.41 3% zhongshengmycin 82.90 ± 3.64 75.74 ± 2.54 29.91 ± 1.20 27.04 ± 2.01 75% chlorothalonil 54.32 ± 3.01 68.67 ± 1.89 48.44 ± 2.89 49.20 ± 2.07 50% sulfur·carbendazim 30.25 ± 1.46 19.27 ± 1.04 7.93 ± 2.44 3.58 ± 2.19 20% bismerthiazol 42.98 ± 1.42 52.10 ± 1.27 24.29 ± 2.21 35.12 ± 2.22 72% agricultural streptomycin sulfate 25.62 ± 2.81 18.96 ± 2.33 11.84 ± 2.77 8.49 ± 2.86 70% thiophanate methyl 13.10 ± 1.16 32.64 ± 2.64 5.35 ± 1.02 5.21 ± 2.57 50% kresoxim methyl 82.47 ± 1.92 68.79 ± 2.63 47.47 ± 1.39 64.79 ± 2.63 430 g/L tebuconazole 94.26 ± 301 95.46 ± 1.18 84.72 ± 3.61 75.46 ± 1.73 45% prochloraz 63.25 ± 2.51 98.79 ± 1.73 10.26 ± 1.85 84.26 ± 1.42 Bacillus cereus preparation (8*10 9 spores/g) 48.00 ± 2.06 42.61 ± 2.94 17.19 ± 2.83 20.69 ± 1.84 Beauveria bassiana preparation (2*10 11 spores/g) 51.76 ± 2.99 78.96 ± 3.01 6.14 ± 2.28 3.73 ± 2.39 Verticillium chlamydosporium preparation (2.5 * 10 9 spores/g) 28.82 ± 1.70 17.13 ± 2.12 6.14 ± 2.64 12.25 ± 2.60 Inhibition Effect of 10 Edible and Medicinal Plant Extracts on the Tested Fungus The ethanol extract was obtained from 10 edible and medicinal plants, namely, Houttuynia cordata Thunb, Mentha haplocalyx , Zanthoxylum bungeanum Maxim, Lonicera japonica Thunb, Dendrobium officinale Kimura et Migo, Piper nigrum, Zingiber officinale Roscoe, Gastrodia elata , Schisandra chinensis , and Illicium verum . Some of the plants are renowned for their microbe-inhibitory activity. The inhibition rate results in this study are shown in Table 3 . All 10 ethanolic extracts prepared at a final concentration of 50 mg/ml showed an inhibitory effect on the two fungal pathogens, A. alternata H4 and F. proliferatum H8. In comparison, 50 mg/ml and 80% ethylicin was included in the test. The inhibition rates of Zanthoxylum bungeanum Maxim against A. alternata H4 and F. proliferatum H8 were 68.75 and 75.12%, those of Zingiber officinale Roscoe were 70.14 and 60.48%, and those of Piper nigrum were 69.84 and 60.93%, respectively. The three extracts showed the highest inhibitory effect against both tested pathogens of dragon fruit. Notably, 80% ethylicin at the same concentration exhibited a 100% inhibition rate in both strains. TABLE 3 Determination of the inhibitory effect of 10 Chinese edible and medicinal plant extracts on A. alternata H4 and F. proliferatum H8. Plant extract Inhibition rate (%) Plant extract Inhibition rate (%) A. alternata H4 F. proliferatum H8 A. alternata H4 F. proliferatum H8 Zanthoxylum bungeanum Maxim 68.75 ± 1.15 75.12 ± 2.38 Mentha haplocalyx 7.24 ± 1.03 32.48 ± 1.54 Lonicera japonica Thunb 11.27 ± 1.72 34.84 ± 1.69 Dendrobium officinale Kimura et Migo 37.94 ± 1.35 26.75 ± 1.82 Zingiber officinale Roscoe 70.14 ± 2.16 60.48 ± 1.37 Piper nigrum 69.84 ± 1.19 60.93 ± 1.79 Illicium verum 37.68 ± 1.27 52.48 ± 1.77 Schisandra chinensis 48.37 ± 2.42 35.17 ± 2.80 Houttuynia cordata Thunb 24.64 ± 1.93 45.11 ± 2.60 Gastrodia elata 15.24 ± 1.91 21.68 ± 2.54 80% ethylicin 100 100 Fungicidal Efficacy Test on Representative Fungi With Promising Candidates Based on previousscreening results, several fungicides were selected for the efficacy test. As shown in Table 4 , 430 g/L tebuconazole exhibited the smallest EC 50 at 0.0133 μg/ml, suggesting the highest potency toward A. alternata H4, while 50% iprodione with EC 50 at 3.6840 μg/ml showed the second highest potency. For the pathogen F. proliferatum H8, 45% prochloraz showed the smallest EC 50 at 0.0122 μg/ml, and 430 g/L tebuconazole is the second smallest with EC 50 at 0.0307 μg/ml. It is worth mentioning that the efficacy of the biological agent Bacillus cereus is also tested, which exhibited EC 50 at 81.3915 μg/ml toward F. proliferatum H8. TABLE 4 Determination of the efficacy of screened fungicides to A. alternata H4 and F. proliferatum H8. Pathogenic strain Control agent Regression equation Correlation coefficient EC50 (µg/ml) A. alternata H4 3% benziothiazolinone y = 0.8324x+4.1566 0.9770 10.3089 10% difenoconazole y = 0.7454x+4.2309 0.9871 10.7595 50% iprodione y = 0.8428x+4.5227 0.9934 3.6840 50% kresoxim-methyl y = 0.5347x+4.3594 0.9813 15.7781 80% ethylicin y = 1.0551x+3.4653 0.9956 28.4809 430 g/L tebuconazole y = 0.2501x+5.4687 0.9254 0.0133 F. proliferatum H8 10% difenoconazole y = 0.3704x+5.1064 0.9903 0.5161 45% prochloraz y = 0.8479x+6.6203 0.9013 0.0122 50% iprodione y = 0.9036x+4.2205 0.9953 7.2888 50% kresoxim-methyl y = 0.3352x+4.7836 0.9766 4.4216 80% ethylicin y = 0.9706x+3.8647 0.9959 14.7804 430 g/L tebuconazole y = 0.7926x+6.1988 0.9111 0.0307 Isolation and Purification of Pathogenic Fungi A number of fungal isolates were separated from diseased dragon fruit tissue, and seven fungal strains were preliminarily established based on colony and mycelium morphology and labeled as H1, H2, H4, H6, H7, H8, and H9. Colony and conidia morphology are shown in Figure 1 . The appearance of the H1 colony on PDA is brown with white slowly growing edges. The spores are elliptical to round with varying sizes but connected in tandem to each other. The H2 colony is whitish gray and grows fast. Its spores are rod-shaped under the microscope and distributed around the mycelia. The H4 strain colony exhibits a white appearance and grows fast with elliptical spores arranged in clusters around the top of the spore stalk. The colony of strain H6 is yellowish-brown with dense hyphae, and the spores appear round when observed under a microscope. The H7 colony is white in appearance and grows slowly on PDA. The spores are irregularly rod-shaped. The H8 colony is white with dense mycelia. The spores are oval under the microscope and densely clustered around spore stalks. The H9 colony is dark green and grows fast; the spores are spherical and connected in tandem to each other to form long chains. FIGURE 1 Morphological characteristics of pathogens and pathogenicity confirmation. (A) Colony morphology on the PDA plate. (B) Microscopic image of conidia taken with ×400 magnification, (C) non-injury inoculation, and (D) injury inoculation; fruits on the right were inoculated with mycelia "cakes," and the ones on the left are blank controls. Pathogenicity of Isolated Microorganisms The isolated strains were re-inoculated onto fruits in two ways, a non-injured inoculation and a stab-injured inoculation. As shown in Figure 1 , all the seven strains can infect dragon fruits with or without an injury. Despite the presence of a stabbing wound, no observable infection occurred in control experiments where blank agar was used for 'inoculation'. Taxonomic Identification The ITS1 segments were PCR-amplified from corresponding genomic DNA extracted from the seven pathogenic fungi. Their nucleotide sequences were used as a query for blast searches, and top hits are listed in Table 1 . According to the search results, the pathogenic strains were roughly assigned as Phoma herbarum H1, Colletotrichum nymphaeae H2, Alternaria alternata H4, Aspergillus oryzae H6, Penicillium spinulosum H7, Fusarium proliferatum H8, and Nemania bipapillata H9. Phylogenetic relationships based on the ITS1 sequence similarity between the strains and selected top hits are illustrated in Figure 2 . TABLE 1 NCBI blast results of samples. Strains Taxonomy Sequence similarity (%) Top hit H1 Phoma herbarum 99 MT367635.1 H2 Colletotrichum nymphaeae 100 MW217266.1 H4 Alternaria alternata 100 MN944587.1 H6 Aspergillus oryzae 99 MH345908.1 H7 Penicillium spinulosum 100 JQ639057.1 H8 Fusarium proliferatum 99 MG543763.1 H9 Nemania bipapillata 99 JQ341104.1 FIGURE 2 Molecular phylogenetic tree based on the rDNA-ITS sequence similarity. Sequence alignment and tree building were performed by MEGA5.0 using the neighbor-joining method, and phylogeny was tested by 500 bootstrap replications. The numbers on branches were calculated as bootstrap values. Accession numbers of ITS1 sequences from isolated strains were, namely, ON514545.1 (H1), ON514546.1 (H2), ON514547.1 (H4), ON514548.1 (H6), ON514549.1 (H7), ON514550.1 (H8), and ON514551.1 (H9). Preliminary Screening of Agro-Agents for Fungicides Against Representative Pathogenic Fungi In consideration of the prevalence of A. alternata and F. proliferatum in fruit disease, the two strains H4 and H8 were chosen as representative pathogenic microorganisms of dragon fruit and tested their susceptibility to a series of agricultural agents. As shown in Table 2 , a total of 20 agro-agents were screened for indoor toxicity at concentrations of 50 μg/ml and 10 μg/ml on A. alternata H4 and F. proliferatum H8. For A. alternata H4, at 50 μg/ml concentration, 10% difenoconazole, 430 g/L tebuconazole, and 3% zhongshengmycin showed the highest inhibition rate, while at 10 μ g/ml concentration, 10% difenoconazole, 430 g/L tebuconazole, 50% iprodione were the three most potent candidates. For F. proliferatum H8, 3% benziothiazolinone, 430 g/L tebuconazole, and 45% prochloraz exhibited the highest toxicity at a concentration of 50 μg/ml, while at the 10 μg/ml level, 10% difenoconazole, 50% iprodione, and 45% prochloraz were the top three candidates. Notably, three microbial preparations were included in the screening; the insecticidal Beauveria bassiana showed the highest inhibition rate at 50 μg/ml concentration, and the fungicidal Bacillus cereus was the most potent at 10 μg/ml. TABLE 2 Inhibitory effects of 20 agro-agents at the concentrations of 50 µg/ml and 10 µg/ml on A. alternata H4 and F. proliferatum H8. Agro-agents Inhibition rate (%) 50 µg/ml 10 µg/ml A. alternata H4 F. proliferatum H8 A. alternata H4 F. proliferatum H8 80% mancozeb 47.39 ± 1.72 33.52 ± 2.34 6.43 ± 1.93 17.96 ± 1.51 10% difenoconazole 92.46 ± 2.85 87.92 ± 3.60 80.76 ± 2.93 82.93 ± 3.37 2% wuyiencin 52.14 ± 2.40 56.36 ± 2.58 46.71 ± 2.91 38.08 ± 2.93 3% benziothiazolinone 73.68 ± 2.78 84.05 ± 3.56 54.44 ± 2.80 46.54 ± 1.38 50% chloroisobromine cyanuric acid 10.61 ± 2.33 12.49 ± 2.56 4.96 ± 2.80 6.01 ± 1.69 50% iprodione 70.25 ± 1.63 74.32 ± 2.79 56.37 ± 3.67 67.54 ± 2.79 50% benzylpenicillin 39.78 ± 1.61 56.12 ± 2.78 5.26 ± 2.70 16.51 ± 1.75 80% ethylicin 76.09 ± 2.85 82.54 ± 2.24 54.11 ± 2.00 61.54 ± 2.41 3% zhongshengmycin 82.90 ± 3.64 75.74 ± 2.54 29.91 ± 1.20 27.04 ± 2.01 75% chlorothalonil 54.32 ± 3.01 68.67 ± 1.89 48.44 ± 2.89 49.20 ± 2.07 50% sulfur·carbendazim 30.25 ± 1.46 19.27 ± 1.04 7.93 ± 2.44 3.58 ± 2.19 20% bismerthiazol 42.98 ± 1.42 52.10 ± 1.27 24.29 ± 2.21 35.12 ± 2.22 72% agricultural streptomycin sulfate 25.62 ± 2.81 18.96 ± 2.33 11.84 ± 2.77 8.49 ± 2.86 70% thiophanate methyl 13.10 ± 1.16 32.64 ± 2.64 5.35 ± 1.02 5.21 ± 2.57 50% kresoxim methyl 82.47 ± 1.92 68.79 ± 2.63 47.47 ± 1.39 64.79 ± 2.63 430 g/L tebuconazole 94.26 ± 301 95.46 ± 1.18 84.72 ± 3.61 75.46 ± 1.73 45% prochloraz 63.25 ± 2.51 98.79 ± 1.73 10.26 ± 1.85 84.26 ± 1.42 Bacillus cereus preparation (8*10 9 spores/g) 48.00 ± 2.06 42.61 ± 2.94 17.19 ± 2.83 20.69 ± 1.84 Beauveria bassiana preparation (2*10 11 spores/g) 51.76 ± 2.99 78.96 ± 3.01 6.14 ± 2.28 3.73 ± 2.39 Verticillium chlamydosporium preparation (2.5 * 10 9 spores/g) 28.82 ± 1.70 17.13 ± 2.12 6.14 ± 2.64 12.25 ± 2.60 Inhibition Effect of 10 Edible and Medicinal Plant Extracts on the Tested Fungus The ethanol extract was obtained from 10 edible and medicinal plants, namely, Houttuynia cordata Thunb, Mentha haplocalyx , Zanthoxylum bungeanum Maxim, Lonicera japonica Thunb, Dendrobium officinale Kimura et Migo, Piper nigrum, Zingiber officinale Roscoe, Gastrodia elata , Schisandra chinensis , and Illicium verum . Some of the plants are renowned for their microbe-inhibitory activity. The inhibition rate results in this study are shown in Table 3 . All 10 ethanolic extracts prepared at a final concentration of 50 mg/ml showed an inhibitory effect on the two fungal pathogens, A. alternata H4 and F. proliferatum H8. In comparison, 50 mg/ml and 80% ethylicin was included in the test. The inhibition rates of Zanthoxylum bungeanum Maxim against A. alternata H4 and F. proliferatum H8 were 68.75 and 75.12%, those of Zingiber officinale Roscoe were 70.14 and 60.48%, and those of Piper nigrum were 69.84 and 60.93%, respectively. The three extracts showed the highest inhibitory effect against both tested pathogens of dragon fruit. Notably, 80% ethylicin at the same concentration exhibited a 100% inhibition rate in both strains. TABLE 3 Determination of the inhibitory effect of 10 Chinese edible and medicinal plant extracts on A. alternata H4 and F. proliferatum H8. Plant extract Inhibition rate (%) Plant extract Inhibition rate (%) A. alternata H4 F. proliferatum H8 A. alternata H4 F. proliferatum H8 Zanthoxylum bungeanum Maxim 68.75 ± 1.15 75.12 ± 2.38 Mentha haplocalyx 7.24 ± 1.03 32.48 ± 1.54 Lonicera japonica Thunb 11.27 ± 1.72 34.84 ± 1.69 Dendrobium officinale Kimura et Migo 37.94 ± 1.35 26.75 ± 1.82 Zingiber officinale Roscoe 70.14 ± 2.16 60.48 ± 1.37 Piper nigrum 69.84 ± 1.19 60.93 ± 1.79 Illicium verum 37.68 ± 1.27 52.48 ± 1.77 Schisandra chinensis 48.37 ± 2.42 35.17 ± 2.80 Houttuynia cordata Thunb 24.64 ± 1.93 45.11 ± 2.60 Gastrodia elata 15.24 ± 1.91 21.68 ± 2.54 80% ethylicin 100 100 Fungicidal Efficacy Test on Representative Fungi With Promising Candidates Based on previousscreening results, several fungicides were selected for the efficacy test. As shown in Table 4 , 430 g/L tebuconazole exhibited the smallest EC 50 at 0.0133 μg/ml, suggesting the highest potency toward A. alternata H4, while 50% iprodione with EC 50 at 3.6840 μg/ml showed the second highest potency. For the pathogen F. proliferatum H8, 45% prochloraz showed the smallest EC 50 at 0.0122 μg/ml, and 430 g/L tebuconazole is the second smallest with EC 50 at 0.0307 μg/ml. It is worth mentioning that the efficacy of the biological agent Bacillus cereus is also tested, which exhibited EC 50 at 81.3915 μg/ml toward F. proliferatum H8. TABLE 4 Determination of the efficacy of screened fungicides to A. alternata H4 and F. proliferatum H8. Pathogenic strain Control agent Regression equation Correlation coefficient EC50 (µg/ml) A. alternata H4 3% benziothiazolinone y = 0.8324x+4.1566 0.9770 10.3089 10% difenoconazole y = 0.7454x+4.2309 0.9871 10.7595 50% iprodione y = 0.8428x+4.5227 0.9934 3.6840 50% kresoxim-methyl y = 0.5347x+4.3594 0.9813 15.7781 80% ethylicin y = 1.0551x+3.4653 0.9956 28.4809 430 g/L tebuconazole y = 0.2501x+5.4687 0.9254 0.0133 F. proliferatum H8 10% difenoconazole y = 0.3704x+5.1064 0.9903 0.5161 45% prochloraz y = 0.8479x+6.6203 0.9013 0.0122 50% iprodione y = 0.9036x+4.2205 0.9953 7.2888 50% kresoxim-methyl y = 0.3352x+4.7836 0.9766 4.4216 80% ethylicin y = 0.9706x+3.8647 0.9959 14.7804 430 g/L tebuconazole y = 0.7926x+6.1988 0.9111 0.0307 Discussion This study descripted the isolation and identification of several pathogenic fungal strains from diseased dragon fruits suffering from soft rot, anthracnose, and black spot based on field observation. The capability of the pathogens to infect healthy dragon fruits was confirmed by re-inoculation. Based on the morphological and molecular characteristics, the isolates were roughly identified to be P. herbarum H1, C. nymphaeae H2, A. alternata H4, A. oryzae H6, P. spinulosum H7, F. proliferatum H8, and N. bipapillata H9. Previous studies have suggested that a variety of Fusarium can cause postharvest soft rot disease to dragon fruit ( Oeurn et al., 2015 ), but few studies report on F. proliferatum , which is found for the first time in the study in Guizhou province, suggesting that F. proliferatum could be a regional pathogen. C. gloeosporioides ( Bordoh et al., 2020 ) and C. truncatum ( Guo et al., 2014 ; Vijaya et al., 2015 ) of Colletotrichum genera were recognized as the pathogenic microorganisms of pitaya anthracnose. C. nymphaeae was reported to cause anthrax in other fruits such as plum ( Chang et al., 2018 ) and blueberry ( Tomoo et al., 2015 ), and this is the first time found in diseased pitaya. A. alternata is the main pathogen causing postharvest disease, which can cause black spot disease in dragon fruit ( Castro et al., 2017 ), pears ( Tian et al., 2006 ), peaches ( Inoue and Nasu, 2000 ), and other fruits ( Prusky et al., 1997 ; Prusky et al., 1999 ; Zhao and Liu, 2012 ) in the post-harvest preservation. This study found first that A. oryzae can cause infection on dragon fruit. Since A. oryzae is a ubiquitous environmental fungus, the infection observed in this study could be opportunistic. This research reported for the first time several pathogenic microorganisms of dragon fruit in Guizhou, indicating that the distribution of pathogenic microorganisms in dragon fruit varies with a geographical environment which on the other hand signifies the importance of the geography-specific plan for prevention and control measures. A. alternata and F. proliferatum are two typical postharvest pathogenic microorganisms that can cause postharvest diseases in a variety of fruits ( Konstantinou et al., 2011 ; Abd Murad et al., 2017 ; Castro et al., 2017 ; Wang et al., 2021 ), so we selected these two to test their susceptibility to the control agents, respectively. The results showed that A. alternata H4 exhibited the highest sensitivity to 430 g/L tebuconazole and the lowest sensitivity to 80% ethylicin, while F. proliferatum H8 showed the highest sensitivity to 45% prochloraz and the lowest sensitivity to Bacillus cereus . The results showed that 10% difenoconazole, 50% iprodione, 50% kresoxim-methyl, 80% ethylicin, and 430 g/L tebuconazole all had inhibitory effects on the two postharvest pathogenic microorganisms of pitaya to a certain degree. Natural antimicrobials obtained from plants can provide alternative materials instead of commonly used fungicides in a more sustainable and environment-friendly way ( Romanazzi et al., 2012 ). This study demonstrated that the 10 edible and medicinal plant extracts all showed some inhibitory effects on the two representative pathogenic fungi, although they are much less potent than chemically synthesized fungicides. Chemical fungicides showing the most efficacy to tested pathogens can be good choices for field application; however, long-term single use of control agents will likely cause drug resistance of pathogenic microorganisms. Therefore, it is recommended to carry out field experiments using a combination of agents which would improve the control effect, but the study only conducted a sensitivity test indoors which may not represent the field effect. Pathogenic microorganisms can infect fruits in many ways. For example, fungal spores can spread with the help of wind or insects to infect fruits. Therefore, cultivation and management should be strengthened at regular times, and attention should be paid to ventilation and light transmission to reduce the reproduction of fungi since the reproduction and accumulation of pathogenic microorganisms would cause more fruits to rot, aggravating microbial infections. At present, the postharvest preservation methods of dragon fruit mainly include low-temperature preservation, film preservation, chemical preservation, thermal treatments, and irradiation ( Jalgaonkar et al., 2020 ). These preservation measures have problems such as being time-consuming and labor-intensive, high cost, chemical residues, environmental pollution, and food safety hazards. Plant extracts, especially those from edible or medicinal plants, with fungal inhibitory effects can find their use in postharvest disease prevention. Conclusion The study isolated seven strains from diseased dragon fruits, and their ability to infect healthy dragon fruits was confirmed. Of them, Penicillium spinulosum , Phoma herbarum , Nemania bipapillata and Aspergillus oryzae were for the first time found to cause dragon fruit disease. In consideration of their prevalence in postharvest fruit diseases, A. alternat a H8 and F. proliferatum H4 were chosen as representative pathogens for the drug susceptibility test. Among the tested agro-agents and plant extracts, 430 g/L tebuconazole and 45% prochloraz were found to be the most potent fungicides against H8 and H4, respectively. Data Availability Statement The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material. Author Contributions WW designed and devised the study, and YL and HC analyzed the data and prepared the original manuscript. HC supervised YA and LM for pathogen isolation and the drug sensitivity test. HW revised the manuscript. All authors discussed, edited, and approved the final version. Funding We acknowledge the funds from the Science and TechnologyFund Project of Guizhou [QKHJC(2020)1Y089], the Young Sci-Tech Talents Growth Program from the Department of Education of Guizhou Province [No. QJHKYZ(2018)297], the Guiyang Science and Technology Bureau and Guiyang University [GYU-KYZ(2019~2020)PT10-06], the Research Foundation of Guiyang University [GYU-KY-(2022)], and the UndergraduateInnovation and Entrepreneurship Training Program (No. 2019520371). Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's Note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7123036/

China’s Institutional Mechanisms for Influenza A (H1N1) Prevention and Control

Innovation in institutional mechanisms is a fundamental issue in effectively dealing with public health emergencies. In the wake of the 2003 SARS Epidemic, China initially established a public health emergency management system and an emergency organization and management network, placing emphasis on "government leading, unified command, local management, responsibility on all levels, management by classifications, and inter-departmental coordination," which strengthened the existing health emergency preparation system. China's Current Public Health Emergency Institutional Mechanisms The unexpected outbreak of the 2003 SARS epidemic revealed some of China's shortcomings in the field of emergency management, including: weak institutional mechanisms, lack of organizational communication, inefficient flow of information, and insufficient preparedness. China's fight against SARS, as it were, not only posed a great challenge to the nation's socialist modernization, but at the same time offered an important opportunity for the country to improve emergency management, especially in the field of public health. Construction of a National Emergency Management System In the wake of the SARS Epidemic, the Chinese government began pushing for a national emergency management system in a systematic, planned and gradual manner, and made remarkable progress in emergency management structured on the "one plan three systems (contingency plans, institutions, mechanisms and legislation)." In regards to contingency planning, the country formed a system consisting of contingency plans at central, local, departmental, and enterprise levels as well as plans for major events, and this system played an important role in dealing with public emergencies. 1 In regards to institution building, a national emergency management system consisting of general emergency management offices as well as of special emergency management bodies were established. The Emergency Management Office of the State Council and emergency management bodies of provincial (regional and municipal) governments were set up in succession, 2 in addition to emergency management systems in specific fields such as health. In comparison with the pre-SARS environment where "departments played a dominant role, and coordination was inadequate," these new institutions showed the permanent, comprehensive, and specialized nature of emergency management, 3 and laid an organizational foundation for the future. In addition, society as a whole began getting involved with the emergency management process, including: further strengthening military emergency system construction and local assistance 4 ; giving full rein to experts in emergency management 5 ; and developing and implementing local emergency management plans that targeted "communities, rural areas, enterprises and schools." Looking at mechanism construction, progress was also made in research on a science-based emergency management system, and an emergency management mechanism characterized by "unified leadership, responsiveness, orderly coordination, and efficient operation"—which enabled the interconnection of early warning, mass mobilization, quick response, and emergency handling—was gradually established to effectively mitigated public health emergencies. In regards to legislation, the Emergency Response Law was took effect on November 1st, 2007. Establishment and Development of China's Public Health Emergency System China bolstered construction on their public health emergency management system according to the general requirements for such a mechanism. On May 12th, 2003, the State Council promulgated the Regulations on Preparedness for and Response to Emergent Public Health Hazards which stressed the need in building a public health emergency management system that could ensure "unobstructed information, rapid response, effective leadership, and definite duties." At the May 14th Executive Meeting of the State Council, chaired by Premier Wen Jiabao, members discussed building a mechanism for national response to public health emergencies, and the principles for such a mechanism were determined: "ensure unified central leadership with levels of responsibility; impose regulations and management in accordance with the law to ensure rapid response capabilities; improve the monitoring system to increase early-warning capabilities; and boost infrastructure to secure sustained operation." On June 4th, at another Executive Meeting of the State Council chaired by Premier Wen Jiabao, he once again stressed, referring to the SARS response and economic impact, that efforts should be made to "accelerate the construction of the public health emergency management mechanism; forge ahead with developing an information network system, the disease prevention and control system, and the emergency rescue system; and ensure readiness and preparedness for any emergency." At the National SARS Work Conference held on July 28th, 2003, Premier Wen Jiabao ensured the construction and improvement of the public health emergency response mechanism, the disease control and prevention system, and the health legislation enforcement and supervision system in three years' time. Building on that, he promised to improve the country's rural health system, urban basic medical service system, environmental health system, and funding security system for the long term. Efforts would be made also to strengthen the disease control and prevention system, increase public health emergency management capabilities, boost rural health development, improve healthcare for the rural population, strengthen environmental health system, and implement national health campaigns. In February 2007, the MOH outline the following overall goals for public health emergency work in the 11th Five-Year Plan: establish and improve health emergency management legislation and the health emergency contingency planning system; build an emergency management mechanism characterized by "unified leadership, responsiveness, orderly coordination, and efficient operation" with "predominantly local management, hierarchical responsibility, and comprehensive coordination;" bolster health emergency management recruitment; improve the public health emergency monitoring and warning system; strengthen capacity for quick and effective response to health emergencies; and shape an environment of health emergency management characterized by inter-departmental coordination, collaboration, and social participation under the leadership of central and local governments. The Formulation of a Top-Down System for Public Health Emergency Planning During and after the 2003 SARS Epidemic, China continued to establish and improve public health emergency legislation, regulations and contingency plans, and initially formed a national system for public health emergencies contingency planning (see Fig. 4.1 ). Fig. 4.1 China's hierarchical system for public health emergencies planning On August 28th, 2004, the Standing Committee's 11th Meeting of the 10th National People's Congress amended the Infectious Disease Prevention and Treatment Law . In January 2005, Premier Wen Jiabao chaired an Executive Meeting of the State Council where the National Overall Contingency Plan for Public Health Emergencies with twenty five special contingency plans and eighty departmental contingency plans were adopted in principle (a total of 106 plans). In 2005–2006, the State Council successively formulated and promulgated the Overall Contingency Plan for Public Health Emergencies (January 8th, 2006), the National Contingency Plan for Public Health Emergencies (February 26th, 2006), and the National Contingency Plan for Medical and Health Rescue in Case of Public Health Emergencies (February 26th, 2006). The State Council issued the National Contingency Plan for Major Animal Disease Emergencies (February 26th, 2006), the National Contingency Plan for Major Food Safety Accidents (February 26th, 2006), the National Contingency Plan for Nuclear and Radiation Accidents, and others. The State Council organized the drafting and revision of such health emergency response plans such as the National Contingency Plan for Medical and Health Rescue, the Contingency Plan for Community - level Public Health Emergencies, and the Contingency Plan for Food Poisoning Emergencies . Local governments and health departments also formulated health emergency response plans in accordance with their local conditions. In preparation for an influenza pandemic, the MOH formulated the Influenza Pandemic Preparedness and Response Plan of the Ministry of Health (Tentative) , which was issued on September 28th, 2005. On July 12th, 2006, the MOH issued the Emergency Response Plan for Highly Pathogenic Human and Avian Influenza , which, as a departmental contingency plan under the National Overall Contingency Plan for Public Health Emergencies , provided systematic organization and leadership against an influenza pandemic, and it outlined important factors for response efforts including division of labor, preparedness, emergency response, and supervision. The Establishment of Public Health Emergency Management Bodies As arranged by the State Council, in March 2004 the MOH set up the Health Emergency Office (Public Health Emergency Operations Center) which became responsible for the organization and coordination for managing emergency preparations and countermeasures. The MOH Health Emergency Office (Public Health Emergency Operations Center) established sub-offices responsible for integrated coordination, monitoring and alert, emergency response guidance, and emergency countermeasure management. The duties for this operations center include: directing and coordinating national health emergency efforts; developing health emergency and medical rescue plans, systems, contingency plans and measures; directing health emergency activities such as public health emergency preparedness, monitoring and alert, response and rescue, and analysis and evaluation; providing guidance on local implementation of prevention, control, and medical rescue measures in response to public health and other emergencies; establishing and improving health emergency information and operations systems; publishing public health emergency response information; directing and organizing health emergency response training and exercises; keeping records and plans on the national stockpile and providing recommendations on their usage; managing the National Expert Advisory Committee on Public Health Emergencies as well as public health emergency experts; directing and organizing preparation and response measures against acute infectious diseases; organizing medical rescue efforts in case of serious natural disasters, terrorist incidents, food safety emergencies, and nuclear radiation accidents; organizing and coordinating health emergency response services for major national events; organizing health emergency research and health education programs; responsible for the organization and coordination of domestic implementation for the International Health Regulations ; coordinating the implementation of the Biological Weapons Convention in the health industry; and carrying out routine work for the Office of the MOH Leading Group for Disaster Relief and Disease Prevention. At the same time, a health emergency management system was established, which included the following seven subsystems—Emergency Response Security, Command & Decision-making, Emergency Response Workforce, Monitoring & Alert, Emergency Response Management, Risk Communication, and Science, Technology & Education (Fig. 4.2 ). Fig. 4.2 Composition of the health emergency systems Currently, health departments (bureaus) in 30 provinces (autonomous regions, and municipalities directly under the central government) have established health emergency response offices and the China CDC along with some provincial CDCs also have established special emergency management departments. For example, Beijing established a public health emergency operations center in March 2006, and Shanghai did as well in early 2009 to strengthen its capacity for pubic health emergency management. See Fig. 4.3 for a detailed chart of the national health emergency response command system. Fig. 4.3 Framework of China's public health emergency response system The Establishment of Cross-Region, Cross-Department Public Health Emergency Coordination Mechanisms The MOH established a public health emergency coordination mechanism with thirty one central and national departments to deal with inter-departmental collaboration, which effectively strengthened communication and coordination between departments dealing with public health emergencies. The national government and the Special Administrative Regions of Hong Kong and Macau entered into a three-party emergency response collaboration agreement and decided upon implementation regulations, and established a linkage mechanism for information communication and health emergency response. Additionally, the MOH established a joint prevention and control mechanism with the Ministry of Agriculture (MOA) to protect against highly pathogenic zoonotic viruses such as avian influenza and Streptococcus suis ; the MOH established a coordination mechanism for joint prevention and control of public health emergencies at ports with the General Administration of Quality Supervision, Inspection and Quarantine (AQSIQ). In collaboration with the Ministry of Railway (MOR), the Ministry of Transport (MOT), and the AQSIQ, the MOH issued notices on the prevention and control of the importation of infectious diseases from abroad; and together with the MOE, issued a document requiring schools to appoint part-time or full-time teachers to identify and report infectious diseases or other health emergencies at the school. This time marked the initial formation of a working inter-departmental mechanism positioned to combat health emergencies through "paying equal attention to both prevention and response, and instilling continued collaboration for any event." The Focus on Team Building for Public Health Emergency Experts In early 2006, the National Expert Advisory Committee on Public Health Emergencies was established in Beijing, with the Vice Health Minister serving as its chairman. Consisting of 105 members, the Committee's routine management work was conducted through the MOH Health Emergency Office. Its main duties included: provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies. Construction of a National Emergency Management System In the wake of the SARS Epidemic, the Chinese government began pushing for a national emergency management system in a systematic, planned and gradual manner, and made remarkable progress in emergency management structured on the "one plan three systems (contingency plans, institutions, mechanisms and legislation)." In regards to contingency planning, the country formed a system consisting of contingency plans at central, local, departmental, and enterprise levels as well as plans for major events, and this system played an important role in dealing with public emergencies. 1 In regards to institution building, a national emergency management system consisting of general emergency management offices as well as of special emergency management bodies were established. The Emergency Management Office of the State Council and emergency management bodies of provincial (regional and municipal) governments were set up in succession, 2 in addition to emergency management systems in specific fields such as health. In comparison with the pre-SARS environment where "departments played a dominant role, and coordination was inadequate," these new institutions showed the permanent, comprehensive, and specialized nature of emergency management, 3 and laid an organizational foundation for the future. In addition, society as a whole began getting involved with the emergency management process, including: further strengthening military emergency system construction and local assistance 4 ; giving full rein to experts in emergency management 5 ; and developing and implementing local emergency management plans that targeted "communities, rural areas, enterprises and schools." Looking at mechanism construction, progress was also made in research on a science-based emergency management system, and an emergency management mechanism characterized by "unified leadership, responsiveness, orderly coordination, and efficient operation"—which enabled the interconnection of early warning, mass mobilization, quick response, and emergency handling—was gradually established to effectively mitigated public health emergencies. In regards to legislation, the Emergency Response Law was took effect on November 1st, 2007. Establishment and Development of China's Public Health Emergency System China bolstered construction on their public health emergency management system according to the general requirements for such a mechanism. On May 12th, 2003, the State Council promulgated the Regulations on Preparedness for and Response to Emergent Public Health Hazards which stressed the need in building a public health emergency management system that could ensure "unobstructed information, rapid response, effective leadership, and definite duties." At the May 14th Executive Meeting of the State Council, chaired by Premier Wen Jiabao, members discussed building a mechanism for national response to public health emergencies, and the principles for such a mechanism were determined: "ensure unified central leadership with levels of responsibility; impose regulations and management in accordance with the law to ensure rapid response capabilities; improve the monitoring system to increase early-warning capabilities; and boost infrastructure to secure sustained operation." On June 4th, at another Executive Meeting of the State Council chaired by Premier Wen Jiabao, he once again stressed, referring to the SARS response and economic impact, that efforts should be made to "accelerate the construction of the public health emergency management mechanism; forge ahead with developing an information network system, the disease prevention and control system, and the emergency rescue system; and ensure readiness and preparedness for any emergency." At the National SARS Work Conference held on July 28th, 2003, Premier Wen Jiabao ensured the construction and improvement of the public health emergency response mechanism, the disease control and prevention system, and the health legislation enforcement and supervision system in three years' time. Building on that, he promised to improve the country's rural health system, urban basic medical service system, environmental health system, and funding security system for the long term. Efforts would be made also to strengthen the disease control and prevention system, increase public health emergency management capabilities, boost rural health development, improve healthcare for the rural population, strengthen environmental health system, and implement national health campaigns. In February 2007, the MOH outline the following overall goals for public health emergency work in the 11th Five-Year Plan: establish and improve health emergency management legislation and the health emergency contingency planning system; build an emergency management mechanism characterized by "unified leadership, responsiveness, orderly coordination, and efficient operation" with "predominantly local management, hierarchical responsibility, and comprehensive coordination;" bolster health emergency management recruitment; improve the public health emergency monitoring and warning system; strengthen capacity for quick and effective response to health emergencies; and shape an environment of health emergency management characterized by inter-departmental coordination, collaboration, and social participation under the leadership of central and local governments. The Formulation of a Top-Down System for Public Health Emergency Planning During and after the 2003 SARS Epidemic, China continued to establish and improve public health emergency legislation, regulations and contingency plans, and initially formed a national system for public health emergencies contingency planning (see Fig. 4.1 ). Fig. 4.1 China's hierarchical system for public health emergencies planning On August 28th, 2004, the Standing Committee's 11th Meeting of the 10th National People's Congress amended the Infectious Disease Prevention and Treatment Law . In January 2005, Premier Wen Jiabao chaired an Executive Meeting of the State Council where the National Overall Contingency Plan for Public Health Emergencies with twenty five special contingency plans and eighty departmental contingency plans were adopted in principle (a total of 106 plans). In 2005–2006, the State Council successively formulated and promulgated the Overall Contingency Plan for Public Health Emergencies (January 8th, 2006), the National Contingency Plan for Public Health Emergencies (February 26th, 2006), and the National Contingency Plan for Medical and Health Rescue in Case of Public Health Emergencies (February 26th, 2006). The State Council issued the National Contingency Plan for Major Animal Disease Emergencies (February 26th, 2006), the National Contingency Plan for Major Food Safety Accidents (February 26th, 2006), the National Contingency Plan for Nuclear and Radiation Accidents, and others. The State Council organized the drafting and revision of such health emergency response plans such as the National Contingency Plan for Medical and Health Rescue, the Contingency Plan for Community - level Public Health Emergencies, and the Contingency Plan for Food Poisoning Emergencies . Local governments and health departments also formulated health emergency response plans in accordance with their local conditions. In preparation for an influenza pandemic, the MOH formulated the Influenza Pandemic Preparedness and Response Plan of the Ministry of Health (Tentative) , which was issued on September 28th, 2005. On July 12th, 2006, the MOH issued the Emergency Response Plan for Highly Pathogenic Human and Avian Influenza , which, as a departmental contingency plan under the National Overall Contingency Plan for Public Health Emergencies , provided systematic organization and leadership against an influenza pandemic, and it outlined important factors for response efforts including division of labor, preparedness, emergency response, and supervision. The Establishment of Public Health Emergency Management Bodies As arranged by the State Council, in March 2004 the MOH set up the Health Emergency Office (Public Health Emergency Operations Center) which became responsible for the organization and coordination for managing emergency preparations and countermeasures. The MOH Health Emergency Office (Public Health Emergency Operations Center) established sub-offices responsible for integrated coordination, monitoring and alert, emergency response guidance, and emergency countermeasure management. The duties for this operations center include: directing and coordinating national health emergency efforts; developing health emergency and medical rescue plans, systems, contingency plans and measures; directing health emergency activities such as public health emergency preparedness, monitoring and alert, response and rescue, and analysis and evaluation; providing guidance on local implementation of prevention, control, and medical rescue measures in response to public health and other emergencies; establishing and improving health emergency information and operations systems; publishing public health emergency response information; directing and organizing health emergency response training and exercises; keeping records and plans on the national stockpile and providing recommendations on their usage; managing the National Expert Advisory Committee on Public Health Emergencies as well as public health emergency experts; directing and organizing preparation and response measures against acute infectious diseases; organizing medical rescue efforts in case of serious natural disasters, terrorist incidents, food safety emergencies, and nuclear radiation accidents; organizing and coordinating health emergency response services for major national events; organizing health emergency research and health education programs; responsible for the organization and coordination of domestic implementation for the International Health Regulations ; coordinating the implementation of the Biological Weapons Convention in the health industry; and carrying out routine work for the Office of the MOH Leading Group for Disaster Relief and Disease Prevention. At the same time, a health emergency management system was established, which included the following seven subsystems—Emergency Response Security, Command & Decision-making, Emergency Response Workforce, Monitoring & Alert, Emergency Response Management, Risk Communication, and Science, Technology & Education (Fig. 4.2 ). Fig. 4.2 Composition of the health emergency systems Currently, health departments (bureaus) in 30 provinces (autonomous regions, and municipalities directly under the central government) have established health emergency response offices and the China CDC along with some provincial CDCs also have established special emergency management departments. For example, Beijing established a public health emergency operations center in March 2006, and Shanghai did as well in early 2009 to strengthen its capacity for pubic health emergency management. See Fig. 4.3 for a detailed chart of the national health emergency response command system. Fig. 4.3 Framework of China's public health emergency response system The Establishment of Cross-Region, Cross-Department Public Health Emergency Coordination Mechanisms The MOH established a public health emergency coordination mechanism with thirty one central and national departments to deal with inter-departmental collaboration, which effectively strengthened communication and coordination between departments dealing with public health emergencies. The national government and the Special Administrative Regions of Hong Kong and Macau entered into a three-party emergency response collaboration agreement and decided upon implementation regulations, and established a linkage mechanism for information communication and health emergency response. Additionally, the MOH established a joint prevention and control mechanism with the Ministry of Agriculture (MOA) to protect against highly pathogenic zoonotic viruses such as avian influenza and Streptococcus suis ; the MOH established a coordination mechanism for joint prevention and control of public health emergencies at ports with the General Administration of Quality Supervision, Inspection and Quarantine (AQSIQ). In collaboration with the Ministry of Railway (MOR), the Ministry of Transport (MOT), and the AQSIQ, the MOH issued notices on the prevention and control of the importation of infectious diseases from abroad; and together with the MOE, issued a document requiring schools to appoint part-time or full-time teachers to identify and report infectious diseases or other health emergencies at the school. This time marked the initial formation of a working inter-departmental mechanism positioned to combat health emergencies through "paying equal attention to both prevention and response, and instilling continued collaboration for any event." The Focus on Team Building for Public Health Emergency Experts In early 2006, the National Expert Advisory Committee on Public Health Emergencies was established in Beijing, with the Vice Health Minister serving as its chairman. Consisting of 105 members, the Committee's routine management work was conducted through the MOH Health Emergency Office. Its main duties included: provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies. The Formulation of a Top-Down System for Public Health Emergency Planning During and after the 2003 SARS Epidemic, China continued to establish and improve public health emergency legislation, regulations and contingency plans, and initially formed a national system for public health emergencies contingency planning (see Fig. 4.1 ). Fig. 4.1 China's hierarchical system for public health emergencies planning On August 28th, 2004, the Standing Committee's 11th Meeting of the 10th National People's Congress amended the Infectious Disease Prevention and Treatment Law . In January 2005, Premier Wen Jiabao chaired an Executive Meeting of the State Council where the National Overall Contingency Plan for Public Health Emergencies with twenty five special contingency plans and eighty departmental contingency plans were adopted in principle (a total of 106 plans). In 2005–2006, the State Council successively formulated and promulgated the Overall Contingency Plan for Public Health Emergencies (January 8th, 2006), the National Contingency Plan for Public Health Emergencies (February 26th, 2006), and the National Contingency Plan for Medical and Health Rescue in Case of Public Health Emergencies (February 26th, 2006). The State Council issued the National Contingency Plan for Major Animal Disease Emergencies (February 26th, 2006), the National Contingency Plan for Major Food Safety Accidents (February 26th, 2006), the National Contingency Plan for Nuclear and Radiation Accidents, and others. The State Council organized the drafting and revision of such health emergency response plans such as the National Contingency Plan for Medical and Health Rescue, the Contingency Plan for Community - level Public Health Emergencies, and the Contingency Plan for Food Poisoning Emergencies . Local governments and health departments also formulated health emergency response plans in accordance with their local conditions. In preparation for an influenza pandemic, the MOH formulated the Influenza Pandemic Preparedness and Response Plan of the Ministry of Health (Tentative) , which was issued on September 28th, 2005. On July 12th, 2006, the MOH issued the Emergency Response Plan for Highly Pathogenic Human and Avian Influenza , which, as a departmental contingency plan under the National Overall Contingency Plan for Public Health Emergencies , provided systematic organization and leadership against an influenza pandemic, and it outlined important factors for response efforts including division of labor, preparedness, emergency response, and supervision. The Establishment of Public Health Emergency Management Bodies As arranged by the State Council, in March 2004 the MOH set up the Health Emergency Office (Public Health Emergency Operations Center) which became responsible for the organization and coordination for managing emergency preparations and countermeasures. The MOH Health Emergency Office (Public Health Emergency Operations Center) established sub-offices responsible for integrated coordination, monitoring and alert, emergency response guidance, and emergency countermeasure management. The duties for this operations center include: directing and coordinating national health emergency efforts; developing health emergency and medical rescue plans, systems, contingency plans and measures; directing health emergency activities such as public health emergency preparedness, monitoring and alert, response and rescue, and analysis and evaluation; providing guidance on local implementation of prevention, control, and medical rescue measures in response to public health and other emergencies; establishing and improving health emergency information and operations systems; publishing public health emergency response information; directing and organizing health emergency response training and exercises; keeping records and plans on the national stockpile and providing recommendations on their usage; managing the National Expert Advisory Committee on Public Health Emergencies as well as public health emergency experts; directing and organizing preparation and response measures against acute infectious diseases; organizing medical rescue efforts in case of serious natural disasters, terrorist incidents, food safety emergencies, and nuclear radiation accidents; organizing and coordinating health emergency response services for major national events; organizing health emergency research and health education programs; responsible for the organization and coordination of domestic implementation for the International Health Regulations ; coordinating the implementation of the Biological Weapons Convention in the health industry; and carrying out routine work for the Office of the MOH Leading Group for Disaster Relief and Disease Prevention. At the same time, a health emergency management system was established, which included the following seven subsystems—Emergency Response Security, Command & Decision-making, Emergency Response Workforce, Monitoring & Alert, Emergency Response Management, Risk Communication, and Science, Technology & Education (Fig. 4.2 ). Fig. 4.2 Composition of the health emergency systems Currently, health departments (bureaus) in 30 provinces (autonomous regions, and municipalities directly under the central government) have established health emergency response offices and the China CDC along with some provincial CDCs also have established special emergency management departments. For example, Beijing established a public health emergency operations center in March 2006, and Shanghai did as well in early 2009 to strengthen its capacity for pubic health emergency management. See Fig. 4.3 for a detailed chart of the national health emergency response command system. Fig. 4.3 Framework of China's public health emergency response system The Establishment of Cross-Region, Cross-Department Public Health Emergency Coordination Mechanisms The MOH established a public health emergency coordination mechanism with thirty one central and national departments to deal with inter-departmental collaboration, which effectively strengthened communication and coordination between departments dealing with public health emergencies. The national government and the Special Administrative Regions of Hong Kong and Macau entered into a three-party emergency response collaboration agreement and decided upon implementation regulations, and established a linkage mechanism for information communication and health emergency response. Additionally, the MOH established a joint prevention and control mechanism with the Ministry of Agriculture (MOA) to protect against highly pathogenic zoonotic viruses such as avian influenza and Streptococcus suis ; the MOH established a coordination mechanism for joint prevention and control of public health emergencies at ports with the General Administration of Quality Supervision, Inspection and Quarantine (AQSIQ). In collaboration with the Ministry of Railway (MOR), the Ministry of Transport (MOT), and the AQSIQ, the MOH issued notices on the prevention and control of the importation of infectious diseases from abroad; and together with the MOE, issued a document requiring schools to appoint part-time or full-time teachers to identify and report infectious diseases or other health emergencies at the school. This time marked the initial formation of a working inter-departmental mechanism positioned to combat health emergencies through "paying equal attention to both prevention and response, and instilling continued collaboration for any event." The Focus on Team Building for Public Health Emergency Experts In early 2006, the National Expert Advisory Committee on Public Health Emergencies was established in Beijing, with the Vice Health Minister serving as its chairman. Consisting of 105 members, the Committee's routine management work was conducted through the MOH Health Emergency Office. Its main duties included: provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies. The Establishment, Composition and Operations of the Joint National Influenza A (H1N1) Prevention and Control Mechanism The Establishment of the Joint National Influenza A (H1N1) Prevention and Control Mechanism The MOH responded immediately upon receiving a disease outbreak notice from the WHO on April 25th, 2009, as the MOH General Office launched the working mechanism comprised of the MOH Leading Group and Expert Panel for Influenza Pandemic Prevention and Control [in accordance with the Influenza Pandemic Preparedness and Response Plan of the Ministry of Health (Tentative) ]. The MOH also issued a Notice on Strengthening Preparedness for and Response to Human Swine Influenza to health departments requiring the following: medical and disease control and prevention institutions at various levels to strengthen cases monitoring and reporting; and prepare for Influenza A (H1N1) in terms of processing, technology, manpower, and material resources. At the same time, the MOH immediately forwarded related information to the MOA and the AQSIQ. On April 26th, Health Minister Chen Zhu convened a meeting of the MOH Leading Group and Expert Panel for Influenza Pandemic Prevention and Control, at which the attendees analyzed swine influenza situations in the United States and Mexico, predicted epidemic trends, and deliberated on domestic strategies and measures to cope with a swine flu pandemic. Health Minister Chen also held an inter-departmental meeting with the MOA, the AQSIQ, and other ministries to analyze epidemic trends and discuss response strategies and measures. Immediately after the meeting was over, the MOH reported in writing that very night to the State Council on the progress of epidemic prevention and control work. On April 27th, following the emergency meeting held in Geneva, the WHO elevated the pandemic alert level from Phase 3 to Phase 4, stating that the "swine flu" was widespread and was being transmitted by humans in different ways. General Secretary Hu Jintao issued instructions to place prevention and control against this virus as the nation's top priority. On the same day, Vice Premier Li Keqiang convened the State Council Meeting regarding the Human-Swine Influenza Prevention Working Mechanism, resulting in the decision to establish a multi-departmental working mechanism for joint prevention and control of the human-swine influenza. As required by the State Council meeting, the MOH called together the Publicity Department of the Communist Party of China (CCPPD), the Ministry of Foreign Affairs (MFA), the NDRC, the MIIT, the MOF, the Ministry of Transport (MOT), the MOA, the MOC, the AQSIQ, the China National Tourism Administration (CNTA), the Civil Aviation Administration of China (CAAC) among other departments on that very night for a meeting to deliberate on Influenza A (H1N1) prevention and control; the meeting established the multi-departmental working mechanism for joint prevention and control of human-swine influenza and the Notice on Strengthening Human - Swine Influenza Prevention and Control was drafted and published on the night of April 27th after State Council review. The MOH issued the Notice of the MOH General Office on Strengthening Preparedness for and Response to Human Swine Influenza . On April 28th, Premier Wen Jiabao convened a State Council Executive Meeting which deliberated on how to strengthen national response to human-swine influenza; at the meeting they defined the overall prevention and control principles and strategies of "taking threats to public health seriously, responding actively, and coping with the epidemic in a scientific manner according to law through joint prevention and control efforts." On April 29th, the WHO raised its pandemic alert level from Phase 4 to Phase 5. On April 30th, at a press conference held at the State Council Information Office, the MOH declared the establishment of a multi-departmental working mechanism for joint prevention and control against the human-swine influenza, which would be spearheaded by the MOH. Under this mechanism, 33 departments and institutions (which later increased to 38) constituted 8 work groups—General Office, Ports, Healthcare, Support, Dissemination and Communication, Foreign Collaboration, Science and Technology, and Animal Husbandry and Veterinary—and an expert committee, forming a "8 + 1" pattern for joint prevention and control efforts. On the afternoon of May 1st, the joint prevention and control mechanism held its second joint conference, renaming the human-swine influenza which was occurring in Mexico and the United States to "Influenza A (H1N1)." The former wording of "multi-departmental work mechanism for joint prevention and control of human-swine influenza" was changed to the "Joint National Influenza A (H1N1) Prevention and Control Mechanism" and roles and responsibilities were outlined for the mechanism, work groups, and the expert committee. At the same time, a meeting system for all members and liaisons was established. Problems in principle which a work group encountered would be solved by the work group itself, and those which the work group struggled with would be settled through coordination under the joint prevention and control mechanism—general affairs would be solved by regular liaison meetings, and major issues decided by plenary meetings. The Composition of the Joint National Influenza A (H1N1) Prevention and Control Mechanism Organizational Structure Health Minister Chen Zhu and MOH Party Group Secretary and Vice Health Minister Zhang Mao chaired the Joint National Influenza A (H1N1) Prevention and Control Mechanism, and the mechanism was comprised of eight working groups—General Office, Ports, Healthcare, Support, Dissemination and Communication, Foreign Collaboration, Science and Technology, and Animal Husbandry and Veterinary—and an Expert Advisory Committee. The heads of the work groups, the leaders of the Health Department of the People's Liberation Army General Logistics Department (GLD), the Logistics Department of the Chinese People's Armed Police Force (PAPF), and the chairman of the Expert Advisory Committee, served as members of the joint prevention and control mechanism. The work groups, the GLD Health Department and the PAPF Logistics Department each designated one or two departmental-level officials as liaisons for routine communication purposes. Main Responsibilities and the Division of Labor The main duties of the Joint National Influenza A (H1N1) Prevention and Control Mechanism included: meet regularly to evaluate epidemic trends and determine prevention and control strategies; formulate prevention and control policies, response plans and major measures; coordinate and provide guidance on the implementation by all related departments and in various regions of prevention and control measures; and organize supervision and inspection activities concerning the implementation of prevention and control measures. See Table 4.1 for the main responsibilities of the work groups and the Expert Advisory Committee. Table 4.1 Composition of the work groups and the expert advisory committee for the joint national influenza A (H1N1) prevention and control mechanism Work group Leading organization Members Duties General office group MOH CCPPD, MFA, NDRC, MOST, MOA, AQSIQ, GLD Health Department, PAPF Logistics Department, Expert Advisory Committee Comprehensively coordinate routine affairs among departments of the joint prevention and control mechanism; organize regular meetings of the joint prevention and control mechanism and oversee the handling of top agendas; collect, sort out, and report to higher-ups about progress in prevention and control efforts; prepare progress reports on the joint prevention and control mechanism; publish information on epidemic situations and response efforts; ensure the consistency in style of writing for documents intended for outside use; and shoulder other assignments from leaders Ports group AQSIQ MFA, Ministry of Public Security (MPS), MOT, MOR, MOA, MOC, MOH, General Administration of Customs (GAC), CNTA, State Council Information Office (SCIO), CAAC, and State Post Bureau (SPB) Carry out health quarantine, surveillance, health supervision and management at entry-exit ports; collect, sort out and report on information regarding epidemic developments, response measures, etc. abroad; ensure the quality and standardization of emergency materials Healthcare group MOH State Administration of Traditional Chinese Medicine (SATCM), AQSIQ, GLD Health Department, PAPF Logistics Department, and China CDC Organize the formulation and revision of technical solutions to the diagnosis, treatment, prevention and control of Influenza A (H1N1), and oversee the implementation of those solutions; provide guidance on nationwide monitoring, reporting, and epidemiological surveys of epidemic situations, specimen collection and testing, epidemic management, etc.; direct national medical treatment plans of Influenza A (H1N1); send experts, as needed for epidemic prevention and control, to go and help with epidemic handling and medical treatment in priority regions; provide recommendations on strategies and measures for improving response efforts; and undertake other assignments from leaders Support group NDRC MIIT, MOF, MOT, MOR, MOC, MOH, AQSIQ, CAAC, SFDA, and GLD Health Department Plan as a whole the support logistics for emergency materials, and deliberate on and coordinate major issues involved in the process; have a full grasp on nationwide support, respond to the demand for emergency materials as well as their production, circulation, storage and allocation, and coordinate affairs concerning supply and demand, production, storage, transportation, etc. of emergency materials; coordinate and arrange funding for production, purchase (import), storage, etc. of emergency materials (including instruments and equipment); oversee the implementation of support measures in regions and by departments concerned, and closely track progress in the process; monitor market situations regarding supply and demand of basic living supplies and cleaning supplies; maintain the market order through market supervision and management, price stabilization and punishment of business violations, and where needed, deliberate on and take related measures; provide feedback on support work and summarize such information in a timely manner; complete other assignments from leaders Dissemination and communication group CCPPD International Communication Office, State Administration of Radio, Film, and Television (SARFT), General Administration of Press and Publication (GAPP), and main central news media Report in a timely fashion on Influenza A (H1N1) situations and on progress made in major efforts under the joint prevention and control mechanism, and guide public opinion positively and correctly; arrange news releases on the joint prevention and control mechanism, and where necessary, organize press conferences; track public opinion at home and abroad, and clarify facts in a timely fashion; strengthen management and guidance on the release of online news; assist related departments to disseminate knowledge about protection against Influenza A (H1N1), to increase the population's self-protection capabilities; collect, sort out and report information on public opinion monitoring, etc. in real time Foreign collaboration group MOF and MOH MOC, AQSIQ, CNTA, and the Hong Kong and Macao Affairs Office, Taiwan Affairs Office, and Overseas Chinese Affairs of the State Council Handle foreign affairs related to Influenza A (H1N1) prevention and control; coordinate the handling of major foreign affairs in terms of Influenza A (H1N1) prevention and control, and coordinate affairs related to Influenza A (H1N1) prevention and control in the regions of Hong Kong, Macao and Taiwan; collect and report information on, and coordinate, affairs of collaboration and communication with international organizations such as the WHO, regional organizations, foreign governments, as well as with Hong Kong, Macao and Taiwan; collect and report information on national organizations and epidemic situations abroad; provide guidance on and oversee influenza response efforts by Chinese staff sent abroad, as well as the protection of overseas Chinese; promote related departments' external communication about China's response efforts and measures; urge related departments to track experience and practices of other countries and international organizations with regard to coping with Influenza A (H1N1); work with related departments to track and survey impacts of the virus on the country's diplomacy, economy and trade, tourism, and people migration, and make policy recommendations accordingly; oversee foreign aid and receive international donations; and complete assignments from the State Council Science and technology group MOST MOH, MOA, MOE, AQSIQ, GLD Health Department, and Chinese Academy of Sciences (CAS) Decide on plans for technical research into Influenza A (H1N1) prevention and treatment; organize research projects aimed to tackle technological difficulties in response to the virus; coordinate and solve technological issues involved in the development and application of monitoring technologies, drugs and vaccines on a unified basis; collect, sort out and report on such information as latest research developments in a timely fashion Animal husbandry and veterinary group MOA NDRC, MOF, MIIT, MOT, MOC, State Administration for Industry and Commerce (SAIC), AQSIQ, GLD Department of Military Supplies, Materials and Oils, and PAPF Logistics Department Closely monitor, test, prevent, and control animal epidemics; provide scientific judgments on epidemic trends; adjust prevention and control measures on a continued basis; and strengthen swine management through integrated epidemic prevention and control measures, with emphasis placed on swine epidemic monitoring and epidemiological surveys and on ensuring stable swine reproduction Expert advisory committee MOH Experts from MOE, MOA, AQSIQ, GLD Health Department, CAS, as well as local health departments Provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies The Operation of the Joint National Influenza A (H1N1) Prevention and Control Mechanism Thirty three meetings were convened under the Joint National Prevention and Control Mechanism, in which regulations were formulated, signed, and issued for local implementation. Implementation issues would be reported in real time to related State Council departments for instructions. The establishment of the Joint National Prevention and Control Mechanism played a crucial role in the scientific and orderly response to Influenza A (H1N1), in that (1) the mechanism raised the priority level for Influenza A (H1N1) prevention and control for related departments and local governments, (2) clarified and divided responsibilities, (3) addressed investment issues, and (4) enhanced inter-departmental cooperation. Working Consultation System A consultation system and a liaison meeting system were established under the Joint National Prevention and Control Mechanism, which were designed to ensure effective implementation of prevention and control measures. Specific issues in prevention and control work would be resolved through consultation at liaisons meetings, and major issues decided by plenary meetings. Each working group established a fixed meeting system where they could provide timely progress reports, discuss and address problems, and push ahead with prevention and control work within their fields. The establishment and improvement of the coordination mechanism remarkably increased the efficiency of inter-departmental coordination and response efforts. This success evinces the importance of a multi-departmental coordination and collaboration mechanism based on risk communication for effective epidemic prevention and control. Firstly, documents were issued jointly by related departments under the new Joint National Prevention and Control Mechanism. For example, on June 22nd, 2009, the MOE and the MOH jointly issued the Work Plan for Influenza A (H1N1) Prevention and Control in Schools (Tentative) ; the two ministries also jointly issued the Urgent Notice on Strengthening the Management of Vaccination to Students against Influenza A (H1N1) on November 4th, and the Notice on Strengthening Influenza A (H1N1) Prevention and Control in Rural Schools on November 26th. Secondly, members were supplemented to the Joint Prevention and Control Mechanism to improve efficiency. On November 23rd, 2009, based on the demands of the nation's prevention and control efforts, the MOH Office of Health Emergency sent letters to the General Office of the Supreme People's Court, the Ministry of Supervision (MOS), the Ministry of Civil Affairs (MCA), the Ministry of Human Resources and Social Security (MOHRSS), and the Legislative Affairs Office (LAO) of the State Council, adding them all as members of the mechanism. Thirdly, horizontal collaboration between related departments was strengthened. For example, on April 29th, 2009, the MOH Office of Health Emergency sent a letter to the MIIT Department of Consumer Goods Industry recommending an increase in the national stockpile of supplies necessary for Influenza A (H1N1) prevention and control including medical supplies and response gear. In another example, the Office of Health Emergency, Department of Medical Affairs, and Bureau of Disease Control and Prevention of the MOH, the China CDC, and the Chinese Medical Association, jointly formulated the Technical Guidance on Prevention and Control of Human - Swine Influenza , and the Plan for Diagnosis and Treatment of Human - Swine Influenza (2009) . One more example occurred on July 30th, when the Foreign Collaboration Group for the Joint Prevention and Control Mechanism, along with the NDRC, the MIIT, and the SFDA met to discuss donating Influenza A (H1N1) prevention and control materials to the WHO and developing countries affected by the pandemic. Information Reporting System During the course of the epidemic response efforts, risk evaluation, and risk management, it was necessary to build unobstructed information exchange channels to ensure the accurate transfer of data and information. Under the Joint Prevention and Control Mechanism, each of the lead departments for the work groups and the Expert Advisory Committee appointed people to collect, sort out, tabulate and analyze their work groups' epidemic information and latest progress on a daily basis, and to report in writing daily data collected by 18:00 to the General Group prior to 20:00 p.m. The General Group would then prioritize epidemic information and progress reports, and submit it representing the entire Joint Prevention and Control Mechanism to the General Duty Office of the State Council. Strengthened coordination and communication between the work groups and their members ensured an unobstructed flow of information as the working groups were informed of major issues as they happened. The General Group was also charged with publishing information crucial for the public's knowledge on epidemic prevention and control. Updates on Influenza A (H1N1) prevention and control were submitted via four reporting systems: the Disease Monitoring Information Reporting Management Subsystem of the China Information System for Disease Control and Prevention, the Public Health Emergency Reported Information Management System, the China Influenza Information Monitoring System, and Administrative Reporting System for Health Departments. Reported information mainly included: ongoing epidemic situations, monitoring results from sentinel hospitals and network laboratories, progress in vaccination and results of side effects monitoring, and ongoing regional and departmental response efforts. With Influenza A (H1N1) cases rising rapidly, issues with epidemic information reporting began to occur, such as overlapping reports, large discrepancies between confirmed reported cases and actual cases, and the circulation of ambiguous epidemic information. To better and more accurately reflect national epidemic situations and trends, the MOH General Office issued the Notice on Strengthening Work of Reporting Deaths from Influenza A (H1N1) on November 4th, 2009, and the Notice on Adjusting the Work of Reporting Information on the Influenza A (H1N1) Epidemic on November 13th, 2009. By April 7th, 2010, the General Group had submitted more than 700 work reports regarding Influenza A (H1N1), and compiled and published over 200 Response Progress to Influenza A (H1N1) under the Joint Prevention and Control Mechanism reports. Information on epidemic trends and response efforts was released in a timely, open, and transparent manner, and by this time eight news conferences and nine press briefings had been held regarding the latest progress in Influenza A (H1N1) prevention and control. Supervision and Inspection System To ensure that prevention and control measures were implemented effectively and efficiently, the work groups each established a supervision and inspection system by which to examine routine work on a regular basis, identify existing deficiencies and problems, and supervise and inspect response contingency plans and procedures, operational capacities, epidemic monitoring, epidemiological investigations, designated hospitals and their isolated areas, medical observation, material supplies, staff training, and information dissemination on prevention and control. The Establishment of the Joint National Influenza A (H1N1) Prevention and Control Mechanism The MOH responded immediately upon receiving a disease outbreak notice from the WHO on April 25th, 2009, as the MOH General Office launched the working mechanism comprised of the MOH Leading Group and Expert Panel for Influenza Pandemic Prevention and Control [in accordance with the Influenza Pandemic Preparedness and Response Plan of the Ministry of Health (Tentative) ]. The MOH also issued a Notice on Strengthening Preparedness for and Response to Human Swine Influenza to health departments requiring the following: medical and disease control and prevention institutions at various levels to strengthen cases monitoring and reporting; and prepare for Influenza A (H1N1) in terms of processing, technology, manpower, and material resources. At the same time, the MOH immediately forwarded related information to the MOA and the AQSIQ. On April 26th, Health Minister Chen Zhu convened a meeting of the MOH Leading Group and Expert Panel for Influenza Pandemic Prevention and Control, at which the attendees analyzed swine influenza situations in the United States and Mexico, predicted epidemic trends, and deliberated on domestic strategies and measures to cope with a swine flu pandemic. Health Minister Chen also held an inter-departmental meeting with the MOA, the AQSIQ, and other ministries to analyze epidemic trends and discuss response strategies and measures. Immediately after the meeting was over, the MOH reported in writing that very night to the State Council on the progress of epidemic prevention and control work. On April 27th, following the emergency meeting held in Geneva, the WHO elevated the pandemic alert level from Phase 3 to Phase 4, stating that the "swine flu" was widespread and was being transmitted by humans in different ways. General Secretary Hu Jintao issued instructions to place prevention and control against this virus as the nation's top priority. On the same day, Vice Premier Li Keqiang convened the State Council Meeting regarding the Human-Swine Influenza Prevention Working Mechanism, resulting in the decision to establish a multi-departmental working mechanism for joint prevention and control of the human-swine influenza. As required by the State Council meeting, the MOH called together the Publicity Department of the Communist Party of China (CCPPD), the Ministry of Foreign Affairs (MFA), the NDRC, the MIIT, the MOF, the Ministry of Transport (MOT), the MOA, the MOC, the AQSIQ, the China National Tourism Administration (CNTA), the Civil Aviation Administration of China (CAAC) among other departments on that very night for a meeting to deliberate on Influenza A (H1N1) prevention and control; the meeting established the multi-departmental working mechanism for joint prevention and control of human-swine influenza and the Notice on Strengthening Human - Swine Influenza Prevention and Control was drafted and published on the night of April 27th after State Council review. The MOH issued the Notice of the MOH General Office on Strengthening Preparedness for and Response to Human Swine Influenza . On April 28th, Premier Wen Jiabao convened a State Council Executive Meeting which deliberated on how to strengthen national response to human-swine influenza; at the meeting they defined the overall prevention and control principles and strategies of "taking threats to public health seriously, responding actively, and coping with the epidemic in a scientific manner according to law through joint prevention and control efforts." On April 29th, the WHO raised its pandemic alert level from Phase 4 to Phase 5. On April 30th, at a press conference held at the State Council Information Office, the MOH declared the establishment of a multi-departmental working mechanism for joint prevention and control against the human-swine influenza, which would be spearheaded by the MOH. Under this mechanism, 33 departments and institutions (which later increased to 38) constituted 8 work groups—General Office, Ports, Healthcare, Support, Dissemination and Communication, Foreign Collaboration, Science and Technology, and Animal Husbandry and Veterinary—and an expert committee, forming a "8 + 1" pattern for joint prevention and control efforts. On the afternoon of May 1st, the joint prevention and control mechanism held its second joint conference, renaming the human-swine influenza which was occurring in Mexico and the United States to "Influenza A (H1N1)." The former wording of "multi-departmental work mechanism for joint prevention and control of human-swine influenza" was changed to the "Joint National Influenza A (H1N1) Prevention and Control Mechanism" and roles and responsibilities were outlined for the mechanism, work groups, and the expert committee. At the same time, a meeting system for all members and liaisons was established. Problems in principle which a work group encountered would be solved by the work group itself, and those which the work group struggled with would be settled through coordination under the joint prevention and control mechanism—general affairs would be solved by regular liaison meetings, and major issues decided by plenary meetings. The Composition of the Joint National Influenza A (H1N1) Prevention and Control Mechanism Organizational Structure Health Minister Chen Zhu and MOH Party Group Secretary and Vice Health Minister Zhang Mao chaired the Joint National Influenza A (H1N1) Prevention and Control Mechanism, and the mechanism was comprised of eight working groups—General Office, Ports, Healthcare, Support, Dissemination and Communication, Foreign Collaboration, Science and Technology, and Animal Husbandry and Veterinary—and an Expert Advisory Committee. The heads of the work groups, the leaders of the Health Department of the People's Liberation Army General Logistics Department (GLD), the Logistics Department of the Chinese People's Armed Police Force (PAPF), and the chairman of the Expert Advisory Committee, served as members of the joint prevention and control mechanism. The work groups, the GLD Health Department and the PAPF Logistics Department each designated one or two departmental-level officials as liaisons for routine communication purposes. Main Responsibilities and the Division of Labor The main duties of the Joint National Influenza A (H1N1) Prevention and Control Mechanism included: meet regularly to evaluate epidemic trends and determine prevention and control strategies; formulate prevention and control policies, response plans and major measures; coordinate and provide guidance on the implementation by all related departments and in various regions of prevention and control measures; and organize supervision and inspection activities concerning the implementation of prevention and control measures. See Table 4.1 for the main responsibilities of the work groups and the Expert Advisory Committee. Table 4.1 Composition of the work groups and the expert advisory committee for the joint national influenza A (H1N1) prevention and control mechanism Work group Leading organization Members Duties General office group MOH CCPPD, MFA, NDRC, MOST, MOA, AQSIQ, GLD Health Department, PAPF Logistics Department, Expert Advisory Committee Comprehensively coordinate routine affairs among departments of the joint prevention and control mechanism; organize regular meetings of the joint prevention and control mechanism and oversee the handling of top agendas; collect, sort out, and report to higher-ups about progress in prevention and control efforts; prepare progress reports on the joint prevention and control mechanism; publish information on epidemic situations and response efforts; ensure the consistency in style of writing for documents intended for outside use; and shoulder other assignments from leaders Ports group AQSIQ MFA, Ministry of Public Security (MPS), MOT, MOR, MOA, MOC, MOH, General Administration of Customs (GAC), CNTA, State Council Information Office (SCIO), CAAC, and State Post Bureau (SPB) Carry out health quarantine, surveillance, health supervision and management at entry-exit ports; collect, sort out and report on information regarding epidemic developments, response measures, etc. abroad; ensure the quality and standardization of emergency materials Healthcare group MOH State Administration of Traditional Chinese Medicine (SATCM), AQSIQ, GLD Health Department, PAPF Logistics Department, and China CDC Organize the formulation and revision of technical solutions to the diagnosis, treatment, prevention and control of Influenza A (H1N1), and oversee the implementation of those solutions; provide guidance on nationwide monitoring, reporting, and epidemiological surveys of epidemic situations, specimen collection and testing, epidemic management, etc.; direct national medical treatment plans of Influenza A (H1N1); send experts, as needed for epidemic prevention and control, to go and help with epidemic handling and medical treatment in priority regions; provide recommendations on strategies and measures for improving response efforts; and undertake other assignments from leaders Support group NDRC MIIT, MOF, MOT, MOR, MOC, MOH, AQSIQ, CAAC, SFDA, and GLD Health Department Plan as a whole the support logistics for emergency materials, and deliberate on and coordinate major issues involved in the process; have a full grasp on nationwide support, respond to the demand for emergency materials as well as their production, circulation, storage and allocation, and coordinate affairs concerning supply and demand, production, storage, transportation, etc. of emergency materials; coordinate and arrange funding for production, purchase (import), storage, etc. of emergency materials (including instruments and equipment); oversee the implementation of support measures in regions and by departments concerned, and closely track progress in the process; monitor market situations regarding supply and demand of basic living supplies and cleaning supplies; maintain the market order through market supervision and management, price stabilization and punishment of business violations, and where needed, deliberate on and take related measures; provide feedback on support work and summarize such information in a timely manner; complete other assignments from leaders Dissemination and communication group CCPPD International Communication Office, State Administration of Radio, Film, and Television (SARFT), General Administration of Press and Publication (GAPP), and main central news media Report in a timely fashion on Influenza A (H1N1) situations and on progress made in major efforts under the joint prevention and control mechanism, and guide public opinion positively and correctly; arrange news releases on the joint prevention and control mechanism, and where necessary, organize press conferences; track public opinion at home and abroad, and clarify facts in a timely fashion; strengthen management and guidance on the release of online news; assist related departments to disseminate knowledge about protection against Influenza A (H1N1), to increase the population's self-protection capabilities; collect, sort out and report information on public opinion monitoring, etc. in real time Foreign collaboration group MOF and MOH MOC, AQSIQ, CNTA, and the Hong Kong and Macao Affairs Office, Taiwan Affairs Office, and Overseas Chinese Affairs of the State Council Handle foreign affairs related to Influenza A (H1N1) prevention and control; coordinate the handling of major foreign affairs in terms of Influenza A (H1N1) prevention and control, and coordinate affairs related to Influenza A (H1N1) prevention and control in the regions of Hong Kong, Macao and Taiwan; collect and report information on, and coordinate, affairs of collaboration and communication with international organizations such as the WHO, regional organizations, foreign governments, as well as with Hong Kong, Macao and Taiwan; collect and report information on national organizations and epidemic situations abroad; provide guidance on and oversee influenza response efforts by Chinese staff sent abroad, as well as the protection of overseas Chinese; promote related departments' external communication about China's response efforts and measures; urge related departments to track experience and practices of other countries and international organizations with regard to coping with Influenza A (H1N1); work with related departments to track and survey impacts of the virus on the country's diplomacy, economy and trade, tourism, and people migration, and make policy recommendations accordingly; oversee foreign aid and receive international donations; and complete assignments from the State Council Science and technology group MOST MOH, MOA, MOE, AQSIQ, GLD Health Department, and Chinese Academy of Sciences (CAS) Decide on plans for technical research into Influenza A (H1N1) prevention and treatment; organize research projects aimed to tackle technological difficulties in response to the virus; coordinate and solve technological issues involved in the development and application of monitoring technologies, drugs and vaccines on a unified basis; collect, sort out and report on such information as latest research developments in a timely fashion Animal husbandry and veterinary group MOA NDRC, MOF, MIIT, MOT, MOC, State Administration for Industry and Commerce (SAIC), AQSIQ, GLD Department of Military Supplies, Materials and Oils, and PAPF Logistics Department Closely monitor, test, prevent, and control animal epidemics; provide scientific judgments on epidemic trends; adjust prevention and control measures on a continued basis; and strengthen swine management through integrated epidemic prevention and control measures, with emphasis placed on swine epidemic monitoring and epidemiological surveys and on ensuring stable swine reproduction Expert advisory committee MOH Experts from MOE, MOA, AQSIQ, GLD Health Department, CAS, as well as local health departments Provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies Organizational Structure Health Minister Chen Zhu and MOH Party Group Secretary and Vice Health Minister Zhang Mao chaired the Joint National Influenza A (H1N1) Prevention and Control Mechanism, and the mechanism was comprised of eight working groups—General Office, Ports, Healthcare, Support, Dissemination and Communication, Foreign Collaboration, Science and Technology, and Animal Husbandry and Veterinary—and an Expert Advisory Committee. The heads of the work groups, the leaders of the Health Department of the People's Liberation Army General Logistics Department (GLD), the Logistics Department of the Chinese People's Armed Police Force (PAPF), and the chairman of the Expert Advisory Committee, served as members of the joint prevention and control mechanism. The work groups, the GLD Health Department and the PAPF Logistics Department each designated one or two departmental-level officials as liaisons for routine communication purposes. Main Responsibilities and the Division of Labor The main duties of the Joint National Influenza A (H1N1) Prevention and Control Mechanism included: meet regularly to evaluate epidemic trends and determine prevention and control strategies; formulate prevention and control policies, response plans and major measures; coordinate and provide guidance on the implementation by all related departments and in various regions of prevention and control measures; and organize supervision and inspection activities concerning the implementation of prevention and control measures. See Table 4.1 for the main responsibilities of the work groups and the Expert Advisory Committee. Table 4.1 Composition of the work groups and the expert advisory committee for the joint national influenza A (H1N1) prevention and control mechanism Work group Leading organization Members Duties General office group MOH CCPPD, MFA, NDRC, MOST, MOA, AQSIQ, GLD Health Department, PAPF Logistics Department, Expert Advisory Committee Comprehensively coordinate routine affairs among departments of the joint prevention and control mechanism; organize regular meetings of the joint prevention and control mechanism and oversee the handling of top agendas; collect, sort out, and report to higher-ups about progress in prevention and control efforts; prepare progress reports on the joint prevention and control mechanism; publish information on epidemic situations and response efforts; ensure the consistency in style of writing for documents intended for outside use; and shoulder other assignments from leaders Ports group AQSIQ MFA, Ministry of Public Security (MPS), MOT, MOR, MOA, MOC, MOH, General Administration of Customs (GAC), CNTA, State Council Information Office (SCIO), CAAC, and State Post Bureau (SPB) Carry out health quarantine, surveillance, health supervision and management at entry-exit ports; collect, sort out and report on information regarding epidemic developments, response measures, etc. abroad; ensure the quality and standardization of emergency materials Healthcare group MOH State Administration of Traditional Chinese Medicine (SATCM), AQSIQ, GLD Health Department, PAPF Logistics Department, and China CDC Organize the formulation and revision of technical solutions to the diagnosis, treatment, prevention and control of Influenza A (H1N1), and oversee the implementation of those solutions; provide guidance on nationwide monitoring, reporting, and epidemiological surveys of epidemic situations, specimen collection and testing, epidemic management, etc.; direct national medical treatment plans of Influenza A (H1N1); send experts, as needed for epidemic prevention and control, to go and help with epidemic handling and medical treatment in priority regions; provide recommendations on strategies and measures for improving response efforts; and undertake other assignments from leaders Support group NDRC MIIT, MOF, MOT, MOR, MOC, MOH, AQSIQ, CAAC, SFDA, and GLD Health Department Plan as a whole the support logistics for emergency materials, and deliberate on and coordinate major issues involved in the process; have a full grasp on nationwide support, respond to the demand for emergency materials as well as their production, circulation, storage and allocation, and coordinate affairs concerning supply and demand, production, storage, transportation, etc. of emergency materials; coordinate and arrange funding for production, purchase (import), storage, etc. of emergency materials (including instruments and equipment); oversee the implementation of support measures in regions and by departments concerned, and closely track progress in the process; monitor market situations regarding supply and demand of basic living supplies and cleaning supplies; maintain the market order through market supervision and management, price stabilization and punishment of business violations, and where needed, deliberate on and take related measures; provide feedback on support work and summarize such information in a timely manner; complete other assignments from leaders Dissemination and communication group CCPPD International Communication Office, State Administration of Radio, Film, and Television (SARFT), General Administration of Press and Publication (GAPP), and main central news media Report in a timely fashion on Influenza A (H1N1) situations and on progress made in major efforts under the joint prevention and control mechanism, and guide public opinion positively and correctly; arrange news releases on the joint prevention and control mechanism, and where necessary, organize press conferences; track public opinion at home and abroad, and clarify facts in a timely fashion; strengthen management and guidance on the release of online news; assist related departments to disseminate knowledge about protection against Influenza A (H1N1), to increase the population's self-protection capabilities; collect, sort out and report information on public opinion monitoring, etc. in real time Foreign collaboration group MOF and MOH MOC, AQSIQ, CNTA, and the Hong Kong and Macao Affairs Office, Taiwan Affairs Office, and Overseas Chinese Affairs of the State Council Handle foreign affairs related to Influenza A (H1N1) prevention and control; coordinate the handling of major foreign affairs in terms of Influenza A (H1N1) prevention and control, and coordinate affairs related to Influenza A (H1N1) prevention and control in the regions of Hong Kong, Macao and Taiwan; collect and report information on, and coordinate, affairs of collaboration and communication with international organizations such as the WHO, regional organizations, foreign governments, as well as with Hong Kong, Macao and Taiwan; collect and report information on national organizations and epidemic situations abroad; provide guidance on and oversee influenza response efforts by Chinese staff sent abroad, as well as the protection of overseas Chinese; promote related departments' external communication about China's response efforts and measures; urge related departments to track experience and practices of other countries and international organizations with regard to coping with Influenza A (H1N1); work with related departments to track and survey impacts of the virus on the country's diplomacy, economy and trade, tourism, and people migration, and make policy recommendations accordingly; oversee foreign aid and receive international donations; and complete assignments from the State Council Science and technology group MOST MOH, MOA, MOE, AQSIQ, GLD Health Department, and Chinese Academy of Sciences (CAS) Decide on plans for technical research into Influenza A (H1N1) prevention and treatment; organize research projects aimed to tackle technological difficulties in response to the virus; coordinate and solve technological issues involved in the development and application of monitoring technologies, drugs and vaccines on a unified basis; collect, sort out and report on such information as latest research developments in a timely fashion Animal husbandry and veterinary group MOA NDRC, MOF, MIIT, MOT, MOC, State Administration for Industry and Commerce (SAIC), AQSIQ, GLD Department of Military Supplies, Materials and Oils, and PAPF Logistics Department Closely monitor, test, prevent, and control animal epidemics; provide scientific judgments on epidemic trends; adjust prevention and control measures on a continued basis; and strengthen swine management through integrated epidemic prevention and control measures, with emphasis placed on swine epidemic monitoring and epidemiological surveys and on ensuring stable swine reproduction Expert advisory committee MOH Experts from MOE, MOA, AQSIQ, GLD Health Department, CAS, as well as local health departments Provide recommendations on appropriate response levels and countermeasures for confirmed public health emergencies; give advice on public health emergency preparation; participate in formulating and revising contingency plans and technical solutions to public health emergencies; provide technical guidance on public health emergency mitigation; advise on the termination of health emergency countermeasures; provide post-emergency evaluations; and undertake other assignments from public health emergency operations and management bodies The Operation of the Joint National Influenza A (H1N1) Prevention and Control Mechanism Thirty three meetings were convened under the Joint National Prevention and Control Mechanism, in which regulations were formulated, signed, and issued for local implementation. Implementation issues would be reported in real time to related State Council departments for instructions. The establishment of the Joint National Prevention and Control Mechanism played a crucial role in the scientific and orderly response to Influenza A (H1N1), in that (1) the mechanism raised the priority level for Influenza A (H1N1) prevention and control for related departments and local governments, (2) clarified and divided responsibilities, (3) addressed investment issues, and (4) enhanced inter-departmental cooperation. Working Consultation System A consultation system and a liaison meeting system were established under the Joint National Prevention and Control Mechanism, which were designed to ensure effective implementation of prevention and control measures. Specific issues in prevention and control work would be resolved through consultation at liaisons meetings, and major issues decided by plenary meetings. Each working group established a fixed meeting system where they could provide timely progress reports, discuss and address problems, and push ahead with prevention and control work within their fields. The establishment and improvement of the coordination mechanism remarkably increased the efficiency of inter-departmental coordination and response efforts. This success evinces the importance of a multi-departmental coordination and collaboration mechanism based on risk communication for effective epidemic prevention and control. Firstly, documents were issued jointly by related departments under the new Joint National Prevention and Control Mechanism. For example, on June 22nd, 2009, the MOE and the MOH jointly issued the Work Plan for Influenza A (H1N1) Prevention and Control in Schools (Tentative) ; the two ministries also jointly issued the Urgent Notice on Strengthening the Management of Vaccination to Students against Influenza A (H1N1) on November 4th, and the Notice on Strengthening Influenza A (H1N1) Prevention and Control in Rural Schools on November 26th. Secondly, members were supplemented to the Joint Prevention and Control Mechanism to improve efficiency. On November 23rd, 2009, based on the demands of the nation's prevention and control efforts, the MOH Office of Health Emergency sent letters to the General Office of the Supreme People's Court, the Ministry of Supervision (MOS), the Ministry of Civil Affairs (MCA), the Ministry of Human Resources and Social Security (MOHRSS), and the Legislative Affairs Office (LAO) of the State Council, adding them all as members of the mechanism. Thirdly, horizontal collaboration between related departments was strengthened. For example, on April 29th, 2009, the MOH Office of Health Emergency sent a letter to the MIIT Department of Consumer Goods Industry recommending an increase in the national stockpile of supplies necessary for Influenza A (H1N1) prevention and control including medical supplies and response gear. In another example, the Office of Health Emergency, Department of Medical Affairs, and Bureau of Disease Control and Prevention of the MOH, the China CDC, and the Chinese Medical Association, jointly formulated the Technical Guidance on Prevention and Control of Human - Swine Influenza , and the Plan for Diagnosis and Treatment of Human - Swine Influenza (2009) . One more example occurred on July 30th, when the Foreign Collaboration Group for the Joint Prevention and Control Mechanism, along with the NDRC, the MIIT, and the SFDA met to discuss donating Influenza A (H1N1) prevention and control materials to the WHO and developing countries affected by the pandemic. Information Reporting System During the course of the epidemic response efforts, risk evaluation, and risk management, it was necessary to build unobstructed information exchange channels to ensure the accurate transfer of data and information. Under the Joint Prevention and Control Mechanism, each of the lead departments for the work groups and the Expert Advisory Committee appointed people to collect, sort out, tabulate and analyze their work groups' epidemic information and latest progress on a daily basis, and to report in writing daily data collected by 18:00 to the General Group prior to 20:00 p.m. The General Group would then prioritize epidemic information and progress reports, and submit it representing the entire Joint Prevention and Control Mechanism to the General Duty Office of the State Council. Strengthened coordination and communication between the work groups and their members ensured an unobstructed flow of information as the working groups were informed of major issues as they happened. The General Group was also charged with publishing information crucial for the public's knowledge on epidemic prevention and control. Updates on Influenza A (H1N1) prevention and control were submitted via four reporting systems: the Disease Monitoring Information Reporting Management Subsystem of the China Information System for Disease Control and Prevention, the Public Health Emergency Reported Information Management System, the China Influenza Information Monitoring System, and Administrative Reporting System for Health Departments. Reported information mainly included: ongoing epidemic situations, monitoring results from sentinel hospitals and network laboratories, progress in vaccination and results of side effects monitoring, and ongoing regional and departmental response efforts. With Influenza A (H1N1) cases rising rapidly, issues with epidemic information reporting began to occur, such as overlapping reports, large discrepancies between confirmed reported cases and actual cases, and the circulation of ambiguous epidemic information. To better and more accurately reflect national epidemic situations and trends, the MOH General Office issued the Notice on Strengthening Work of Reporting Deaths from Influenza A (H1N1) on November 4th, 2009, and the Notice on Adjusting the Work of Reporting Information on the Influenza A (H1N1) Epidemic on November 13th, 2009. By April 7th, 2010, the General Group had submitted more than 700 work reports regarding Influenza A (H1N1), and compiled and published over 200 Response Progress to Influenza A (H1N1) under the Joint Prevention and Control Mechanism reports. Information on epidemic trends and response efforts was released in a timely, open, and transparent manner, and by this time eight news conferences and nine press briefings had been held regarding the latest progress in Influenza A (H1N1) prevention and control. Supervision and Inspection System To ensure that prevention and control measures were implemented effectively and efficiently, the work groups each established a supervision and inspection system by which to examine routine work on a regular basis, identify existing deficiencies and problems, and supervise and inspect response contingency plans and procedures, operational capacities, epidemic monitoring, epidemiological investigations, designated hospitals and their isolated areas, medical observation, material supplies, staff training, and information dissemination on prevention and control. Working Consultation System A consultation system and a liaison meeting system were established under the Joint National Prevention and Control Mechanism, which were designed to ensure effective implementation of prevention and control measures. Specific issues in prevention and control work would be resolved through consultation at liaisons meetings, and major issues decided by plenary meetings. Each working group established a fixed meeting system where they could provide timely progress reports, discuss and address problems, and push ahead with prevention and control work within their fields. The establishment and improvement of the coordination mechanism remarkably increased the efficiency of inter-departmental coordination and response efforts. This success evinces the importance of a multi-departmental coordination and collaboration mechanism based on risk communication for effective epidemic prevention and control. Firstly, documents were issued jointly by related departments under the new Joint National Prevention and Control Mechanism. For example, on June 22nd, 2009, the MOE and the MOH jointly issued the Work Plan for Influenza A (H1N1) Prevention and Control in Schools (Tentative) ; the two ministries also jointly issued the Urgent Notice on Strengthening the Management of Vaccination to Students against Influenza A (H1N1) on November 4th, and the Notice on Strengthening Influenza A (H1N1) Prevention and Control in Rural Schools on November 26th. Secondly, members were supplemented to the Joint Prevention and Control Mechanism to improve efficiency. On November 23rd, 2009, based on the demands of the nation's prevention and control efforts, the MOH Office of Health Emergency sent letters to the General Office of the Supreme People's Court, the Ministry of Supervision (MOS), the Ministry of Civil Affairs (MCA), the Ministry of Human Resources and Social Security (MOHRSS), and the Legislative Affairs Office (LAO) of the State Council, adding them all as members of the mechanism. Thirdly, horizontal collaboration between related departments was strengthened. For example, on April 29th, 2009, the MOH Office of Health Emergency sent a letter to the MIIT Department of Consumer Goods Industry recommending an increase in the national stockpile of supplies necessary for Influenza A (H1N1) prevention and control including medical supplies and response gear. In another example, the Office of Health Emergency, Department of Medical Affairs, and Bureau of Disease Control and Prevention of the MOH, the China CDC, and the Chinese Medical Association, jointly formulated the Technical Guidance on Prevention and Control of Human - Swine Influenza , and the Plan for Diagnosis and Treatment of Human - Swine Influenza (2009) . One more example occurred on July 30th, when the Foreign Collaboration Group for the Joint Prevention and Control Mechanism, along with the NDRC, the MIIT, and the SFDA met to discuss donating Influenza A (H1N1) prevention and control materials to the WHO and developing countries affected by the pandemic. Information Reporting System During the course of the epidemic response efforts, risk evaluation, and risk management, it was necessary to build unobstructed information exchange channels to ensure the accurate transfer of data and information. Under the Joint Prevention and Control Mechanism, each of the lead departments for the work groups and the Expert Advisory Committee appointed people to collect, sort out, tabulate and analyze their work groups' epidemic information and latest progress on a daily basis, and to report in writing daily data collected by 18:00 to the General Group prior to 20:00 p.m. The General Group would then prioritize epidemic information and progress reports, and submit it representing the entire Joint Prevention and Control Mechanism to the General Duty Office of the State Council. Strengthened coordination and communication between the work groups and their members ensured an unobstructed flow of information as the working groups were informed of major issues as they happened. The General Group was also charged with publishing information crucial for the public's knowledge on epidemic prevention and control. Updates on Influenza A (H1N1) prevention and control were submitted via four reporting systems: the Disease Monitoring Information Reporting Management Subsystem of the China Information System for Disease Control and Prevention, the Public Health Emergency Reported Information Management System, the China Influenza Information Monitoring System, and Administrative Reporting System for Health Departments. Reported information mainly included: ongoing epidemic situations, monitoring results from sentinel hospitals and network laboratories, progress in vaccination and results of side effects monitoring, and ongoing regional and departmental response efforts. With Influenza A (H1N1) cases rising rapidly, issues with epidemic information reporting began to occur, such as overlapping reports, large discrepancies between confirmed reported cases and actual cases, and the circulation of ambiguous epidemic information. To better and more accurately reflect national epidemic situations and trends, the MOH General Office issued the Notice on Strengthening Work of Reporting Deaths from Influenza A (H1N1) on November 4th, 2009, and the Notice on Adjusting the Work of Reporting Information on the Influenza A (H1N1) Epidemic on November 13th, 2009. By April 7th, 2010, the General Group had submitted more than 700 work reports regarding Influenza A (H1N1), and compiled and published over 200 Response Progress to Influenza A (H1N1) under the Joint Prevention and Control Mechanism reports. Information on epidemic trends and response efforts was released in a timely, open, and transparent manner, and by this time eight news conferences and nine press briefings had been held regarding the latest progress in Influenza A (H1N1) prevention and control. Supervision and Inspection System To ensure that prevention and control measures were implemented effectively and efficiently, the work groups each established a supervision and inspection system by which to examine routine work on a regular basis, identify existing deficiencies and problems, and supervise and inspect response contingency plans and procedures, operational capacities, epidemic monitoring, epidemiological investigations, designated hospitals and their isolated areas, medical observation, material supplies, staff training, and information dissemination on prevention and control. The Establishment, Composition, and Operation of Local Influenza A (H1N1) Prevention and Control Mechanisms In the course of Influenza A(H1N1) prevention and control efforts, local governments, as instructed by the central government, examined their own conditions and established local bodies to command and coordinate response measures. The local departments worked together to implement disease prevention and control measures in priority areas and among targeted groups, and ensured continued epidemic monitoring and treatment. The Establishment of Local Influenza A (H1N1) Prevention and Control Mechanisms The following three modes mainly represent actual prevention and control measures adopted by local governments. The first work mode was similar to the National Joint Prevention and Control Mechanism. For example, Shaanxi set up a leading group for Influenza A (H1N1) prevention and control, whose office was located inside the provincial Department of Health and their local structure followed the "8 + 1" joint prevention and control mechanism model. Guangdong established a joint prevention and control mechanism with the participation of thirty two departments, and nine work groups and three panels of clinical, disease prevention and control, and etiological experts functioned under the mechanism. On April 30th, Fujian established an Influenza A (H1N1) prevention and control work group, headed by a provincial government official; the office was located inside the building of the provincial Department of Health whose emergency management office was charged with performing routine work for the group. The Emergency Guidance on Influenza A (H1N1) Prevention and Control in Fujian (Tentative) , issued on May 10th, 2009, outlined response guidance as "prevention first through joint prevention and control, timely management, and level-by-level responsibility." The second work mode was the emergency operations center or leading group. After discovering their first confirmed Influenza A (H1N1) case, some provinces and cities upgraded their existing disease prevention and control mechanisms and established an Influenza A (H1N1) response leading group or emergency operations center. Beijing was the first in the country to establish a municipal-level public health emergency operations center in May 2006, and had earlier (April 25th, 2009) launched a public health emergency response mechanism after the WHO declared the outbreak of swine influenza in Mexico; an Influenza Prevention and Control Office (at the general office of the municipal government before it relocated to the Municipal Bureau of Health) was established under the emergency operations center, whose members included twenty two committees, eighteen district and county governments, and the GLD Health Department. This Control Office established a public health emergency response and medical rescue collaboration mechanism with the China CDC, the Academy of Military Medical Sciences, and other institutions. It established a mutual fixed epidemic communication system with local agricultural, educational, industrial, and commercial departments. A joint command response mechanism was also created with other special operations centers in Beijing. On May 13th, Shandong established a provincial public health emergency leading group and started Level-II response measures after discovering its first—and the country's second—confirmed imported case of Influenza A (H1N1). On June 2nd, Hubei established an Influenza A (H1N1) emergency operations center after the province's first case was confirmed. Immediately after the country's first Influenza A (H1N1) case was confirmed on May 11th in Sichuan, Sichuan initiated Level-II public health emergency response measures and set up a provincial response leading group as per the State Council's requirement of handling the virus as a Category B infectious disease. The third work mode was the joint conference system. Some provinces and cities established a joint conference system in response to the Influenza A (H1N1) outbreak. For example, Henan established an Influenza A (H1N1) joint conference system on April 30th, and on the same day Guangxi established a 12-department joint conference system for its own prevention and control efforts. While establishing provincial-level joint prevention and control mechanisms, health departments also set up internal expert panels. For example, Fujian Provincial Department of Health set up a provincial-level Influenza A (H1N1) prevention and control expert supervision panel; Guangdong Provincial Department of Health established three expert panels for clinics, disease prevention and control, and etiology; Sichuan Provincial Department of Health established a leading group, a technical guidance expert panel, and a medical rescue panel on April 30th. The Composition of Local Influenza A (H1N1) Prevention and Control Mechanisms The composition of local mechanisms for Influenza A (H1N1) prevention and control basically followed the framework of the National Joint Prevention and Control Mechanism, with an office and several work groups collaborating under a leading group or operations center. For example, Beijing's public health emergency operations center was responsible for the city's influenza prevention and control, and it was comprised of eight work groups plus an office, these groups included: immigration inspection, healthcare, epidemiological survey, material security, dissemination and communication, information, animal husbandry and veterinary, and social prevention and control supervision (referred to as "one office and eight groups"). Fujian's Influenza A (H1N1) prevention and control leading group consisted of thirty one departments and organizations, including the Provincial Department of Health, a press office, and a development and reform commission. Sichuan's Influenza A (H1N1) prevention and control leading group (operations center) included departments from emergency management, public security, development and reform, transportation, immigration inspection and quarantine, tourism, civil aviation, foreign affairs, and publicity. The office of Sichuan's Influenza A (H1N1) prevention and control leading group was originally located in the Provincial Department of Health, which was then moved to the provincial government's General Office Building as the epidemic worsened. Its work groups consisted of emergency coordination, general support, information secretaries, epidemic prevention and control, medical rescue, supervision and inspection, press and communication, and health education. The Operation of Local Influenza A (H1N1) Prevention and Control Mechanisms Local Influenza A (H1N1) prevention and control mechanisms adopted a similar communication and coordination mechanism to the Joint National Prevention and Control Mechanism, and operated in with joint offices, conferences, etc. Sichuan is one example of this. The provincial operations center and the provincial leading group (headquarters) shared offices in order to strengthen the province's joint prevention and control against Influenza A (H1N1), and its seven work groups were comprised of highly capable professionals from the emergency management office, third secretariat office of the general office, and the Provincial Health Department. The Health Department met regularly with the departments of public security, civil aviation, immigration quarantine, economy and trade, animal husbandry, as well as PLA and People's Armed Police troops stationed in the province. Adjustments were made in real time in according with latest local epidemic situations. Latest information on epidemic updates across the province were reported daily to members and related departments, and where cases were discovered, the departments of health, public security, foreign affairs, railway, transport, and others worked closely to track close contacts and ensure they were medically observed. Through close collaboration between departments at various levels on joint prevention and control, Sichuan ensured that cases were discovered, reported, isolated and treated at the earliest possible time, which delayed the spread of the virus and lowered epidemic intensity. The Establishment of Local Influenza A (H1N1) Prevention and Control Mechanisms The following three modes mainly represent actual prevention and control measures adopted by local governments. The first work mode was similar to the National Joint Prevention and Control Mechanism. For example, Shaanxi set up a leading group for Influenza A (H1N1) prevention and control, whose office was located inside the provincial Department of Health and their local structure followed the "8 + 1" joint prevention and control mechanism model. Guangdong established a joint prevention and control mechanism with the participation of thirty two departments, and nine work groups and three panels of clinical, disease prevention and control, and etiological experts functioned under the mechanism. On April 30th, Fujian established an Influenza A (H1N1) prevention and control work group, headed by a provincial government official; the office was located inside the building of the provincial Department of Health whose emergency management office was charged with performing routine work for the group. The Emergency Guidance on Influenza A (H1N1) Prevention and Control in Fujian (Tentative) , issued on May 10th, 2009, outlined response guidance as "prevention first through joint prevention and control, timely management, and level-by-level responsibility." The second work mode was the emergency operations center or leading group. After discovering their first confirmed Influenza A (H1N1) case, some provinces and cities upgraded their existing disease prevention and control mechanisms and established an Influenza A (H1N1) response leading group or emergency operations center. Beijing was the first in the country to establish a municipal-level public health emergency operations center in May 2006, and had earlier (April 25th, 2009) launched a public health emergency response mechanism after the WHO declared the outbreak of swine influenza in Mexico; an Influenza Prevention and Control Office (at the general office of the municipal government before it relocated to the Municipal Bureau of Health) was established under the emergency operations center, whose members included twenty two committees, eighteen district and county governments, and the GLD Health Department. This Control Office established a public health emergency response and medical rescue collaboration mechanism with the China CDC, the Academy of Military Medical Sciences, and other institutions. It established a mutual fixed epidemic communication system with local agricultural, educational, industrial, and commercial departments. A joint command response mechanism was also created with other special operations centers in Beijing. On May 13th, Shandong established a provincial public health emergency leading group and started Level-II response measures after discovering its first—and the country's second—confirmed imported case of Influenza A (H1N1). On June 2nd, Hubei established an Influenza A (H1N1) emergency operations center after the province's first case was confirmed. Immediately after the country's first Influenza A (H1N1) case was confirmed on May 11th in Sichuan, Sichuan initiated Level-II public health emergency response measures and set up a provincial response leading group as per the State Council's requirement of handling the virus as a Category B infectious disease. The third work mode was the joint conference system. Some provinces and cities established a joint conference system in response to the Influenza A (H1N1) outbreak. For example, Henan established an Influenza A (H1N1) joint conference system on April 30th, and on the same day Guangxi established a 12-department joint conference system for its own prevention and control efforts. While establishing provincial-level joint prevention and control mechanisms, health departments also set up internal expert panels. For example, Fujian Provincial Department of Health set up a provincial-level Influenza A (H1N1) prevention and control expert supervision panel; Guangdong Provincial Department of Health established three expert panels for clinics, disease prevention and control, and etiology; Sichuan Provincial Department of Health established a leading group, a technical guidance expert panel, and a medical rescue panel on April 30th. The Composition of Local Influenza A (H1N1) Prevention and Control Mechanisms The composition of local mechanisms for Influenza A (H1N1) prevention and control basically followed the framework of the National Joint Prevention and Control Mechanism, with an office and several work groups collaborating under a leading group or operations center. For example, Beijing's public health emergency operations center was responsible for the city's influenza prevention and control, and it was comprised of eight work groups plus an office, these groups included: immigration inspection, healthcare, epidemiological survey, material security, dissemination and communication, information, animal husbandry and veterinary, and social prevention and control supervision (referred to as "one office and eight groups"). Fujian's Influenza A (H1N1) prevention and control leading group consisted of thirty one departments and organizations, including the Provincial Department of Health, a press office, and a development and reform commission. Sichuan's Influenza A (H1N1) prevention and control leading group (operations center) included departments from emergency management, public security, development and reform, transportation, immigration inspection and quarantine, tourism, civil aviation, foreign affairs, and publicity. The office of Sichuan's Influenza A (H1N1) prevention and control leading group was originally located in the Provincial Department of Health, which was then moved to the provincial government's General Office Building as the epidemic worsened. Its work groups consisted of emergency coordination, general support, information secretaries, epidemic prevention and control, medical rescue, supervision and inspection, press and communication, and health education. The Operation of Local Influenza A (H1N1) Prevention and Control Mechanisms Local Influenza A (H1N1) prevention and control mechanisms adopted a similar communication and coordination mechanism to the Joint National Prevention and Control Mechanism, and operated in with joint offices, conferences, etc. Sichuan is one example of this. The provincial operations center and the provincial leading group (headquarters) shared offices in order to strengthen the province's joint prevention and control against Influenza A (H1N1), and its seven work groups were comprised of highly capable professionals from the emergency management office, third secretariat office of the general office, and the Provincial Health Department. The Health Department met regularly with the departments of public security, civil aviation, immigration quarantine, economy and trade, animal husbandry, as well as PLA and People's Armed Police troops stationed in the province. Adjustments were made in real time in according with latest local epidemic situations. Latest information on epidemic updates across the province were reported daily to members and related departments, and where cases were discovered, the departments of health, public security, foreign affairs, railway, transport, and others worked closely to track close contacts and ensure they were medically observed. Through close collaboration between departments at various levels on joint prevention and control, Sichuan ensured that cases were discovered, reported, isolated and treated at the earliest possible time, which delayed the spread of the virus and lowered epidemic intensity. Social Participation in Local Influenza A (H1N1) Prevention and Control Mechanisms Government departments, enterprises, institutions, communities, and nonprofit organizations (NPOs) all play important roles in prevention and control of an infectious disease. With the country's response efforts entering its second phase, the 13th Meeting of the National Joint Prevention and Control Mechanism, held on June 10th, 2006, proposed further improvements of existing mechanisms, in particular establishing accountability systems and mass prevention and control mechanisms with participation from urban communities, schools, enterprises and villages. These mechanisms could better implement tailored measures, disseminate self-protection knowledge for families and individuals, and improve measures that maintain the status quo and normal economic operations. Community Participation When confronted with a public health emergency, under the guidance of the government, the society can effectively avoid or reduce potential damage by achieving preliminary prevention and control targets at local levels through community involvement and solidarity along with raising public awareness in self-protection. In its 1989 Work Report, the WHO mentioned two types of community participation, i.e. participation as a means, and participation as a goal, and analyzed effects of the two. In the course of China's Influenza A (H1N1) prevention and control efforts, communities, the most basic social units, played an important role in knowledge dissemination and health education, the tracking and isolation of close contacts, and epidemic supervision. When uncertainties still surrounded the pathology and virulence of Influenza A (H1N1) in the early days of the epidemic, local communities launched information dissemination and health education campaigns, playing a crucial role in stabilizing public opinion and raising awareness of scientific disease prevention and treatment methods. In the case of community-level outbreaks, affected communities generally adopted comprehensive response measures, which emphasized managing the sources of infection in order to contain and control the transmission of the influenza virus. Measures mainly included the following: (1) Sub-district offices or town governments mobilized social forces—as per laws, rules, and regulations—to provide support for isolated cases, including logistical service to personnel engaging in medical observation; (2) Close contacts were medically observed centrally or at home, and healthcare workers reported daily on patients' progress; (3) Patients with influenza-like symptoms were recommended to rest at home and not participate in unnecessary public gatherings or travel; (4) Schools, nurseries and kindergartens, nursing homes, and construction sites were required to conduct health inspections, and enterprises with a concentrated amount of personnel or those who provided social services were required to perform morning health inspections; (5) Information on epidemic trends and response measures were published in real time, and efforts were made to strengthen information disclosure within communities; and (6) Health education and risk communication were carried out through multiple channels. When the virus broke out in communities, healthcare departments managed cases categorically and adopted comprehensive measures for strengthening the treatment of severely ill cases, lowering case fatality rates, and mitigating epidemic damage. With prevention and control measures in place against community-level outbreaks, response measures for priority areas mainly included the following: (1) As per related laws, rules, and regulations, local governments mobilized social forces to ensure the logistical support of measures like home-based treatment of cases with influenza-like symptoms; (2) Migration was cut or restricted, recreational areas were temporarily shut down, and large-scale gatherings were canceled or postponed. Enterprises and institutions within communities were permitted to grant time off for all or some of their employees; (3) Schools, nurseries, and kindergartens were closed per related regulations; (4) Enterprises and institutions as social services providers with large workforces implemented a health reporting system, management was enhanced where there were large flows of people, and people with influenza-like symptoms were recommended to rest and receive treatment at home; (5) When necessary, outbreak points were put under isolated control, and quarantine measures were taken in epidemic areas. At the same time, measures were taken to organize and encourage volunteers to participate in prevention and control activities, to help maintain the normal operation within communities, and provide mental health interventions to avoid adverse effects on public health. These measures, which were designed based on real community conditions, effectively guaranteed the protection of the status quo, and laid a strong foundation for local Influenza A (H1N1) prevention and control efforts. Enterprise Participation Drug Stockpiling Enterprises In the course of Influenza A (H1N1) prevention and control efforts, drug stockpiling enterprises responded actively to the government's call for material reserves and production. Because influenza drugs weren't prevalent in clinical use, they were traditionally stockpiled through loans, government subsidies, business opportunities and moderate enterprise compensation. Problems arose during the implementation of this mechanism such as subsidy inaccessibility and unreasonable compensation. For example, the current 10% subsidy policies regarding corporate loans and government subsidies hardly met the needs of enterprises, and the problem of unreasonable compensation to pharmaceutical enterprises still existed, partly because specific mechanisms were lacking. At the same time, more than 80% of emergency response drugs were not on the standing list of medications, and so it was necessary to build a long-term relationship between the government and enterprises to specify respective duties, and link stockpile funding with corporate social responsibility to balance compensation. Reagent Manufacturers Reagent manufacturing was one of the government's top priorities during the entire course of its Influenza A (H1N1) prevention and control efforts. Some reagent manufacturers which had developed and produced reagents for biological agents such as anthrax during the 2008 Olympic Games already had experience in emergency response. For example, Beijing Kinghawk Pharmaceutical Co., Ltd., the country's first to obtain approval for an Influenza A (H1N1) testing kit, signed a strategic alliance agreement with the China CDC during the Influenza A (H1N1) Epidemic. The enterprise also provided its laboratories voluntarily when there was no clear policy on state funding, doing its best for society as a corporate citizen. With its technology reserve, seven production platforms and ninety approved products, Kinghawk was able to perform research and development on product standardization during a critical time of the epidemic. The China CDC had access to international resources for preliminary research and development and successfully obtained information and strains from the WHO. This collaboration between the two parties made it possible to develop preliminary products in 72 h and thus ensured that considerable demand for clinical diagnosis was met. This played a positive role in case diagnosis during the early phases of the Influenza A (H1N1) Epidemic and was also quite meaningful in terms of drug use guidance. The government provided active support to research and development efforts. Take Kinghawk as an example. After Kinghawk signed the agreement with the government on May 4th, 2009, both the Beijing Economic-Technological Development Food and Drug Administration and the Beijing Food and Drug Administration provided recommendations. Kinghawk had developed a rapid test kit by May 11th, received approval from the CFDA on June 17th to launch the emergency response system, and got approved for manufacturing the reagent for 250,000 people on September 25th. Kinghawk had collaborated with the China CDC in the past and had experience in reagent development and manufacturing. The development of the reagent fully demonstrated the efficiency in collaboration between the government and a commercial enterprise, and also guaranteed the timeliness of preliminary disease diagnosis. Areas Provided for Isolated Cases In the containment phase, hotels and other requisitioned enterprises across the country showed full support for the response measures by providing isolation zones of Influenza A (H1N1) cases. Many hotels suitable for isolation purposes were private firms and thus could not be requisitioned through administrative orders, which put a certain amount of pressure on local governments. However, coordination efforts by local governments did earn support and assistance from these hotels. Infrastructure Enterprises Transport enterprises shouldered the heavy task of implementing disease prevention and control for the floating population. Civil aviation, railways, road and related enterprises implemented strict prevention and control measures, including disseminating knowledge about disease prevention and control, and providing necessary infrastructure support for emergency response efforts targeting the floating population. All in all, a healthy transportation environment helped lower the transmission of the disease. Facing the unexpected onset of Influenza A (H1N1), these transportation enterprises all established prevention and control groups. For example, the CAAC North China Regional Administration established a Capital Airport Influenza A (H1N1) Prevention and Control Leading Group, which was based on the former Capital Airport Public Health Emergency Leading Group; Beijing Capital International Airport Company Limited, Air China Limited, China Southern Beijing Company, China Eastern Beijing Company, Hainan Airlines Beijing Company, CAAC Air Traffic Management Bureau all set up their own epidemic response teams to ensure the orderly implementation of prevention and control measures. Telecommunications enterprises did their duties as corporate citizens and actively cooperated in Influenza A (H1N1) prevention and control efforts. China Mobile Group Beijing Company Limited, China Telecom Group Beijing Company Limited, China Unicom Beijing Company Limited, among other telecommunications operators, suspended their normal user notification group-messaging services and mustered network resources to send messages on epidemic updates while increasing maintenance staff, strengthening network monitoring, and closely watching the impact of group messaging on their systems. Participation of NPOs Nonprofit organizations, or NPOs, are organizations that fulfill particular social causes or missions without seeking profit for their efforts, and NPOs represented a crucial social force in the course of the nation's Influenza A (H1N1) prevention and control. For example, the Beijing Red Cross established a public health emergency operations center which consisted of a general information group, a rescue response group, a fundraising and aid group, a publicity group, and a public relations group. This organization actively engaged in response efforts in accordance with the Beijing Red Cross emergency contingency plans for public emergencies . Public Participation The entire nation was involved in epidemic prevention and control, including its citizens. During the response to the virus, volunteers played an important role when multiple departments suffered emergency manpower shortages. For example, medical and healthcare students in colleges and universities volunteered to work on the front lines of epidemic prevention and control. In Beijing, 170 student volunteers from Capital Medical University assisted with response efforts at Capital Airport, and similar volunteering also occurred in Fujian. The public actively supported the government's Influenza A (H1N1) prevention and control measures, and voluntarily took part in the process via the Internet and other media channels; for example volunteers called those who had just returned from abroad and informed them of potential isolation measures. At the same time, increased public health awareness was also instrumental in successfully dealing with the disease. Community Participation When confronted with a public health emergency, under the guidance of the government, the society can effectively avoid or reduce potential damage by achieving preliminary prevention and control targets at local levels through community involvement and solidarity along with raising public awareness in self-protection. In its 1989 Work Report, the WHO mentioned two types of community participation, i.e. participation as a means, and participation as a goal, and analyzed effects of the two. In the course of China's Influenza A (H1N1) prevention and control efforts, communities, the most basic social units, played an important role in knowledge dissemination and health education, the tracking and isolation of close contacts, and epidemic supervision. When uncertainties still surrounded the pathology and virulence of Influenza A (H1N1) in the early days of the epidemic, local communities launched information dissemination and health education campaigns, playing a crucial role in stabilizing public opinion and raising awareness of scientific disease prevention and treatment methods. In the case of community-level outbreaks, affected communities generally adopted comprehensive response measures, which emphasized managing the sources of infection in order to contain and control the transmission of the influenza virus. Measures mainly included the following: (1) Sub-district offices or town governments mobilized social forces—as per laws, rules, and regulations—to provide support for isolated cases, including logistical service to personnel engaging in medical observation; (2) Close contacts were medically observed centrally or at home, and healthcare workers reported daily on patients' progress; (3) Patients with influenza-like symptoms were recommended to rest at home and not participate in unnecessary public gatherings or travel; (4) Schools, nurseries and kindergartens, nursing homes, and construction sites were required to conduct health inspections, and enterprises with a concentrated amount of personnel or those who provided social services were required to perform morning health inspections; (5) Information on epidemic trends and response measures were published in real time, and efforts were made to strengthen information disclosure within communities; and (6) Health education and risk communication were carried out through multiple channels. When the virus broke out in communities, healthcare departments managed cases categorically and adopted comprehensive measures for strengthening the treatment of severely ill cases, lowering case fatality rates, and mitigating epidemic damage. With prevention and control measures in place against community-level outbreaks, response measures for priority areas mainly included the following: (1) As per related laws, rules, and regulations, local governments mobilized social forces to ensure the logistical support of measures like home-based treatment of cases with influenza-like symptoms; (2) Migration was cut or restricted, recreational areas were temporarily shut down, and large-scale gatherings were canceled or postponed. Enterprises and institutions within communities were permitted to grant time off for all or some of their employees; (3) Schools, nurseries, and kindergartens were closed per related regulations; (4) Enterprises and institutions as social services providers with large workforces implemented a health reporting system, management was enhanced where there were large flows of people, and people with influenza-like symptoms were recommended to rest and receive treatment at home; (5) When necessary, outbreak points were put under isolated control, and quarantine measures were taken in epidemic areas. At the same time, measures were taken to organize and encourage volunteers to participate in prevention and control activities, to help maintain the normal operation within communities, and provide mental health interventions to avoid adverse effects on public health. These measures, which were designed based on real community conditions, effectively guaranteed the protection of the status quo, and laid a strong foundation for local Influenza A (H1N1) prevention and control efforts. Enterprise Participation Drug Stockpiling Enterprises In the course of Influenza A (H1N1) prevention and control efforts, drug stockpiling enterprises responded actively to the government's call for material reserves and production. Because influenza drugs weren't prevalent in clinical use, they were traditionally stockpiled through loans, government subsidies, business opportunities and moderate enterprise compensation. Problems arose during the implementation of this mechanism such as subsidy inaccessibility and unreasonable compensation. For example, the current 10% subsidy policies regarding corporate loans and government subsidies hardly met the needs of enterprises, and the problem of unreasonable compensation to pharmaceutical enterprises still existed, partly because specific mechanisms were lacking. At the same time, more than 80% of emergency response drugs were not on the standing list of medications, and so it was necessary to build a long-term relationship between the government and enterprises to specify respective duties, and link stockpile funding with corporate social responsibility to balance compensation. Reagent Manufacturers Reagent manufacturing was one of the government's top priorities during the entire course of its Influenza A (H1N1) prevention and control efforts. Some reagent manufacturers which had developed and produced reagents for biological agents such as anthrax during the 2008 Olympic Games already had experience in emergency response. For example, Beijing Kinghawk Pharmaceutical Co., Ltd., the country's first to obtain approval for an Influenza A (H1N1) testing kit, signed a strategic alliance agreement with the China CDC during the Influenza A (H1N1) Epidemic. The enterprise also provided its laboratories voluntarily when there was no clear policy on state funding, doing its best for society as a corporate citizen. With its technology reserve, seven production platforms and ninety approved products, Kinghawk was able to perform research and development on product standardization during a critical time of the epidemic. The China CDC had access to international resources for preliminary research and development and successfully obtained information and strains from the WHO. This collaboration between the two parties made it possible to develop preliminary products in 72 h and thus ensured that considerable demand for clinical diagnosis was met. This played a positive role in case diagnosis during the early phases of the Influenza A (H1N1) Epidemic and was also quite meaningful in terms of drug use guidance. The government provided active support to research and development efforts. Take Kinghawk as an example. After Kinghawk signed the agreement with the government on May 4th, 2009, both the Beijing Economic-Technological Development Food and Drug Administration and the Beijing Food and Drug Administration provided recommendations. Kinghawk had developed a rapid test kit by May 11th, received approval from the CFDA on June 17th to launch the emergency response system, and got approved for manufacturing the reagent for 250,000 people on September 25th. Kinghawk had collaborated with the China CDC in the past and had experience in reagent development and manufacturing. The development of the reagent fully demonstrated the efficiency in collaboration between the government and a commercial enterprise, and also guaranteed the timeliness of preliminary disease diagnosis. Areas Provided for Isolated Cases In the containment phase, hotels and other requisitioned enterprises across the country showed full support for the response measures by providing isolation zones of Influenza A (H1N1) cases. Many hotels suitable for isolation purposes were private firms and thus could not be requisitioned through administrative orders, which put a certain amount of pressure on local governments. However, coordination efforts by local governments did earn support and assistance from these hotels. Infrastructure Enterprises Transport enterprises shouldered the heavy task of implementing disease prevention and control for the floating population. Civil aviation, railways, road and related enterprises implemented strict prevention and control measures, including disseminating knowledge about disease prevention and control, and providing necessary infrastructure support for emergency response efforts targeting the floating population. All in all, a healthy transportation environment helped lower the transmission of the disease. Facing the unexpected onset of Influenza A (H1N1), these transportation enterprises all established prevention and control groups. For example, the CAAC North China Regional Administration established a Capital Airport Influenza A (H1N1) Prevention and Control Leading Group, which was based on the former Capital Airport Public Health Emergency Leading Group; Beijing Capital International Airport Company Limited, Air China Limited, China Southern Beijing Company, China Eastern Beijing Company, Hainan Airlines Beijing Company, CAAC Air Traffic Management Bureau all set up their own epidemic response teams to ensure the orderly implementation of prevention and control measures. Telecommunications enterprises did their duties as corporate citizens and actively cooperated in Influenza A (H1N1) prevention and control efforts. China Mobile Group Beijing Company Limited, China Telecom Group Beijing Company Limited, China Unicom Beijing Company Limited, among other telecommunications operators, suspended their normal user notification group-messaging services and mustered network resources to send messages on epidemic updates while increasing maintenance staff, strengthening network monitoring, and closely watching the impact of group messaging on their systems. Drug Stockpiling Enterprises In the course of Influenza A (H1N1) prevention and control efforts, drug stockpiling enterprises responded actively to the government's call for material reserves and production. Because influenza drugs weren't prevalent in clinical use, they were traditionally stockpiled through loans, government subsidies, business opportunities and moderate enterprise compensation. Problems arose during the implementation of this mechanism such as subsidy inaccessibility and unreasonable compensation. For example, the current 10% subsidy policies regarding corporate loans and government subsidies hardly met the needs of enterprises, and the problem of unreasonable compensation to pharmaceutical enterprises still existed, partly because specific mechanisms were lacking. At the same time, more than 80% of emergency response drugs were not on the standing list of medications, and so it was necessary to build a long-term relationship between the government and enterprises to specify respective duties, and link stockpile funding with corporate social responsibility to balance compensation. Reagent Manufacturers Reagent manufacturing was one of the government's top priorities during the entire course of its Influenza A (H1N1) prevention and control efforts. Some reagent manufacturers which had developed and produced reagents for biological agents such as anthrax during the 2008 Olympic Games already had experience in emergency response. For example, Beijing Kinghawk Pharmaceutical Co., Ltd., the country's first to obtain approval for an Influenza A (H1N1) testing kit, signed a strategic alliance agreement with the China CDC during the Influenza A (H1N1) Epidemic. The enterprise also provided its laboratories voluntarily when there was no clear policy on state funding, doing its best for society as a corporate citizen. With its technology reserve, seven production platforms and ninety approved products, Kinghawk was able to perform research and development on product standardization during a critical time of the epidemic. The China CDC had access to international resources for preliminary research and development and successfully obtained information and strains from the WHO. This collaboration between the two parties made it possible to develop preliminary products in 72 h and thus ensured that considerable demand for clinical diagnosis was met. This played a positive role in case diagnosis during the early phases of the Influenza A (H1N1) Epidemic and was also quite meaningful in terms of drug use guidance. The government provided active support to research and development efforts. Take Kinghawk as an example. After Kinghawk signed the agreement with the government on May 4th, 2009, both the Beijing Economic-Technological Development Food and Drug Administration and the Beijing Food and Drug Administration provided recommendations. Kinghawk had developed a rapid test kit by May 11th, received approval from the CFDA on June 17th to launch the emergency response system, and got approved for manufacturing the reagent for 250,000 people on September 25th. Kinghawk had collaborated with the China CDC in the past and had experience in reagent development and manufacturing. The development of the reagent fully demonstrated the efficiency in collaboration between the government and a commercial enterprise, and also guaranteed the timeliness of preliminary disease diagnosis. Areas Provided for Isolated Cases In the containment phase, hotels and other requisitioned enterprises across the country showed full support for the response measures by providing isolation zones of Influenza A (H1N1) cases. Many hotels suitable for isolation purposes were private firms and thus could not be requisitioned through administrative orders, which put a certain amount of pressure on local governments. However, coordination efforts by local governments did earn support and assistance from these hotels. Infrastructure Enterprises Transport enterprises shouldered the heavy task of implementing disease prevention and control for the floating population. Civil aviation, railways, road and related enterprises implemented strict prevention and control measures, including disseminating knowledge about disease prevention and control, and providing necessary infrastructure support for emergency response efforts targeting the floating population. All in all, a healthy transportation environment helped lower the transmission of the disease. Facing the unexpected onset of Influenza A (H1N1), these transportation enterprises all established prevention and control groups. For example, the CAAC North China Regional Administration established a Capital Airport Influenza A (H1N1) Prevention and Control Leading Group, which was based on the former Capital Airport Public Health Emergency Leading Group; Beijing Capital International Airport Company Limited, Air China Limited, China Southern Beijing Company, China Eastern Beijing Company, Hainan Airlines Beijing Company, CAAC Air Traffic Management Bureau all set up their own epidemic response teams to ensure the orderly implementation of prevention and control measures. Telecommunications enterprises did their duties as corporate citizens and actively cooperated in Influenza A (H1N1) prevention and control efforts. China Mobile Group Beijing Company Limited, China Telecom Group Beijing Company Limited, China Unicom Beijing Company Limited, among other telecommunications operators, suspended their normal user notification group-messaging services and mustered network resources to send messages on epidemic updates while increasing maintenance staff, strengthening network monitoring, and closely watching the impact of group messaging on their systems. Participation of NPOs Nonprofit organizations, or NPOs, are organizations that fulfill particular social causes or missions without seeking profit for their efforts, and NPOs represented a crucial social force in the course of the nation's Influenza A (H1N1) prevention and control. For example, the Beijing Red Cross established a public health emergency operations center which consisted of a general information group, a rescue response group, a fundraising and aid group, a publicity group, and a public relations group. This organization actively engaged in response efforts in accordance with the Beijing Red Cross emergency contingency plans for public emergencies . Public Participation The entire nation was involved in epidemic prevention and control, including its citizens. During the response to the virus, volunteers played an important role when multiple departments suffered emergency manpower shortages. For example, medical and healthcare students in colleges and universities volunteered to work on the front lines of epidemic prevention and control. In Beijing, 170 student volunteers from Capital Medical University assisted with response efforts at Capital Airport, and similar volunteering also occurred in Fujian. The public actively supported the government's Influenza A (H1N1) prevention and control measures, and voluntarily took part in the process via the Internet and other media channels; for example volunteers called those who had just returned from abroad and informed them of potential isolation measures. At the same time, increased public health awareness was also instrumental in successfully dealing with the disease. Analysis and Reflections on Influenza A (H1N1) Prevention and Control Mechanisms Experience in Mechanism Building for Influenza A (H1N1) Prevention and Control The Timely Establishment of Influenza A (H1N1) Prevention and Control Mechanisms At the beginning of the Influenza A (H1N1) Epidemic, China established a national level emergency management mechanism directly under the leadership of the State Council that enabled cross-departmental joint prevention and control collaboration, which provided an effective organizational support and operation mechanism for the response efforts. Though the MOH had formulated the Ministry of Health's Influenza Pandemic Preparedness and Response Plan (Tentative) before the epidemic broke out, this document focused only on the duties of the MOH and didn't encompass more complex coordination and collaboration with related government departments. The joint prevention and control mechanism remedied this flaw by providing a platform for coordination and collaboration between the MOH and other related departments. Also, because this mechanism was not like the State Council's operations center, it allowed some space for strengthening the State Council's leadership and collaboration once the epidemic worsened. During the prevention and control efforts, local governments adapted and innovated central policies and their implementation in light of local epidemic situations, public health trends, and demographic and economic conditions. Some areas established prevention and control mechanisms with local characteristics. The main features of these mechanisms are as follows: The first was the establishment of a strong leadership system. In the process of prevention and control, local governments established their respective public health emergency leadership systems based on local epidemic situations, geographic features, and public health resources. The second was the innovation in ideas and methods. Local epidemic prevention and control bodies closely monitored trends and reengineered their methods based on existing departmental systems in order to better target obstacles encountered in operations. For example, in the early days of the epidemic, the Beijing government issued a Notice on Further Specifying Duties and Prioritizing Operations to Strengthen Influenza A (H1N1) Prevention and Control , which articulated the new public health notion of "responsibility of four sides" (government, departments, enterprises, and individuals). This clarification brought about effective collaboration between the government and the society in public health emergency management. The Beijing Immigration Inspection and Quarantine Bureau employed risk analysis methods in its prevention and control efforts and ensured electronic transfer of information on inbound passengers, which not only increased quarantine and inspection efficiency but also scientifically and efficiently improved response measures. Fujian was the country's first province to implement temporary isolation measures through its local Health Department. Henan created an epidemic prevention and control network of "three horizontal fronts"—arrangements at a government level, measures at enterprise (institution) level, and protection at a local level; and "three vertical fronts"—government supervision, inter-departmental collaboration, and public opinion guidance. These institutional innovations proved very effective in the response efforts. The third was the establishment of an inter-provincial support mechanism. On November 13th, 2009, the MOH General Office issued the Notice on Strengthening Medical Treatment of Influenza A (H1N1) Patients (No. 245, 2009), announcing the decision to establish an inter-provincial support mechanism for medical treatment of Influenza A (H1N1) patients as per the Notice of the State Council on Strengthening the Ongoing Work on Influenza A (H1N1) Prevention and Control (No. 23, 2009) and as needed for patients. The form of assistance was technical support, especially in regards to medical treatment of seriously and critically ill patients. The Active Implementation of an Expert Decision-Making Mechanism Throughout the entire duration of the prevention and control efforts, governments greatly heeded experts in various fields, which aided governments in creating more scientific policy adjustments and technical plans, and consequently reduced blindness and uncertainty in policy implementation. Experts from CDCs, hospitals, publicity departments, and other departments took part in the decision-making process, and their input was adopted in real time. Some experts even took the initiative to provide police recommendations directly to decision makers. At the same time, governments sought out expert opinions through different methods and channels, i.e., consultation at joint prevention and control meetings or direct consultation with the experts. Expert recommendations ensured scientific policies and more targeted and effective policy formulation. Policy Adjustments Based on a Local Context In regards to policymaking, some local governments formulated policies and adjustments based on local conditions and epidemic trends. For example, Jiangmen experimented with a home-based isolation policy, while Shenzhen created corridors at ports specifically for foreigners and a separate one for students commuting between Shenzhen and Hong Kong for school. Also, in terms of policy adjustment, some local departments were able to adjust related policies in time to better suit local epidemic situations. As for issues that necessitated policy coordination, local departments also made strategic adjustments as early as possible. For example, the Guangdong Immigration Inspection and Quarantine Bureau, at experts' suggestion, transferred persons who required isolation and medical observation to health departments for categorical management, which thus ensured the efficient use of epidemic prevention and control resources. The Gradual Realization of Widespread and Diverse Societal Participation in Disease Prevention and Control Over the course of Influenza A (H1N1) prevention and control efforts, the government cultivated an environment of widespread social participation under the leadership of the party and government, with enterprises, communities, volunteers and other social actors playing crucial roles in the response efforts. Reflections on Influenza A (H1N1) Prevention and Control Mechanisms The Legal Status of the Joint National Prevention and Control Mechanism The Joint National Prevention and Control Mechanism was essentially a command and decision-making mechanism established according to the potential amount of damage Influenza A (H1N1) could inflict upon society. On the one hand, Influenza A (H1N1) response required inter-departmental collaboration, and relying solely upon health departments for countermeasures wouldn't be enough; on the other hand, because the virus was not as virulent as to merit the establishment of a State Council Operations Center (or Headquarters), the State Council instead instructed the MOH to establish a multi-departmental joint prevention and control mechanism; and this new organization represented a relatively flexible and effective response mechanism. Although local governments were already aware of the epidemic at its onset and were actively engaging with different departments in their response efforts, because there were no explicit provisions in related laws and contingency plans for the Joint National Prevention and Control Mechanism at the central level, no corresponding normative documents were available for its implementation at a local level. No unified standards on the name, content, form of establishment, and system structure for local governments' prevention and control bodies existed. Although local governments adapted as they went, it was still an environment that incited disorder and confusion. Issues with Emergency Command Responsibilities, Authority and Administrative Levels On the one hand, participating departments fully endorsed the Joint National Prevention and Control Mechanism. This mechanism, they thought, possessed several advantages: Firstly, the joint consultation system made it possible to directly formulate and sign policies at joint prevention and control conferences, which saved time for everyone; Secondly, the joint briefing system required the work groups to send daily reports to other units and departments, thus facilitating both inter-group and inter-departmental communication; And finally, internal collaboration within groups was solid, and the briefing system allowed an unobstructed flow of information. However, the Joint National Prevention and Control Mechanism based upon consultation and communication had its limitations. On issues involving departmental interest, division of duty, and so on, this horizontal collaboration was less efficient than regulation and control by a single, high level leadership department. One contested issue dealt with the location of the local joint prevention and control office: should it be set up in the comprehensive emergency management office of the local government or in the emergency management office of a local specialized department. Some provincial emergency management offices insisted that for an emergency event like the ongoing Influenza A (H1N1) epidemic, a joint prevention and control office should be located in a specialized department so as to leverage the department's expertise and increase response flexibility, convenience and efficiency. In this scenario, the provincial emergency management office would be tasked with solving issues that the specialized department could not. On the other hand, some provincial health department's emergency offices argued that if the office was located in the local government, the joint prevention and control office would enjoy greater authority and more efficient collaboration. The Transition Between Peacetime Mechanisms and Emergency Response Mechanisms Achieving a smooth and effective transition between peacetime and public emergency, and establishing mechanisms that combined crucial components from both systems, was a new challenge that arose in the Influenza A (H1N1) Epidemic. After the 2003 SARS Epidemic, local governments established permanent public health emergency response departments and corresponding working mechanisms to deal with future public health emergencies. These departments and mechanisms should have been employed upon the onset of the Influenza A (H1N1) Epidemic. However, most provinces established completely new leading groups only after the central government established Joint National Prevention and Control Mechanism. In one example, a provincial health department already had a permanent public health emergency operations center, but, after the central government established the Influenza A (H1N1) Joint National Prevention and Control Mechanism, this province created an entirely new prevention and control leading group and a port leading group. At the same time, the Health Department also established new eternal mechanisms, including: the provincial CDC established an emergency response department with leaders from major sections like emergency management and vaccination planning (starting in 2005, this provincial CDC implemented a "3 in 1" meeting system with participation from emergency management, disease control, and the disease monitoring department). The main reason for this redundancy was because the central government did not provide specific conditions or qualifications for contingency planning and management for the transition period between peacetime to emergency. Thus, local governments lacked a clear transition mechanism that they could utilize. It was the reason that many local governments chose to re-establish emergency management bodies when Influenza A (H1N1) broke out. Problems with Inter-departmental Coordination As public health emergency management involved multiple collaboration systems from the central government down to local governments, regions, and departments, inter-departmental collaboration in the response efforts was intrinsically complicated. The response to this epidemic revealed problems that existed both in horizontal and vertical coordination. In regards to horizontal coordination between central departments, the health, education, security, transportation and many other departments were involved in the Influenza A (H1N1) response efforts, which created an environment where responsibilities could easily overlap and grey areas would occur in management. There was also a lack of coordination and standardization between central-level ministries' policy documents for Influenza A (H1N1) countermeasures. For example, in regards to content standardization, the health authorities felt that using the temperature of 37.3 °C as the sole standard for sending people to the hospital was unreasonable and would cause an unnecessary burden on hospitals. In regards to time standardization, on December 2nd, 2009, one province stipulated that only patients with a temperature of 38 °C or higher must be sent to a hospital, and it took the country two more weeks to follow suit. In regards to inter-departmental work, port laboratories in some provinces had begun testing in the early days of epidemic, but stopped after provincial health departments decided that ports were not fit for such work. Obstacles also arose in horizontal coordination and collaboration between local departments. A lack of information communication between local departments due to the unavailability of complete information in the early stages of prevention and control made it nearly impossible for effective collaboration. In regards to the division of labor and coordination between the central and local governments for disease prevention and control, some local governments held that the central government should have presented broader goals and authorized provinces and cities greater autonomy in their response measures. Some felt that the central government should not have made Influenza A (H1N1) prevention and control an issue of political significance but should have been objective in understanding the differences between executive leadership and scientists' opinions. While the main duty of administrative leaders should have been to organize and mobilize social resources needed to cope with the epidemic, scientists should have been the ones to handle technical issues such as epidemic analysis and response measures. At the same time, more efficient communication should have been present between central and local departments tasked with specific operations. For example, some local management departments felt that the entire process was quite political, making some documents difficult to fully implement; in the two most volatile months that lasted from April 28th to June, documents were issued frequently, and in some cases were in conflict with one another and lacked integrity and continuity. In regards to adjustment of prevention and control strategies, some regions' health departments reported the following issues: higher-level departments frequently adjusted technical guidance and strategies for prevention and control, there was a wide variety of information reporting methods and they were constantly in flux, different departments formulated their own response requirements, and differences occurred in measures and standards; all of which greatly complicated local response operations. Certain communication and coordination issues also existed within the Health Department's internal system. The MOH internal horizontal collaboration needs to be strengthened epidemiological investigations, clinical diagnoses, and laboratory testing to combine the medical treatment and disease prevention. For example, the China CDC played a crucial role as a central technical support body of Influenza A (H1N1) prevention and control in epidemiological information collection, monitoring, analysis, and judgment, but at the same time it also had a lot of administrative duties, and its services and duties overlapped with those of the MOH's Bureau of Disease Control and Prevention. There should be unified leadership and coordination between higher and lower-level health departments within the national epidemic prevention and control system. A certain degree of flexibility is also necessary as provinces differ in epidemic situations, medical resources, geographic features, and so on. In regards to information reporting within the health system, though the China CDC and the MOH had established information systems relating to epidemic surveillance, including an epidemic direct reporting system, no information sharing mechanism was created between the China CDC and medical institutions; in particular, some county level medical institutions didn't even have sound data collection and reporting systems. This resulted in a single point of decision making and command, and their lack of network and information technology weakened the support they could've had in implementing response measures. Roles of NPOs Have Yet to Be Leveraged and Improved NPOs such as the Red Cross Society of China played important roles during the Influenza A (H1N1) prevention and control efforts. However, by comparison with developed countries, China still lags behind in terms of public participation in public health emergencies. There still remain limitations in skill and knowledge, as no emergency volunteer systems or working mechanisms were formed, and no leveraging of NPO resources really occurred. Experience in Mechanism Building for Influenza A (H1N1) Prevention and Control The Timely Establishment of Influenza A (H1N1) Prevention and Control Mechanisms At the beginning of the Influenza A (H1N1) Epidemic, China established a national level emergency management mechanism directly under the leadership of the State Council that enabled cross-departmental joint prevention and control collaboration, which provided an effective organizational support and operation mechanism for the response efforts. Though the MOH had formulated the Ministry of Health's Influenza Pandemic Preparedness and Response Plan (Tentative) before the epidemic broke out, this document focused only on the duties of the MOH and didn't encompass more complex coordination and collaboration with related government departments. The joint prevention and control mechanism remedied this flaw by providing a platform for coordination and collaboration between the MOH and other related departments. Also, because this mechanism was not like the State Council's operations center, it allowed some space for strengthening the State Council's leadership and collaboration once the epidemic worsened. During the prevention and control efforts, local governments adapted and innovated central policies and their implementation in light of local epidemic situations, public health trends, and demographic and economic conditions. Some areas established prevention and control mechanisms with local characteristics. The main features of these mechanisms are as follows: The first was the establishment of a strong leadership system. In the process of prevention and control, local governments established their respective public health emergency leadership systems based on local epidemic situations, geographic features, and public health resources. The second was the innovation in ideas and methods. Local epidemic prevention and control bodies closely monitored trends and reengineered their methods based on existing departmental systems in order to better target obstacles encountered in operations. For example, in the early days of the epidemic, the Beijing government issued a Notice on Further Specifying Duties and Prioritizing Operations to Strengthen Influenza A (H1N1) Prevention and Control , which articulated the new public health notion of "responsibility of four sides" (government, departments, enterprises, and individuals). This clarification brought about effective collaboration between the government and the society in public health emergency management. The Beijing Immigration Inspection and Quarantine Bureau employed risk analysis methods in its prevention and control efforts and ensured electronic transfer of information on inbound passengers, which not only increased quarantine and inspection efficiency but also scientifically and efficiently improved response measures. Fujian was the country's first province to implement temporary isolation measures through its local Health Department. Henan created an epidemic prevention and control network of "three horizontal fronts"—arrangements at a government level, measures at enterprise (institution) level, and protection at a local level; and "three vertical fronts"—government supervision, inter-departmental collaboration, and public opinion guidance. These institutional innovations proved very effective in the response efforts. The third was the establishment of an inter-provincial support mechanism. On November 13th, 2009, the MOH General Office issued the Notice on Strengthening Medical Treatment of Influenza A (H1N1) Patients (No. 245, 2009), announcing the decision to establish an inter-provincial support mechanism for medical treatment of Influenza A (H1N1) patients as per the Notice of the State Council on Strengthening the Ongoing Work on Influenza A (H1N1) Prevention and Control (No. 23, 2009) and as needed for patients. The form of assistance was technical support, especially in regards to medical treatment of seriously and critically ill patients. The Active Implementation of an Expert Decision-Making Mechanism Throughout the entire duration of the prevention and control efforts, governments greatly heeded experts in various fields, which aided governments in creating more scientific policy adjustments and technical plans, and consequently reduced blindness and uncertainty in policy implementation. Experts from CDCs, hospitals, publicity departments, and other departments took part in the decision-making process, and their input was adopted in real time. Some experts even took the initiative to provide police recommendations directly to decision makers. At the same time, governments sought out expert opinions through different methods and channels, i.e., consultation at joint prevention and control meetings or direct consultation with the experts. Expert recommendations ensured scientific policies and more targeted and effective policy formulation. Policy Adjustments Based on a Local Context In regards to policymaking, some local governments formulated policies and adjustments based on local conditions and epidemic trends. For example, Jiangmen experimented with a home-based isolation policy, while Shenzhen created corridors at ports specifically for foreigners and a separate one for students commuting between Shenzhen and Hong Kong for school. Also, in terms of policy adjustment, some local departments were able to adjust related policies in time to better suit local epidemic situations. As for issues that necessitated policy coordination, local departments also made strategic adjustments as early as possible. For example, the Guangdong Immigration Inspection and Quarantine Bureau, at experts' suggestion, transferred persons who required isolation and medical observation to health departments for categorical management, which thus ensured the efficient use of epidemic prevention and control resources. The Gradual Realization of Widespread and Diverse Societal Participation in Disease Prevention and Control Over the course of Influenza A (H1N1) prevention and control efforts, the government cultivated an environment of widespread social participation under the leadership of the party and government, with enterprises, communities, volunteers and other social actors playing crucial roles in the response efforts. The Timely Establishment of Influenza A (H1N1) Prevention and Control Mechanisms At the beginning of the Influenza A (H1N1) Epidemic, China established a national level emergency management mechanism directly under the leadership of the State Council that enabled cross-departmental joint prevention and control collaboration, which provided an effective organizational support and operation mechanism for the response efforts. Though the MOH had formulated the Ministry of Health's Influenza Pandemic Preparedness and Response Plan (Tentative) before the epidemic broke out, this document focused only on the duties of the MOH and didn't encompass more complex coordination and collaboration with related government departments. The joint prevention and control mechanism remedied this flaw by providing a platform for coordination and collaboration between the MOH and other related departments. Also, because this mechanism was not like the State Council's operations center, it allowed some space for strengthening the State Council's leadership and collaboration once the epidemic worsened. During the prevention and control efforts, local governments adapted and innovated central policies and their implementation in light of local epidemic situations, public health trends, and demographic and economic conditions. Some areas established prevention and control mechanisms with local characteristics. The main features of these mechanisms are as follows: The first was the establishment of a strong leadership system. In the process of prevention and control, local governments established their respective public health emergency leadership systems based on local epidemic situations, geographic features, and public health resources. The second was the innovation in ideas and methods. Local epidemic prevention and control bodies closely monitored trends and reengineered their methods based on existing departmental systems in order to better target obstacles encountered in operations. For example, in the early days of the epidemic, the Beijing government issued a Notice on Further Specifying Duties and Prioritizing Operations to Strengthen Influenza A (H1N1) Prevention and Control , which articulated the new public health notion of "responsibility of four sides" (government, departments, enterprises, and individuals). This clarification brought about effective collaboration between the government and the society in public health emergency management. The Beijing Immigration Inspection and Quarantine Bureau employed risk analysis methods in its prevention and control efforts and ensured electronic transfer of information on inbound passengers, which not only increased quarantine and inspection efficiency but also scientifically and efficiently improved response measures. Fujian was the country's first province to implement temporary isolation measures through its local Health Department. Henan created an epidemic prevention and control network of "three horizontal fronts"—arrangements at a government level, measures at enterprise (institution) level, and protection at a local level; and "three vertical fronts"—government supervision, inter-departmental collaboration, and public opinion guidance. These institutional innovations proved very effective in the response efforts. The third was the establishment of an inter-provincial support mechanism. On November 13th, 2009, the MOH General Office issued the Notice on Strengthening Medical Treatment of Influenza A (H1N1) Patients (No. 245, 2009), announcing the decision to establish an inter-provincial support mechanism for medical treatment of Influenza A (H1N1) patients as per the Notice of the State Council on Strengthening the Ongoing Work on Influenza A (H1N1) Prevention and Control (No. 23, 2009) and as needed for patients. The form of assistance was technical support, especially in regards to medical treatment of seriously and critically ill patients. The Active Implementation of an Expert Decision-Making Mechanism Throughout the entire duration of the prevention and control efforts, governments greatly heeded experts in various fields, which aided governments in creating more scientific policy adjustments and technical plans, and consequently reduced blindness and uncertainty in policy implementation. Experts from CDCs, hospitals, publicity departments, and other departments took part in the decision-making process, and their input was adopted in real time. Some experts even took the initiative to provide police recommendations directly to decision makers. At the same time, governments sought out expert opinions through different methods and channels, i.e., consultation at joint prevention and control meetings or direct consultation with the experts. Expert recommendations ensured scientific policies and more targeted and effective policy formulation. Policy Adjustments Based on a Local Context In regards to policymaking, some local governments formulated policies and adjustments based on local conditions and epidemic trends. For example, Jiangmen experimented with a home-based isolation policy, while Shenzhen created corridors at ports specifically for foreigners and a separate one for students commuting between Shenzhen and Hong Kong for school. Also, in terms of policy adjustment, some local departments were able to adjust related policies in time to better suit local epidemic situations. As for issues that necessitated policy coordination, local departments also made strategic adjustments as early as possible. For example, the Guangdong Immigration Inspection and Quarantine Bureau, at experts' suggestion, transferred persons who required isolation and medical observation to health departments for categorical management, which thus ensured the efficient use of epidemic prevention and control resources. The Gradual Realization of Widespread and Diverse Societal Participation in Disease Prevention and Control Over the course of Influenza A (H1N1) prevention and control efforts, the government cultivated an environment of widespread social participation under the leadership of the party and government, with enterprises, communities, volunteers and other social actors playing crucial roles in the response efforts. Reflections on Influenza A (H1N1) Prevention and Control Mechanisms The Legal Status of the Joint National Prevention and Control Mechanism The Joint National Prevention and Control Mechanism was essentially a command and decision-making mechanism established according to the potential amount of damage Influenza A (H1N1) could inflict upon society. On the one hand, Influenza A (H1N1) response required inter-departmental collaboration, and relying solely upon health departments for countermeasures wouldn't be enough; on the other hand, because the virus was not as virulent as to merit the establishment of a State Council Operations Center (or Headquarters), the State Council instead instructed the MOH to establish a multi-departmental joint prevention and control mechanism; and this new organization represented a relatively flexible and effective response mechanism. Although local governments were already aware of the epidemic at its onset and were actively engaging with different departments in their response efforts, because there were no explicit provisions in related laws and contingency plans for the Joint National Prevention and Control Mechanism at the central level, no corresponding normative documents were available for its implementation at a local level. No unified standards on the name, content, form of establishment, and system structure for local governments' prevention and control bodies existed. Although local governments adapted as they went, it was still an environment that incited disorder and confusion. Issues with Emergency Command Responsibilities, Authority and Administrative Levels On the one hand, participating departments fully endorsed the Joint National Prevention and Control Mechanism. This mechanism, they thought, possessed several advantages: Firstly, the joint consultation system made it possible to directly formulate and sign policies at joint prevention and control conferences, which saved time for everyone; Secondly, the joint briefing system required the work groups to send daily reports to other units and departments, thus facilitating both inter-group and inter-departmental communication; And finally, internal collaboration within groups was solid, and the briefing system allowed an unobstructed flow of information. However, the Joint National Prevention and Control Mechanism based upon consultation and communication had its limitations. On issues involving departmental interest, division of duty, and so on, this horizontal collaboration was less efficient than regulation and control by a single, high level leadership department. One contested issue dealt with the location of the local joint prevention and control office: should it be set up in the comprehensive emergency management office of the local government or in the emergency management office of a local specialized department. Some provincial emergency management offices insisted that for an emergency event like the ongoing Influenza A (H1N1) epidemic, a joint prevention and control office should be located in a specialized department so as to leverage the department's expertise and increase response flexibility, convenience and efficiency. In this scenario, the provincial emergency management office would be tasked with solving issues that the specialized department could not. On the other hand, some provincial health department's emergency offices argued that if the office was located in the local government, the joint prevention and control office would enjoy greater authority and more efficient collaboration. The Transition Between Peacetime Mechanisms and Emergency Response Mechanisms Achieving a smooth and effective transition between peacetime and public emergency, and establishing mechanisms that combined crucial components from both systems, was a new challenge that arose in the Influenza A (H1N1) Epidemic. After the 2003 SARS Epidemic, local governments established permanent public health emergency response departments and corresponding working mechanisms to deal with future public health emergencies. These departments and mechanisms should have been employed upon the onset of the Influenza A (H1N1) Epidemic. However, most provinces established completely new leading groups only after the central government established Joint National Prevention and Control Mechanism. In one example, a provincial health department already had a permanent public health emergency operations center, but, after the central government established the Influenza A (H1N1) Joint National Prevention and Control Mechanism, this province created an entirely new prevention and control leading group and a port leading group. At the same time, the Health Department also established new eternal mechanisms, including: the provincial CDC established an emergency response department with leaders from major sections like emergency management and vaccination planning (starting in 2005, this provincial CDC implemented a "3 in 1" meeting system with participation from emergency management, disease control, and the disease monitoring department). The main reason for this redundancy was because the central government did not provide specific conditions or qualifications for contingency planning and management for the transition period between peacetime to emergency. Thus, local governments lacked a clear transition mechanism that they could utilize. It was the reason that many local governments chose to re-establish emergency management bodies when Influenza A (H1N1) broke out. Problems with Inter-departmental Coordination As public health emergency management involved multiple collaboration systems from the central government down to local governments, regions, and departments, inter-departmental collaboration in the response efforts was intrinsically complicated. The response to this epidemic revealed problems that existed both in horizontal and vertical coordination. In regards to horizontal coordination between central departments, the health, education, security, transportation and many other departments were involved in the Influenza A (H1N1) response efforts, which created an environment where responsibilities could easily overlap and grey areas would occur in management. There was also a lack of coordination and standardization between central-level ministries' policy documents for Influenza A (H1N1) countermeasures. For example, in regards to content standardization, the health authorities felt that using the temperature of 37.3 °C as the sole standard for sending people to the hospital was unreasonable and would cause an unnecessary burden on hospitals. In regards to time standardization, on December 2nd, 2009, one province stipulated that only patients with a temperature of 38 °C or higher must be sent to a hospital, and it took the country two more weeks to follow suit. In regards to inter-departmental work, port laboratories in some provinces had begun testing in the early days of epidemic, but stopped after provincial health departments decided that ports were not fit for such work. Obstacles also arose in horizontal coordination and collaboration between local departments. A lack of information communication between local departments due to the unavailability of complete information in the early stages of prevention and control made it nearly impossible for effective collaboration. In regards to the division of labor and coordination between the central and local governments for disease prevention and control, some local governments held that the central government should have presented broader goals and authorized provinces and cities greater autonomy in their response measures. Some felt that the central government should not have made Influenza A (H1N1) prevention and control an issue of political significance but should have been objective in understanding the differences between executive leadership and scientists' opinions. While the main duty of administrative leaders should have been to organize and mobilize social resources needed to cope with the epidemic, scientists should have been the ones to handle technical issues such as epidemic analysis and response measures. At the same time, more efficient communication should have been present between central and local departments tasked with specific operations. For example, some local management departments felt that the entire process was quite political, making some documents difficult to fully implement; in the two most volatile months that lasted from April 28th to June, documents were issued frequently, and in some cases were in conflict with one another and lacked integrity and continuity. In regards to adjustment of prevention and control strategies, some regions' health departments reported the following issues: higher-level departments frequently adjusted technical guidance and strategies for prevention and control, there was a wide variety of information reporting methods and they were constantly in flux, different departments formulated their own response requirements, and differences occurred in measures and standards; all of which greatly complicated local response operations. Certain communication and coordination issues also existed within the Health Department's internal system. The MOH internal horizontal collaboration needs to be strengthened epidemiological investigations, clinical diagnoses, and laboratory testing to combine the medical treatment and disease prevention. For example, the China CDC played a crucial role as a central technical support body of Influenza A (H1N1) prevention and control in epidemiological information collection, monitoring, analysis, and judgment, but at the same time it also had a lot of administrative duties, and its services and duties overlapped with those of the MOH's Bureau of Disease Control and Prevention. There should be unified leadership and coordination between higher and lower-level health departments within the national epidemic prevention and control system. A certain degree of flexibility is also necessary as provinces differ in epidemic situations, medical resources, geographic features, and so on. In regards to information reporting within the health system, though the China CDC and the MOH had established information systems relating to epidemic surveillance, including an epidemic direct reporting system, no information sharing mechanism was created between the China CDC and medical institutions; in particular, some county level medical institutions didn't even have sound data collection and reporting systems. This resulted in a single point of decision making and command, and their lack of network and information technology weakened the support they could've had in implementing response measures. Roles of NPOs Have Yet to Be Leveraged and Improved NPOs such as the Red Cross Society of China played important roles during the Influenza A (H1N1) prevention and control efforts. However, by comparison with developed countries, China still lags behind in terms of public participation in public health emergencies. There still remain limitations in skill and knowledge, as no emergency volunteer systems or working mechanisms were formed, and no leveraging of NPO resources really occurred. The Legal Status of the Joint National Prevention and Control Mechanism The Joint National Prevention and Control Mechanism was essentially a command and decision-making mechanism established according to the potential amount of damage Influenza A (H1N1) could inflict upon society. On the one hand, Influenza A (H1N1) response required inter-departmental collaboration, and relying solely upon health departments for countermeasures wouldn't be enough; on the other hand, because the virus was not as virulent as to merit the establishment of a State Council Operations Center (or Headquarters), the State Council instead instructed the MOH to establish a multi-departmental joint prevention and control mechanism; and this new organization represented a relatively flexible and effective response mechanism. Although local governments were already aware of the epidemic at its onset and were actively engaging with different departments in their response efforts, because there were no explicit provisions in related laws and contingency plans for the Joint National Prevention and Control Mechanism at the central level, no corresponding normative documents were available for its implementation at a local level. No unified standards on the name, content, form of establishment, and system structure for local governments' prevention and control bodies existed. Although local governments adapted as they went, it was still an environment that incited disorder and confusion. Issues with Emergency Command Responsibilities, Authority and Administrative Levels On the one hand, participating departments fully endorsed the Joint National Prevention and Control Mechanism. This mechanism, they thought, possessed several advantages: Firstly, the joint consultation system made it possible to directly formulate and sign policies at joint prevention and control conferences, which saved time for everyone; Secondly, the joint briefing system required the work groups to send daily reports to other units and departments, thus facilitating both inter-group and inter-departmental communication; And finally, internal collaboration within groups was solid, and the briefing system allowed an unobstructed flow of information. However, the Joint National Prevention and Control Mechanism based upon consultation and communication had its limitations. On issues involving departmental interest, division of duty, and so on, this horizontal collaboration was less efficient than regulation and control by a single, high level leadership department. One contested issue dealt with the location of the local joint prevention and control office: should it be set up in the comprehensive emergency management office of the local government or in the emergency management office of a local specialized department. Some provincial emergency management offices insisted that for an emergency event like the ongoing Influenza A (H1N1) epidemic, a joint prevention and control office should be located in a specialized department so as to leverage the department's expertise and increase response flexibility, convenience and efficiency. In this scenario, the provincial emergency management office would be tasked with solving issues that the specialized department could not. On the other hand, some provincial health department's emergency offices argued that if the office was located in the local government, the joint prevention and control office would enjoy greater authority and more efficient collaboration. The Transition Between Peacetime Mechanisms and Emergency Response Mechanisms Achieving a smooth and effective transition between peacetime and public emergency, and establishing mechanisms that combined crucial components from both systems, was a new challenge that arose in the Influenza A (H1N1) Epidemic. After the 2003 SARS Epidemic, local governments established permanent public health emergency response departments and corresponding working mechanisms to deal with future public health emergencies. These departments and mechanisms should have been employed upon the onset of the Influenza A (H1N1) Epidemic. However, most provinces established completely new leading groups only after the central government established Joint National Prevention and Control Mechanism. In one example, a provincial health department already had a permanent public health emergency operations center, but, after the central government established the Influenza A (H1N1) Joint National Prevention and Control Mechanism, this province created an entirely new prevention and control leading group and a port leading group. At the same time, the Health Department also established new eternal mechanisms, including: the provincial CDC established an emergency response department with leaders from major sections like emergency management and vaccination planning (starting in 2005, this provincial CDC implemented a "3 in 1" meeting system with participation from emergency management, disease control, and the disease monitoring department). The main reason for this redundancy was because the central government did not provide specific conditions or qualifications for contingency planning and management for the transition period between peacetime to emergency. Thus, local governments lacked a clear transition mechanism that they could utilize. It was the reason that many local governments chose to re-establish emergency management bodies when Influenza A (H1N1) broke out. Problems with Inter-departmental Coordination As public health emergency management involved multiple collaboration systems from the central government down to local governments, regions, and departments, inter-departmental collaboration in the response efforts was intrinsically complicated. The response to this epidemic revealed problems that existed both in horizontal and vertical coordination. In regards to horizontal coordination between central departments, the health, education, security, transportation and many other departments were involved in the Influenza A (H1N1) response efforts, which created an environment where responsibilities could easily overlap and grey areas would occur in management. There was also a lack of coordination and standardization between central-level ministries' policy documents for Influenza A (H1N1) countermeasures. For example, in regards to content standardization, the health authorities felt that using the temperature of 37.3 °C as the sole standard for sending people to the hospital was unreasonable and would cause an unnecessary burden on hospitals. In regards to time standardization, on December 2nd, 2009, one province stipulated that only patients with a temperature of 38 °C or higher must be sent to a hospital, and it took the country two more weeks to follow suit. In regards to inter-departmental work, port laboratories in some provinces had begun testing in the early days of epidemic, but stopped after provincial health departments decided that ports were not fit for such work. Obstacles also arose in horizontal coordination and collaboration between local departments. A lack of information communication between local departments due to the unavailability of complete information in the early stages of prevention and control made it nearly impossible for effective collaboration. In regards to the division of labor and coordination between the central and local governments for disease prevention and control, some local governments held that the central government should have presented broader goals and authorized provinces and cities greater autonomy in their response measures. Some felt that the central government should not have made Influenza A (H1N1) prevention and control an issue of political significance but should have been objective in understanding the differences between executive leadership and scientists' opinions. While the main duty of administrative leaders should have been to organize and mobilize social resources needed to cope with the epidemic, scientists should have been the ones to handle technical issues such as epidemic analysis and response measures. At the same time, more efficient communication should have been present between central and local departments tasked with specific operations. For example, some local management departments felt that the entire process was quite political, making some documents difficult to fully implement; in the two most volatile months that lasted from April 28th to June, documents were issued frequently, and in some cases were in conflict with one another and lacked integrity and continuity. In regards to adjustment of prevention and control strategies, some regions' health departments reported the following issues: higher-level departments frequently adjusted technical guidance and strategies for prevention and control, there was a wide variety of information reporting methods and they were constantly in flux, different departments formulated their own response requirements, and differences occurred in measures and standards; all of which greatly complicated local response operations. Certain communication and coordination issues also existed within the Health Department's internal system. The MOH internal horizontal collaboration needs to be strengthened epidemiological investigations, clinical diagnoses, and laboratory testing to combine the medical treatment and disease prevention. For example, the China CDC played a crucial role as a central technical support body of Influenza A (H1N1) prevention and control in epidemiological information collection, monitoring, analysis, and judgment, but at the same time it also had a lot of administrative duties, and its services and duties overlapped with those of the MOH's Bureau of Disease Control and Prevention. There should be unified leadership and coordination between higher and lower-level health departments within the national epidemic prevention and control system. A certain degree of flexibility is also necessary as provinces differ in epidemic situations, medical resources, geographic features, and so on. In regards to information reporting within the health system, though the China CDC and the MOH had established information systems relating to epidemic surveillance, including an epidemic direct reporting system, no information sharing mechanism was created between the China CDC and medical institutions; in particular, some county level medical institutions didn't even have sound data collection and reporting systems. This resulted in a single point of decision making and command, and their lack of network and information technology weakened the support they could've had in implementing response measures. Roles of NPOs Have Yet to Be Leveraged and Improved NPOs such as the Red Cross Society of China played important roles during the Influenza A (H1N1) prevention and control efforts. However, by comparison with developed countries, China still lags behind in terms of public participation in public health emergencies. There still remain limitations in skill and knowledge, as no emergency volunteer systems or working mechanisms were formed, and no leveraging of NPO resources really occurred.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5659385/

Nanotechnology in Glycomics: Applications in Diagnostics, Therapy, Imaging, and Separation Processes

Abstract This review comprehensively covers the most recent achievements (from 2013) in the successful integration of nanomaterials in the field of glycomics. The first part of the paper addresses the beneficial properties of nanomaterials for the construction of biosensors, bioanalytical devices, and protocols for the detection of various analytes, including viruses and whole cells, together with their key characteristics. The second part of the review focuses on the application of nanomaterials integrated with glycans for various biomedical applications, that is, vaccines against viral and bacterial infections and cancer cells, as therapeutic agents, for in vivo imaging and nuclear magnetic resonance imaging, and for selective drug delivery. The final part of the review describes various ways in which glycan enrichment can be effectively done using nanomaterials, molecularly imprinted polymers with polymer thickness controlled at the nanoscale, with a subsequent analysis of glycans by mass spectrometry. A short section describing an active glycoprofiling by microengines (microrockets) is covered as well. 1. NANOTECHNOLOGY The term nanotechnology was first used in 1974 by Professor Nario Taniguchi, 1 but the idea and concepts behind nanoscience began more than a decade earlier with a talk by Professor Richard Feynman, "There´s Plenty of Room at the Bottom." 2 In this talk, the process by which science could control and manipulate single atoms was described. The real breakthrough in nanoscience came in 1981, 3 when the scanning tunneling microscope was developed, enabling the observation of individual atoms, and after the invention of atomic force microscopy (AFM), nanotechnology as a scientific discipline was born. Currently, nanotechnology covers processes for the design, preparation, and application of extremely small things. Materials and structures could be designated as "nano" only if their size (at least one dimension) is within the range of 1 to 100 nm. 4 The discovery of nanomaterials, such as fullerenes 5 and graphene, 6 awarded the Nobel Prize to theirs discoverers. Nanomaterials now form a large family of materials, including metal/semiconducting nanoparticles (NPs), quantum dots (QDs), nanowires, fullerenes, graphene and its derivatives, graphene QDs, and carbon nanotubes (CNTs, Fig. 1 ). 7 , 8 It is not only the size of nanomaterials that matters, but their remarkable physical and chemical properties are gaining increasing attention from scientists both from fundamental and application points of view, using nanomaterials in biology, chemistry, and applied physics. 9 – 14 An interesting feature of nanomaterials is that their dimension is similar to that of biomolecules, such as DNA/RNA, proteins, lipids, and carbohydrates, with numerous applications in biology and biomedicine as well. 9 – 14 Figure 1 Various forms of nanomaterials, which can be integrated with biomolecules for applications in glycomics with their size related to other objects of our world, micro‐ and nano‐world. 2. GLYCOMICS For quite a long time, carbohydrates were considered only as reservoirs of energy and as building blocks providing the organisms strength, that is, cellulose and chitin, which are the most abundant polymers on Earth. 15 Glycans are complex carbohydrates consisting of saccharide units that link together and are attached to proteins and lipids to form glycoproteins and glycolipids, respectively (Fig. 2 ). 16 Glycans are bound to proteins during co‐ and posttranslational modifications via a multistep enzymatic process in the endoplasmic reticulum and Golgi apparatus. 17 Glycans can be classified into several categories based on the bond between a glycan and a protein: N ‐glycans (via –NH 2 group to asparagine), O ‐glycans (via –OH group to serine, threonine, or hydroxylated amino acids), and less‐abundant forms of glycans, such as C ‐glycans (C—C bond via tryptophan) and the quite unusual S ‐glycans (C—S bond via cysteine). 18 , 19 Moreover, glycans can be branched by the formation of biantennary, triantennary, and more complex antennary structures. 20 , 21 Figure 2 The challenges in glycan analysis and separation due to complex and heterogeneous structure of glycans as a result of a variable monosaccharide composition, branching, and multiple glycosylation sites of glycoconjugates. Structure of the most common carbohydrates moieties within glycans is shown as well. Glycomics as a scientific discipline studying the structure and function of glycans is a younger sister of more developed genomics and proteomics. There are several reasons why glycomics is still behind genomics and proteomics: (i) glycans and glycoconjugates are more structurally complex than proteins and DNA/RNA; (ii) it is quite challenging to determine glycan identity/sequences using traditional instrumental techniques; and (iii) glycan biosynthesis cannot be predicted from a template as in the case of DNA and proteins. In addition, the chemical synthesis of oligosaccharides is very challenging (protection of all functional groups with a variety of protecting groups in order to generate the site‐specific deprotection of those which are intended to form a chemical bond) and represents yet another obstacle to obtain valuable intermediate or incomplete structures. Despite intensive research in genomics and proteomics, there are still many questions that cannot be answered by analyzing genome and proteome alone, and glycomics has to be added into the equation. Currently, it is estimated that 70% of human cytosolic proteins and 80% of membrane proteins are glycosylated, underlining the important involvement of glycans within the human body. 16 The function of glycans in living organisms has been revealed at an amazing pace with involvement of glycans in cell—cell and host—pathogen interactions, the binding of cells to the extracellular matrix, the immune response, the differentiation of cells, and other physiological and pathological processes. 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 Since glycosylation involves numerous proteins during synthesis, including transport proteins, glycosyltransferases/glycosidases, and other glycan‐processing enzymes, depending on the actual physiological state of the cell, distinct glycoforms of the same protein within a particular cell can be formed. 34 Each form of such protein can possess different properties, such as distribution in the cell, folding, or stability; thus, glycan alterations can influence the physiological functions of proteins and can indicate pathological processes as well. 35 The correlation between diverse forms of cancer and altered glycan structures has been intensively studied. An increasing amount of evidence suggests that the occurrence of a certain glycan structure correlates with cancer invasiveness, the presence of tumor circulating cells, and the ability to metastasize distant organs. 35 , 36 , 37 , 38 , 39 Thus, glycoproteins can be used as cancer biomarkers for screening, diagnostic, predictive, and monitoring purposes. 40 Changes in protein glycosylation have also been observed in other pathological processes, such as inflammatory and autoimmune diseases, and in processes of aging. 41 , 42 Changes in the glycosylation of immunoglobulin G (IgG) are influenced by pregnancy and gender, 43 which should to be considered in future studies. Several methods have been developed for structural glycan analysis. The most common method is mass spectrometry (MS) in combination with other separation techniques, including a battery of electrophoretic and chromatographic methods and nuclear magnetic resonance. Instrumental techniques can be effectively applied for glycan analysis, but the identification of glycan isoforms and the linkages between carbohydrates within a glycan structure are not trivial issues. 44 Moreover, such an analysis requires sophisticated instrumentation and skilled operators to correctly interpret data with low analysis throughput. Moreover, various pretreatment steps (glycan release, enrichment, and modification) are needed prior to analysis. Lectins are carbohydrate‐binding proteins other than enzymes or antibodies and other than carbohydrate sensor/transport proteins 45 with potential therapeutic applications. 46 Lectins as natural glycan‐recognizing proteins can be effectively applied to analyze glycans because they can detect intact glycans still attached to proteins or even cells. 47 , 48 , 49 , 50 The first step toward the effective utilization of lectins in glycan analysis and for diagnostic purposes was the introduction of lectin microarrays/biochips. 51 , 52 , 53 , 54 Glycan microarrays are a valuable tool in glycomics for the identification of glycan‐binding proteins, 55 , 56 and together with lectin microarrays, such highly parallel analyses with a minute consumption of reagents had led to numerous important discoveries regarding glycan involvement in various cellular processes. However, the main disadvantages of microarray‐based analysis are the low sensitivity of detection with a rather high limit of detection (LOD) and the need to fluorescently label either the ligand or sample. 51 , 52 This is why alternatives to glycan/lectin microarrays with a biosensor detection platform as a viable option have been intensively sought. 57 , 58 , 59 Because of the above‐mentioned importance of glycans in pathological processes, there is substantial interest in visualizing them directly within intact cells and tissues. Numerous strategies have been developed to achieve this, using glycan interactions to allow for a targeted delivery of an imaging probe to a chosen cell/tissue; hence, a precise and accurate detection of, for example, tumors or other cells/tissues with specific receptors was possible. Moreover, when a therapeutic agent is selectively delivered to a targeted cell/tissue, a significant reduction in adverse side effects can be achieved. Chemotherapy is good example of this approach—cytotoxic agents are delivered to and released only in tumor tissues, while healthy cells are not affected. Recently, a theranostic approach has been frequently used, relying on particles selectively delivering both an imaging probe and a drug to a targeted cell/tissue. Furthermore, the specificity of these interactions can be used to design vaccines and novel therapeutic approaches, as glycans bound to receptors that are naturally targeted by pathogens prevent the first step of infection (e.g., pathogen binding). Finally, a partial functional restoration of glycosidases used to cure "liposomal storage diseases" (e.g., Gaucher disease) mediated via glycan interactions has been described as well. 3. NANOTECHNOLOGY IN GLYCOMICS One of the earliest efforts to describe the beneficial properties of nanomaterials in the field of glycomics was published by Reichardt et al. in 2013, applying nanomaterials as support for the delivery of vaccines, cellular imaging, biosensors/diagnostics, glycan isolation, and enrichment. 9 Since this review was published, there has been an explosion of papers focusing on the integration of nanomaterials in the field of glycomics (Scheme 1 ), and in this article, we would like to provide a summary of the achievements in this field in the last 3 years (from 2013). Scheme 1 Overview of different nanomaterials (outside the blue circle) used in combination with biomaterials/biomolecules (inside the blue circle) for various applications (in yellow circle) reviewed in this study. The figure was redrawn from 7 . Copyright (2014), with permission from Elsevier. In the sections describing the integration of nanomaterials within various types of transducers, you will find a comprehensive review of recently developed biosensors, bioanalytical devices, and protocols for the detection of various analytes, including viruses and whole cells, together with their main characteristics (Section 4. ). It is divided into subsections according to nanomaterials used, that is, metal NPs ( Section 4.A ), hybrid nanomaterials ( Section 4.B ), carbon nanomaterials ( Section 4.C ), QDs ( Section 4.G ), and others—mostly glycopolymers ( Section 4.D ) and synthetic receptors ( Section 4.E ), used as an immobilization platform or as a part of an amplification strategy. Some of these applications have been summarized in recent reviews. 7 , 60 Sections describing the application of nanomaterials for various biomedical applications will basically cover the progress in the fabrication of glycan‐based NPs that can either induce an immune response (vaccines against viral and bacterial infections and, more recently, even against cancer cells) or possess another therapeutic effect, including antibacterial effect or enzyme inhibition (Section 5. ). An extensive section covering glycan‐targeted nanocarriers for drug delivery (Section 6.A); therapeutic purposes prepared from carbonaceous, metallic, or polymeric nanomaterials ( Section 6.B ); and advances in in vivo imaging ( Section 6.C ), that is, the selective delivery of imaging probes, is provided as well. There are also sections and subsections dealing with the application of carbon, metal, polymer, and magnetic NPs (MNPs) for nuclear magnetic resonance imaging (NMRi) and another subsection focused on other imaging methods besides fluorescence and NMRi ( Section 6.C ). There are only two review papers either not comprehensively covering the application of a diverse range of NPs—covering biomedical applications only of gold‐ and polymer‐based glyconanoparticles 61 —or not focused on recent developments in this field. 62 The final part of the paper will describe various ways for glycan release ( Section 7.A ) and enrichment ( Section 7.C ), which can occur in a timely way with a minute consumption of precious samples using various nanomaterials ( Sections 7.D – 7.K ). Glycan‐enrichment procedures performed directly on matrix‐assisted laser desorption/ionization (MALDI) plates (Section 8. ) will be provided as well. The application of nanomaterials or nanoporous materials for matrix‐free MS (Section 8. ) will also be described. This section will also discuss the preparation of molecularly imprinted polymers (MIP) with polymer thickness controlled at the nanoscale for the selective affinity‐like enrichment of specific glycoproteins ( Section 7.E ). Moreover, active glycoprofiling by microengines/microrockets (Section 9. ) will be covered as well. Only one review paper has focused on the application of nanotechnology in glycan enrichment; it was published in 2014 and only covers advancements in the field until the end of 2013. 63 4. NANOGLYCOSENSING Glycan moieties attached to protein backbones or lipids are destined mostly for the cell surface. Since glycosylation is the most common co‐ and posttranslational modification of proteins, it is also increasingly recognized as a phenotype modulator of various pathological changes on cell surfaces, mostly during cancer development, 35 , 64 and glycomics is thus believed to soon provide relevant information and detection strategies for glycan‐based diagnostics. 65 A relatively small amount of glycans within a sample of interest could be present (e.g., a prostate‐specific antigen [PSA] contains only one glycan per molecule). 66 Thus, for the structural analysis and identification of glycans using mainly MS, capillary electrophoresis, and liquid chromatography, 67 enrichment methods must be performed prior to analysis. 63 For the analysis alone, the use of glycosylated nanomaterials has gained increasing attention in recent years because NPs decorated with suitable biorecognition ligands, that is, glycan structures in a multivalent form similar to the glycocalyx structures on cell surfaces, 68 could be used for the detailed study of glycan—lectin (Latin: legere —to choose, to read) interactions. 69 The following section describes in detail the most recent progress in the area of glycosensing using various nanostructures, including (but not limited to) metal, carbon, and polymer nanomaterials with respect to the analytical performance of such devices for the analysis of glycoproteins, lectins, viruses, and bacteria or for dynamic N ‐glycan evaluation on cell surfaces. A. Metal Nanomaterials Different lectins, which are ubiquitously found in nature, often provide a model system for studying glycan—protein interactions. These interactions play a pivotal role in viral infections, cell adhesion, and the differentiation and progression of various diseases and are thus of great interest for many clinical applications. Despite this fact, they have attracted much less attention than antibodies. 70 The most commonly used NPs are gold NPs (AuNPs) because they can be easily prepared and modified using standard thiol chemistry (a covalent‐like modification). 4 , 71 The synthesis, modification, and engineering of different (glyco)NPs, including metal, magnetic, and carbon NPs and QDs with a focus on biosensing applications, have been reviewed elsewhere. 60 , 72 , 73 1. Optical Biosensors Moreover, using precisely engineered inter‐NP spacers (e.g., polymer grafts), it was possible to control the degree of plasmonic coupling between NPs. 74 This principle was demonstrated with poly[(lactose) m ‐ b ‐(pyridine) n ] and bivalent galactose‐binding lectin from Ricinus communis (RCA 120 , M w = 120 kDa) on gold nanorods (AuNRs), where the authors claimed to develop a protein assay with an LOD for lectin down to pictogram per milliliter (fM range). 74 Other study 75 focused on the preparation of a sandwich configuration with graphene oxide (GO) and phenoxy‐derivatized dextran (conjugated together through π—π stacking interactions) deposited on gold substrate for specific Concanavalin A (Con A) binding with a final amplification of the signal using dextran‐capped AuNPs. This system was, however, less sensitive than a previous one, offering an LOD of 0.39 μg mL −1 (3.75 nM), but still with a 28.7‐fold increase in sensitivity compared to a direct assay (without any amplification reagent). 75 The application of different graphene‐based interfaces for the surface plasmon resonance (SPR) analysis of biospecific interactions between different ligand—receptor couples was recently reviewed. 76 Since the change in the localized SPR (LSPR) signal of metal NPs is not very sensitive and provides only a little information about the nature of interacting molecules, Craig et al. aimed to develop an on/off surface‐enhanced Raman scattering (SERS) aggregation system for picomolar Con A detection. The authors took advantage of the multivalent character of the interaction (since Con A is a homotetrameric molecule at a physiological pH with four glycan‐binding sites) with aggregated NPs producing a strong electromagnetic field between their interfaces, resulting in an increased SERS intensity. 77 The LOD for this system was as low as 40 pM for the chosen silver NPs (AgNPs) because they provided a tenfold increase in scattering efficiency compared to AuNPs. 77 Plasmonic and, in general, label‐free methods are of a particular interest for glycan‐lectin‐based biosensing because of a negative effect of lectin/glycan labeling on the biorecognition. In 2015, a novel approach called plasmon waveguide resonance was introduced based on glass prisms coated with 50 nm silver and 460 nm silica layers derivatized with mannose and lactose using Cu I ‐catalyzed Huisgen azide‐alkyne cycloaddition (CuAAC). 78 In addition to "click chemistry" based techniques, photocoupling reactions with underivatized glycans represent an alternative approach comparable to CuAAC. 79 AuNPs can be easily covalently attached on a plane gold surface via thiol linkers containing an –SH or –NH 2 head group for sensitive glycoprotein detection using immobilized lectin molecules, 80 providing an increased surface area for immobilization, as well as for the suppression of steric hindrance between lectin molecules because of the spherical shape of AuNPs. This effect proved to be so crucial that prepared impedimetric, a Sambucus nigra agglutinin (SNA) based biosensor, was more sensitive, with an LOD 3 orders of magnitude lower (shift from fM to aM range for the detection of sialylated glycoprotein fetuin) even though the total amount of lectin on the surface was lower in the NP‐based 3D configuration compared to the 2D (planar gold) configuration. 81 Most often, metal NPs are used as signal amplifiers. NPs decorated with saccharide structures ( d ‐maltoheptaose as the largest carbon source among maltodextrins that can be transported to Escherichia coli cytoplasm) promote particle internalization for silica, magnetic, silica‐coated MNPs and silica‐coated QDs. 82 Other optical detection platforms are based on colorimetric bioassays. The possibilities for the improvement of the selectivity and sensitivity of such assays for the detection of lectins, toxins, and viruses were recently reviewed. 83 Besides lectins, different toxins possess the ability to specifically bind various saccharide residues. Within the so‐called AB 5 bacterial toxin family (which also includes cholera toxin, shiga and shiga‐like toxins, and pertussis toxin), a protein is produced by enterotoxic E. coli causing "traveller´s diarrhoea": heat‐labile enterotoxin. Poonthiyil et al. prepared an efficient colorimetric sensor using 12 nm AuNPs with attached thiol‐modified galactose moieties binding to the B‐subunit of heat‐labile enterotoxin (while the A‐subunit is typically the one causing a particular disease among all members of this family) that is able to detect the toxin down to 100 nM. 84 Glycan‐decorated AuNPs may also be used for the rapid evaluation of viral hemagglutinin´s specificity because viral pathogens use lectins encoded by their own or host genome to replicate and spread 85 ; these AuNPs can also be used in antiviral drugs strategies. In general, avian‐adapted influenza viruses prefer α2,3‐linked sialic acid (ligands commonly found in the intestinal epithelia of birds), and human‐adapted viruses bind preferentially to α2,6‐linked sialic acid. Using host cell receptors as biorecognition elements immobilized on AuNPs, it was possible to easily and rapidly evaluate a potential threat to the human host without using complex immunoanalytical strategies. Since influenza hemagglutinin molecules contain several binding sites (Fig. 3 ), the aggregation of NPs occurred with a red‐to‐purple shift. 86 Figure 3 A hemagglutinin from H1N1 Influenza A virus (complexed to three molecules of 3´‐sialyllactose, denoted with yellow arrows), an interaction used to bound virus to glycan epitopes of host cells ( http://pdb.org , code 3HTO). This is an alternative to using highly specific monoclonal antibodies, where an LOD of 7.8 hemagglutinin units (i.e., 7.8×10 6 particles) was achieved recently for human H3N2 influenza A virus without any additional amplification. 87 By using the epitope glycan structures, however, knowing the exact influenza subtype is not required. Moreover, glycans‐based biosensors can benefit from the multivalent character of the protein—glycan bond. 88 , 89 , 90 For instance, NPs modified with thiolated monovalent and trivalent α2,6‐linked sialic acid and thiolated polyethyleneglycol (PEG) were successfully used to discriminate between human influenza virus X31 (H3N2) and avian RG14 (α2,3‐sialic acid binding H5N1 strain) virus with an LOD of 2.55 μg mL −1 . In addition, the trivalent configuration provided more rapid results and a greater sensitivity for plasmonic detection relying on glycoNPs' aggregation compared to a monovalent configuration. 91 As shown in this and other studies, when self‐assembled monolayers (SAMs) on gold surfaces are used, the concentration of a diluting thiol as well as its functional head group and length are crucial for assay performance. Finally, using glycan‐modified nanostructured surfaces, it is also possible to reliably detect any new and, for people, potentially pathogenic viral strains without needing any other reagents just by investigating the specificity of the viral strains toward the human glycan epitopes present in exposed epithelia. SAMs as a form of nanotechnology 4 allow the fine‐tuning of other surface properties, such as wettability and adsorption processes, 92 in addition to tuning interfacial chemical reactivity toward the subsequent covalent immobilization of biomolecules. 2. Electrochemical Biosensors Among others, electrochemical methods (e.g., increasingly popular electrochemical impedance spectroscopy [EIS], suitable for single‐molecule to whole‐cell detection) 93 are highly sensitive for providing an alternative platform for viral hemagglutinins or whole‐particle detection down to attomolar detection limits or single‐viral particles, respectively. 94 , 95 Tung et al. prepared a nanostructured (gold nanohemisphere‐modified) biosensor surface. 96 The impedimetric biosensor with enhanced sensitivity due to increased surface reaction area was used for the ultra‐weak binding of C‐type lectin domain family 5, member A and mosquito‐borne dengue virus particles, causing hemorrhagic fever and shock syndrome, with tens of millions infected people every year. 96 Another ultrasensitive attomolar detection of human H1 and avian H5 viral hemagglutinin was successfully performed on a field‐effect transistor (FET) based device (Fig. 4 ), modified with 3‐aminooxypropyltriethoxysilane (using standard silane coupling) and the simple glycan blotting of two different trisaccharide receptors (α2,3‐ and α2,6‐sialyllactose). 97 It is worth mentioning here that the above‐mentioned PEG molecules have become a gold standard for surface modification (even in a microarray format) 98 for the suppression of nonspecific interactions and for the stabilization of NPs when used for biosensing applications. 99 However, the introduction of glycan moieties, such as N ‐acetylglucosamine (GlcNAc) and lactose, on the surface of AuNPs and AuNRs was proposed to be an alternative to PEG stabilization, ensuring colloidal stability in protein‐rich media and preventing phagocytosis by macrophages, but at the same time exhibiting an excellent sensitivity toward carbohydrate‐binding proteins. 100 In addition to PEG, zwitterionic thiol derivatives (e.g., carboxy‐ or sulfobetaine derivatives in single‐component SAMs and mixed binary SAMs) 101 , 102 are promising ways to prepare nanostructured interfaces resisting nonspecific interactions. 103 Figure 4 A field‐effect transistor (FET) biosensor device for detection of viral hemagglutinins down to attomolar level using glycoblotting protocol; (A) typical scheme and picture of the described device and (B) modification scheme of FET device surface using silane chemistry. The slight difference in the glycan epitope (3´‐ vs. 6´‐sialyllactose) is enough to distinguish between Influenza viruses able to infect a human and other types of Influenza . Reprinted with permission from 97. Copyright 2013 American Chemical Society. The multivalent amplification effects of glycan—receptor interactions are naturally present in biological systems, for example, complexly branched glycans on glycoproteins or densely packed rafts of glycolipids. Therefore, synthetic multivalent saccharides (among others) have often been used to mimic nature. 104 Good examples are synthetic polymers, such as poly(acrylamidophenyl‐α‐mannose‐ co ‐acrylamide) attached to AuNPs, 105 layer‐by‐layer modifications of MNPs coated with polysaccharide shells of hyaluronan and chitosan. 106 These polymers were applied for the selective enrichment of glycoconjugates in more complex biological samples 106 and for the non‐covalent modification of hydrophobic, dodecanethiol‐stabilized AuNPs (synthesized by Brust—Schiffrin method 107 ) by multivalent glycocalixarenes with four mannose units to improve targeting efficiency toward intact cells. 108 Glycan parts of different enzymes frequently serve as models to study interactions with lectins. A practical application for the food industry and agriculture is an electrochemical device based on palladium NPs (PdNPs) as catalysts for the 3,3´,5,5´‐tetramethylbenzidine sulfate/H 2 O 2 system with immobilized mannose‐binding jacalin‐related lectin from rice ( Oryza sativa , a bioprobe) for the electrochemical detection of Magnaporthe oryzae (also M. grisea ) chitinase (a biochemical marker) during early rice infections. The use of a magnetic‐controllable electrode together with magnetic bead based PdNPs allows the detection of chitinase down to 17 pg mL −1 (approximately 420 fM). Moreover, using chronopotentiometry, it was possible to detect chitinase 2 days after rice infection, while a standard enzyme‐linked immunosorbent assay (ELISA) could detect the chitinase only 4 days after infection. 109 In addition to NPs and nanorods, nanoporous materials, for example, nanoporous gold in combination with square‐wave voltammetry (SWV) 110 and nanoporous gold monoliths in combination with thermogravimetric analysis, 111 were used to effectively and sensitively detect glycan—protein interactions, for example, to detect high‐mannose glycan‐containing ovalbumin molecules. 112 In situ glycosensing on the cell surface may play an important role in determining the physiological status of the whole cells in the biopsy samples acquired from patients suffering from various diseases. EIS was used to detect human colon cancer DLD‐1 cells with an LOD of 40 cells mL −1 by bovine serum albumin (BSA) incorporated Ag nanoflowers on a glassy carbon electrode (GCE) with a 3D porous architecture and a large surface area and the retention of immobilized cells activity after binding (attributed to the presence of BSA as a biocompatible support). 113 After the conjugation of the cells with the selected lectin (SNA in this case), the average number of sialic acid molecules on a single living cell was counted as approximately 2.16 × 10 12 . 113 Su et al. developed a novel lab‐on‐a‐paper device for the electrochemical sensing of K562 cancer cells with an LOD of 400 cells mL −1 and a wide linear range spanning 5 orders of magnitude based on a macroporous Au‐paper electrode. 114 Given that the volume used for incubation was as low as 10 μL of cell suspension, the LOD of the device was approximately 4 cells. An in situ monitoring of multiglycan expression in response to drug treatment was achieved using differential pulse voltammetry (DPV) and horseradish peroxidase (HRP) labeled wheat germ agglutinin (WGA), peanut agglutinin, Dolichos biflorus agglutinin, and Con A. A similar device designed by the same group later that year based on an aptamer modified 3D macroporous Au‐paper electrode was developed in a microfluidic format to screen anticancer drugs (Fig. 5 ). 115 This device could also detect as little as 350 cells mL −1 (very similar to a previous case, e.g., approximately four cells in 10 μL) using the biosensor signal generated using an HRP‐labeled annexin V bioprobe. This bioprobe specifically interacts with membrane phosphatidylserine molecules (in the presence of Ca 2+ ions), whose externalization cannot proceed in healthy and necrotic cells, thus providing a highly specific response toward apoptotic cells (translocation from the inner to the outer leaflet of the membrane is an important indicator of apoptosis). 115 Recent advances in electrochemical cytosensing amplified by nanostructures and nanocrystals were reviewed by Hasanzadeh et al. 116 Figure 5 Schematic presentation of a macroporous Au paper electrode (μPECD); (A) cells are incubated with HRP‐folic acid conjugate, (B) folding of μPECD and clamping the μPECD between circuit boards (front (C) and reverse side (D)) and finally the detection principle (E), where a is o‐ phenylenediamine and H 2 O 2 , b is 2,2´‐diaminoazobenzene and c is the final product of the reaction. The real size of the device is compared to a Chinese 1 yuan coin. Reprinted from 115 . Copyright 2014, with permission from Elsevier. No doubt the most commonly investigated and the most commercially successful biosensors are those for glucose detection based mostly on glucose oxidase (GOx) enzyme, despite the fact that even enzyme‐free nanostructured sensors have been described. 117 However, the synergistic effect of an NP catalyst with a conducting hydrogel heterostructure‐based interface 118 and the dependence of the sensor performance on NP shape were demonstrated only recently. 119 Li et al. not only described a simple glucose detection, but went even further with the imaging of intracellular glucose consumption in living cancer cells. 120 Their system is based on apo‐GOx (an inactive form of GOx) modified AuNPs and fluorescein isothiocyanate (FITC) dextran. In the presence of glucose in the environment, the quenched fluorescence of FITC‐dextran is recovered as glucose exhibits greater affinity than dextran to apo‐GOx . The LOD of 5 nM, along with the introduction of apo‐GOx instead of GOx, means no consumption of O 2 with subsequent H 2 O 2 production (causing cellular damage) and makes this assay a simple, sensitive, and "biofriendly" method for various disease diagnoses and metabolomics studies. 120 Even though glucose is an important molecule for diagnostics of different conditions and cell metabolism and its importance is highlighted in many sections throughout this review, glucose as a monosaccharide is not considered as a glycan and thus glucose‐based biosensors are not discussed here in more details. It is quite difficult to analyze electrochemically inactive glycans and oligo‐ and polysaccharides using electrochemical methods. 121 However, Paleček and his team recently published several papers about the deacetylation of N ‐acetylated glycans to make the –NH 2 groups free and thus electrochemically active 122 and about the modification of glycans with Os(VI)L complexes for subpicomolar detection. 123 Most recently, a paper about the label‐free electrochemical detection of interaction between Con A lectin and a glycoprotein on an atomically smooth mercury electrode was described. 124 B. Hybrid Nanomaterials and Nanocomposites Very often, not only single nanomaterials but also combinations of two materials where at least one is a nanomaterial are used for bio‐ and cytosensing. 125 AuNPs, forming novel nanocomposites with other materials, are often used as signal amplifiers. Poly(ethylenimine) (PEI)‐reduced GO (rGO) and hollow AuNPs deposited on GCE with immobilized GOx as recognition elements significantly improved the signal intensity of the luminol/H 2 O 2 electrochemiluminiscent (ECL) system for Con A detection down to 0.31 ng mL −1 (approximately 3 pM). 126 Chen et al. developed a sandwich electrochemiluminiscent biosensor with a Con A‐integrating AuNP‐modified Ru(bpy) 3 2+ ‐doped silica nanoprobe and a multiwalled CNTs (MWCNTs) modified electrode with another Con A on the surface (Fig. 6 ). 127 They were able to detect myelogenous leukemia K562 cells with an LOD of 600 cells mL −1 and, more importantly, to dynamically observe the cell surface glycoprofile during different phases of growth in vitro in response to external stimuli—glycan release by peptide‐ N ‐glycosidase (PNGase) F or incubation with the N‐ glycan inhibitor tunicamycin. 127 Figure 6 A schematic illustration of electrochemiluminiscent ECL biosensor for dynamic evaluation of cell surface N ‐glycan expression; (A) fabrication procedures of Con A‐modified NPs presenting the lectin molecules in a multivalent manner and (B) ECL biosensor for cytosensing and evaluating cell surface N ‐glycans, while the signal is reflecting the action of various inhibitors or glycosidases compared to untreated cells. Reprinted with permission from 127 . Copyright 2013 American Chemical Society. The same team later that year published another study describing competitive recognition and a signal amplification strategy using AuNPs modified with GOx. 128 They counted all of the mannose moieties on a single K562 cell (1.8 × 10 10 ) and again demonstrated the importance of the multivalent character of glycan—protein interactions, as the apparent dissociation constant between GOx‐Au and Con A nanoprobes was 1.64 nM—approximately 5 orders of magnitude lower than in the interaction of Con A with mannose. 128 The low LOD for K562 (50 cells mL −1 with a working volume of 200 μL and a linear response of up to 800 cells mL −1 ) was achieved using a graphene‐hemin‐AuNRs ternary composite as a peroxidase mimetic. 129 For influenza detection, a nanohybrid of the Pt NPs (PtNPs), porous ZnO spheres and hemin was synthesized for an amplified electrochemical immunosensor (Fig. 7 ). 130 Briefly, by the in situ generation of a redox probe by alkaline phosphatase (i.e., the release of 1‐naphthol from inactive 1‐naphthyl phosphate) and the excellent behavior of the Pt‐pZnO‐hemin nanocomposite applied as a signal enhancer, the influenza antigen was successfully detected on an antibody‐modified electrode in a sandwich configuration using DPV with an LOD of 0.76 pg mL −1 and with a linear range spanning 4 orders of magnitude. 130 Figure 7 (A) A scheme for preparation of a Pt‐pZnO‐hemin conjugate with a secondary antibody (green) and alkaline phosphatase (yellow). (B) A working principle of the proposed biosensor for influenza antigen detection using a primary antibody (purple) immobilized on an AuNPs modified electrode. Reprinted from 130 . Copyright 2016, with permission from Elsevier. More recently, He et al. introduced a novel sandwich strategy for a dual‐potential responsive, ECL biosensor for simultaneous cytosensing and surface N ‐glycan evaluation. 131 At a potential of 1.25 V, chemiluminescence was generated by Ru(phen) 3 2+ ECL probes intercalated in the grooves of double‐stranded DNA consisting of a DNA aptamer for MCF‐7 (breast cancer) cell recognition and a complementary capture DNA strand and immobilized on electrochemically reduced MoS 2 nanosheets. In the presence of cells, the capture DNA and the ECL probe were released from the electrode interface. The sandwich was then completed by a Con A‐conjugated AuNP‐modified graphite‐C 3 N 4 to detect cell surface mannose units at a negative potential of –1.6 V. 131 Zhang et al. published a paper on an ECL biosensor based on PEI‐rGO and hollow AuNPs. 126 The interaction between AuNPs and the—NH 2 groups of PEI was used for AuNP and GOx immobilization in this case, where GOx served as a producer of H 2 O 2 for the luminol/H 2 O 2 ECL system. In the presence of Con A, a decrease in the ECL intensity was observed, with an LOD down to 310 pg mL −1 (approximately 3 pM) and with a linear range from 1 to 20 ng mL −1 . The authors claimed that they developed an assay with a nearly 1000‐fold improved detection limit for Con A compared to previously published methods. 126 The same team prepared a similar device using a nanocomposite consisting of C 60 fullerene and rGO as a detection interface and hollow Au nanosphere‐conjugated GOx as a label. 132 The interaction of GOx with the electrode interface was mediated by phenoxy‐derivatized dextran, which served as a recognition element for Con A. Using a luminol/H 2 O 2 based ECL system, the LOD for Con A was estimated to be 30 pg mL −1 (approximately 288 fM) with a linear range spanning 3 orders of magnitude (from 0.1 to 100 ng mL −1 ). 132 Multivalent recognition and dual‐signal amplification strategies with Con A‐conjugated poly(amidoamine) (PAMAM) on a chemically rGO interface and HRP‐aptamer‐AuNPs nanoprobes were reported to detect CCRF‐CEM (human acute lymphoblastic leukemia) cells down to 10 cells mL −1 with excellent selectivity and could dynamically evaluate surface N ‐glycans. 133 Graphene could also be used as a support in combination with other metal NPs for biosensing applications. For instance, an rGO‐ and AgNP‐based nanocomposite was used as a redox probe together with phenoxy‐derivatized dextran and GOx as biorecognition elements for sensitive Con A detection on GCE. 134 Different electrochemical techniques, that is, cyclic voltammetry, DPV, and EIS, were used for signal generation with an LOD for Con A as low as 0.67 ng mL −1 (approximately 6.44 pM) and with a linear range from 2.0 to 322 ng mL −1 . Furthermore, the device was successfully used in diluted real human sera with recoveries from 92 to 108% and showed no major interference from BSA, cytochrome c, or phytohemagglutinin, suggesting possible applications for rapid and reliable clinical diagnostics. 134 Another example of the application of a nanocomposite for biosensing purposes is an rGO‐tetraethylene pentamine‐1‐butyl‐3‐methylimidazolium hexafluorophosphate hybrid composite. 135 A dense adsorption of bimetallic AuPtNPs, subsequently used for SNA lectin immobilization, was achieved by free –NH 2 groups from tetraethylene pentamine. This biosensor was used for the electrochemical detection of α2,6‐sialylated glycan down to 3 fg mL −1 and showed a wide linear range covering 8 orders of magnitude. As in a previous case, the recovery when analyzing real human sera was very similar, that is, with a range from 100.8 to 101.4% for Neu5Ac‐α2,6‐Gal‐β‐MP glycoside (4‐methoxyphenyl group via O ‐glycosidic linkage) spiked to a final concentration of 1 pg mL −1 to 100 ng mL −1 . 135 Because α2,6‐sialylated glycans might play an important role in clinical diagnostics using various biomarkers (e.g., on PSA), 136 new methods for their ultrasensitive detection are still emerging. 137 For example, the nanocomposite composed of graphite oxide, Prussian blue, and PTC‐NH 2 (ammonolysis product of 3,4,9,10‐perylenetetracarboxylic dianhydride) was used on a GCE to immobilize AuNPs through free –NH 2 groups and SNA‐I lectin for the DPV analysis of α2,6‐bound sialic acid on serum glycoproteins down to 0.03 pg mL −1 with a linear range spanning 5 orders of magnitude. 138 Although many different and successful strategies and protocols have been proposed relaying on different electroanalytical approaches for the analysis of complex glycan structures, the main challenge to be addressed for the application of affinity biosensors for real sample analysis is the efficient blocking of nonspecific interactions on the biorecognition interface, although BSA could be effectively applied as a blocking agent in some cases. Additionally, it is necessary to note that despite the fact that SNA lectin is routinely used to detect α2,6‐bound sialic acid, according to some specialized vendors, there is a minor α2,3‐sialic acid binding activity present as well. 139 C. Carbon Nanomaterials Engineered carbon nanomaterials (nanotubes, graphite, fullerenes, and graphene as main matrices for the conjugation of biomolecules) can be successfully applied for the preparation of biosensors. 140 There are different conjugation techniques for the functionalization of carbon structures with carbohydrates; the most common is the use of a carboxylic group formed on carbon surfaces using strong acidic oxidation with the subsequent conversion of –COOH groups to acyl chlorides, direct (carbodiimide activated) amidation, or ligation with an azide (Staudinger ligation). 141 Single‐walled CNTs (SWCNTs) were also functionalized through microwave‐assisted functionalization using perfluorophenyl azides with mannose and galactose. 142 This interface provided a reliable platform for agglomeration studies using FITC‐Con A lectin, specifically binding to α‐ d ‐mannopyranoside and, to a lower extent, to α‐ d ‐glucopyranoside residues, but not to Gal‐modified SWCNTs. 142 Through phenylacetylene‐SWCNTs and few‐layer graphene flakes, Ragoussi et al. prepared a carbohydrate‐modified (α‐ d ‐mannosyl glycodendron‐bearing) carbon nanostructures for Con A detection (using AFM, fluorescence, and UV/VIS studies), connected by means of CuAAC "click reaction" mechanism. 143 This group even managed to capture and observe the same object for their AFM study, leading to a reliable height profile analysis of the nanostructures before and after treatment. 143 Other detection platforms may be useful for studying glycan—lectin interactions using graphene‐modified surfaces. For SPR‐based experiments, graphene was grown through chemical vapor deposition (CVD) on polycrystalline Cu foils in a five‐step process on a 50‐nm‐thick Au film as single and double layers. 144 A simple immersion of this interface into a 100 nM mannose solution was sufficient for the mannose modification of the interface through the interaction of carbohydrate with the aromatic ring structure of graphene. Using mannose‐specific Lens culinaris agglutinin and GlcNAc‐ and sialic acid‐specific Triticum vulgaris agglutinin, it was shown that noncovalent surface modification by a simple mannose adsorption allowed the tuning of surface selectivity towards a specific receptor in a simple manner, allowing an LOD of approximately 1 μg mL −1 (low nM range) for lectins with a linear range of up to 1000 μg mL −1 . 144 However, electrochemical and ECL‐based devices are more sensitive than optical platforms. Since the Nobel Prize in physics in 2010, graphene has become an increasingly popular material and has recently been applied for biosensing purposes, as well as for glycomics, mainly in cytosensing applications using different detection strategies. As already discussed in Section B. .B regarding nanohybrids, the combination of different materials, the preparation of nanocomposites, and the very frequent utilization of a sandwich format analysis is common for biosensor construction. It has been previously reported that polymer dendrimers (e.g., PAMAM in this case) provide excellent support for the immobilization of glycans, allowing them to interact with proteins involving a multivalency effect. Molecular recognition using so‐called corona‐phase complexes consisting of synthetic polymers and CNTs, where the two components show affinity toward a selected analyte only if they stick together via surface forces stabilizing them and giving the polymer its final configuration (with a possibility to predict recognition specificity in advance), was also reported. 145 SPR gold surface coated with rGO (electrophoretically deposited) can be easily modified by a simple immersion in a solution of a particular polymer to prepare strongly negative (poly(sodium 4‐styrenesulfonate)) or positive (PEI) surfaces, as well as a surface modified by different saccharide moieties (mannose, lactose) through π–π stacking and electrostatic interactions. Subramanian et al. modified SPR chips with rGO to study the affinity of three different pathogenic E. coli strains to surfaces mediated by the presence of different adhesins on a bacterial cell membrane because those are responsible for the colonization of different epithelial structures and surfaces. 146 The modified SPR interface interacted strongly with highly pathogenic E. coli 107/86 strain in a quantitative manner with a linear response spanning 7 orders of magnitude and with an LOD of ∼100 cfu mL −1 (cfu—colony‐forming units) for bacterial strains 107/86 and UTI89. 146 ECL methods based on carbon nanomaterials could be used for cytosensing applications, as well. The simplest sandwich configuration, in which GO‐modified GCE served to immobilize the antibodies of interest (anti‐PSA in this case for the specific biorecognition of membrane PSA) and with a subsequent surface blocking by the use of BSA, was used to detect PC‐3 (prostate cancer) cells down to 260 cells mL −1 . 147 With a linear range spanning almost 2 orders of magnitude, ruthenium complex‐labeled WGA served as a signal probe. 147 A similar concept used for the impedimetric detection of HL‐60 (human promyelocytic leukemia) cells down to 500 cells mL −1 was based on a graphene surface modified by carboxymethyl chitosan. 148 This composite served to support the layer‐by‐layer assembly of PEI and folic acid for the fabrication of a label‐free cytosensor. Folic acid served as a biorecognition element because the overexpression of folate receptors often occurs in some tumor cell lines. 148 Graphene may be successfully utilized in various forms in biosensing technologies using not only plane graphene sheets but also monolithic and macroporous graphene foam. Such a 3D matrix (grown by CVD) was used to prepare an immunosensor for carcinoembryonic antigen (CEA, a tumor biomarker). 149 Briefly, a graphene substrate was used for the polymerization of dopamine, which subsequently served as a matrix for noncovalent Con A immobilization and interaction with HRP‐labeled anti‐CEA as a biorecognition element bound to Con A via a glycan part of HRP. After the surface was blocked by HRP, various electrochemical methods (mainly DPV using an electrochemical mediator) were used to detect CEA down to 90 pg mL −1 (approximately 500 fM), and the biosensor did not show any response toward other biomolecules, such as BSA, PSA, HRP, or glucose. Noncovalent graphene modification could be achieved not only by unmodified saccharides, 144 but as well as using "clickable" monosaccharide derivatives, such as azido galactosides immobilized on an alkynyl anthraquinone‐modified graphene electrode for the label‐free EIS detection of cancer cells. 150 D. Other Nanostructures, Glycopolymers, and Boronic Acid Derivatives Commonly used nanomaterials in glycomics in addition to metal and carbon nanostructures are glycopolymer‐based micelles, vesicles, or nonspherical NPs that are able to interact with lectins as multivalent ligands in a manner similar to natural glycoproteins. 48 Block copolymers often self‐assemble into diverse morphologies in solution depending on their properties, providing a promising bottom‐up engineering strategy for different applications of such nanostructures. 61 Polymer scaffolds may also be effectively glycosylated in vitro using a wide variety of available glycosyltransferases to prepare glycan structures mimicking those present in nature for the biorecognition of multivalent glycans by their specific lectin receptors. 151 Any information contained in the "sugar code" must be controlled with an extreme precision during glycopolymer preparation in laboratories for diagnostic and other purposes because every small difference in vivo may significantly affect a biorecognition event and lead to structural and functional abnormalities in the organism. Therefore, structural control of carbohydrate sequences during the synthesis of glycomimetics and multivalent glycopolymers is of highest importance to obtain reliable data. 152 , 153 Such synthetic glycopolymers are promising tools for use in emerging biomedical applications and research, including biosensing, biomolecular recognition, and vaccine development. 154 , 155 Multivalency and complexity of lectin—glycan interactions are applied in numerous processes in nature. For the study of such a complex combination of binding mechanisms in real time, dendrimers may serve as useful tools to evaluate the binding capacity of lectin receptors and the effect of avidity. Mannosylated gallic acid‐triethylene glycol‐based dendrimers in combination with SPR provided important structural data for studying biorecognition between Con A and mannose‐modified dendrimers. 156 An amphiphilic block copolymer consisting of hydrophilic lactose and hydrophobic pyridine was synthesized via reversible addition‐fragmentation chain transfer polymerization. 157 Glycosurface prepared on Au quartz crystal microbalance (QCM) chips was used to reliably detect RCA 120 in the nanomolar concentration range without any significant binding of BSA as a nonspecific probe. 157 Moreover, the authors in this study calculated the K A for the system, obtaining a value of 6.3 × 10 6 M −1 ; this value is normally in the range of 10 3 M −1 for monovalent lactose and its receptor. A similar value of K A (same order, 2.3 × 10 6 M −1 ) was obtained in another study using QCM and RCA 120 lectin binding to galactose‐containing gradient glycopolymer synthesized by RAFT polymerization. 158 By synchronizing enzymatic monomer transformation with polymerization, the authors obtained a gradient sugar distribution in a final amphiphilic polymer. 158 The lowest detectable concentration (5 μg mL −1 ) was again in the low nanomolar region. Superior lectin binding was achieved for the gradient polymer compared to the statistical glycopolymer, underlining the relevance of multivalency in the case of lectin—glycan interactions. 158 Glycoconjugated amphiphilic polymers can also be used for the encapsulation of fluorescent QDs. 159 Prior to its encapsulation, amphiphilic poly(isoprene)‐ b ‐poly(ethylene glycol) diblock copolymer was covalently modified by a carbohydrate moiety ( d ‐manno‐heptulose, d ‐glucose, d ‐galactose, bis(nitroso)‐streptozotocin, or d ‐maltose) using Huisgen‐type click chemistry, and interaction with Con A was studied again using the SPR method, showing enhanced affinity constants due to multivalent binding effects. 159 Supramolecular structures, which are of high importance in nanotechnology these days, may also be prepared by click chemistry reactions, as in the work published by Assali et al., in which the authors managed to synthesize poly(diacetylene)‐based nanomaterials with different morphologies. 160 Neoglycolipids with an amide bond between the hydrophilic and hydrophobic parts of the amphiphilic molecule formed 3D micelles, while triazole‐containing ones (obtained by "click‐reaction") allowed 1D nanotube formation. 160 Block glyco‐copolymers may also be used for cell imaging and as an effective drug delivery system (see Section 6. ). They may also enhance the uptake of drug‐loaded micelles by cells, as in the case of the increased uptake of doxorubicin‐loaded sugar (glucose or maltose as a biorecognizable hydrophilic block modification) and poly(4‐substituted‐ε‐caprolactone) copolymer micelles by HeLa cells, compared to free doxorubicin. 161 Since glycans are the most complex biomolecules, there is a need for high‐throughput methods for their analysis. In addition to commercial microarrays, a novel super‐microarray (containing many microarrays on the same slide) for lectin glycan sensing was recently developed. Such arrays use glycan‐labeled dye‐doped silica NPs (SiNPs) and a set of lectins immobilized on epoxy slides with poly(dimethylsiloxane) as an insulator, allowing the generation of many individual lectin microarrays, which significantly increase the assay throughput and, due to the multivalency of glycan‐modified NPs, also increase the affinity (over the free glycan and corresponding lectin) by 4–7 orders of magnitude. 162 Moreover, fluorescently labeled NPs offer higher stability and fluorescence compared to free organic dyes. Although glycans‐modified NPs have previously been prepared, the first attempt to prepare carbohydrate‐modified SiNPs was published recently by Ahire et al. 163 d ‐mannose‐capped SiNPs (prepared from amine‐terminated NPs using N,N´‐dicyclohexylcarbodiimide) were used to detect Con A when the interaction caused the aggregation of NPs. To show their biochemical activity, the photoluminescence of these NPs after interacting with MCF‐7 human breast cancer cells was also investigated. 163 For electrochemical analysis, conductive polymers are highly relevant for the enhanced sensitivity of detection when used as a solid‐state redox probe. Thiophene containing fused quinone moieties were electrochemically polymerized on a gold electrode surface to couple thiol‐modified mannose. 164 Such electropolymerization created a thin film on a solid surface with the ability to control its thickness very precisely up to several nanometers with subsequent application to construct microsensors. This new glycosurface allowed the detection of two major bacterial cell surface biomarkers—namely, fimbriae proteins on bacterial pili and lipopolysaccharides (LPSs) on G ‐ bacteria (by Con A‐mediated binding), using SWV and QCM methods down to 25 and 50 cells mL −1 , respectively. 164 Moreover, it was quite simple using this method to selectively distinguish between G‐ and G+ bacteria. 164 E. Synthetic Receptors for Glycosensing Common biorecognition elements for the sensitive detection of various analytes (biomolecules, viruses, or even bacteria) include antibodies and less common nucleic acid aptamers. In the past decade, carbohydrates have been increasingly studied due to their presence on the surfaces of proteins and cells. For the purpose of glycocode deciphering, lectins from various sources are commonly used. 22 Boronic acids also bind saccharides via reversible interactions, mostly with linear diols or even cis ‐1,2‐diols on five‐membered rings or 1,3‐diols to form five‐ or six‐membered rings. 165 , 166 Fluorescent diboronic acid compounds with dipeptide linkers were synthesized to discriminate cell‐surface Lewis X (Le x ) trisaccharide present on Chinese hamster ovary (CHO) CHOFUT4 cells at micromolar concentrations. 167 The control cells (without glycan expression, HEP3B cells predominantly expressing Le y , B16FUT3 cells expressing sialyl Lewis a (Le a ) and COLO205 cells expressing sLe x and sLe a but no Le x ) were not labeled, 167 suggesting the possibility of preparing compounds with a specificity toward glycans comparable to that of naturally occurring lectins. As previously mentioned ( Section 4.A.2 ), pathogenic agents, such as viruses and bacteria, use their envelope proteins (agglutinins) and adhesin lectins to recognize and attach themselves to host cells and tissues via glycans. This principle was used to prepare a novel electrochemical displacement sensor based on three different boronic acid derivative tracers (containing a ferrocene molecule). 168 The displacement of tracers by Con A lectin molecules or E. coli cells led to a decrease in the electrochemical signal monitored by SWV. Moreover, the use of thiolated mannose‐OEG conjugate ensured low nonspecific interactions. Con A could be detected with an LOD of 1 μg mL −1 (approximately 9.6 nM, with a linear range spanning ∼2 orders of magnitude), and E. coli cells could be counted down to 600 cells mL −1 . 168 The novel tracer used in this study, 2‐hydroxymethyl phenyl boronic acid derivative, binds to mannose even at a neutral pH, expanding the application of the system toward real biological samples (e.g., urine). 168 To date, many synthetically prepared "boronolectins" showed only a moderate fluorescence enhancement with a requirement of significant amount of co‐solvents in aqueous solution (i.e., dimethylsulfoxide and ethanol). A newly engineered boronolectin derived from tricarbocyanine combined with a boronic acid fragment linked by a piperazine unit exhibited improved certain properties, such as excellent water solubility and sensitive fluorogenicity, upon binding to carbohydrate moieties under a physiological pH. 169 To conclude, because boronic acid derivatives are able to successfully mimic lectins as natural glycan decipherers, they may be used not only to detect various analytes but also to selectively bind to free viral particles to inhibit their progression and surface adhesion, as in the case of lipid nanocapsules functionalized with amphiphilic boronic acid for hepatitis C virus inhibition, similar to cyanovirin‐N or griffithsin (both potent HIV inhibitors). 170 Their use as ultrasensitive solid‐phase microextraction probes for in vivo and in vitro sensing purposes in biofluids and even semisolid biotissues was also demonstrated. 171 F. Naked‐Eye Detection Using Nanostructures AuNPs modified with different saccharide moieties (lactose, arabinose, cellobiose, sucrose, mannose, glucose, and galactose) were applied by Jayawardena et al. to successfully distinguish among four different lectins with different specificities. 172 Con A, soybean agglutinin, Griffonia simplicifolia agglutinin, and Arachis hypogaea peanut agglutinin were detected by observing a red shift in the λ max of the LSPR absorption (LSPR on NPs, as opposed to propagating SPR biosensors). 172 Such a library‐oriented approach of glycan‐decorated NPs was later used to prepare polymer‐stabilized glyco‐AuNPs for a rapid, high‐throughput, and 96‐well microplate‐compatible evaluation and identification of pathogenic lectins without a need for any infrastructure because the output of these measurements (red‐to‐blue color shift upon AuNP aggregation) was monitored by a digital camera (Fig. 8 ). 173 Plasmonic metal NPs thus have great potential for their use in biosensor technology due to their sensitive spectral response to the local environment of NPs. 174 Figure 8 An overview of the colorimetric detection principle of lectin–glycan interactions with naked eye by glyco‐AuNPs aggregation due to lectin interaction. In case of an aggregation of NPs (in presence of lectin molecules), red‐to‐blue shift in color occurred. Reproduced from 173 , with permission of the Royal Society of Chemistry. SPR, however, lacks the higher throughput capability compared to lectin microarrays. This drawback was overcome recently 175 by establishing a lectin microarray based on a multiplexed SPR interface for the simultaneous measurement of up to 96 interactions by the immobilization of 18 different unmodified lectins (at different dilutions), including controls. A microarray GOAL (Glyco‐gold NP‐based Oriented immobilized Antibody microarray for Lectin) assay was also introduced as a novel approach for the naked‐eye detection of lectin‐carbohydrate interactions after silver enhancement using oriented, surface‐immobilized anti‐lectin antibodies. 176 Moreover, these modified AuNPs were highly stable and resistive to any nonspecific protein adsorption. 176 Human IgGs are extremely important markers of various diseases, which can be applied in a quantitative and qualitative manner because these glycoproteins are responsible for an effective immune response. The glycan part of human IgG was shown to be associated with autoimmune disease progression, mainly rheumatoid arthritis, 16 , 101 where the N ‐linked biantennary complex glycan in the Fc region is terminated with galactose or even GlcNAc, while in healthy individuals, IgG´s glycan can be terminated with sialic acid. 177 The GalNAc biosensor based on poly(diacetylene) nanovesicles developed by Hao et al. was applied for a noninvasive and real‐time colorimetric analysis of galactose‐deficient IgA1 (playing an important role in the pathogenesis of glomerulonephritis—IgA nephropathy) using nanovesicles modified with Helix aspersa agglutinin for naked‐eye detection. 178 In addition to glycoproteins, other glycoconjugates, such as glycolipids, were recently used for sensing applications. A fluorescent glycolipid monomer was synthesized using conjugation between 1‐pyreneboronic acid and a glycolipid based on a condensation reaction between d ‐glucose and oleic acid for the qualitative and chiral sensing of 80 nmol of amino acids ( l ‐ and d ‐tryptophan and phenylalanine) by the naked eye (Fig. 9 ). 179 In order to study carbohydrate–carbohydrate or carbohydrate–protein interactions using glycolipids, three different strategies could be utilized: (i) insertion of a synthetically prepared glycolipid into a lipid matrix, (ii) preparation of glycolipids that aggregate to form liposomes or micelles, and (iii) modification of a hydrophobic surface by a desired sugar derivative. 180 Figure 9 Photographs of the aqueous dispersions of the D‐vesicles (self‐assembled morphology of synthesized fluorescent glycolipid monomer c = 1.5 mmol, average of 48 nm in diameter) in the presence and absence of amino acids (80 nmol) for the naked‐eye detection. Reproduced from 179 with permission of the Royal Society of Chemistry. G. Quantum Dots (QDs) QDs have also attracted considerable attention in many different fields, including bioimaging and the detection of various analytes, mainly because of their tunable optical size‐dependent properties. 181 In recent years, the importance of detecting various forms of viruses has emerged with a focus on the identification of various glycoforms present on viral surfaces. A two‐step procedure was developed for virus detection, including the isolation of viral hemagglutinins by glycan‐modified paramagnetic beads, labeling hemagglutinins with CdS QDs with their subsequent electrochemical detection by voltammetry using 3D printed microfluidic chips. 182 The other detection principle employed by Chen and Neethirajan was based on a homogenous fluorescence quenching principle. 183 They used a sandwich configuration (antibody‐modified AuNPs and glycan‐conjugated QDs) to entrap influenza A hemagglutinins in between these two probes. As a result, a fluorescence decrease due to a nonradiative energy transfer between these two probes was observed. Of course, QDs could be conjugated with a diverse range of different structures in the same way as for the other above‐mentioned nanomaterials. The conjugation of CdSeTe@ZnS‐SiO 2 QDs modified with 3‐aminophenylboronic acid was used to monitor changes in the relative amount of sialic acid on K562 cell surfaces after a 3´‐azido‐3´‐deoxythymidine treatment, showing a significant increase in sialic acid expression. 183 Another paper 184 aimed to develop a photoelectrochemical biosensor using low‐toxic Ag 2 S QDs for glucose detection as well as for the detection of MCF‐7 breast cancer cells down to 32 μM and 98 cells mL −1 . A. Metal Nanomaterials Different lectins, which are ubiquitously found in nature, often provide a model system for studying glycan—protein interactions. These interactions play a pivotal role in viral infections, cell adhesion, and the differentiation and progression of various diseases and are thus of great interest for many clinical applications. Despite this fact, they have attracted much less attention than antibodies. 70 The most commonly used NPs are gold NPs (AuNPs) because they can be easily prepared and modified using standard thiol chemistry (a covalent‐like modification). 4 , 71 The synthesis, modification, and engineering of different (glyco)NPs, including metal, magnetic, and carbon NPs and QDs with a focus on biosensing applications, have been reviewed elsewhere. 60 , 72 , 73 1. Optical Biosensors Moreover, using precisely engineered inter‐NP spacers (e.g., polymer grafts), it was possible to control the degree of plasmonic coupling between NPs. 74 This principle was demonstrated with poly[(lactose) m ‐ b ‐(pyridine) n ] and bivalent galactose‐binding lectin from Ricinus communis (RCA 120 , M w = 120 kDa) on gold nanorods (AuNRs), where the authors claimed to develop a protein assay with an LOD for lectin down to pictogram per milliliter (fM range). 74 Other study 75 focused on the preparation of a sandwich configuration with graphene oxide (GO) and phenoxy‐derivatized dextran (conjugated together through π—π stacking interactions) deposited on gold substrate for specific Concanavalin A (Con A) binding with a final amplification of the signal using dextran‐capped AuNPs. This system was, however, less sensitive than a previous one, offering an LOD of 0.39 μg mL −1 (3.75 nM), but still with a 28.7‐fold increase in sensitivity compared to a direct assay (without any amplification reagent). 75 The application of different graphene‐based interfaces for the surface plasmon resonance (SPR) analysis of biospecific interactions between different ligand—receptor couples was recently reviewed. 76 Since the change in the localized SPR (LSPR) signal of metal NPs is not very sensitive and provides only a little information about the nature of interacting molecules, Craig et al. aimed to develop an on/off surface‐enhanced Raman scattering (SERS) aggregation system for picomolar Con A detection. The authors took advantage of the multivalent character of the interaction (since Con A is a homotetrameric molecule at a physiological pH with four glycan‐binding sites) with aggregated NPs producing a strong electromagnetic field between their interfaces, resulting in an increased SERS intensity. 77 The LOD for this system was as low as 40 pM for the chosen silver NPs (AgNPs) because they provided a tenfold increase in scattering efficiency compared to AuNPs. 77 Plasmonic and, in general, label‐free methods are of a particular interest for glycan‐lectin‐based biosensing because of a negative effect of lectin/glycan labeling on the biorecognition. In 2015, a novel approach called plasmon waveguide resonance was introduced based on glass prisms coated with 50 nm silver and 460 nm silica layers derivatized with mannose and lactose using Cu I ‐catalyzed Huisgen azide‐alkyne cycloaddition (CuAAC). 78 In addition to "click chemistry" based techniques, photocoupling reactions with underivatized glycans represent an alternative approach comparable to CuAAC. 79 AuNPs can be easily covalently attached on a plane gold surface via thiol linkers containing an –SH or –NH 2 head group for sensitive glycoprotein detection using immobilized lectin molecules, 80 providing an increased surface area for immobilization, as well as for the suppression of steric hindrance between lectin molecules because of the spherical shape of AuNPs. This effect proved to be so crucial that prepared impedimetric, a Sambucus nigra agglutinin (SNA) based biosensor, was more sensitive, with an LOD 3 orders of magnitude lower (shift from fM to aM range for the detection of sialylated glycoprotein fetuin) even though the total amount of lectin on the surface was lower in the NP‐based 3D configuration compared to the 2D (planar gold) configuration. 81 Most often, metal NPs are used as signal amplifiers. NPs decorated with saccharide structures ( d ‐maltoheptaose as the largest carbon source among maltodextrins that can be transported to Escherichia coli cytoplasm) promote particle internalization for silica, magnetic, silica‐coated MNPs and silica‐coated QDs. 82 Other optical detection platforms are based on colorimetric bioassays. The possibilities for the improvement of the selectivity and sensitivity of such assays for the detection of lectins, toxins, and viruses were recently reviewed. 83 Besides lectins, different toxins possess the ability to specifically bind various saccharide residues. Within the so‐called AB 5 bacterial toxin family (which also includes cholera toxin, shiga and shiga‐like toxins, and pertussis toxin), a protein is produced by enterotoxic E. coli causing "traveller´s diarrhoea": heat‐labile enterotoxin. Poonthiyil et al. prepared an efficient colorimetric sensor using 12 nm AuNPs with attached thiol‐modified galactose moieties binding to the B‐subunit of heat‐labile enterotoxin (while the A‐subunit is typically the one causing a particular disease among all members of this family) that is able to detect the toxin down to 100 nM. 84 Glycan‐decorated AuNPs may also be used for the rapid evaluation of viral hemagglutinin´s specificity because viral pathogens use lectins encoded by their own or host genome to replicate and spread 85 ; these AuNPs can also be used in antiviral drugs strategies. In general, avian‐adapted influenza viruses prefer α2,3‐linked sialic acid (ligands commonly found in the intestinal epithelia of birds), and human‐adapted viruses bind preferentially to α2,6‐linked sialic acid. Using host cell receptors as biorecognition elements immobilized on AuNPs, it was possible to easily and rapidly evaluate a potential threat to the human host without using complex immunoanalytical strategies. Since influenza hemagglutinin molecules contain several binding sites (Fig. 3 ), the aggregation of NPs occurred with a red‐to‐purple shift. 86 Figure 3 A hemagglutinin from H1N1 Influenza A virus (complexed to three molecules of 3´‐sialyllactose, denoted with yellow arrows), an interaction used to bound virus to glycan epitopes of host cells ( http://pdb.org , code 3HTO). This is an alternative to using highly specific monoclonal antibodies, where an LOD of 7.8 hemagglutinin units (i.e., 7.8×10 6 particles) was achieved recently for human H3N2 influenza A virus without any additional amplification. 87 By using the epitope glycan structures, however, knowing the exact influenza subtype is not required. Moreover, glycans‐based biosensors can benefit from the multivalent character of the protein—glycan bond. 88 , 89 , 90 For instance, NPs modified with thiolated monovalent and trivalent α2,6‐linked sialic acid and thiolated polyethyleneglycol (PEG) were successfully used to discriminate between human influenza virus X31 (H3N2) and avian RG14 (α2,3‐sialic acid binding H5N1 strain) virus with an LOD of 2.55 μg mL −1 . In addition, the trivalent configuration provided more rapid results and a greater sensitivity for plasmonic detection relying on glycoNPs' aggregation compared to a monovalent configuration. 91 As shown in this and other studies, when self‐assembled monolayers (SAMs) on gold surfaces are used, the concentration of a diluting thiol as well as its functional head group and length are crucial for assay performance. Finally, using glycan‐modified nanostructured surfaces, it is also possible to reliably detect any new and, for people, potentially pathogenic viral strains without needing any other reagents just by investigating the specificity of the viral strains toward the human glycan epitopes present in exposed epithelia. SAMs as a form of nanotechnology 4 allow the fine‐tuning of other surface properties, such as wettability and adsorption processes, 92 in addition to tuning interfacial chemical reactivity toward the subsequent covalent immobilization of biomolecules. 2. Electrochemical Biosensors Among others, electrochemical methods (e.g., increasingly popular electrochemical impedance spectroscopy [EIS], suitable for single‐molecule to whole‐cell detection) 93 are highly sensitive for providing an alternative platform for viral hemagglutinins or whole‐particle detection down to attomolar detection limits or single‐viral particles, respectively. 94 , 95 Tung et al. prepared a nanostructured (gold nanohemisphere‐modified) biosensor surface. 96 The impedimetric biosensor with enhanced sensitivity due to increased surface reaction area was used for the ultra‐weak binding of C‐type lectin domain family 5, member A and mosquito‐borne dengue virus particles, causing hemorrhagic fever and shock syndrome, with tens of millions infected people every year. 96 Another ultrasensitive attomolar detection of human H1 and avian H5 viral hemagglutinin was successfully performed on a field‐effect transistor (FET) based device (Fig. 4 ), modified with 3‐aminooxypropyltriethoxysilane (using standard silane coupling) and the simple glycan blotting of two different trisaccharide receptors (α2,3‐ and α2,6‐sialyllactose). 97 It is worth mentioning here that the above‐mentioned PEG molecules have become a gold standard for surface modification (even in a microarray format) 98 for the suppression of nonspecific interactions and for the stabilization of NPs when used for biosensing applications. 99 However, the introduction of glycan moieties, such as N ‐acetylglucosamine (GlcNAc) and lactose, on the surface of AuNPs and AuNRs was proposed to be an alternative to PEG stabilization, ensuring colloidal stability in protein‐rich media and preventing phagocytosis by macrophages, but at the same time exhibiting an excellent sensitivity toward carbohydrate‐binding proteins. 100 In addition to PEG, zwitterionic thiol derivatives (e.g., carboxy‐ or sulfobetaine derivatives in single‐component SAMs and mixed binary SAMs) 101 , 102 are promising ways to prepare nanostructured interfaces resisting nonspecific interactions. 103 Figure 4 A field‐effect transistor (FET) biosensor device for detection of viral hemagglutinins down to attomolar level using glycoblotting protocol; (A) typical scheme and picture of the described device and (B) modification scheme of FET device surface using silane chemistry. The slight difference in the glycan epitope (3´‐ vs. 6´‐sialyllactose) is enough to distinguish between Influenza viruses able to infect a human and other types of Influenza . Reprinted with permission from 97. Copyright 2013 American Chemical Society. The multivalent amplification effects of glycan—receptor interactions are naturally present in biological systems, for example, complexly branched glycans on glycoproteins or densely packed rafts of glycolipids. Therefore, synthetic multivalent saccharides (among others) have often been used to mimic nature. 104 Good examples are synthetic polymers, such as poly(acrylamidophenyl‐α‐mannose‐ co ‐acrylamide) attached to AuNPs, 105 layer‐by‐layer modifications of MNPs coated with polysaccharide shells of hyaluronan and chitosan. 106 These polymers were applied for the selective enrichment of glycoconjugates in more complex biological samples 106 and for the non‐covalent modification of hydrophobic, dodecanethiol‐stabilized AuNPs (synthesized by Brust—Schiffrin method 107 ) by multivalent glycocalixarenes with four mannose units to improve targeting efficiency toward intact cells. 108 Glycan parts of different enzymes frequently serve as models to study interactions with lectins. A practical application for the food industry and agriculture is an electrochemical device based on palladium NPs (PdNPs) as catalysts for the 3,3´,5,5´‐tetramethylbenzidine sulfate/H 2 O 2 system with immobilized mannose‐binding jacalin‐related lectin from rice ( Oryza sativa , a bioprobe) for the electrochemical detection of Magnaporthe oryzae (also M. grisea ) chitinase (a biochemical marker) during early rice infections. The use of a magnetic‐controllable electrode together with magnetic bead based PdNPs allows the detection of chitinase down to 17 pg mL −1 (approximately 420 fM). Moreover, using chronopotentiometry, it was possible to detect chitinase 2 days after rice infection, while a standard enzyme‐linked immunosorbent assay (ELISA) could detect the chitinase only 4 days after infection. 109 In addition to NPs and nanorods, nanoporous materials, for example, nanoporous gold in combination with square‐wave voltammetry (SWV) 110 and nanoporous gold monoliths in combination with thermogravimetric analysis, 111 were used to effectively and sensitively detect glycan—protein interactions, for example, to detect high‐mannose glycan‐containing ovalbumin molecules. 112 In situ glycosensing on the cell surface may play an important role in determining the physiological status of the whole cells in the biopsy samples acquired from patients suffering from various diseases. EIS was used to detect human colon cancer DLD‐1 cells with an LOD of 40 cells mL −1 by bovine serum albumin (BSA) incorporated Ag nanoflowers on a glassy carbon electrode (GCE) with a 3D porous architecture and a large surface area and the retention of immobilized cells activity after binding (attributed to the presence of BSA as a biocompatible support). 113 After the conjugation of the cells with the selected lectin (SNA in this case), the average number of sialic acid molecules on a single living cell was counted as approximately 2.16 × 10 12 . 113 Su et al. developed a novel lab‐on‐a‐paper device for the electrochemical sensing of K562 cancer cells with an LOD of 400 cells mL −1 and a wide linear range spanning 5 orders of magnitude based on a macroporous Au‐paper electrode. 114 Given that the volume used for incubation was as low as 10 μL of cell suspension, the LOD of the device was approximately 4 cells. An in situ monitoring of multiglycan expression in response to drug treatment was achieved using differential pulse voltammetry (DPV) and horseradish peroxidase (HRP) labeled wheat germ agglutinin (WGA), peanut agglutinin, Dolichos biflorus agglutinin, and Con A. A similar device designed by the same group later that year based on an aptamer modified 3D macroporous Au‐paper electrode was developed in a microfluidic format to screen anticancer drugs (Fig. 5 ). 115 This device could also detect as little as 350 cells mL −1 (very similar to a previous case, e.g., approximately four cells in 10 μL) using the biosensor signal generated using an HRP‐labeled annexin V bioprobe. This bioprobe specifically interacts with membrane phosphatidylserine molecules (in the presence of Ca 2+ ions), whose externalization cannot proceed in healthy and necrotic cells, thus providing a highly specific response toward apoptotic cells (translocation from the inner to the outer leaflet of the membrane is an important indicator of apoptosis). 115 Recent advances in electrochemical cytosensing amplified by nanostructures and nanocrystals were reviewed by Hasanzadeh et al. 116 Figure 5 Schematic presentation of a macroporous Au paper electrode (μPECD); (A) cells are incubated with HRP‐folic acid conjugate, (B) folding of μPECD and clamping the μPECD between circuit boards (front (C) and reverse side (D)) and finally the detection principle (E), where a is o‐ phenylenediamine and H 2 O 2 , b is 2,2´‐diaminoazobenzene and c is the final product of the reaction. The real size of the device is compared to a Chinese 1 yuan coin. Reprinted from 115 . Copyright 2014, with permission from Elsevier. No doubt the most commonly investigated and the most commercially successful biosensors are those for glucose detection based mostly on glucose oxidase (GOx) enzyme, despite the fact that even enzyme‐free nanostructured sensors have been described. 117 However, the synergistic effect of an NP catalyst with a conducting hydrogel heterostructure‐based interface 118 and the dependence of the sensor performance on NP shape were demonstrated only recently. 119 Li et al. not only described a simple glucose detection, but went even further with the imaging of intracellular glucose consumption in living cancer cells. 120 Their system is based on apo‐GOx (an inactive form of GOx) modified AuNPs and fluorescein isothiocyanate (FITC) dextran. In the presence of glucose in the environment, the quenched fluorescence of FITC‐dextran is recovered as glucose exhibits greater affinity than dextran to apo‐GOx . The LOD of 5 nM, along with the introduction of apo‐GOx instead of GOx, means no consumption of O 2 with subsequent H 2 O 2 production (causing cellular damage) and makes this assay a simple, sensitive, and "biofriendly" method for various disease diagnoses and metabolomics studies. 120 Even though glucose is an important molecule for diagnostics of different conditions and cell metabolism and its importance is highlighted in many sections throughout this review, glucose as a monosaccharide is not considered as a glycan and thus glucose‐based biosensors are not discussed here in more details. It is quite difficult to analyze electrochemically inactive glycans and oligo‐ and polysaccharides using electrochemical methods. 121 However, Paleček and his team recently published several papers about the deacetylation of N ‐acetylated glycans to make the –NH 2 groups free and thus electrochemically active 122 and about the modification of glycans with Os(VI)L complexes for subpicomolar detection. 123 Most recently, a paper about the label‐free electrochemical detection of interaction between Con A lectin and a glycoprotein on an atomically smooth mercury electrode was described. 124 1. Optical Biosensors Moreover, using precisely engineered inter‐NP spacers (e.g., polymer grafts), it was possible to control the degree of plasmonic coupling between NPs. 74 This principle was demonstrated with poly[(lactose) m ‐ b ‐(pyridine) n ] and bivalent galactose‐binding lectin from Ricinus communis (RCA 120 , M w = 120 kDa) on gold nanorods (AuNRs), where the authors claimed to develop a protein assay with an LOD for lectin down to pictogram per milliliter (fM range). 74 Other study 75 focused on the preparation of a sandwich configuration with graphene oxide (GO) and phenoxy‐derivatized dextran (conjugated together through π—π stacking interactions) deposited on gold substrate for specific Concanavalin A (Con A) binding with a final amplification of the signal using dextran‐capped AuNPs. This system was, however, less sensitive than a previous one, offering an LOD of 0.39 μg mL −1 (3.75 nM), but still with a 28.7‐fold increase in sensitivity compared to a direct assay (without any amplification reagent). 75 The application of different graphene‐based interfaces for the surface plasmon resonance (SPR) analysis of biospecific interactions between different ligand—receptor couples was recently reviewed. 76 Since the change in the localized SPR (LSPR) signal of metal NPs is not very sensitive and provides only a little information about the nature of interacting molecules, Craig et al. aimed to develop an on/off surface‐enhanced Raman scattering (SERS) aggregation system for picomolar Con A detection. The authors took advantage of the multivalent character of the interaction (since Con A is a homotetrameric molecule at a physiological pH with four glycan‐binding sites) with aggregated NPs producing a strong electromagnetic field between their interfaces, resulting in an increased SERS intensity. 77 The LOD for this system was as low as 40 pM for the chosen silver NPs (AgNPs) because they provided a tenfold increase in scattering efficiency compared to AuNPs. 77 Plasmonic and, in general, label‐free methods are of a particular interest for glycan‐lectin‐based biosensing because of a negative effect of lectin/glycan labeling on the biorecognition. In 2015, a novel approach called plasmon waveguide resonance was introduced based on glass prisms coated with 50 nm silver and 460 nm silica layers derivatized with mannose and lactose using Cu I ‐catalyzed Huisgen azide‐alkyne cycloaddition (CuAAC). 78 In addition to "click chemistry" based techniques, photocoupling reactions with underivatized glycans represent an alternative approach comparable to CuAAC. 79 AuNPs can be easily covalently attached on a plane gold surface via thiol linkers containing an –SH or –NH 2 head group for sensitive glycoprotein detection using immobilized lectin molecules, 80 providing an increased surface area for immobilization, as well as for the suppression of steric hindrance between lectin molecules because of the spherical shape of AuNPs. This effect proved to be so crucial that prepared impedimetric, a Sambucus nigra agglutinin (SNA) based biosensor, was more sensitive, with an LOD 3 orders of magnitude lower (shift from fM to aM range for the detection of sialylated glycoprotein fetuin) even though the total amount of lectin on the surface was lower in the NP‐based 3D configuration compared to the 2D (planar gold) configuration. 81 Most often, metal NPs are used as signal amplifiers. NPs decorated with saccharide structures ( d ‐maltoheptaose as the largest carbon source among maltodextrins that can be transported to Escherichia coli cytoplasm) promote particle internalization for silica, magnetic, silica‐coated MNPs and silica‐coated QDs. 82 Other optical detection platforms are based on colorimetric bioassays. The possibilities for the improvement of the selectivity and sensitivity of such assays for the detection of lectins, toxins, and viruses were recently reviewed. 83 Besides lectins, different toxins possess the ability to specifically bind various saccharide residues. Within the so‐called AB 5 bacterial toxin family (which also includes cholera toxin, shiga and shiga‐like toxins, and pertussis toxin), a protein is produced by enterotoxic E. coli causing "traveller´s diarrhoea": heat‐labile enterotoxin. Poonthiyil et al. prepared an efficient colorimetric sensor using 12 nm AuNPs with attached thiol‐modified galactose moieties binding to the B‐subunit of heat‐labile enterotoxin (while the A‐subunit is typically the one causing a particular disease among all members of this family) that is able to detect the toxin down to 100 nM. 84 Glycan‐decorated AuNPs may also be used for the rapid evaluation of viral hemagglutinin´s specificity because viral pathogens use lectins encoded by their own or host genome to replicate and spread 85 ; these AuNPs can also be used in antiviral drugs strategies. In general, avian‐adapted influenza viruses prefer α2,3‐linked sialic acid (ligands commonly found in the intestinal epithelia of birds), and human‐adapted viruses bind preferentially to α2,6‐linked sialic acid. Using host cell receptors as biorecognition elements immobilized on AuNPs, it was possible to easily and rapidly evaluate a potential threat to the human host without using complex immunoanalytical strategies. Since influenza hemagglutinin molecules contain several binding sites (Fig. 3 ), the aggregation of NPs occurred with a red‐to‐purple shift. 86 Figure 3 A hemagglutinin from H1N1 Influenza A virus (complexed to three molecules of 3´‐sialyllactose, denoted with yellow arrows), an interaction used to bound virus to glycan epitopes of host cells ( http://pdb.org , code 3HTO). This is an alternative to using highly specific monoclonal antibodies, where an LOD of 7.8 hemagglutinin units (i.e., 7.8×10 6 particles) was achieved recently for human H3N2 influenza A virus without any additional amplification. 87 By using the epitope glycan structures, however, knowing the exact influenza subtype is not required. Moreover, glycans‐based biosensors can benefit from the multivalent character of the protein—glycan bond. 88 , 89 , 90 For instance, NPs modified with thiolated monovalent and trivalent α2,6‐linked sialic acid and thiolated polyethyleneglycol (PEG) were successfully used to discriminate between human influenza virus X31 (H3N2) and avian RG14 (α2,3‐sialic acid binding H5N1 strain) virus with an LOD of 2.55 μg mL −1 . In addition, the trivalent configuration provided more rapid results and a greater sensitivity for plasmonic detection relying on glycoNPs' aggregation compared to a monovalent configuration. 91 As shown in this and other studies, when self‐assembled monolayers (SAMs) on gold surfaces are used, the concentration of a diluting thiol as well as its functional head group and length are crucial for assay performance. Finally, using glycan‐modified nanostructured surfaces, it is also possible to reliably detect any new and, for people, potentially pathogenic viral strains without needing any other reagents just by investigating the specificity of the viral strains toward the human glycan epitopes present in exposed epithelia. SAMs as a form of nanotechnology 4 allow the fine‐tuning of other surface properties, such as wettability and adsorption processes, 92 in addition to tuning interfacial chemical reactivity toward the subsequent covalent immobilization of biomolecules. 2. Electrochemical Biosensors Among others, electrochemical methods (e.g., increasingly popular electrochemical impedance spectroscopy [EIS], suitable for single‐molecule to whole‐cell detection) 93 are highly sensitive for providing an alternative platform for viral hemagglutinins or whole‐particle detection down to attomolar detection limits or single‐viral particles, respectively. 94 , 95 Tung et al. prepared a nanostructured (gold nanohemisphere‐modified) biosensor surface. 96 The impedimetric biosensor with enhanced sensitivity due to increased surface reaction area was used for the ultra‐weak binding of C‐type lectin domain family 5, member A and mosquito‐borne dengue virus particles, causing hemorrhagic fever and shock syndrome, with tens of millions infected people every year. 96 Another ultrasensitive attomolar detection of human H1 and avian H5 viral hemagglutinin was successfully performed on a field‐effect transistor (FET) based device (Fig. 4 ), modified with 3‐aminooxypropyltriethoxysilane (using standard silane coupling) and the simple glycan blotting of two different trisaccharide receptors (α2,3‐ and α2,6‐sialyllactose). 97 It is worth mentioning here that the above‐mentioned PEG molecules have become a gold standard for surface modification (even in a microarray format) 98 for the suppression of nonspecific interactions and for the stabilization of NPs when used for biosensing applications. 99 However, the introduction of glycan moieties, such as N ‐acetylglucosamine (GlcNAc) and lactose, on the surface of AuNPs and AuNRs was proposed to be an alternative to PEG stabilization, ensuring colloidal stability in protein‐rich media and preventing phagocytosis by macrophages, but at the same time exhibiting an excellent sensitivity toward carbohydrate‐binding proteins. 100 In addition to PEG, zwitterionic thiol derivatives (e.g., carboxy‐ or sulfobetaine derivatives in single‐component SAMs and mixed binary SAMs) 101 , 102 are promising ways to prepare nanostructured interfaces resisting nonspecific interactions. 103 Figure 4 A field‐effect transistor (FET) biosensor device for detection of viral hemagglutinins down to attomolar level using glycoblotting protocol; (A) typical scheme and picture of the described device and (B) modification scheme of FET device surface using silane chemistry. The slight difference in the glycan epitope (3´‐ vs. 6´‐sialyllactose) is enough to distinguish between Influenza viruses able to infect a human and other types of Influenza . Reprinted with permission from 97. Copyright 2013 American Chemical Society. The multivalent amplification effects of glycan—receptor interactions are naturally present in biological systems, for example, complexly branched glycans on glycoproteins or densely packed rafts of glycolipids. Therefore, synthetic multivalent saccharides (among others) have often been used to mimic nature. 104 Good examples are synthetic polymers, such as poly(acrylamidophenyl‐α‐mannose‐ co ‐acrylamide) attached to AuNPs, 105 layer‐by‐layer modifications of MNPs coated with polysaccharide shells of hyaluronan and chitosan. 106 These polymers were applied for the selective enrichment of glycoconjugates in more complex biological samples 106 and for the non‐covalent modification of hydrophobic, dodecanethiol‐stabilized AuNPs (synthesized by Brust—Schiffrin method 107 ) by multivalent glycocalixarenes with four mannose units to improve targeting efficiency toward intact cells. 108 Glycan parts of different enzymes frequently serve as models to study interactions with lectins. A practical application for the food industry and agriculture is an electrochemical device based on palladium NPs (PdNPs) as catalysts for the 3,3´,5,5´‐tetramethylbenzidine sulfate/H 2 O 2 system with immobilized mannose‐binding jacalin‐related lectin from rice ( Oryza sativa , a bioprobe) for the electrochemical detection of Magnaporthe oryzae (also M. grisea ) chitinase (a biochemical marker) during early rice infections. The use of a magnetic‐controllable electrode together with magnetic bead based PdNPs allows the detection of chitinase down to 17 pg mL −1 (approximately 420 fM). Moreover, using chronopotentiometry, it was possible to detect chitinase 2 days after rice infection, while a standard enzyme‐linked immunosorbent assay (ELISA) could detect the chitinase only 4 days after infection. 109 In addition to NPs and nanorods, nanoporous materials, for example, nanoporous gold in combination with square‐wave voltammetry (SWV) 110 and nanoporous gold monoliths in combination with thermogravimetric analysis, 111 were used to effectively and sensitively detect glycan—protein interactions, for example, to detect high‐mannose glycan‐containing ovalbumin molecules. 112 In situ glycosensing on the cell surface may play an important role in determining the physiological status of the whole cells in the biopsy samples acquired from patients suffering from various diseases. EIS was used to detect human colon cancer DLD‐1 cells with an LOD of 40 cells mL −1 by bovine serum albumin (BSA) incorporated Ag nanoflowers on a glassy carbon electrode (GCE) with a 3D porous architecture and a large surface area and the retention of immobilized cells activity after binding (attributed to the presence of BSA as a biocompatible support). 113 After the conjugation of the cells with the selected lectin (SNA in this case), the average number of sialic acid molecules on a single living cell was counted as approximately 2.16 × 10 12 . 113 Su et al. developed a novel lab‐on‐a‐paper device for the electrochemical sensing of K562 cancer cells with an LOD of 400 cells mL −1 and a wide linear range spanning 5 orders of magnitude based on a macroporous Au‐paper electrode. 114 Given that the volume used for incubation was as low as 10 μL of cell suspension, the LOD of the device was approximately 4 cells. An in situ monitoring of multiglycan expression in response to drug treatment was achieved using differential pulse voltammetry (DPV) and horseradish peroxidase (HRP) labeled wheat germ agglutinin (WGA), peanut agglutinin, Dolichos biflorus agglutinin, and Con A. A similar device designed by the same group later that year based on an aptamer modified 3D macroporous Au‐paper electrode was developed in a microfluidic format to screen anticancer drugs (Fig. 5 ). 115 This device could also detect as little as 350 cells mL −1 (very similar to a previous case, e.g., approximately four cells in 10 μL) using the biosensor signal generated using an HRP‐labeled annexin V bioprobe. This bioprobe specifically interacts with membrane phosphatidylserine molecules (in the presence of Ca 2+ ions), whose externalization cannot proceed in healthy and necrotic cells, thus providing a highly specific response toward apoptotic cells (translocation from the inner to the outer leaflet of the membrane is an important indicator of apoptosis). 115 Recent advances in electrochemical cytosensing amplified by nanostructures and nanocrystals were reviewed by Hasanzadeh et al. 116 Figure 5 Schematic presentation of a macroporous Au paper electrode (μPECD); (A) cells are incubated with HRP‐folic acid conjugate, (B) folding of μPECD and clamping the μPECD between circuit boards (front (C) and reverse side (D)) and finally the detection principle (E), where a is o‐ phenylenediamine and H 2 O 2 , b is 2,2´‐diaminoazobenzene and c is the final product of the reaction. The real size of the device is compared to a Chinese 1 yuan coin. Reprinted from 115 . Copyright 2014, with permission from Elsevier. No doubt the most commonly investigated and the most commercially successful biosensors are those for glucose detection based mostly on glucose oxidase (GOx) enzyme, despite the fact that even enzyme‐free nanostructured sensors have been described. 117 However, the synergistic effect of an NP catalyst with a conducting hydrogel heterostructure‐based interface 118 and the dependence of the sensor performance on NP shape were demonstrated only recently. 119 Li et al. not only described a simple glucose detection, but went even further with the imaging of intracellular glucose consumption in living cancer cells. 120 Their system is based on apo‐GOx (an inactive form of GOx) modified AuNPs and fluorescein isothiocyanate (FITC) dextran. In the presence of glucose in the environment, the quenched fluorescence of FITC‐dextran is recovered as glucose exhibits greater affinity than dextran to apo‐GOx . The LOD of 5 nM, along with the introduction of apo‐GOx instead of GOx, means no consumption of O 2 with subsequent H 2 O 2 production (causing cellular damage) and makes this assay a simple, sensitive, and "biofriendly" method for various disease diagnoses and metabolomics studies. 120 Even though glucose is an important molecule for diagnostics of different conditions and cell metabolism and its importance is highlighted in many sections throughout this review, glucose as a monosaccharide is not considered as a glycan and thus glucose‐based biosensors are not discussed here in more details. It is quite difficult to analyze electrochemically inactive glycans and oligo‐ and polysaccharides using electrochemical methods. 121 However, Paleček and his team recently published several papers about the deacetylation of N ‐acetylated glycans to make the –NH 2 groups free and thus electrochemically active 122 and about the modification of glycans with Os(VI)L complexes for subpicomolar detection. 123 Most recently, a paper about the label‐free electrochemical detection of interaction between Con A lectin and a glycoprotein on an atomically smooth mercury electrode was described. 124 B. Hybrid Nanomaterials and Nanocomposites Very often, not only single nanomaterials but also combinations of two materials where at least one is a nanomaterial are used for bio‐ and cytosensing. 125 AuNPs, forming novel nanocomposites with other materials, are often used as signal amplifiers. Poly(ethylenimine) (PEI)‐reduced GO (rGO) and hollow AuNPs deposited on GCE with immobilized GOx as recognition elements significantly improved the signal intensity of the luminol/H 2 O 2 electrochemiluminiscent (ECL) system for Con A detection down to 0.31 ng mL −1 (approximately 3 pM). 126 Chen et al. developed a sandwich electrochemiluminiscent biosensor with a Con A‐integrating AuNP‐modified Ru(bpy) 3 2+ ‐doped silica nanoprobe and a multiwalled CNTs (MWCNTs) modified electrode with another Con A on the surface (Fig. 6 ). 127 They were able to detect myelogenous leukemia K562 cells with an LOD of 600 cells mL −1 and, more importantly, to dynamically observe the cell surface glycoprofile during different phases of growth in vitro in response to external stimuli—glycan release by peptide‐ N ‐glycosidase (PNGase) F or incubation with the N‐ glycan inhibitor tunicamycin. 127 Figure 6 A schematic illustration of electrochemiluminiscent ECL biosensor for dynamic evaluation of cell surface N ‐glycan expression; (A) fabrication procedures of Con A‐modified NPs presenting the lectin molecules in a multivalent manner and (B) ECL biosensor for cytosensing and evaluating cell surface N ‐glycans, while the signal is reflecting the action of various inhibitors or glycosidases compared to untreated cells. Reprinted with permission from 127 . Copyright 2013 American Chemical Society. The same team later that year published another study describing competitive recognition and a signal amplification strategy using AuNPs modified with GOx. 128 They counted all of the mannose moieties on a single K562 cell (1.8 × 10 10 ) and again demonstrated the importance of the multivalent character of glycan—protein interactions, as the apparent dissociation constant between GOx‐Au and Con A nanoprobes was 1.64 nM—approximately 5 orders of magnitude lower than in the interaction of Con A with mannose. 128 The low LOD for K562 (50 cells mL −1 with a working volume of 200 μL and a linear response of up to 800 cells mL −1 ) was achieved using a graphene‐hemin‐AuNRs ternary composite as a peroxidase mimetic. 129 For influenza detection, a nanohybrid of the Pt NPs (PtNPs), porous ZnO spheres and hemin was synthesized for an amplified electrochemical immunosensor (Fig. 7 ). 130 Briefly, by the in situ generation of a redox probe by alkaline phosphatase (i.e., the release of 1‐naphthol from inactive 1‐naphthyl phosphate) and the excellent behavior of the Pt‐pZnO‐hemin nanocomposite applied as a signal enhancer, the influenza antigen was successfully detected on an antibody‐modified electrode in a sandwich configuration using DPV with an LOD of 0.76 pg mL −1 and with a linear range spanning 4 orders of magnitude. 130 Figure 7 (A) A scheme for preparation of a Pt‐pZnO‐hemin conjugate with a secondary antibody (green) and alkaline phosphatase (yellow). (B) A working principle of the proposed biosensor for influenza antigen detection using a primary antibody (purple) immobilized on an AuNPs modified electrode. Reprinted from 130 . Copyright 2016, with permission from Elsevier. More recently, He et al. introduced a novel sandwich strategy for a dual‐potential responsive, ECL biosensor for simultaneous cytosensing and surface N ‐glycan evaluation. 131 At a potential of 1.25 V, chemiluminescence was generated by Ru(phen) 3 2+ ECL probes intercalated in the grooves of double‐stranded DNA consisting of a DNA aptamer for MCF‐7 (breast cancer) cell recognition and a complementary capture DNA strand and immobilized on electrochemically reduced MoS 2 nanosheets. In the presence of cells, the capture DNA and the ECL probe were released from the electrode interface. The sandwich was then completed by a Con A‐conjugated AuNP‐modified graphite‐C 3 N 4 to detect cell surface mannose units at a negative potential of –1.6 V. 131 Zhang et al. published a paper on an ECL biosensor based on PEI‐rGO and hollow AuNPs. 126 The interaction between AuNPs and the—NH 2 groups of PEI was used for AuNP and GOx immobilization in this case, where GOx served as a producer of H 2 O 2 for the luminol/H 2 O 2 ECL system. In the presence of Con A, a decrease in the ECL intensity was observed, with an LOD down to 310 pg mL −1 (approximately 3 pM) and with a linear range from 1 to 20 ng mL −1 . The authors claimed that they developed an assay with a nearly 1000‐fold improved detection limit for Con A compared to previously published methods. 126 The same team prepared a similar device using a nanocomposite consisting of C 60 fullerene and rGO as a detection interface and hollow Au nanosphere‐conjugated GOx as a label. 132 The interaction of GOx with the electrode interface was mediated by phenoxy‐derivatized dextran, which served as a recognition element for Con A. Using a luminol/H 2 O 2 based ECL system, the LOD for Con A was estimated to be 30 pg mL −1 (approximately 288 fM) with a linear range spanning 3 orders of magnitude (from 0.1 to 100 ng mL −1 ). 132 Multivalent recognition and dual‐signal amplification strategies with Con A‐conjugated poly(amidoamine) (PAMAM) on a chemically rGO interface and HRP‐aptamer‐AuNPs nanoprobes were reported to detect CCRF‐CEM (human acute lymphoblastic leukemia) cells down to 10 cells mL −1 with excellent selectivity and could dynamically evaluate surface N ‐glycans. 133 Graphene could also be used as a support in combination with other metal NPs for biosensing applications. For instance, an rGO‐ and AgNP‐based nanocomposite was used as a redox probe together with phenoxy‐derivatized dextran and GOx as biorecognition elements for sensitive Con A detection on GCE. 134 Different electrochemical techniques, that is, cyclic voltammetry, DPV, and EIS, were used for signal generation with an LOD for Con A as low as 0.67 ng mL −1 (approximately 6.44 pM) and with a linear range from 2.0 to 322 ng mL −1 . Furthermore, the device was successfully used in diluted real human sera with recoveries from 92 to 108% and showed no major interference from BSA, cytochrome c, or phytohemagglutinin, suggesting possible applications for rapid and reliable clinical diagnostics. 134 Another example of the application of a nanocomposite for biosensing purposes is an rGO‐tetraethylene pentamine‐1‐butyl‐3‐methylimidazolium hexafluorophosphate hybrid composite. 135 A dense adsorption of bimetallic AuPtNPs, subsequently used for SNA lectin immobilization, was achieved by free –NH 2 groups from tetraethylene pentamine. This biosensor was used for the electrochemical detection of α2,6‐sialylated glycan down to 3 fg mL −1 and showed a wide linear range covering 8 orders of magnitude. As in a previous case, the recovery when analyzing real human sera was very similar, that is, with a range from 100.8 to 101.4% for Neu5Ac‐α2,6‐Gal‐β‐MP glycoside (4‐methoxyphenyl group via O ‐glycosidic linkage) spiked to a final concentration of 1 pg mL −1 to 100 ng mL −1 . 135 Because α2,6‐sialylated glycans might play an important role in clinical diagnostics using various biomarkers (e.g., on PSA), 136 new methods for their ultrasensitive detection are still emerging. 137 For example, the nanocomposite composed of graphite oxide, Prussian blue, and PTC‐NH 2 (ammonolysis product of 3,4,9,10‐perylenetetracarboxylic dianhydride) was used on a GCE to immobilize AuNPs through free –NH 2 groups and SNA‐I lectin for the DPV analysis of α2,6‐bound sialic acid on serum glycoproteins down to 0.03 pg mL −1 with a linear range spanning 5 orders of magnitude. 138 Although many different and successful strategies and protocols have been proposed relaying on different electroanalytical approaches for the analysis of complex glycan structures, the main challenge to be addressed for the application of affinity biosensors for real sample analysis is the efficient blocking of nonspecific interactions on the biorecognition interface, although BSA could be effectively applied as a blocking agent in some cases. Additionally, it is necessary to note that despite the fact that SNA lectin is routinely used to detect α2,6‐bound sialic acid, according to some specialized vendors, there is a minor α2,3‐sialic acid binding activity present as well. 139 C. Carbon Nanomaterials Engineered carbon nanomaterials (nanotubes, graphite, fullerenes, and graphene as main matrices for the conjugation of biomolecules) can be successfully applied for the preparation of biosensors. 140 There are different conjugation techniques for the functionalization of carbon structures with carbohydrates; the most common is the use of a carboxylic group formed on carbon surfaces using strong acidic oxidation with the subsequent conversion of –COOH groups to acyl chlorides, direct (carbodiimide activated) amidation, or ligation with an azide (Staudinger ligation). 141 Single‐walled CNTs (SWCNTs) were also functionalized through microwave‐assisted functionalization using perfluorophenyl azides with mannose and galactose. 142 This interface provided a reliable platform for agglomeration studies using FITC‐Con A lectin, specifically binding to α‐ d ‐mannopyranoside and, to a lower extent, to α‐ d ‐glucopyranoside residues, but not to Gal‐modified SWCNTs. 142 Through phenylacetylene‐SWCNTs and few‐layer graphene flakes, Ragoussi et al. prepared a carbohydrate‐modified (α‐ d ‐mannosyl glycodendron‐bearing) carbon nanostructures for Con A detection (using AFM, fluorescence, and UV/VIS studies), connected by means of CuAAC "click reaction" mechanism. 143 This group even managed to capture and observe the same object for their AFM study, leading to a reliable height profile analysis of the nanostructures before and after treatment. 143 Other detection platforms may be useful for studying glycan—lectin interactions using graphene‐modified surfaces. For SPR‐based experiments, graphene was grown through chemical vapor deposition (CVD) on polycrystalline Cu foils in a five‐step process on a 50‐nm‐thick Au film as single and double layers. 144 A simple immersion of this interface into a 100 nM mannose solution was sufficient for the mannose modification of the interface through the interaction of carbohydrate with the aromatic ring structure of graphene. Using mannose‐specific Lens culinaris agglutinin and GlcNAc‐ and sialic acid‐specific Triticum vulgaris agglutinin, it was shown that noncovalent surface modification by a simple mannose adsorption allowed the tuning of surface selectivity towards a specific receptor in a simple manner, allowing an LOD of approximately 1 μg mL −1 (low nM range) for lectins with a linear range of up to 1000 μg mL −1 . 144 However, electrochemical and ECL‐based devices are more sensitive than optical platforms. Since the Nobel Prize in physics in 2010, graphene has become an increasingly popular material and has recently been applied for biosensing purposes, as well as for glycomics, mainly in cytosensing applications using different detection strategies. As already discussed in Section B. .B regarding nanohybrids, the combination of different materials, the preparation of nanocomposites, and the very frequent utilization of a sandwich format analysis is common for biosensor construction. It has been previously reported that polymer dendrimers (e.g., PAMAM in this case) provide excellent support for the immobilization of glycans, allowing them to interact with proteins involving a multivalency effect. Molecular recognition using so‐called corona‐phase complexes consisting of synthetic polymers and CNTs, where the two components show affinity toward a selected analyte only if they stick together via surface forces stabilizing them and giving the polymer its final configuration (with a possibility to predict recognition specificity in advance), was also reported. 145 SPR gold surface coated with rGO (electrophoretically deposited) can be easily modified by a simple immersion in a solution of a particular polymer to prepare strongly negative (poly(sodium 4‐styrenesulfonate)) or positive (PEI) surfaces, as well as a surface modified by different saccharide moieties (mannose, lactose) through π–π stacking and electrostatic interactions. Subramanian et al. modified SPR chips with rGO to study the affinity of three different pathogenic E. coli strains to surfaces mediated by the presence of different adhesins on a bacterial cell membrane because those are responsible for the colonization of different epithelial structures and surfaces. 146 The modified SPR interface interacted strongly with highly pathogenic E. coli 107/86 strain in a quantitative manner with a linear response spanning 7 orders of magnitude and with an LOD of ∼100 cfu mL −1 (cfu—colony‐forming units) for bacterial strains 107/86 and UTI89. 146 ECL methods based on carbon nanomaterials could be used for cytosensing applications, as well. The simplest sandwich configuration, in which GO‐modified GCE served to immobilize the antibodies of interest (anti‐PSA in this case for the specific biorecognition of membrane PSA) and with a subsequent surface blocking by the use of BSA, was used to detect PC‐3 (prostate cancer) cells down to 260 cells mL −1 . 147 With a linear range spanning almost 2 orders of magnitude, ruthenium complex‐labeled WGA served as a signal probe. 147 A similar concept used for the impedimetric detection of HL‐60 (human promyelocytic leukemia) cells down to 500 cells mL −1 was based on a graphene surface modified by carboxymethyl chitosan. 148 This composite served to support the layer‐by‐layer assembly of PEI and folic acid for the fabrication of a label‐free cytosensor. Folic acid served as a biorecognition element because the overexpression of folate receptors often occurs in some tumor cell lines. 148 Graphene may be successfully utilized in various forms in biosensing technologies using not only plane graphene sheets but also monolithic and macroporous graphene foam. Such a 3D matrix (grown by CVD) was used to prepare an immunosensor for carcinoembryonic antigen (CEA, a tumor biomarker). 149 Briefly, a graphene substrate was used for the polymerization of dopamine, which subsequently served as a matrix for noncovalent Con A immobilization and interaction with HRP‐labeled anti‐CEA as a biorecognition element bound to Con A via a glycan part of HRP. After the surface was blocked by HRP, various electrochemical methods (mainly DPV using an electrochemical mediator) were used to detect CEA down to 90 pg mL −1 (approximately 500 fM), and the biosensor did not show any response toward other biomolecules, such as BSA, PSA, HRP, or glucose. Noncovalent graphene modification could be achieved not only by unmodified saccharides, 144 but as well as using "clickable" monosaccharide derivatives, such as azido galactosides immobilized on an alkynyl anthraquinone‐modified graphene electrode for the label‐free EIS detection of cancer cells. 150 D. Other Nanostructures, Glycopolymers, and Boronic Acid Derivatives Commonly used nanomaterials in glycomics in addition to metal and carbon nanostructures are glycopolymer‐based micelles, vesicles, or nonspherical NPs that are able to interact with lectins as multivalent ligands in a manner similar to natural glycoproteins. 48 Block copolymers often self‐assemble into diverse morphologies in solution depending on their properties, providing a promising bottom‐up engineering strategy for different applications of such nanostructures. 61 Polymer scaffolds may also be effectively glycosylated in vitro using a wide variety of available glycosyltransferases to prepare glycan structures mimicking those present in nature for the biorecognition of multivalent glycans by their specific lectin receptors. 151 Any information contained in the "sugar code" must be controlled with an extreme precision during glycopolymer preparation in laboratories for diagnostic and other purposes because every small difference in vivo may significantly affect a biorecognition event and lead to structural and functional abnormalities in the organism. Therefore, structural control of carbohydrate sequences during the synthesis of glycomimetics and multivalent glycopolymers is of highest importance to obtain reliable data. 152 , 153 Such synthetic glycopolymers are promising tools for use in emerging biomedical applications and research, including biosensing, biomolecular recognition, and vaccine development. 154 , 155 Multivalency and complexity of lectin—glycan interactions are applied in numerous processes in nature. For the study of such a complex combination of binding mechanisms in real time, dendrimers may serve as useful tools to evaluate the binding capacity of lectin receptors and the effect of avidity. Mannosylated gallic acid‐triethylene glycol‐based dendrimers in combination with SPR provided important structural data for studying biorecognition between Con A and mannose‐modified dendrimers. 156 An amphiphilic block copolymer consisting of hydrophilic lactose and hydrophobic pyridine was synthesized via reversible addition‐fragmentation chain transfer polymerization. 157 Glycosurface prepared on Au quartz crystal microbalance (QCM) chips was used to reliably detect RCA 120 in the nanomolar concentration range without any significant binding of BSA as a nonspecific probe. 157 Moreover, the authors in this study calculated the K A for the system, obtaining a value of 6.3 × 10 6 M −1 ; this value is normally in the range of 10 3 M −1 for monovalent lactose and its receptor. A similar value of K A (same order, 2.3 × 10 6 M −1 ) was obtained in another study using QCM and RCA 120 lectin binding to galactose‐containing gradient glycopolymer synthesized by RAFT polymerization. 158 By synchronizing enzymatic monomer transformation with polymerization, the authors obtained a gradient sugar distribution in a final amphiphilic polymer. 158 The lowest detectable concentration (5 μg mL −1 ) was again in the low nanomolar region. Superior lectin binding was achieved for the gradient polymer compared to the statistical glycopolymer, underlining the relevance of multivalency in the case of lectin—glycan interactions. 158 Glycoconjugated amphiphilic polymers can also be used for the encapsulation of fluorescent QDs. 159 Prior to its encapsulation, amphiphilic poly(isoprene)‐ b ‐poly(ethylene glycol) diblock copolymer was covalently modified by a carbohydrate moiety ( d ‐manno‐heptulose, d ‐glucose, d ‐galactose, bis(nitroso)‐streptozotocin, or d ‐maltose) using Huisgen‐type click chemistry, and interaction with Con A was studied again using the SPR method, showing enhanced affinity constants due to multivalent binding effects. 159 Supramolecular structures, which are of high importance in nanotechnology these days, may also be prepared by click chemistry reactions, as in the work published by Assali et al., in which the authors managed to synthesize poly(diacetylene)‐based nanomaterials with different morphologies. 160 Neoglycolipids with an amide bond between the hydrophilic and hydrophobic parts of the amphiphilic molecule formed 3D micelles, while triazole‐containing ones (obtained by "click‐reaction") allowed 1D nanotube formation. 160 Block glyco‐copolymers may also be used for cell imaging and as an effective drug delivery system (see Section 6. ). They may also enhance the uptake of drug‐loaded micelles by cells, as in the case of the increased uptake of doxorubicin‐loaded sugar (glucose or maltose as a biorecognizable hydrophilic block modification) and poly(4‐substituted‐ε‐caprolactone) copolymer micelles by HeLa cells, compared to free doxorubicin. 161 Since glycans are the most complex biomolecules, there is a need for high‐throughput methods for their analysis. In addition to commercial microarrays, a novel super‐microarray (containing many microarrays on the same slide) for lectin glycan sensing was recently developed. Such arrays use glycan‐labeled dye‐doped silica NPs (SiNPs) and a set of lectins immobilized on epoxy slides with poly(dimethylsiloxane) as an insulator, allowing the generation of many individual lectin microarrays, which significantly increase the assay throughput and, due to the multivalency of glycan‐modified NPs, also increase the affinity (over the free glycan and corresponding lectin) by 4–7 orders of magnitude. 162 Moreover, fluorescently labeled NPs offer higher stability and fluorescence compared to free organic dyes. Although glycans‐modified NPs have previously been prepared, the first attempt to prepare carbohydrate‐modified SiNPs was published recently by Ahire et al. 163 d ‐mannose‐capped SiNPs (prepared from amine‐terminated NPs using N,N´‐dicyclohexylcarbodiimide) were used to detect Con A when the interaction caused the aggregation of NPs. To show their biochemical activity, the photoluminescence of these NPs after interacting with MCF‐7 human breast cancer cells was also investigated. 163 For electrochemical analysis, conductive polymers are highly relevant for the enhanced sensitivity of detection when used as a solid‐state redox probe. Thiophene containing fused quinone moieties were electrochemically polymerized on a gold electrode surface to couple thiol‐modified mannose. 164 Such electropolymerization created a thin film on a solid surface with the ability to control its thickness very precisely up to several nanometers with subsequent application to construct microsensors. This new glycosurface allowed the detection of two major bacterial cell surface biomarkers—namely, fimbriae proteins on bacterial pili and lipopolysaccharides (LPSs) on G ‐ bacteria (by Con A‐mediated binding), using SWV and QCM methods down to 25 and 50 cells mL −1 , respectively. 164 Moreover, it was quite simple using this method to selectively distinguish between G‐ and G+ bacteria. 164 E. Synthetic Receptors for Glycosensing Common biorecognition elements for the sensitive detection of various analytes (biomolecules, viruses, or even bacteria) include antibodies and less common nucleic acid aptamers. In the past decade, carbohydrates have been increasingly studied due to their presence on the surfaces of proteins and cells. For the purpose of glycocode deciphering, lectins from various sources are commonly used. 22 Boronic acids also bind saccharides via reversible interactions, mostly with linear diols or even cis ‐1,2‐diols on five‐membered rings or 1,3‐diols to form five‐ or six‐membered rings. 165 , 166 Fluorescent diboronic acid compounds with dipeptide linkers were synthesized to discriminate cell‐surface Lewis X (Le x ) trisaccharide present on Chinese hamster ovary (CHO) CHOFUT4 cells at micromolar concentrations. 167 The control cells (without glycan expression, HEP3B cells predominantly expressing Le y , B16FUT3 cells expressing sialyl Lewis a (Le a ) and COLO205 cells expressing sLe x and sLe a but no Le x ) were not labeled, 167 suggesting the possibility of preparing compounds with a specificity toward glycans comparable to that of naturally occurring lectins. As previously mentioned ( Section 4.A.2 ), pathogenic agents, such as viruses and bacteria, use their envelope proteins (agglutinins) and adhesin lectins to recognize and attach themselves to host cells and tissues via glycans. This principle was used to prepare a novel electrochemical displacement sensor based on three different boronic acid derivative tracers (containing a ferrocene molecule). 168 The displacement of tracers by Con A lectin molecules or E. coli cells led to a decrease in the electrochemical signal monitored by SWV. Moreover, the use of thiolated mannose‐OEG conjugate ensured low nonspecific interactions. Con A could be detected with an LOD of 1 μg mL −1 (approximately 9.6 nM, with a linear range spanning ∼2 orders of magnitude), and E. coli cells could be counted down to 600 cells mL −1 . 168 The novel tracer used in this study, 2‐hydroxymethyl phenyl boronic acid derivative, binds to mannose even at a neutral pH, expanding the application of the system toward real biological samples (e.g., urine). 168 To date, many synthetically prepared "boronolectins" showed only a moderate fluorescence enhancement with a requirement of significant amount of co‐solvents in aqueous solution (i.e., dimethylsulfoxide and ethanol). A newly engineered boronolectin derived from tricarbocyanine combined with a boronic acid fragment linked by a piperazine unit exhibited improved certain properties, such as excellent water solubility and sensitive fluorogenicity, upon binding to carbohydrate moieties under a physiological pH. 169 To conclude, because boronic acid derivatives are able to successfully mimic lectins as natural glycan decipherers, they may be used not only to detect various analytes but also to selectively bind to free viral particles to inhibit their progression and surface adhesion, as in the case of lipid nanocapsules functionalized with amphiphilic boronic acid for hepatitis C virus inhibition, similar to cyanovirin‐N or griffithsin (both potent HIV inhibitors). 170 Their use as ultrasensitive solid‐phase microextraction probes for in vivo and in vitro sensing purposes in biofluids and even semisolid biotissues was also demonstrated. 171 F. Naked‐Eye Detection Using Nanostructures AuNPs modified with different saccharide moieties (lactose, arabinose, cellobiose, sucrose, mannose, glucose, and galactose) were applied by Jayawardena et al. to successfully distinguish among four different lectins with different specificities. 172 Con A, soybean agglutinin, Griffonia simplicifolia agglutinin, and Arachis hypogaea peanut agglutinin were detected by observing a red shift in the λ max of the LSPR absorption (LSPR on NPs, as opposed to propagating SPR biosensors). 172 Such a library‐oriented approach of glycan‐decorated NPs was later used to prepare polymer‐stabilized glyco‐AuNPs for a rapid, high‐throughput, and 96‐well microplate‐compatible evaluation and identification of pathogenic lectins without a need for any infrastructure because the output of these measurements (red‐to‐blue color shift upon AuNP aggregation) was monitored by a digital camera (Fig. 8 ). 173 Plasmonic metal NPs thus have great potential for their use in biosensor technology due to their sensitive spectral response to the local environment of NPs. 174 Figure 8 An overview of the colorimetric detection principle of lectin–glycan interactions with naked eye by glyco‐AuNPs aggregation due to lectin interaction. In case of an aggregation of NPs (in presence of lectin molecules), red‐to‐blue shift in color occurred. Reproduced from 173 , with permission of the Royal Society of Chemistry. SPR, however, lacks the higher throughput capability compared to lectin microarrays. This drawback was overcome recently 175 by establishing a lectin microarray based on a multiplexed SPR interface for the simultaneous measurement of up to 96 interactions by the immobilization of 18 different unmodified lectins (at different dilutions), including controls. A microarray GOAL (Glyco‐gold NP‐based Oriented immobilized Antibody microarray for Lectin) assay was also introduced as a novel approach for the naked‐eye detection of lectin‐carbohydrate interactions after silver enhancement using oriented, surface‐immobilized anti‐lectin antibodies. 176 Moreover, these modified AuNPs were highly stable and resistive to any nonspecific protein adsorption. 176 Human IgGs are extremely important markers of various diseases, which can be applied in a quantitative and qualitative manner because these glycoproteins are responsible for an effective immune response. The glycan part of human IgG was shown to be associated with autoimmune disease progression, mainly rheumatoid arthritis, 16 , 101 where the N ‐linked biantennary complex glycan in the Fc region is terminated with galactose or even GlcNAc, while in healthy individuals, IgG´s glycan can be terminated with sialic acid. 177 The GalNAc biosensor based on poly(diacetylene) nanovesicles developed by Hao et al. was applied for a noninvasive and real‐time colorimetric analysis of galactose‐deficient IgA1 (playing an important role in the pathogenesis of glomerulonephritis—IgA nephropathy) using nanovesicles modified with Helix aspersa agglutinin for naked‐eye detection. 178 In addition to glycoproteins, other glycoconjugates, such as glycolipids, were recently used for sensing applications. A fluorescent glycolipid monomer was synthesized using conjugation between 1‐pyreneboronic acid and a glycolipid based on a condensation reaction between d ‐glucose and oleic acid for the qualitative and chiral sensing of 80 nmol of amino acids ( l ‐ and d ‐tryptophan and phenylalanine) by the naked eye (Fig. 9 ). 179 In order to study carbohydrate–carbohydrate or carbohydrate–protein interactions using glycolipids, three different strategies could be utilized: (i) insertion of a synthetically prepared glycolipid into a lipid matrix, (ii) preparation of glycolipids that aggregate to form liposomes or micelles, and (iii) modification of a hydrophobic surface by a desired sugar derivative. 180 Figure 9 Photographs of the aqueous dispersions of the D‐vesicles (self‐assembled morphology of synthesized fluorescent glycolipid monomer c = 1.5 mmol, average of 48 nm in diameter) in the presence and absence of amino acids (80 nmol) for the naked‐eye detection. Reproduced from 179 with permission of the Royal Society of Chemistry. G. Quantum Dots (QDs) QDs have also attracted considerable attention in many different fields, including bioimaging and the detection of various analytes, mainly because of their tunable optical size‐dependent properties. 181 In recent years, the importance of detecting various forms of viruses has emerged with a focus on the identification of various glycoforms present on viral surfaces. A two‐step procedure was developed for virus detection, including the isolation of viral hemagglutinins by glycan‐modified paramagnetic beads, labeling hemagglutinins with CdS QDs with their subsequent electrochemical detection by voltammetry using 3D printed microfluidic chips. 182 The other detection principle employed by Chen and Neethirajan was based on a homogenous fluorescence quenching principle. 183 They used a sandwich configuration (antibody‐modified AuNPs and glycan‐conjugated QDs) to entrap influenza A hemagglutinins in between these two probes. As a result, a fluorescence decrease due to a nonradiative energy transfer between these two probes was observed. Of course, QDs could be conjugated with a diverse range of different structures in the same way as for the other above‐mentioned nanomaterials. The conjugation of CdSeTe@ZnS‐SiO 2 QDs modified with 3‐aminophenylboronic acid was used to monitor changes in the relative amount of sialic acid on K562 cell surfaces after a 3´‐azido‐3´‐deoxythymidine treatment, showing a significant increase in sialic acid expression. 183 Another paper 184 aimed to develop a photoelectrochemical biosensor using low‐toxic Ag 2 S QDs for glucose detection as well as for the detection of MCF‐7 breast cancer cells down to 32 μM and 98 cells mL −1 . 5. CARBOHYDRATE‐BASED VACCINES, ADJUVANTS, AND THERAPEUTICS With ongoing progress in glycomics, NPs displaying glycan moieties have been gradually recognized as potential therapeutic agents. There are numerous reviews mapping this broad research field. Hence, the commercialization and real application of developed glyco‐based vaccines are well documented. For example, in a recent study, 185 a list of glycoconjugates that were already approved or used in clinical trials as vaccines is provided, while other authors have focused on the application of nanomaterials in vaccine design. 186 , 187 A. Vaccines and Adjuvants When saccharide‐based vaccines are developed using nanomaterials, the latter components are primarily applied as glycan carriers, acting as effective immunogenic moieties. The first glycan‐based vaccine was Pneumo Vax produced by Merck (Darmstadt, Germany)—an unconjugated capsular polysaccharide isolated from Pneumonia serotypes. 188 Capsular antigenic glycans are in fact the only choice in the development of glycan‐based antibacterial vaccines, but the application potential of glycans given by their combinatoric diversity is very large. The immunogenic effect of these vaccines increases significantly by the increased valency of glycans employed, that is, typically by their conjugation to proteins, polymeric scaffolds, or other NPs. As already summarized, 188 carbohydrate‐based vaccines were developed against viruses, some prokaryotes and, quite recently but with ever‐growing interest and success, even against cancer cells. Although noncarbohydrate immunogens (i.e., protein‐based antibodies) have been more broadly used for vaccine development, carbohydrate immunogens are similarly important and useful. Recent achievements in the efficient conjugation of the latter immunogens with NPs are discussed in the following sections. In addition, recently reported studies where carbohydrates are used as building blocks (not immunogens) for the development of vaccines are covered as well. 1. Carbohydrate Immunogens Conjugated with Polymers In order to overcome the low immunogenicity of cancer cells, tumor‐specific or tumor‐associated antigens can be multiplied on the surface of NPs. The administration of such NPs helps the immune system to fight disease in a more efficient way. A multivalent display of antigens allows their better recognition by respective receptors with the subsequent induction of the immune response. It should be noted that natural antigenic carbohydrates (except for zwitterionic glycans) are believed to mostly lack the ability to be displayed on the B‐cell surface via the major histocompatibility complex, resulting in their incapability to activate T cells. This activation was, instead, achieved by displaying parts of carrier proteins co‐delivered in a vaccine. In the following sections, however, glycan‐nano vaccines without peptide adjuvants are also shown to be fabricable. Using "classical" adjuvants, the conjugation of immunogenic synthetic mucin (MUC1, a surface‐displayed glycopeptide typical for tumor epithelial cells) with different NPs was tested intensively. In a recent review paper, a study describing MUC1 conjugated with tetanus toxoid carrier protein was described. 189 This combination provided the best performance as judged from the highest production of specific IgGs (see Fig. 10 for the structure of the glycoprotein conjugate) among other reviewed MUC1 vaccines. 189 More recently, the same antigen was conjugated with a bacterial lipoprotein—a Toll‐like receptor ligand to boost the elicited immune reaction (Fig. 11 ). 190 The same composition (i.e., MUC1 antigen conjugated to a Toll‐like receptor) amended with a T‐helper‐cell epitope was employed by Abdel‐Aal et al. 191 Their conjugate, however, was not multivalent, but it was incorporated into small lamellar vesicles. These vesicles were used to immunize mice that were subsequently exposed to tumor induction. Animals that had been immunized with the best‐performing conjugate exhibited an approximately twofold smaller tumor after 2 weeks compared to other vaccines or a blank. 191 Another cancer antigen is the breast cancer cell‐specific hexasaccharide "Globo H," and its conjugation with diphtheria toxoid CRM 197 and α‐galactosylceramide C34 as an adjuvant provided the highest immune response amongst the conjugates tested using different carrier proteins and adjuvants. This effect observed was even higher than for a similar clinical trial phase III vaccine. 192 In order to synthesize well‐defined nanoconjugates, a method using the tyrosine‐specific binding of immunogenic glycans present on the surface of CRM 197 was developed 193 , 194 with consequent application in the development of a glycan‐based vaccine eliciting an immune response against Streptococcus infection. 195 Similarly, McCarthy et al. developed a chemoenzymatic synthesis of a well‐defined poly(sialic acid)‐tetanus toxoid glycoconjugate. 196 These studies have underlined how important it is to develop novel progressive methods for the precise and well‐defined synthesis of glycoconjugates. Figure 10 Synthesis and structure of a carrier protein‐mucin vaccine. Conjugation of a sialyl‐Tn MUC1 glycopeptide antigen (19) terminated via a coupling agent (diethyl squarate; 57) with a carrier protein (Tetanus toxoid or BSA) into a final multivalent carbohydrate‐bearing vaccination conjugate (58). Reproduced from 189 with permission of the Royal Society of Chemistry. Figure 11 A schematic structure of a vaccine consisting of 4‐valent glycopeptide MUC1 conjugated to a bacterial lipopeptide (Pam3CSK4), a ligand of Toll‐like receptor (TLR2) which helps to elicitate immune reaction. Reproduced from 190 with permission from John Wiley & Sons. In addition to discovering new possibilities in the synthesis of well‐defined glycoconjugates consisting of already known components, some studies, more biochemical than nanotechnological, have focused on immunogenic efficiency and the synthesis of new poly‐/oligo‐saccharides, mainly derivatives of bacterial cell wall epitopes. 197 , 198 , 199 , 200 Similarly, Johannes et al. investigated the increased immunogenicity of fluorinated analogues of tumor‐associated carbohydrate antigens conjugated to a support applicable as a vaccine against human breast cancer. 201 Alternatively, carbohydrates were used as building blocks or monomers for supramolecular scaffolds forming vaccine NPs. Since polysaccharides can form polyvalent ions, they can form self‐assembled nanoscaffolds upon mixing with relevant (polyvalent) counterions. Derivatives of chitin, the second most abundant polysaccharide on Earth and a very promising material in this field of research, can be used as very potent vaccine adjuvants. The mucoadhesion of chitin derivatives is helpful in vaccine administration and allows for an increased rate of chitosan‐based NP internalization via mannose‐binding receptors, inducing both the humoral and cellular immune response. 202 The latter was found to induce the degradation of the polysaccharide NPs, and chitosan dissolved in cytosol was found to be a more efficient immunogen than whole, nondestructed chitosan NPs. 203 The stability of prepared vaccine NPs is also an important issue. For example, a DNA‐based antitumor vaccine was effectively protected against degradation at a low pH when present in alginic acid based NPs. 204 Similarly, the coencapsulation of an antigen to be delivered with a commercially available adjuvant (C48/80) into chitosan NPs significantly decreased the dosage needed for the induction of a strong immune response compared to vaccination with a nonencapsulated antigen and an adjuvant. 205 , 206 A similar effect was observed for dendritic glucan, 207 and immunogenic properties have also been assigned to β‐glucans applied for the preparation of antifungal vaccines 208 and to polysaccharide based on δ‐inuline, which was utilized as an adjuvant. 209 Cholesteryl pullulan was used to entrap an additional immunogen—tumor necrosis factor α. 210 The nasal coadministration of such a prepared adjuvant combined with commercial anti‐influenza vaccine significantly elevated the resistance of mice treated against the influenza virus compared to control groups vaccinated without the adjuvant. 210 A better administration of tumor‐associated carbohydrate antigens was developed by Zhou et al., who investigated different derivatives of synthetic lipid A. 211 These compounds have been previously found as promising adjuvants, but a more efficient synthesis is needed for their wider utilization. 211 The efficacy of polysaccharide vaccines can be further improved by the decoration of NPs, with additional glycan moieties with mannose‐coated chitosan particles being a good example. 212 , 213 Another approach was presented by Fagan et al., who used tetra‐ O ‐acetyl‐α‐ d ‐glucopyranosyl bromide as a core structure for the synthesis of dendritic nanoscaffolds displaying multivalent immunogens—streptococcal B‐cell epitope. 214 It is also possible to prepare effective conjugates based on polymers and glycans where carbohydrate moieties are not primarily acting as immunogens but rather promote selective delivery and administration. A good example is mannosylated liposomes delivering albumin as a model antigen efficiently to dendritic cells responsible for the induction of humoral and immune responses. 215 The presence of mannose receptors on the surface of DC has also been widely used in drug delivery, imaging, and other biomedical applications that are discussed later. The overall therapeutic efficacy of mannosylated liposomes loaded with antitumor antigen and lipidic adjuvant was significantly boosted by the codelivery of small interfering RNA (siRNA) responsible for the downregulation of expression of immunosuppressive interleukins in tumor cells. 216 In another approach, Kim et al. developed fiber‐like supramolecular assemblies coated with mannose‐tethered lectin Con A. 217 Such protein coating was responsible for the immunogenicity of the fabricated conjugates and confirmed by observed interleukin production after the treatment of T cells with Con A‐coated NPs. 217 Carbohydrate immunogens playing a role as adjuvants, that is, helping to elicit a less‐specific immunoresponse, could also be delivered by diverse nanopharmaceutical systems. For example, a synthetic Mycobacterium tuberculosis epitope derivative (a fusion protein) and a glucopyranosyl lipid moiety were tested in liposomes, nanoemulsions, and adjuvants, with a promising immune response induction observed in mice with nanoemulsions selected as the best option for further vaccine approval and trials. 218 Controlled selective delivery was also achieved by the conjugation of a bioactive molecule (inhibitor of a transforming growth factor‐β receptor) with mannose‐6‐phosphate‐human serum albumin. 219 This carrier reacted selectively with receptors of hepatic stellate cells, while the conjugated inhibitor prevented transforming growth factor‐β‐induced activation, which is a key factor in the development of liver fibrosis. 219 2. Metal and Metal Oxide NPs in Recent Vaccine Development Due to their versatile and reproducible preparation and modification, AuNPs have attracted researchers' interest as promising therapeutic agents 220 , 221 with potential especially in vaccine development. 186 Although older studies have demonstrated that AuNPs decorated with analogues of viral or tumor cell polysaccharide epitopes were sufficiently immunogenic, 221 there is only one recent study describing vaccine design based on AuNPs coated with a carrier protein and an LPS from a nonvirulent bacterial strain Burkholderia thailandensis . 222 This glycoconjugate significantly increased the production of LPS‐specific antibodies in nonhuman primates exposed to the virulent bacterial strain B. mallei . In animals vaccinated with the AuNP‐LPS‐based conjugate no signs of bacteria were found, while in non‐vaccinated animals the pathogen cells were detected ‐ 102 (in animals that survived the test) and 104 (not survived animals) cfu per mg of tissue. Such results can be considered promising for the development of an efficient protective antiglanders (glanders = contagious and highly fatal disease, which can affect humans) vaccine for humans. 222 Parry et al. showed that an AuNP‐glycan‐based vaccine efficiently induced the immune response even in the absence of peptide or protein adjuvants. 223 These authors synthesized polymers displaying glycan units mimicking tumor tissue specific mucin (see Fig. 12 ). These immunogens were then conjugated in one step with in situ prepared AuNPs. Contrary to unmodified polymers, all glycan conjugates induced an immune response, as evidenced by IgG titres. The authors found out that the optimal number of glycan units per AuNPs‐conjugated polymer chain is 20–25, regardless of the chain length. 223 Figure 12 Preparation and characterization of a carbohydrate Tn‐antigen/AuNPs vaccine (Tn = N ‐α‐acetylgalactosamine linked to serine or threonine). Inset figures show representative dynamic light scattering data (top) and TEM image (bottom) of glycoNPs (scale bar = 20 nm). Reprinted with permission from 223 . Copyright 2013 American Chemical Society. Fallarini et al. investigated simple AuNP–glycan conjugates with the aim of understanding the essential components to design an effective vaccine. 225 In this study, they compared the effect of NPs decorated with nonimmunogenic mono‐ and disaccharides mimicking parts of the capsular polysaccharides of Neisseria meningitidis bacterium. The conjugate induced an immune cell response, contrary to nonconjugated forms of carbohydrates, and due to the possible intracellular degradation of the glycoconjugate, a disaccharide‐modified conjugate was more efficient. 225 Interesting behavior was observed for AuNPs grafted with fucose‐ended linkers. Such AuNP conjugates with a specific glycan surface density were internalized by dendritic cells via their Dendritic Cell‐Specific Intercellular adhesion molecule‐3‐Grabbing Non‐integrin (DC‐SIGN) receptors (see Fig. 13 ), but without an expected induction of subsequent interleukin production. Observed targeted internalization is a focus for antigen delivery and for the induction of a desired DC‐SIGN‐mediated signaling cascade. 224 Figure 13 A schematic depiction of fucosylated AuNPs interacting with dendritic cell‐specific intercellular adhesion molecule‐3‐grabbing nonintegrin (DC‐SIGN) receptors. Reprinted with permission from 224 . Copyright 2014 American Chemical Society. In order to induce an anticancer immune response, a tumor‐specific antigen (a phospholipid‐functionalized glycopeptide) was used for the formation of a scaffold with an iron oxide core. 226 The highest amount of elicited IgG (mean titres ∼81,402) was confirmed after the vaccination of mice by the conjugate having only one carbohydrate antigen per phospholipid‐modified glycopeptide chain while surprisingly, mean titres of only ∼7530 were observed when two glycosyl units were bound to the chain. It should also be noted that the formation of NPs with an iron oxide core significantly increased the IgG titres for free glycosylated phospholipid‐modified peptide chains from ∼5032 to ∼36,600. Furthermore, a complex induction of the immune response leading to tumor cell degradation was observed, 226 suggesting that iron oxide NPs can be an alternative to AuNPs for vaccine design. Carbohydrates were used also as adjuvants in AuNP‐based vaccines, with chitosan‐coated AuNPs decorated with plant saponins applied as tetanus toxoid carriers, but the authors did not describe any specific role of chitosan in the fabricated conjugates. 227 B. Nonimmunogenic Therapeutic GlycoNPs In addition to the activation of the immune system with a consequent therapeutic biochemical cascade, a few other therapeutic effects of glycoconjugates were discovered. Such efforts rely on (i) the competitive binding of lectins/glycans on receptors, thus preventing the successful binding of pathogenic viruses/bacteria, and (ii) the ability of glycoconjugates to selectively agglutinate pathogenic particles, thus eliminating their adverse effect. Since the interaction between a carbohydrate‐binding protein and a carbohydrate is not very strong, multivalent recognition entities must be present on a therapeutic NP surface. There have been numerous reports describing the fabrication of such therapeutics, with the main achievements comprehensively summarized in recent excellent reviews. 90 , 188 , 228 , 229 , 230 , 231 In summary, these papers referred to diverse kinds of multivalent glycoconjugates, that is, based on carbon and metallic NPs, organic supramolecular scaffolds or proteins with many of such conjugates having therapeutic potential. Moreover, novel synthetic and conjugation protocols have been described as well. Interesting conclusions are provided in a review by Jiménez Blanco et al., who suggested that heteroglycoconjugates, compared to homoglycoconjugates, provide additional regulation possibilities, in addition to the already known protein–carbohydrate binding mechanism. 232 These findings have been considered in recent developments in the design of nonimmunogenic therapeutic glycoconjugates. 1. Metallic NPs Few recent studies have focused on the antibacterial properties of AuNPs 233 or AgNPs 234 coated with 6‐ O ‐chitosan sulfate. Importantly, these glycoconjugates are not only antibacterial due to the presence of Ag or Au but also anticoagulant, making them almost perfect candidates for the surface coating of, for example, medical devices that need to be kept sterile. Similar, although offering only antibacterial properties, AgNPs and AuNPs have been conjugated with other polysaccharides, for example, starch 235 and aminocellulose. 236 A similar antibacterial effect has been observed with AgNPs coated with 12‐C‐monosaccharide‐dodecanoic acid 237 or with MNPs stabilized by poly(ethylene oxide) and functionalized with a sialic acid derivative. 238 While the aforementioned glycoNPs were tested mainly against E. coli , magnetic glycoNPs displaying fucose‐bearing oligosaccharides efficiently block the adhesion of Helicobacter pylori . 239 2. Carbon NPs Carbon NPs are another choice in the fabrication of glycoconjugates with antiadhesive properties. Recent studies include mannose‐bearing diamond NPs exhibiting an excellent inhibition of E. coli type 1 FimH‐mediated adhesion. 240 , 241 Ragoussi et al. have described similar mannose modification performed on CNTs and graphene sheets but provided only the results of a selective lectin binding test 242 ; the inhibition of bacterial adhesion by such NPs has yet to be assessed. On the other hand, Luczkowiak et al. reported an inhibition of Ebola pseudotype virus binding to DC‐SIGN receptor‐displaying cells. 243 In their study, fullerene with 12 mannoses (Fig. 14 ) was the most effective virus inhibitor, while increasing mannose valency led to decreased inhibition efficiency, 243 confirming a need for the controlled synthesis of conjugates with their extensive assessment. Figure 14 Structure of multivalent mannosylated fullerene used for inhibition of Ebola‐like virus binding. Reprinted with permission from 243 . Copyright 2013 American Chemical Society. 3. Synthetic Polymers Despite the relative ease of controlled chemical ways of glycosylation of carbon and metallic NPs, polymer scaffolds seem to have been explored the most intensively in this field. One reason may be a wider diversity of synthesized polymeric glycoconjugates available, as shown in the extensive study by Percec et al. 244 Recently reported examples of fully synthesized scaffolds include PAMAM dendrimers displaying carbohydrate Le x moieties capable of competitively blocking DC‐SIGN receptors, thus inhibiting the first phase of HIV infection. 245 More importantly, the glycodendrimers exhibited a negligible inhibition of langerin, which, contrary to DC‐SIGN, helped to internalize and destroy HIV particles. From the testing of a mini library comprising of third‐, fourth‐, and fifth‐generation glycodendrimers, the latter was found to be approximately twice as efficient in inhibiting the binding of model virus‐like particles compared to the other ones (third and fourth generation). Finally, the conjugate was tested for the inhibition of the real virus influencing DC‐SIGN‐mediated binding and trans‐infection. 245 Fucosylated, branched phosphodiester‐based scaffolds were reported to bind efficiently to Burkholderia ambifaria lectin with a promising application as inhibitors of bacterial infection. 246 Cholera toxin has been successfully ligated with newly synthesized five‐armed molecular scaffolds, bearing a galactose‐oligomer recognition unit on the end of each arm. 247 Pseudomonas aeruginosa lectin A recognition was suppressed by α‐ l ‐fucoside and by a β‐ d ‐galacto‐pyranoside dendrimer hybrid, which may be further exploited for the treatment of Pseudomonas aeruginosa ‐based infections 248 , 249 (for Pseudomonas aeruginosa lectin A‐glycan binding AFM images, see 250 ). Yan et al. investigated the influence of multivalency of n ‐heptyl‐α‐ d ‐mannose on the strength of binding to FimH adhesin displayed on type 1 fimbriae of E. coli causing Crohn's disease. 251 They found that the linear polymer decorated with numerous n ‐heptyl‐α‐ d ‐mannose moieties was more efficient in the coagulation of bacteria (Fig. 15 ) compared to monovalent glycoconjugates or polyvalent star‐like particles. 251 The potential of multivalent synthetic glycodendrimers was further demonstrated by Ghirardello et al., who reported a 109×106‐fold stronger binding potential (relative potency) to WGA of a 48‐valent GlcNAc dendrimer compared to an unconjugated GlcNAc moiety. 252 A relative potency of only 1168 was reported for a 30‐valent N ‐acetyl‐ d ‐lactosamine (LacNAc) "onion peel" like dendrimer compared to the LacNAc monomer, binding specifically to leguminous lectin from Erythrina cristagalli . 253 Figure 15 Glycopolymer‐induced agglutination of type 1 fimbriaeted E. coli (strain UTI89). (A) Fluorescence microscopy pictures of Katushka‐expressing type 1 fimbriaeted E. coli UTI89; (B) exposed to 1 μM of L188; (C) to 3 μM of L188 and resulting in a "bacterial egg" agglomerate with approximate dimensions of 98 × 51 μm. Reprinted with permission from 251 . Copyright 2015 American Chemical Society. In order to protect against eukaryotic parasites, Campo et al. synthesized cyclic triazole‐linked oligomers with pseudo‐glycosidic units with an affinity toward the trans‐sialidase enzyme from Trypanosoma cruzi . 254 As a result, the inhibition of the parasite development after their invasion into macrophage cells was observed, but a quite high concentration (250 μM) of the therapeutic oligomer was needed. Nevertheless, these results are only preliminary, and further optimization of oligomer formation (i.e., a pseudo‐glycosidic unit density and its precise location) is needed to increase the inhibition activity. 254 4. Biopolymers and Proteins Galactose units displayed on a branched oligopeptide were efficiently bound to the surface lectin of P. aeruginosa , inhibiting film formation and disrupting existing biofilms. 255 Bouckaert et al. explored cyclodextran NPs conjugated with n ‐heptyl‐α‐ d ‐mannose using various spacers as potential FimH agonists. 256 They did not observe improved adhesion to the FimH adhesin of E. coli ; however, their results suggest that differences in the length of the spacer arm may account for the low undesired affinity for human mannose‐binding lectins. 256 A further step in the development of antivirotics and antibacterial drugs that are not based on recently used synthetic antibiotics, to which an increased number of bugs are resistant, was described by the preparation of a cyclopeptide‐based dendrimer decorated by glycan moieties using an oxime ligation. 257 The best‐performing, 64‐valent glycodendrimer from the reported library of compounds (Fig. 16 ) exhibited a 40,000‐fold increased lectin‐binding potency compared to monovalent methyl α‐ l ‐fucopyranoside. 257 Cholera toxin inhibitor with an IC 50 value of 100 pM based on a tetravalent neoglycoprotein prepared by a simple chemical modification of an inactive B subunit of cholera toxin was reported by Branson et al. 258 The use of cholera toxin protein as a template secured a "… precise fit of the ligand groups with the spacing and configuration of binding sites on wild‐type CTB." 258 Figure 16 Synthesis and structure of cyclopeptide‐based glycodendrimers. Reproduced from 257 with permission of the Royal Society of Chemistry. Self‐assembled polymeric particles are the third type of organic nanocarriers covered in this section. For example, vesicles self‐assembled from poly(ethylene oxide) and polycaprolactone and further modified with a sialodendrimer block the cellular recognition of influenza virus via ligation with viral hemagglutinins. 259 Authors have reported a decrease in IC 50 from 4 mM to 240 μM (per glycan unit) when mono‐ and eight‐valent glycodendrimers were tested, respectively, and a further decrease to approximately 2 μM when the glycodendrimers were conjugated to the polymer vesicles. 259 Another type of conjugate was prepared in the form of nanodroplets by crosslinking polymer shells that were ready for subsequent covalent modification. 251 Yan et al. decorated such particles with multiple n ‐heptyl‐α‐ d ‐mannose moieties to achieve the efficient coagulation of pathogenic E. coli type 1. 251 E. coli 's type 1 fimbrial adhesin (FimH) was also targeted by Wu at al. using mannose‐terminated DNA oligomers. 260 Complementary DNA strands conjugated with a second‐generation dendrimer were then applied to assemble mannosylated fibers with the capability of agglutinating E. coli strain ORN178. 260 Interestingly, the authors claimed that agglutination ability was independent of mannosyl density on the fiber surface, most likely because of flexible tube‐like NPs. 260 Yu et al. prepared nanotubes that were self‐assembled from amphiphilic pillar[5]arene, with glucose acting as the hydrophilic part. 261 These glyco‐nanotubes were able to agglutinate E. coli cells more effectively with an increased number of glucose units with the highest agglutination index of 54 (an average number of bacteria connected to each other after successful agglutination). 261 Therapeutic potential also can be predicted for self‐assembled oligopeptide‐LacNAc conjugates capable of moderating the activity of galectin, a lectin‐type protein with signaling and other properties. 262 5. GlycoNPs for Enzyme Inhibition and Other Therapeutic Functions Applications of iminosugars for the treatment of lysosome storage diseases (e.g., Gaucher disease or glycosphingolipid lysosomal storage disorder, characterized by a mutation‐based impairment of different glycosidases with a subsequent pathological accumulation of glycolipids) have been investigated in recent years. Iminosugars are selective and reversible inhibitors of glycosidases due to selective binding to their active sites. This binding can help to fold impaired enzymes, thus partially restoring their hydrolytic activity at the higher pH present in lysosomes. Furthermore, the misfolded enzymes "protected" by the bound iminosugar can avoid degradation otherwise induced by a cellular quality control. 263 Using this mechanism, iminosugars could be applied for so‐called "pharmacological chaperone therapy," an emerging way to treat lysosomal storage diseases. The nature of recognition between iminosugar and respective glycosidase differs significantly from that of lectin–glycan binding. This fact raises the question of whether multivalency can increase the inhibition potency of investigated glycomimetics. In fact, previous studies have suggested that multivalent iminosugars lost their inhibition ability, while recent studies using well‐defined and controlled synthesis have reported the opposite. In the study of Garcia Fernandez et al., a mini library of iminosugar‐based glycomimetics (either monovalent or conjugated to fullerenes) was tested for its potency to inhibit a variety of glycosidases. 264 Their results suggested that in some cases, multivalency can "switch on" the inhibition by switching the binding mode, rather than by the sheer presence of a higher amount of recognition moieties. 264 In a more recent work, several scaffolds and conjugates displaying deoxynojirimycin, a broad glycosidase inhibitor, were tested. 265 Interestingly, the strongest inhibitor (800‐fold stronger than a monovalent analogue) was just a tetravalent iminosugar present on a porphyrin‐based scaffold. The authors suggested that the valency itself did not cause an increased inhibition potential, but rather the spatial conformation of monosaccharides did. 265 Moreno‐Clavio et al. tested several l ‐fucosidase iminosugar‐based inhibitors with less success. 266 Only a sevenfold‐higher inhibition effect on α‐ l ‐fucosidases was observed when a trivalent scaffold was used compared to a monovalent one. 266 On the other side, a relative α‐mannosidase inhibition potency as high as 3000 was reported for the N ‐alkyl analogue of 1‐deoxynojirimycin conjugated on a synthetic glycoprotein scaffold (Fig. 17 ), 267 almost 10,000 for the 21‐valent 1‐deoxynojirimycin scaffold 268 and higher than 500 for glycocyloclopeptide with seven copies of 1‐deoxynojirimycin. 269 In the latter case, the large inhibition potential was obtained only on glycoconjugates with a longer (C 9 ) spacer, while C 6 spacers of the same glycoconjugate resulted only in a relative inhibition potency of 42. 269 Figure 17 1‐Deoxynojirimycin‐based glycopolypeptides: synthesis, self‐assembly, and glycosidase inhibition model. CuAAC = Cu I catalyzed Huisgen azide‐alkyne cycloaddition. Reproduced from 267 with permission of the Royal Society of Chemistry. The real therapeutic effect of different 3‐ to 14‐valent iminosugar conjugates was investigated by Joosten et al. through the in vitro measurement of β‐glucocerebrosidase activities in N370S Gaucher fibroblasts. 270 The authors reported a "mild, but significant" effect of multivalency. Importantly, the highest inhibition rate was not correlated with the highest chaperoning (therapeutic) effect. 270 This report clearly suggests that similar systematic investigations will be necessary to develop efficient NP‐based drugs against lysosome storage diseases. A similar approach applied for the treatment of lysosome storage diseases was also investigated for the treatment of cystic fibrosis. 1‐Deoxynojirimycin and its derivatives correct a misfolded cystic fibrosis transmembrane conductance regulator protein in a similar manner as the aforementioned chaperone effect of iminosugars towards glycosidases, with a 1000‐fold increased inhibition activity of 3‐valent iminosugar compared to a monovalent analogue control. 271 The concept of pharmacological chaperones may sound very promising, but although the activation of glycosidases has been observed, it hardly reached therapeutically sufficient levels, with typically an approximately twofold increase in the activity. 272 From this point of view, the work of Brissonnet et al. is noteworthy. 273 These authors reported an activity increase as high as 70‐fold of a bacterial mannoside‐phosphorylase using cyclodextran‐based NPs displaying deoxymannojirimycin (more than 100‐valent). Nevertheless, further studies are needed to elucidate the nature of the enzyme activation gain. 273 The selective interaction of a disaccharide called Thomas Friedrich antigen (TF ag ) with specific tumor cells displaying Gal‐3 lectin had a cytotoxic effect on the targeted cells: a conjugate of small AuPs coated with TF ag via an amino acid linker was synthesized, and an approximately 100‐fold higher cytotoxicity toward the Gal‐3‐positive cells was achieved compared to monomeric units. 274 The main cytotoxic effect of the TF ag relied on apoptosis induced by an inhibited Gal‐3 signaling, 274 but it is anticipated that the TF ag targeting may also work well in drug delivery. Another glycan with a therapeutic effect is heparan sulfate, a glycosaminoglycan‐based polysaccharide capable of inhibiting protease β‐secretase responsible for the accumulation of plaques causing Alzheimer's disease. Nevertheless, carbohydrate synthesis is rather complicated and expensive. Therefore, a mini library of dendritic polymers bearing multiple copies of different heparan sulfate monomers was recently introduced. 275 Some of these conjugates were found to possess inhibition potential equivalent to that of previously tested oligomers, which is an important result for the future development and production of anti‐Alzheimer's disease drugs. 275 A. Vaccines and Adjuvants When saccharide‐based vaccines are developed using nanomaterials, the latter components are primarily applied as glycan carriers, acting as effective immunogenic moieties. The first glycan‐based vaccine was Pneumo Vax produced by Merck (Darmstadt, Germany)—an unconjugated capsular polysaccharide isolated from Pneumonia serotypes. 188 Capsular antigenic glycans are in fact the only choice in the development of glycan‐based antibacterial vaccines, but the application potential of glycans given by their combinatoric diversity is very large. The immunogenic effect of these vaccines increases significantly by the increased valency of glycans employed, that is, typically by their conjugation to proteins, polymeric scaffolds, or other NPs. As already summarized, 188 carbohydrate‐based vaccines were developed against viruses, some prokaryotes and, quite recently but with ever‐growing interest and success, even against cancer cells. Although noncarbohydrate immunogens (i.e., protein‐based antibodies) have been more broadly used for vaccine development, carbohydrate immunogens are similarly important and useful. Recent achievements in the efficient conjugation of the latter immunogens with NPs are discussed in the following sections. In addition, recently reported studies where carbohydrates are used as building blocks (not immunogens) for the development of vaccines are covered as well. 1. Carbohydrate Immunogens Conjugated with Polymers In order to overcome the low immunogenicity of cancer cells, tumor‐specific or tumor‐associated antigens can be multiplied on the surface of NPs. The administration of such NPs helps the immune system to fight disease in a more efficient way. A multivalent display of antigens allows their better recognition by respective receptors with the subsequent induction of the immune response. It should be noted that natural antigenic carbohydrates (except for zwitterionic glycans) are believed to mostly lack the ability to be displayed on the B‐cell surface via the major histocompatibility complex, resulting in their incapability to activate T cells. This activation was, instead, achieved by displaying parts of carrier proteins co‐delivered in a vaccine. In the following sections, however, glycan‐nano vaccines without peptide adjuvants are also shown to be fabricable. Using "classical" adjuvants, the conjugation of immunogenic synthetic mucin (MUC1, a surface‐displayed glycopeptide typical for tumor epithelial cells) with different NPs was tested intensively. In a recent review paper, a study describing MUC1 conjugated with tetanus toxoid carrier protein was described. 189 This combination provided the best performance as judged from the highest production of specific IgGs (see Fig. 10 for the structure of the glycoprotein conjugate) among other reviewed MUC1 vaccines. 189 More recently, the same antigen was conjugated with a bacterial lipoprotein—a Toll‐like receptor ligand to boost the elicited immune reaction (Fig. 11 ). 190 The same composition (i.e., MUC1 antigen conjugated to a Toll‐like receptor) amended with a T‐helper‐cell epitope was employed by Abdel‐Aal et al. 191 Their conjugate, however, was not multivalent, but it was incorporated into small lamellar vesicles. These vesicles were used to immunize mice that were subsequently exposed to tumor induction. Animals that had been immunized with the best‐performing conjugate exhibited an approximately twofold smaller tumor after 2 weeks compared to other vaccines or a blank. 191 Another cancer antigen is the breast cancer cell‐specific hexasaccharide "Globo H," and its conjugation with diphtheria toxoid CRM 197 and α‐galactosylceramide C34 as an adjuvant provided the highest immune response amongst the conjugates tested using different carrier proteins and adjuvants. This effect observed was even higher than for a similar clinical trial phase III vaccine. 192 In order to synthesize well‐defined nanoconjugates, a method using the tyrosine‐specific binding of immunogenic glycans present on the surface of CRM 197 was developed 193 , 194 with consequent application in the development of a glycan‐based vaccine eliciting an immune response against Streptococcus infection. 195 Similarly, McCarthy et al. developed a chemoenzymatic synthesis of a well‐defined poly(sialic acid)‐tetanus toxoid glycoconjugate. 196 These studies have underlined how important it is to develop novel progressive methods for the precise and well‐defined synthesis of glycoconjugates. Figure 10 Synthesis and structure of a carrier protein‐mucin vaccine. Conjugation of a sialyl‐Tn MUC1 glycopeptide antigen (19) terminated via a coupling agent (diethyl squarate; 57) with a carrier protein (Tetanus toxoid or BSA) into a final multivalent carbohydrate‐bearing vaccination conjugate (58). Reproduced from 189 with permission of the Royal Society of Chemistry. Figure 11 A schematic structure of a vaccine consisting of 4‐valent glycopeptide MUC1 conjugated to a bacterial lipopeptide (Pam3CSK4), a ligand of Toll‐like receptor (TLR2) which helps to elicitate immune reaction. Reproduced from 190 with permission from John Wiley & Sons. In addition to discovering new possibilities in the synthesis of well‐defined glycoconjugates consisting of already known components, some studies, more biochemical than nanotechnological, have focused on immunogenic efficiency and the synthesis of new poly‐/oligo‐saccharides, mainly derivatives of bacterial cell wall epitopes. 197 , 198 , 199 , 200 Similarly, Johannes et al. investigated the increased immunogenicity of fluorinated analogues of tumor‐associated carbohydrate antigens conjugated to a support applicable as a vaccine against human breast cancer. 201 Alternatively, carbohydrates were used as building blocks or monomers for supramolecular scaffolds forming vaccine NPs. Since polysaccharides can form polyvalent ions, they can form self‐assembled nanoscaffolds upon mixing with relevant (polyvalent) counterions. Derivatives of chitin, the second most abundant polysaccharide on Earth and a very promising material in this field of research, can be used as very potent vaccine adjuvants. The mucoadhesion of chitin derivatives is helpful in vaccine administration and allows for an increased rate of chitosan‐based NP internalization via mannose‐binding receptors, inducing both the humoral and cellular immune response. 202 The latter was found to induce the degradation of the polysaccharide NPs, and chitosan dissolved in cytosol was found to be a more efficient immunogen than whole, nondestructed chitosan NPs. 203 The stability of prepared vaccine NPs is also an important issue. For example, a DNA‐based antitumor vaccine was effectively protected against degradation at a low pH when present in alginic acid based NPs. 204 Similarly, the coencapsulation of an antigen to be delivered with a commercially available adjuvant (C48/80) into chitosan NPs significantly decreased the dosage needed for the induction of a strong immune response compared to vaccination with a nonencapsulated antigen and an adjuvant. 205 , 206 A similar effect was observed for dendritic glucan, 207 and immunogenic properties have also been assigned to β‐glucans applied for the preparation of antifungal vaccines 208 and to polysaccharide based on δ‐inuline, which was utilized as an adjuvant. 209 Cholesteryl pullulan was used to entrap an additional immunogen—tumor necrosis factor α. 210 The nasal coadministration of such a prepared adjuvant combined with commercial anti‐influenza vaccine significantly elevated the resistance of mice treated against the influenza virus compared to control groups vaccinated without the adjuvant. 210 A better administration of tumor‐associated carbohydrate antigens was developed by Zhou et al., who investigated different derivatives of synthetic lipid A. 211 These compounds have been previously found as promising adjuvants, but a more efficient synthesis is needed for their wider utilization. 211 The efficacy of polysaccharide vaccines can be further improved by the decoration of NPs, with additional glycan moieties with mannose‐coated chitosan particles being a good example. 212 , 213 Another approach was presented by Fagan et al., who used tetra‐ O ‐acetyl‐α‐ d ‐glucopyranosyl bromide as a core structure for the synthesis of dendritic nanoscaffolds displaying multivalent immunogens—streptococcal B‐cell epitope. 214 It is also possible to prepare effective conjugates based on polymers and glycans where carbohydrate moieties are not primarily acting as immunogens but rather promote selective delivery and administration. A good example is mannosylated liposomes delivering albumin as a model antigen efficiently to dendritic cells responsible for the induction of humoral and immune responses. 215 The presence of mannose receptors on the surface of DC has also been widely used in drug delivery, imaging, and other biomedical applications that are discussed later. The overall therapeutic efficacy of mannosylated liposomes loaded with antitumor antigen and lipidic adjuvant was significantly boosted by the codelivery of small interfering RNA (siRNA) responsible for the downregulation of expression of immunosuppressive interleukins in tumor cells. 216 In another approach, Kim et al. developed fiber‐like supramolecular assemblies coated with mannose‐tethered lectin Con A. 217 Such protein coating was responsible for the immunogenicity of the fabricated conjugates and confirmed by observed interleukin production after the treatment of T cells with Con A‐coated NPs. 217 Carbohydrate immunogens playing a role as adjuvants, that is, helping to elicit a less‐specific immunoresponse, could also be delivered by diverse nanopharmaceutical systems. For example, a synthetic Mycobacterium tuberculosis epitope derivative (a fusion protein) and a glucopyranosyl lipid moiety were tested in liposomes, nanoemulsions, and adjuvants, with a promising immune response induction observed in mice with nanoemulsions selected as the best option for further vaccine approval and trials. 218 Controlled selective delivery was also achieved by the conjugation of a bioactive molecule (inhibitor of a transforming growth factor‐β receptor) with mannose‐6‐phosphate‐human serum albumin. 219 This carrier reacted selectively with receptors of hepatic stellate cells, while the conjugated inhibitor prevented transforming growth factor‐β‐induced activation, which is a key factor in the development of liver fibrosis. 219 2. Metal and Metal Oxide NPs in Recent Vaccine Development Due to their versatile and reproducible preparation and modification, AuNPs have attracted researchers' interest as promising therapeutic agents 220 , 221 with potential especially in vaccine development. 186 Although older studies have demonstrated that AuNPs decorated with analogues of viral or tumor cell polysaccharide epitopes were sufficiently immunogenic, 221 there is only one recent study describing vaccine design based on AuNPs coated with a carrier protein and an LPS from a nonvirulent bacterial strain Burkholderia thailandensis . 222 This glycoconjugate significantly increased the production of LPS‐specific antibodies in nonhuman primates exposed to the virulent bacterial strain B. mallei . In animals vaccinated with the AuNP‐LPS‐based conjugate no signs of bacteria were found, while in non‐vaccinated animals the pathogen cells were detected ‐ 102 (in animals that survived the test) and 104 (not survived animals) cfu per mg of tissue. Such results can be considered promising for the development of an efficient protective antiglanders (glanders = contagious and highly fatal disease, which can affect humans) vaccine for humans. 222 Parry et al. showed that an AuNP‐glycan‐based vaccine efficiently induced the immune response even in the absence of peptide or protein adjuvants. 223 These authors synthesized polymers displaying glycan units mimicking tumor tissue specific mucin (see Fig. 12 ). These immunogens were then conjugated in one step with in situ prepared AuNPs. Contrary to unmodified polymers, all glycan conjugates induced an immune response, as evidenced by IgG titres. The authors found out that the optimal number of glycan units per AuNPs‐conjugated polymer chain is 20–25, regardless of the chain length. 223 Figure 12 Preparation and characterization of a carbohydrate Tn‐antigen/AuNPs vaccine (Tn = N ‐α‐acetylgalactosamine linked to serine or threonine). Inset figures show representative dynamic light scattering data (top) and TEM image (bottom) of glycoNPs (scale bar = 20 nm). Reprinted with permission from 223 . Copyright 2013 American Chemical Society. Fallarini et al. investigated simple AuNP–glycan conjugates with the aim of understanding the essential components to design an effective vaccine. 225 In this study, they compared the effect of NPs decorated with nonimmunogenic mono‐ and disaccharides mimicking parts of the capsular polysaccharides of Neisseria meningitidis bacterium. The conjugate induced an immune cell response, contrary to nonconjugated forms of carbohydrates, and due to the possible intracellular degradation of the glycoconjugate, a disaccharide‐modified conjugate was more efficient. 225 Interesting behavior was observed for AuNPs grafted with fucose‐ended linkers. Such AuNP conjugates with a specific glycan surface density were internalized by dendritic cells via their Dendritic Cell‐Specific Intercellular adhesion molecule‐3‐Grabbing Non‐integrin (DC‐SIGN) receptors (see Fig. 13 ), but without an expected induction of subsequent interleukin production. Observed targeted internalization is a focus for antigen delivery and for the induction of a desired DC‐SIGN‐mediated signaling cascade. 224 Figure 13 A schematic depiction of fucosylated AuNPs interacting with dendritic cell‐specific intercellular adhesion molecule‐3‐grabbing nonintegrin (DC‐SIGN) receptors. Reprinted with permission from 224 . Copyright 2014 American Chemical Society. In order to induce an anticancer immune response, a tumor‐specific antigen (a phospholipid‐functionalized glycopeptide) was used for the formation of a scaffold with an iron oxide core. 226 The highest amount of elicited IgG (mean titres ∼81,402) was confirmed after the vaccination of mice by the conjugate having only one carbohydrate antigen per phospholipid‐modified glycopeptide chain while surprisingly, mean titres of only ∼7530 were observed when two glycosyl units were bound to the chain. It should also be noted that the formation of NPs with an iron oxide core significantly increased the IgG titres for free glycosylated phospholipid‐modified peptide chains from ∼5032 to ∼36,600. Furthermore, a complex induction of the immune response leading to tumor cell degradation was observed, 226 suggesting that iron oxide NPs can be an alternative to AuNPs for vaccine design. Carbohydrates were used also as adjuvants in AuNP‐based vaccines, with chitosan‐coated AuNPs decorated with plant saponins applied as tetanus toxoid carriers, but the authors did not describe any specific role of chitosan in the fabricated conjugates. 227 1. Carbohydrate Immunogens Conjugated with Polymers In order to overcome the low immunogenicity of cancer cells, tumor‐specific or tumor‐associated antigens can be multiplied on the surface of NPs. The administration of such NPs helps the immune system to fight disease in a more efficient way. A multivalent display of antigens allows their better recognition by respective receptors with the subsequent induction of the immune response. It should be noted that natural antigenic carbohydrates (except for zwitterionic glycans) are believed to mostly lack the ability to be displayed on the B‐cell surface via the major histocompatibility complex, resulting in their incapability to activate T cells. This activation was, instead, achieved by displaying parts of carrier proteins co‐delivered in a vaccine. In the following sections, however, glycan‐nano vaccines without peptide adjuvants are also shown to be fabricable. Using "classical" adjuvants, the conjugation of immunogenic synthetic mucin (MUC1, a surface‐displayed glycopeptide typical for tumor epithelial cells) with different NPs was tested intensively. In a recent review paper, a study describing MUC1 conjugated with tetanus toxoid carrier protein was described. 189 This combination provided the best performance as judged from the highest production of specific IgGs (see Fig. 10 for the structure of the glycoprotein conjugate) among other reviewed MUC1 vaccines. 189 More recently, the same antigen was conjugated with a bacterial lipoprotein—a Toll‐like receptor ligand to boost the elicited immune reaction (Fig. 11 ). 190 The same composition (i.e., MUC1 antigen conjugated to a Toll‐like receptor) amended with a T‐helper‐cell epitope was employed by Abdel‐Aal et al. 191 Their conjugate, however, was not multivalent, but it was incorporated into small lamellar vesicles. These vesicles were used to immunize mice that were subsequently exposed to tumor induction. Animals that had been immunized with the best‐performing conjugate exhibited an approximately twofold smaller tumor after 2 weeks compared to other vaccines or a blank. 191 Another cancer antigen is the breast cancer cell‐specific hexasaccharide "Globo H," and its conjugation with diphtheria toxoid CRM 197 and α‐galactosylceramide C34 as an adjuvant provided the highest immune response amongst the conjugates tested using different carrier proteins and adjuvants. This effect observed was even higher than for a similar clinical trial phase III vaccine. 192 In order to synthesize well‐defined nanoconjugates, a method using the tyrosine‐specific binding of immunogenic glycans present on the surface of CRM 197 was developed 193 , 194 with consequent application in the development of a glycan‐based vaccine eliciting an immune response against Streptococcus infection. 195 Similarly, McCarthy et al. developed a chemoenzymatic synthesis of a well‐defined poly(sialic acid)‐tetanus toxoid glycoconjugate. 196 These studies have underlined how important it is to develop novel progressive methods for the precise and well‐defined synthesis of glycoconjugates. Figure 10 Synthesis and structure of a carrier protein‐mucin vaccine. Conjugation of a sialyl‐Tn MUC1 glycopeptide antigen (19) terminated via a coupling agent (diethyl squarate; 57) with a carrier protein (Tetanus toxoid or BSA) into a final multivalent carbohydrate‐bearing vaccination conjugate (58). Reproduced from 189 with permission of the Royal Society of Chemistry. Figure 11 A schematic structure of a vaccine consisting of 4‐valent glycopeptide MUC1 conjugated to a bacterial lipopeptide (Pam3CSK4), a ligand of Toll‐like receptor (TLR2) which helps to elicitate immune reaction. Reproduced from 190 with permission from John Wiley & Sons. In addition to discovering new possibilities in the synthesis of well‐defined glycoconjugates consisting of already known components, some studies, more biochemical than nanotechnological, have focused on immunogenic efficiency and the synthesis of new poly‐/oligo‐saccharides, mainly derivatives of bacterial cell wall epitopes. 197 , 198 , 199 , 200 Similarly, Johannes et al. investigated the increased immunogenicity of fluorinated analogues of tumor‐associated carbohydrate antigens conjugated to a support applicable as a vaccine against human breast cancer. 201 Alternatively, carbohydrates were used as building blocks or monomers for supramolecular scaffolds forming vaccine NPs. Since polysaccharides can form polyvalent ions, they can form self‐assembled nanoscaffolds upon mixing with relevant (polyvalent) counterions. Derivatives of chitin, the second most abundant polysaccharide on Earth and a very promising material in this field of research, can be used as very potent vaccine adjuvants. The mucoadhesion of chitin derivatives is helpful in vaccine administration and allows for an increased rate of chitosan‐based NP internalization via mannose‐binding receptors, inducing both the humoral and cellular immune response. 202 The latter was found to induce the degradation of the polysaccharide NPs, and chitosan dissolved in cytosol was found to be a more efficient immunogen than whole, nondestructed chitosan NPs. 203 The stability of prepared vaccine NPs is also an important issue. For example, a DNA‐based antitumor vaccine was effectively protected against degradation at a low pH when present in alginic acid based NPs. 204 Similarly, the coencapsulation of an antigen to be delivered with a commercially available adjuvant (C48/80) into chitosan NPs significantly decreased the dosage needed for the induction of a strong immune response compared to vaccination with a nonencapsulated antigen and an adjuvant. 205 , 206 A similar effect was observed for dendritic glucan, 207 and immunogenic properties have also been assigned to β‐glucans applied for the preparation of antifungal vaccines 208 and to polysaccharide based on δ‐inuline, which was utilized as an adjuvant. 209 Cholesteryl pullulan was used to entrap an additional immunogen—tumor necrosis factor α. 210 The nasal coadministration of such a prepared adjuvant combined with commercial anti‐influenza vaccine significantly elevated the resistance of mice treated against the influenza virus compared to control groups vaccinated without the adjuvant. 210 A better administration of tumor‐associated carbohydrate antigens was developed by Zhou et al., who investigated different derivatives of synthetic lipid A. 211 These compounds have been previously found as promising adjuvants, but a more efficient synthesis is needed for their wider utilization. 211 The efficacy of polysaccharide vaccines can be further improved by the decoration of NPs, with additional glycan moieties with mannose‐coated chitosan particles being a good example. 212 , 213 Another approach was presented by Fagan et al., who used tetra‐ O ‐acetyl‐α‐ d ‐glucopyranosyl bromide as a core structure for the synthesis of dendritic nanoscaffolds displaying multivalent immunogens—streptococcal B‐cell epitope. 214 It is also possible to prepare effective conjugates based on polymers and glycans where carbohydrate moieties are not primarily acting as immunogens but rather promote selective delivery and administration. A good example is mannosylated liposomes delivering albumin as a model antigen efficiently to dendritic cells responsible for the induction of humoral and immune responses. 215 The presence of mannose receptors on the surface of DC has also been widely used in drug delivery, imaging, and other biomedical applications that are discussed later. The overall therapeutic efficacy of mannosylated liposomes loaded with antitumor antigen and lipidic adjuvant was significantly boosted by the codelivery of small interfering RNA (siRNA) responsible for the downregulation of expression of immunosuppressive interleukins in tumor cells. 216 In another approach, Kim et al. developed fiber‐like supramolecular assemblies coated with mannose‐tethered lectin Con A. 217 Such protein coating was responsible for the immunogenicity of the fabricated conjugates and confirmed by observed interleukin production after the treatment of T cells with Con A‐coated NPs. 217 Carbohydrate immunogens playing a role as adjuvants, that is, helping to elicit a less‐specific immunoresponse, could also be delivered by diverse nanopharmaceutical systems. For example, a synthetic Mycobacterium tuberculosis epitope derivative (a fusion protein) and a glucopyranosyl lipid moiety were tested in liposomes, nanoemulsions, and adjuvants, with a promising immune response induction observed in mice with nanoemulsions selected as the best option for further vaccine approval and trials. 218 Controlled selective delivery was also achieved by the conjugation of a bioactive molecule (inhibitor of a transforming growth factor‐β receptor) with mannose‐6‐phosphate‐human serum albumin. 219 This carrier reacted selectively with receptors of hepatic stellate cells, while the conjugated inhibitor prevented transforming growth factor‐β‐induced activation, which is a key factor in the development of liver fibrosis. 219 2. Metal and Metal Oxide NPs in Recent Vaccine Development Due to their versatile and reproducible preparation and modification, AuNPs have attracted researchers' interest as promising therapeutic agents 220 , 221 with potential especially in vaccine development. 186 Although older studies have demonstrated that AuNPs decorated with analogues of viral or tumor cell polysaccharide epitopes were sufficiently immunogenic, 221 there is only one recent study describing vaccine design based on AuNPs coated with a carrier protein and an LPS from a nonvirulent bacterial strain Burkholderia thailandensis . 222 This glycoconjugate significantly increased the production of LPS‐specific antibodies in nonhuman primates exposed to the virulent bacterial strain B. mallei . In animals vaccinated with the AuNP‐LPS‐based conjugate no signs of bacteria were found, while in non‐vaccinated animals the pathogen cells were detected ‐ 102 (in animals that survived the test) and 104 (not survived animals) cfu per mg of tissue. Such results can be considered promising for the development of an efficient protective antiglanders (glanders = contagious and highly fatal disease, which can affect humans) vaccine for humans. 222 Parry et al. showed that an AuNP‐glycan‐based vaccine efficiently induced the immune response even in the absence of peptide or protein adjuvants. 223 These authors synthesized polymers displaying glycan units mimicking tumor tissue specific mucin (see Fig. 12 ). These immunogens were then conjugated in one step with in situ prepared AuNPs. Contrary to unmodified polymers, all glycan conjugates induced an immune response, as evidenced by IgG titres. The authors found out that the optimal number of glycan units per AuNPs‐conjugated polymer chain is 20–25, regardless of the chain length. 223 Figure 12 Preparation and characterization of a carbohydrate Tn‐antigen/AuNPs vaccine (Tn = N ‐α‐acetylgalactosamine linked to serine or threonine). Inset figures show representative dynamic light scattering data (top) and TEM image (bottom) of glycoNPs (scale bar = 20 nm). Reprinted with permission from 223 . Copyright 2013 American Chemical Society. Fallarini et al. investigated simple AuNP–glycan conjugates with the aim of understanding the essential components to design an effective vaccine. 225 In this study, they compared the effect of NPs decorated with nonimmunogenic mono‐ and disaccharides mimicking parts of the capsular polysaccharides of Neisseria meningitidis bacterium. The conjugate induced an immune cell response, contrary to nonconjugated forms of carbohydrates, and due to the possible intracellular degradation of the glycoconjugate, a disaccharide‐modified conjugate was more efficient. 225 Interesting behavior was observed for AuNPs grafted with fucose‐ended linkers. Such AuNP conjugates with a specific glycan surface density were internalized by dendritic cells via their Dendritic Cell‐Specific Intercellular adhesion molecule‐3‐Grabbing Non‐integrin (DC‐SIGN) receptors (see Fig. 13 ), but without an expected induction of subsequent interleukin production. Observed targeted internalization is a focus for antigen delivery and for the induction of a desired DC‐SIGN‐mediated signaling cascade. 224 Figure 13 A schematic depiction of fucosylated AuNPs interacting with dendritic cell‐specific intercellular adhesion molecule‐3‐grabbing nonintegrin (DC‐SIGN) receptors. Reprinted with permission from 224 . Copyright 2014 American Chemical Society. In order to induce an anticancer immune response, a tumor‐specific antigen (a phospholipid‐functionalized glycopeptide) was used for the formation of a scaffold with an iron oxide core. 226 The highest amount of elicited IgG (mean titres ∼81,402) was confirmed after the vaccination of mice by the conjugate having only one carbohydrate antigen per phospholipid‐modified glycopeptide chain while surprisingly, mean titres of only ∼7530 were observed when two glycosyl units were bound to the chain. It should also be noted that the formation of NPs with an iron oxide core significantly increased the IgG titres for free glycosylated phospholipid‐modified peptide chains from ∼5032 to ∼36,600. Furthermore, a complex induction of the immune response leading to tumor cell degradation was observed, 226 suggesting that iron oxide NPs can be an alternative to AuNPs for vaccine design. Carbohydrates were used also as adjuvants in AuNP‐based vaccines, with chitosan‐coated AuNPs decorated with plant saponins applied as tetanus toxoid carriers, but the authors did not describe any specific role of chitosan in the fabricated conjugates. 227 B. Nonimmunogenic Therapeutic GlycoNPs In addition to the activation of the immune system with a consequent therapeutic biochemical cascade, a few other therapeutic effects of glycoconjugates were discovered. Such efforts rely on (i) the competitive binding of lectins/glycans on receptors, thus preventing the successful binding of pathogenic viruses/bacteria, and (ii) the ability of glycoconjugates to selectively agglutinate pathogenic particles, thus eliminating their adverse effect. Since the interaction between a carbohydrate‐binding protein and a carbohydrate is not very strong, multivalent recognition entities must be present on a therapeutic NP surface. There have been numerous reports describing the fabrication of such therapeutics, with the main achievements comprehensively summarized in recent excellent reviews. 90 , 188 , 228 , 229 , 230 , 231 In summary, these papers referred to diverse kinds of multivalent glycoconjugates, that is, based on carbon and metallic NPs, organic supramolecular scaffolds or proteins with many of such conjugates having therapeutic potential. Moreover, novel synthetic and conjugation protocols have been described as well. Interesting conclusions are provided in a review by Jiménez Blanco et al., who suggested that heteroglycoconjugates, compared to homoglycoconjugates, provide additional regulation possibilities, in addition to the already known protein–carbohydrate binding mechanism. 232 These findings have been considered in recent developments in the design of nonimmunogenic therapeutic glycoconjugates. 1. Metallic NPs Few recent studies have focused on the antibacterial properties of AuNPs 233 or AgNPs 234 coated with 6‐ O ‐chitosan sulfate. Importantly, these glycoconjugates are not only antibacterial due to the presence of Ag or Au but also anticoagulant, making them almost perfect candidates for the surface coating of, for example, medical devices that need to be kept sterile. Similar, although offering only antibacterial properties, AgNPs and AuNPs have been conjugated with other polysaccharides, for example, starch 235 and aminocellulose. 236 A similar antibacterial effect has been observed with AgNPs coated with 12‐C‐monosaccharide‐dodecanoic acid 237 or with MNPs stabilized by poly(ethylene oxide) and functionalized with a sialic acid derivative. 238 While the aforementioned glycoNPs were tested mainly against E. coli , magnetic glycoNPs displaying fucose‐bearing oligosaccharides efficiently block the adhesion of Helicobacter pylori . 239 2. Carbon NPs Carbon NPs are another choice in the fabrication of glycoconjugates with antiadhesive properties. Recent studies include mannose‐bearing diamond NPs exhibiting an excellent inhibition of E. coli type 1 FimH‐mediated adhesion. 240 , 241 Ragoussi et al. have described similar mannose modification performed on CNTs and graphene sheets but provided only the results of a selective lectin binding test 242 ; the inhibition of bacterial adhesion by such NPs has yet to be assessed. On the other hand, Luczkowiak et al. reported an inhibition of Ebola pseudotype virus binding to DC‐SIGN receptor‐displaying cells. 243 In their study, fullerene with 12 mannoses (Fig. 14 ) was the most effective virus inhibitor, while increasing mannose valency led to decreased inhibition efficiency, 243 confirming a need for the controlled synthesis of conjugates with their extensive assessment. Figure 14 Structure of multivalent mannosylated fullerene used for inhibition of Ebola‐like virus binding. Reprinted with permission from 243 . Copyright 2013 American Chemical Society. 3. Synthetic Polymers Despite the relative ease of controlled chemical ways of glycosylation of carbon and metallic NPs, polymer scaffolds seem to have been explored the most intensively in this field. One reason may be a wider diversity of synthesized polymeric glycoconjugates available, as shown in the extensive study by Percec et al. 244 Recently reported examples of fully synthesized scaffolds include PAMAM dendrimers displaying carbohydrate Le x moieties capable of competitively blocking DC‐SIGN receptors, thus inhibiting the first phase of HIV infection. 245 More importantly, the glycodendrimers exhibited a negligible inhibition of langerin, which, contrary to DC‐SIGN, helped to internalize and destroy HIV particles. From the testing of a mini library comprising of third‐, fourth‐, and fifth‐generation glycodendrimers, the latter was found to be approximately twice as efficient in inhibiting the binding of model virus‐like particles compared to the other ones (third and fourth generation). Finally, the conjugate was tested for the inhibition of the real virus influencing DC‐SIGN‐mediated binding and trans‐infection. 245 Fucosylated, branched phosphodiester‐based scaffolds were reported to bind efficiently to Burkholderia ambifaria lectin with a promising application as inhibitors of bacterial infection. 246 Cholera toxin has been successfully ligated with newly synthesized five‐armed molecular scaffolds, bearing a galactose‐oligomer recognition unit on the end of each arm. 247 Pseudomonas aeruginosa lectin A recognition was suppressed by α‐ l ‐fucoside and by a β‐ d ‐galacto‐pyranoside dendrimer hybrid, which may be further exploited for the treatment of Pseudomonas aeruginosa ‐based infections 248 , 249 (for Pseudomonas aeruginosa lectin A‐glycan binding AFM images, see 250 ). Yan et al. investigated the influence of multivalency of n ‐heptyl‐α‐ d ‐mannose on the strength of binding to FimH adhesin displayed on type 1 fimbriae of E. coli causing Crohn's disease. 251 They found that the linear polymer decorated with numerous n ‐heptyl‐α‐ d ‐mannose moieties was more efficient in the coagulation of bacteria (Fig. 15 ) compared to monovalent glycoconjugates or polyvalent star‐like particles. 251 The potential of multivalent synthetic glycodendrimers was further demonstrated by Ghirardello et al., who reported a 109×106‐fold stronger binding potential (relative potency) to WGA of a 48‐valent GlcNAc dendrimer compared to an unconjugated GlcNAc moiety. 252 A relative potency of only 1168 was reported for a 30‐valent N ‐acetyl‐ d ‐lactosamine (LacNAc) "onion peel" like dendrimer compared to the LacNAc monomer, binding specifically to leguminous lectin from Erythrina cristagalli . 253 Figure 15 Glycopolymer‐induced agglutination of type 1 fimbriaeted E. coli (strain UTI89). (A) Fluorescence microscopy pictures of Katushka‐expressing type 1 fimbriaeted E. coli UTI89; (B) exposed to 1 μM of L188; (C) to 3 μM of L188 and resulting in a "bacterial egg" agglomerate with approximate dimensions of 98 × 51 μm. Reprinted with permission from 251 . Copyright 2015 American Chemical Society. In order to protect against eukaryotic parasites, Campo et al. synthesized cyclic triazole‐linked oligomers with pseudo‐glycosidic units with an affinity toward the trans‐sialidase enzyme from Trypanosoma cruzi . 254 As a result, the inhibition of the parasite development after their invasion into macrophage cells was observed, but a quite high concentration (250 μM) of the therapeutic oligomer was needed. Nevertheless, these results are only preliminary, and further optimization of oligomer formation (i.e., a pseudo‐glycosidic unit density and its precise location) is needed to increase the inhibition activity. 254 4. Biopolymers and Proteins Galactose units displayed on a branched oligopeptide were efficiently bound to the surface lectin of P. aeruginosa , inhibiting film formation and disrupting existing biofilms. 255 Bouckaert et al. explored cyclodextran NPs conjugated with n ‐heptyl‐α‐ d ‐mannose using various spacers as potential FimH agonists. 256 They did not observe improved adhesion to the FimH adhesin of E. coli ; however, their results suggest that differences in the length of the spacer arm may account for the low undesired affinity for human mannose‐binding lectins. 256 A further step in the development of antivirotics and antibacterial drugs that are not based on recently used synthetic antibiotics, to which an increased number of bugs are resistant, was described by the preparation of a cyclopeptide‐based dendrimer decorated by glycan moieties using an oxime ligation. 257 The best‐performing, 64‐valent glycodendrimer from the reported library of compounds (Fig. 16 ) exhibited a 40,000‐fold increased lectin‐binding potency compared to monovalent methyl α‐ l ‐fucopyranoside. 257 Cholera toxin inhibitor with an IC 50 value of 100 pM based on a tetravalent neoglycoprotein prepared by a simple chemical modification of an inactive B subunit of cholera toxin was reported by Branson et al. 258 The use of cholera toxin protein as a template secured a "… precise fit of the ligand groups with the spacing and configuration of binding sites on wild‐type CTB." 258 Figure 16 Synthesis and structure of cyclopeptide‐based glycodendrimers. Reproduced from 257 with permission of the Royal Society of Chemistry. Self‐assembled polymeric particles are the third type of organic nanocarriers covered in this section. For example, vesicles self‐assembled from poly(ethylene oxide) and polycaprolactone and further modified with a sialodendrimer block the cellular recognition of influenza virus via ligation with viral hemagglutinins. 259 Authors have reported a decrease in IC 50 from 4 mM to 240 μM (per glycan unit) when mono‐ and eight‐valent glycodendrimers were tested, respectively, and a further decrease to approximately 2 μM when the glycodendrimers were conjugated to the polymer vesicles. 259 Another type of conjugate was prepared in the form of nanodroplets by crosslinking polymer shells that were ready for subsequent covalent modification. 251 Yan et al. decorated such particles with multiple n ‐heptyl‐α‐ d ‐mannose moieties to achieve the efficient coagulation of pathogenic E. coli type 1. 251 E. coli 's type 1 fimbrial adhesin (FimH) was also targeted by Wu at al. using mannose‐terminated DNA oligomers. 260 Complementary DNA strands conjugated with a second‐generation dendrimer were then applied to assemble mannosylated fibers with the capability of agglutinating E. coli strain ORN178. 260 Interestingly, the authors claimed that agglutination ability was independent of mannosyl density on the fiber surface, most likely because of flexible tube‐like NPs. 260 Yu et al. prepared nanotubes that were self‐assembled from amphiphilic pillar[5]arene, with glucose acting as the hydrophilic part. 261 These glyco‐nanotubes were able to agglutinate E. coli cells more effectively with an increased number of glucose units with the highest agglutination index of 54 (an average number of bacteria connected to each other after successful agglutination). 261 Therapeutic potential also can be predicted for self‐assembled oligopeptide‐LacNAc conjugates capable of moderating the activity of galectin, a lectin‐type protein with signaling and other properties. 262 5. GlycoNPs for Enzyme Inhibition and Other Therapeutic Functions Applications of iminosugars for the treatment of lysosome storage diseases (e.g., Gaucher disease or glycosphingolipid lysosomal storage disorder, characterized by a mutation‐based impairment of different glycosidases with a subsequent pathological accumulation of glycolipids) have been investigated in recent years. Iminosugars are selective and reversible inhibitors of glycosidases due to selective binding to their active sites. This binding can help to fold impaired enzymes, thus partially restoring their hydrolytic activity at the higher pH present in lysosomes. Furthermore, the misfolded enzymes "protected" by the bound iminosugar can avoid degradation otherwise induced by a cellular quality control. 263 Using this mechanism, iminosugars could be applied for so‐called "pharmacological chaperone therapy," an emerging way to treat lysosomal storage diseases. The nature of recognition between iminosugar and respective glycosidase differs significantly from that of lectin–glycan binding. This fact raises the question of whether multivalency can increase the inhibition potency of investigated glycomimetics. In fact, previous studies have suggested that multivalent iminosugars lost their inhibition ability, while recent studies using well‐defined and controlled synthesis have reported the opposite. In the study of Garcia Fernandez et al., a mini library of iminosugar‐based glycomimetics (either monovalent or conjugated to fullerenes) was tested for its potency to inhibit a variety of glycosidases. 264 Their results suggested that in some cases, multivalency can "switch on" the inhibition by switching the binding mode, rather than by the sheer presence of a higher amount of recognition moieties. 264 In a more recent work, several scaffolds and conjugates displaying deoxynojirimycin, a broad glycosidase inhibitor, were tested. 265 Interestingly, the strongest inhibitor (800‐fold stronger than a monovalent analogue) was just a tetravalent iminosugar present on a porphyrin‐based scaffold. The authors suggested that the valency itself did not cause an increased inhibition potential, but rather the spatial conformation of monosaccharides did. 265 Moreno‐Clavio et al. tested several l ‐fucosidase iminosugar‐based inhibitors with less success. 266 Only a sevenfold‐higher inhibition effect on α‐ l ‐fucosidases was observed when a trivalent scaffold was used compared to a monovalent one. 266 On the other side, a relative α‐mannosidase inhibition potency as high as 3000 was reported for the N ‐alkyl analogue of 1‐deoxynojirimycin conjugated on a synthetic glycoprotein scaffold (Fig. 17 ), 267 almost 10,000 for the 21‐valent 1‐deoxynojirimycin scaffold 268 and higher than 500 for glycocyloclopeptide with seven copies of 1‐deoxynojirimycin. 269 In the latter case, the large inhibition potential was obtained only on glycoconjugates with a longer (C 9 ) spacer, while C 6 spacers of the same glycoconjugate resulted only in a relative inhibition potency of 42. 269 Figure 17 1‐Deoxynojirimycin‐based glycopolypeptides: synthesis, self‐assembly, and glycosidase inhibition model. CuAAC = Cu I catalyzed Huisgen azide‐alkyne cycloaddition. Reproduced from 267 with permission of the Royal Society of Chemistry. The real therapeutic effect of different 3‐ to 14‐valent iminosugar conjugates was investigated by Joosten et al. through the in vitro measurement of β‐glucocerebrosidase activities in N370S Gaucher fibroblasts. 270 The authors reported a "mild, but significant" effect of multivalency. Importantly, the highest inhibition rate was not correlated with the highest chaperoning (therapeutic) effect. 270 This report clearly suggests that similar systematic investigations will be necessary to develop efficient NP‐based drugs against lysosome storage diseases. A similar approach applied for the treatment of lysosome storage diseases was also investigated for the treatment of cystic fibrosis. 1‐Deoxynojirimycin and its derivatives correct a misfolded cystic fibrosis transmembrane conductance regulator protein in a similar manner as the aforementioned chaperone effect of iminosugars towards glycosidases, with a 1000‐fold increased inhibition activity of 3‐valent iminosugar compared to a monovalent analogue control. 271 The concept of pharmacological chaperones may sound very promising, but although the activation of glycosidases has been observed, it hardly reached therapeutically sufficient levels, with typically an approximately twofold increase in the activity. 272 From this point of view, the work of Brissonnet et al. is noteworthy. 273 These authors reported an activity increase as high as 70‐fold of a bacterial mannoside‐phosphorylase using cyclodextran‐based NPs displaying deoxymannojirimycin (more than 100‐valent). Nevertheless, further studies are needed to elucidate the nature of the enzyme activation gain. 273 The selective interaction of a disaccharide called Thomas Friedrich antigen (TF ag ) with specific tumor cells displaying Gal‐3 lectin had a cytotoxic effect on the targeted cells: a conjugate of small AuPs coated with TF ag via an amino acid linker was synthesized, and an approximately 100‐fold higher cytotoxicity toward the Gal‐3‐positive cells was achieved compared to monomeric units. 274 The main cytotoxic effect of the TF ag relied on apoptosis induced by an inhibited Gal‐3 signaling, 274 but it is anticipated that the TF ag targeting may also work well in drug delivery. Another glycan with a therapeutic effect is heparan sulfate, a glycosaminoglycan‐based polysaccharide capable of inhibiting protease β‐secretase responsible for the accumulation of plaques causing Alzheimer's disease. Nevertheless, carbohydrate synthesis is rather complicated and expensive. Therefore, a mini library of dendritic polymers bearing multiple copies of different heparan sulfate monomers was recently introduced. 275 Some of these conjugates were found to possess inhibition potential equivalent to that of previously tested oligomers, which is an important result for the future development and production of anti‐Alzheimer's disease drugs. 275 1. Metallic NPs Few recent studies have focused on the antibacterial properties of AuNPs 233 or AgNPs 234 coated with 6‐ O ‐chitosan sulfate. Importantly, these glycoconjugates are not only antibacterial due to the presence of Ag or Au but also anticoagulant, making them almost perfect candidates for the surface coating of, for example, medical devices that need to be kept sterile. Similar, although offering only antibacterial properties, AgNPs and AuNPs have been conjugated with other polysaccharides, for example, starch 235 and aminocellulose. 236 A similar antibacterial effect has been observed with AgNPs coated with 12‐C‐monosaccharide‐dodecanoic acid 237 or with MNPs stabilized by poly(ethylene oxide) and functionalized with a sialic acid derivative. 238 While the aforementioned glycoNPs were tested mainly against E. coli , magnetic glycoNPs displaying fucose‐bearing oligosaccharides efficiently block the adhesion of Helicobacter pylori . 239 2. Carbon NPs Carbon NPs are another choice in the fabrication of glycoconjugates with antiadhesive properties. Recent studies include mannose‐bearing diamond NPs exhibiting an excellent inhibition of E. coli type 1 FimH‐mediated adhesion. 240 , 241 Ragoussi et al. have described similar mannose modification performed on CNTs and graphene sheets but provided only the results of a selective lectin binding test 242 ; the inhibition of bacterial adhesion by such NPs has yet to be assessed. On the other hand, Luczkowiak et al. reported an inhibition of Ebola pseudotype virus binding to DC‐SIGN receptor‐displaying cells. 243 In their study, fullerene with 12 mannoses (Fig. 14 ) was the most effective virus inhibitor, while increasing mannose valency led to decreased inhibition efficiency, 243 confirming a need for the controlled synthesis of conjugates with their extensive assessment. Figure 14 Structure of multivalent mannosylated fullerene used for inhibition of Ebola‐like virus binding. Reprinted with permission from 243 . Copyright 2013 American Chemical Society. 3. Synthetic Polymers Despite the relative ease of controlled chemical ways of glycosylation of carbon and metallic NPs, polymer scaffolds seem to have been explored the most intensively in this field. One reason may be a wider diversity of synthesized polymeric glycoconjugates available, as shown in the extensive study by Percec et al. 244 Recently reported examples of fully synthesized scaffolds include PAMAM dendrimers displaying carbohydrate Le x moieties capable of competitively blocking DC‐SIGN receptors, thus inhibiting the first phase of HIV infection. 245 More importantly, the glycodendrimers exhibited a negligible inhibition of langerin, which, contrary to DC‐SIGN, helped to internalize and destroy HIV particles. From the testing of a mini library comprising of third‐, fourth‐, and fifth‐generation glycodendrimers, the latter was found to be approximately twice as efficient in inhibiting the binding of model virus‐like particles compared to the other ones (third and fourth generation). Finally, the conjugate was tested for the inhibition of the real virus influencing DC‐SIGN‐mediated binding and trans‐infection. 245 Fucosylated, branched phosphodiester‐based scaffolds were reported to bind efficiently to Burkholderia ambifaria lectin with a promising application as inhibitors of bacterial infection. 246 Cholera toxin has been successfully ligated with newly synthesized five‐armed molecular scaffolds, bearing a galactose‐oligomer recognition unit on the end of each arm. 247 Pseudomonas aeruginosa lectin A recognition was suppressed by α‐ l ‐fucoside and by a β‐ d ‐galacto‐pyranoside dendrimer hybrid, which may be further exploited for the treatment of Pseudomonas aeruginosa ‐based infections 248 , 249 (for Pseudomonas aeruginosa lectin A‐glycan binding AFM images, see 250 ). Yan et al. investigated the influence of multivalency of n ‐heptyl‐α‐ d ‐mannose on the strength of binding to FimH adhesin displayed on type 1 fimbriae of E. coli causing Crohn's disease. 251 They found that the linear polymer decorated with numerous n ‐heptyl‐α‐ d ‐mannose moieties was more efficient in the coagulation of bacteria (Fig. 15 ) compared to monovalent glycoconjugates or polyvalent star‐like particles. 251 The potential of multivalent synthetic glycodendrimers was further demonstrated by Ghirardello et al., who reported a 109×106‐fold stronger binding potential (relative potency) to WGA of a 48‐valent GlcNAc dendrimer compared to an unconjugated GlcNAc moiety. 252 A relative potency of only 1168 was reported for a 30‐valent N ‐acetyl‐ d ‐lactosamine (LacNAc) "onion peel" like dendrimer compared to the LacNAc monomer, binding specifically to leguminous lectin from Erythrina cristagalli . 253 Figure 15 Glycopolymer‐induced agglutination of type 1 fimbriaeted E. coli (strain UTI89). (A) Fluorescence microscopy pictures of Katushka‐expressing type 1 fimbriaeted E. coli UTI89; (B) exposed to 1 μM of L188; (C) to 3 μM of L188 and resulting in a "bacterial egg" agglomerate with approximate dimensions of 98 × 51 μm. Reprinted with permission from 251 . Copyright 2015 American Chemical Society. In order to protect against eukaryotic parasites, Campo et al. synthesized cyclic triazole‐linked oligomers with pseudo‐glycosidic units with an affinity toward the trans‐sialidase enzyme from Trypanosoma cruzi . 254 As a result, the inhibition of the parasite development after their invasion into macrophage cells was observed, but a quite high concentration (250 μM) of the therapeutic oligomer was needed. Nevertheless, these results are only preliminary, and further optimization of oligomer formation (i.e., a pseudo‐glycosidic unit density and its precise location) is needed to increase the inhibition activity. 254 4. Biopolymers and Proteins Galactose units displayed on a branched oligopeptide were efficiently bound to the surface lectin of P. aeruginosa , inhibiting film formation and disrupting existing biofilms. 255 Bouckaert et al. explored cyclodextran NPs conjugated with n ‐heptyl‐α‐ d ‐mannose using various spacers as potential FimH agonists. 256 They did not observe improved adhesion to the FimH adhesin of E. coli ; however, their results suggest that differences in the length of the spacer arm may account for the low undesired affinity for human mannose‐binding lectins. 256 A further step in the development of antivirotics and antibacterial drugs that are not based on recently used synthetic antibiotics, to which an increased number of bugs are resistant, was described by the preparation of a cyclopeptide‐based dendrimer decorated by glycan moieties using an oxime ligation. 257 The best‐performing, 64‐valent glycodendrimer from the reported library of compounds (Fig. 16 ) exhibited a 40,000‐fold increased lectin‐binding potency compared to monovalent methyl α‐ l ‐fucopyranoside. 257 Cholera toxin inhibitor with an IC 50 value of 100 pM based on a tetravalent neoglycoprotein prepared by a simple chemical modification of an inactive B subunit of cholera toxin was reported by Branson et al. 258 The use of cholera toxin protein as a template secured a "… precise fit of the ligand groups with the spacing and configuration of binding sites on wild‐type CTB." 258 Figure 16 Synthesis and structure of cyclopeptide‐based glycodendrimers. Reproduced from 257 with permission of the Royal Society of Chemistry. Self‐assembled polymeric particles are the third type of organic nanocarriers covered in this section. For example, vesicles self‐assembled from poly(ethylene oxide) and polycaprolactone and further modified with a sialodendrimer block the cellular recognition of influenza virus via ligation with viral hemagglutinins. 259 Authors have reported a decrease in IC 50 from 4 mM to 240 μM (per glycan unit) when mono‐ and eight‐valent glycodendrimers were tested, respectively, and a further decrease to approximately 2 μM when the glycodendrimers were conjugated to the polymer vesicles. 259 Another type of conjugate was prepared in the form of nanodroplets by crosslinking polymer shells that were ready for subsequent covalent modification. 251 Yan et al. decorated such particles with multiple n ‐heptyl‐α‐ d ‐mannose moieties to achieve the efficient coagulation of pathogenic E. coli type 1. 251 E. coli 's type 1 fimbrial adhesin (FimH) was also targeted by Wu at al. using mannose‐terminated DNA oligomers. 260 Complementary DNA strands conjugated with a second‐generation dendrimer were then applied to assemble mannosylated fibers with the capability of agglutinating E. coli strain ORN178. 260 Interestingly, the authors claimed that agglutination ability was independent of mannosyl density on the fiber surface, most likely because of flexible tube‐like NPs. 260 Yu et al. prepared nanotubes that were self‐assembled from amphiphilic pillar[5]arene, with glucose acting as the hydrophilic part. 261 These glyco‐nanotubes were able to agglutinate E. coli cells more effectively with an increased number of glucose units with the highest agglutination index of 54 (an average number of bacteria connected to each other after successful agglutination). 261 Therapeutic potential also can be predicted for self‐assembled oligopeptide‐LacNAc conjugates capable of moderating the activity of galectin, a lectin‐type protein with signaling and other properties. 262 5. GlycoNPs for Enzyme Inhibition and Other Therapeutic Functions Applications of iminosugars for the treatment of lysosome storage diseases (e.g., Gaucher disease or glycosphingolipid lysosomal storage disorder, characterized by a mutation‐based impairment of different glycosidases with a subsequent pathological accumulation of glycolipids) have been investigated in recent years. Iminosugars are selective and reversible inhibitors of glycosidases due to selective binding to their active sites. This binding can help to fold impaired enzymes, thus partially restoring their hydrolytic activity at the higher pH present in lysosomes. Furthermore, the misfolded enzymes "protected" by the bound iminosugar can avoid degradation otherwise induced by a cellular quality control. 263 Using this mechanism, iminosugars could be applied for so‐called "pharmacological chaperone therapy," an emerging way to treat lysosomal storage diseases. The nature of recognition between iminosugar and respective glycosidase differs significantly from that of lectin–glycan binding. This fact raises the question of whether multivalency can increase the inhibition potency of investigated glycomimetics. In fact, previous studies have suggested that multivalent iminosugars lost their inhibition ability, while recent studies using well‐defined and controlled synthesis have reported the opposite. In the study of Garcia Fernandez et al., a mini library of iminosugar‐based glycomimetics (either monovalent or conjugated to fullerenes) was tested for its potency to inhibit a variety of glycosidases. 264 Their results suggested that in some cases, multivalency can "switch on" the inhibition by switching the binding mode, rather than by the sheer presence of a higher amount of recognition moieties. 264 In a more recent work, several scaffolds and conjugates displaying deoxynojirimycin, a broad glycosidase inhibitor, were tested. 265 Interestingly, the strongest inhibitor (800‐fold stronger than a monovalent analogue) was just a tetravalent iminosugar present on a porphyrin‐based scaffold. The authors suggested that the valency itself did not cause an increased inhibition potential, but rather the spatial conformation of monosaccharides did. 265 Moreno‐Clavio et al. tested several l ‐fucosidase iminosugar‐based inhibitors with less success. 266 Only a sevenfold‐higher inhibition effect on α‐ l ‐fucosidases was observed when a trivalent scaffold was used compared to a monovalent one. 266 On the other side, a relative α‐mannosidase inhibition potency as high as 3000 was reported for the N ‐alkyl analogue of 1‐deoxynojirimycin conjugated on a synthetic glycoprotein scaffold (Fig. 17 ), 267 almost 10,000 for the 21‐valent 1‐deoxynojirimycin scaffold 268 and higher than 500 for glycocyloclopeptide with seven copies of 1‐deoxynojirimycin. 269 In the latter case, the large inhibition potential was obtained only on glycoconjugates with a longer (C 9 ) spacer, while C 6 spacers of the same glycoconjugate resulted only in a relative inhibition potency of 42. 269 Figure 17 1‐Deoxynojirimycin‐based glycopolypeptides: synthesis, self‐assembly, and glycosidase inhibition model. CuAAC = Cu I catalyzed Huisgen azide‐alkyne cycloaddition. Reproduced from 267 with permission of the Royal Society of Chemistry. The real therapeutic effect of different 3‐ to 14‐valent iminosugar conjugates was investigated by Joosten et al. through the in vitro measurement of β‐glucocerebrosidase activities in N370S Gaucher fibroblasts. 270 The authors reported a "mild, but significant" effect of multivalency. Importantly, the highest inhibition rate was not correlated with the highest chaperoning (therapeutic) effect. 270 This report clearly suggests that similar systematic investigations will be necessary to develop efficient NP‐based drugs against lysosome storage diseases. A similar approach applied for the treatment of lysosome storage diseases was also investigated for the treatment of cystic fibrosis. 1‐Deoxynojirimycin and its derivatives correct a misfolded cystic fibrosis transmembrane conductance regulator protein in a similar manner as the aforementioned chaperone effect of iminosugars towards glycosidases, with a 1000‐fold increased inhibition activity of 3‐valent iminosugar compared to a monovalent analogue control. 271 The concept of pharmacological chaperones may sound very promising, but although the activation of glycosidases has been observed, it hardly reached therapeutically sufficient levels, with typically an approximately twofold increase in the activity. 272 From this point of view, the work of Brissonnet et al. is noteworthy. 273 These authors reported an activity increase as high as 70‐fold of a bacterial mannoside‐phosphorylase using cyclodextran‐based NPs displaying deoxymannojirimycin (more than 100‐valent). Nevertheless, further studies are needed to elucidate the nature of the enzyme activation gain. 273 The selective interaction of a disaccharide called Thomas Friedrich antigen (TF ag ) with specific tumor cells displaying Gal‐3 lectin had a cytotoxic effect on the targeted cells: a conjugate of small AuPs coated with TF ag via an amino acid linker was synthesized, and an approximately 100‐fold higher cytotoxicity toward the Gal‐3‐positive cells was achieved compared to monomeric units. 274 The main cytotoxic effect of the TF ag relied on apoptosis induced by an inhibited Gal‐3 signaling, 274 but it is anticipated that the TF ag targeting may also work well in drug delivery. Another glycan with a therapeutic effect is heparan sulfate, a glycosaminoglycan‐based polysaccharide capable of inhibiting protease β‐secretase responsible for the accumulation of plaques causing Alzheimer's disease. Nevertheless, carbohydrate synthesis is rather complicated and expensive. Therefore, a mini library of dendritic polymers bearing multiple copies of different heparan sulfate monomers was recently introduced. 275 Some of these conjugates were found to possess inhibition potential equivalent to that of previously tested oligomers, which is an important result for the future development and production of anti‐Alzheimer's disease drugs. 275 6. CELL/TISSUE TARGETING While the previous section covered the therapeutic effects of NPs displaying carbohydrate moieties on their surface, there are even more studies investigating glycoconjugates where a carbohydrate moiety plays a different role—targeting specific tissues or cells based on the selectivity of protein–carbohydrate interactions. This selective targeting has been mostly used for the delivery of (noncarbohydrate) drugs, particles, or compounds suitable for thermodynamic therapy or fluorophores/particles, allowing selective tissue/cell imaging using spectrophotometric and other methods (e.g., magnetic resonance). In fact, selective cell imaging is based on efficient selective targeting; thus, in many studies, both aspects are overlapped. A. Drug Delivery From a material point of view, diverse metallic NPs, carbon NPs, and SiNPs conjugated with carbohydrate moieties and drugs to be delivered have been reported in the past. A recent trend (since 2013), however, is the application of polymeric scaffolds or biopolymers, that is, either abundant polysaccharides based on glucose, chitosan, dextrans, or pullulans or protein‐based scaffolds for such purposes. 276 , 277 , 278 Saccharides were applied for the selective internalization of NPs by targeted cells due to receptors displayed on their surfaces with subsequent intracellular drug release. Saccharides were also used to substantially suppress adverse side effects of the treatment by the improved biocompatibility of otherwise hydrophobic, insoluble, or instable substances. After targeted delivery, drugs must be released from the conjugate to perform their therapeutic effect. Most nanocarriers have been prepared as pH‐sensitive conjugates, taking advantage of different pH values inside many tumor cells, where NP degradation occurs with subsequent drug release. Enzymatic nanocarrier degradation has also been described, as well as the degradation of the conjugate triggered by the presence of specific small molecules (glucose, glutathione, etc.). Irradiation‐sensitive nanocarriers have also been described where on‐demand and "remotely controlled" drug release occurred after near‐infrared (NIR) or UV irradiation. 279 In addition to a "model" drug (doxorubicin being one of the most frequently loaded cargoes in nanocarrier studies) or therapeutic proteins, carbohydrate‐based nanocarriers have also been used as nonviral vectors. 277 1. Carbon Nanomaterials Carbon NPs can be divided into four main groups regarding their dimensionality, that is, 0D fullerenes and spherical NPs, 1D nanotubes and nanofibers, 2D graphene sheets and 3D cryogels, aerogels, or foams. Recently, Li et al. compared of all types of carbon NPs for the simple encapsulation of an enzyme with its subsequent intracellular release. 280 While three‐dimensional graphene aerogel was suitable for the long‐term release of a cargo (3% in 24 hr), oxidized CNTs and GO provided fast cargo release (15–20% in 4 hr). 280 Without any doubts, a comparison of different types of NPs is needed for the development of ultraefficient drug carriers. Unfortunately, there is no recent comparative study focusing on the performance of different types of NPs modified with glycans applied in drug delivery. Many recent studies have explored the potential of graphene‐based NPs, that is, GO 281 , 282 , 283 or rGO. 284 , 285 Flakes of the former form of graphene derivative display many oxygen groups, which make GO hydrophilic and easily dispersible in aqueous solutions. 286 Furthermore, surface‐displayed oxygen moieties can be used for the covalent conjugation of various ligands. Most recent studies have investigated only slightly different variations of a basic scheme relying on GO nanosheets decorated with hyaluronic acid (HA) securing the selective binding of nanocarriers by CD44 receptors present on the surface of various kinds of tumor cells. Song et al. described the improved solubility of doxorubicin adsorbed on GO sheets and the "shielding" of a GO's surface charge by increasing its hydrophobicity. 281 In this study, a complex of GO‐doxorubicin with modified HA was reported to occur via H‐bonds between HA's amine and GO's epoxide group. 281 In another study, HA was covalently grafted onto the GO surface via adipic acid dihydrazide. 287 This GO–HA conjugate was treated with doxorubicin solution, and the drug was adsorbed onto a GO surface, most likely by hydrophobic and π–π stacking interactions. 287 HA could be attached to a GO surface via introduced sulfhydryl groups, forming disulfide bonds that were cleaved by glutathione inside targeted cells. 288 A similar approach was tested by Yang et al., who modified the GO surface with carboxymethylated chitosan. 289 Amine groups present in chitosan were used for covalent linkage to GO's hydroxyls and helped to sequentially attach HA with the subsequent loading of doxorubicin. 289 All of the authors declared the in vivo inhibition of tumor growth after injection of the prepared conjugates; however, the inhibition rate hardly exceeded 50% compared to the control. Thus, such relatively low efficiency of cancer growth inhibition complicates the further commercial exploration of doxorubicin‐based delivery systems. Interestingly, an almost 80% decrease in the size of epidermal tumors in mice was observed 28 days after the injection of the conjugate prepared from rGO (Fig. 18 ). 284 Most probably, higher hydrophobicity of rGO compared to GO allowed the adsorption of a higher amount of doxorubicin. Xu et al. combined the GO‐HA‐doxorubicin delivery system with a photothermal therapy by wrapping AuNPs in GO sheets with the subsequent loading of HA and doxorubicin. 290 Drug‐induced cytotoxicity based on a nanoconjugate was comparable to a treatment with the same concentration of unconjugated doxorubicin, but upon laser irradiation, the increased temperature of AuNPs decreased the cell viability of hepatoma cell line Huh7 from 44 to 18%. 290 Figure 18 Schematic depiction of reduced graphene oxide (black support) conjugated with cholesteryl (green), doxorubicin (red), and hyaluronic acid (blue) including hydrophobic and π–π stacking interactions. Reprinted from 284 . Copyright 2013, with permission from Elsevier. GO hydrogel modified with heparin‐mimicking sodium styrene sulfonate was perfectly biocompatible with blood. Such a nanoconjugate could have a broad range of therapeutic applications, including drug delivery. 291 Nevertheless, only a nonselective antimicrobial activity of particles loaded with gentamycin sulfate was tested. 291 CNTs offer similar surface properties as graphene‐based NPs, and their complexes with HA were also tested for doxorubicin delivery. However, when the drug was complexed with CNTs covalently grafted with PEI and HA, an in vitro study revealed a quite similar viability (approx. 50%) of model tumor cells treated with conjugated or unconjugated doxorubicin (24 hr, ∼4 μM doxorubicin). 292 More promising results were obtained with covalently grafted polyethylene glycol derivative to CNTs, crosslinked to a gel form by adding β‐cyclodextrin and treated with camptothecin (a hydrophobic cytotoxic drug). 293 An in vitro study showed the significant growth inhibition of a cancer cell line (∼80% inhibition after 7 days compared to cells treated with blank CNT‐polymer gel). The gelation process relayed on the natural ability of cyclodextrins to include certain molecules (PEG moieties, in this case) into their cavity via host–guest interaction. Such an absorption capacity also has great potential for hydrophobic drug loading. Furthermore, the feasibility of precise chemical modification makes cyclodextrins a great tool not only in smart drug development but generally in the molecular fabrication of diverse mechanisms. 293 These possibilities have been reviewed, for example, by Martínez et al. 294 In addition to carbon, boron nitride nanotubes have also been exploited for drug delivery. Such nanotubes with covalently grafted glycosylated chitosan exhibited low cytotoxicity; therefore, such NPs could be used as nonviral vectors. 295 A plasmid gene was more efficiently expressed in human carcinoma epithelial cells (A549) than in control cell lines even though these NPs did not exhibit any specific receptor ligand selectivity. 295 Enhanced plasmid expression could be assigned to more efficient endocytosis rather than to the presence of specific receptors in A549 cells. 295 An intriguing approach was reported by Zhou et al. by loading doxorubicin into pores of mesoporous carbon NPs. 296 These NPs were subsequently grafted by HA, which guided selective drug targeting. Furthermore, HA acted as a "gatekeeper" because it was specifically recognized and degraded by hyaluronidase‐1 after internalization and allowed drug release. In addition to enzymatic degradation, HA coating could also be removed by glutathione due to the nature of a covalent bond between HA and NPs, making drug release more efficient and controllable. The authors reported an IC 50 of 8.9 μM for doxorubicin carried by the nanocarrier. 296 2. Metal and Metal Oxide NP‐Glycan Conjugates for Drug Delivery AgNPs able to generate oxidative stress have strong antimicrobial properties. Their modification with glycans represents very simple carbohydrate‐guided therapy applicable in bactericidal surface fabrication (vide supra) and possibly also in cancer treatment. Kennedy et al. revealed that AgNPs modified with a monosaccharide galactose were internalized by model neural and cancer cells more efficiently than mannose and glucose conjugates, but interestingly, the former exhibited lower cell toxicity (approx. fourfold lower EC 50 against model cancer cells) compared to less‐internalized glycoconjugates, suggesting an important role of terminal saccharide in glycan moieties. 297 Similarly, Shahbazi et al. observed a significant level of cytotoxicity (approx. 80%) when the HT1080 sarcoma cell line was treated with iron oxide NPs (100 or 200 nm in diameter) coated with glucose, but a similar effect was observed with NPs coated by polyethylene glycol, suggesting the low targeting capability of such conjugates. 298 Apparently, AuNPs have not been considered preferential drug carriers, as discussed in a previous section, despite the fact that there are studies reporting the simple decoration of AuNPs with carbohydrate‐based targeting moieties (mannose modified 108 ) and even loaded with therapeutic agents. The latter was achieved by the conjugation of moieties capable of enhanced drug ligation to AuNPs, with cyclodextrins being very sound examples of such approach. Cyclodextrins (and their derivatives 299 ) can host hydrophobic, insoluble, drugs inside their structure 294 and can be easily conjugated to AuNPs. 300 Alternatively, AuNPs can be conjugated directly with cytotoxic ligands, for example, gold(I) triphenylphosphine. 301 On the other side, similar to the above‐mentioned carbon NP‐based nanocarriers, HA derivative coated on gold nanocages was used as a vehicle for the targeted delivery of doxorubicin into CD44 receptor‐expressing cells. 302 The intracellular hyaluronase‐1‐induced degradation of the conjugate resulted in drug release and selective cytotoxicity. In contrast to carbon NPs, gold nanocages could also be applied for thermodynamic therapy. This synergy of cancer cell treatment led to the complete elimination of a solid tumor in mice in 9 days by combined photo‐ and chemotherapy. 302 Terpen (borneol) stabilized selenium NPs decorated with glucose were also reported to deliver an anticancer drug selectively into Hep2 cells (a hepatocarcinomic cell line). 303 In addition to an efficient pH‐sensitive system of drug release from the conjugate, an intracellular, selenium NP‐directed production of reactive oxygen species (ROS) was also observed, making the fabricated glycoconjugate cytotoxic for drug resistant hepatocarcinomic cells. 303 Similarly, TiO 2 NPs coated with HA and loaded with cisplatin containing ligand were found to carry their payload selectively into an ovarian cell line (A2780) with pH‐induced drug release from these glycoconjugates. 304 Nevertheless, only an approximately twofold increased drug‐induced cytotoxicity against model cancer cells was observed upon treatment with conjugated versus unconjugated drug (a decrease to approx. 40 vs. 70% of the original viability, respectively, using 10 μM drug). 304 Polysaccharide carboxymethyl dextran was also used as a building block for the synthesis of magnetoliposomes loaded with doxorubicin. 305 Drug release was enhanced upon the exposure of the conjugates to a low‐frequency magnetic field in synergy with low pH induced structural changes leading to drug release. In vitro experiments confirmed the low cytotoxicity of the loaded magnetoliposomes to brain cells, while after incubation with 2.5 μg mL −1 of conjugated and free doxorubicin equivalents, the viability decreased to approx. 40 and 70% of original viability, respectively. 305 3. Silica‐Based Glycan‐Targeted Drug Carrier Systems In a rather "traditional" approach, small SiNPs were covalently modified with spacers; each one ended with trivalent glucose as a targeting unit. 306 The spacers provided a space for the inclusion of molecules of the anticancer drug paclitaxel. The drug delivery system was the same as described before, that is, whole conjugates were selectively internalized by cancer cell line HepG2, and the drug was released either by intracellular esterases or by low pH. Due to the good pharmacokinetics of the fabricated conjugate, a much higher toxicity (IC 50 = 0.7 μM paclitaxel equivalent) was induced than with a treatment by the unconjugated drug upon prolonged incubation (no viability change using 0.7 μM paclitaxel). 306 Polysaccharide moieties were also used to increase the biocompatibility of a mesoporous silica doxorubicin carrier. 307 Yu et al. adapted an already developed protocol to prepare HA‐coated mesoporous SiNPs transporting doxorubicin selectively into human colon cancer cells. 308 A decrease of targeted cell viability by approximately 50 and 20%, respectively, was observed upon incubation with 0.25 μM drug equivalent carried by HA‐coated NPs and uncoated NPs or free drug. 308 HA‐targeting mesoporous SiNPs with both a therapeutic agent and HA attached to NP's surface by disulfide bonds were investigated by Zhao et al. 309 Disulfide bonds were cleaved by glutathione, causing efficient drug release (Fig. 19 ). Cytotoxicity assessment showed that after 72 hr of treatment with the drug‐loaded conjugate, a cell line with CD44 receptors retained a viability of 39%, while the viability of a control cell line without CD44 was 49%, suggesting rather strong nonspecific internalization by the NIH 3T3 cell line. 309 The same group also prepared mesoporous SiNPs with pores loaded with doxorubicin and capped by HA attached to the surface by disulfide bonds. 310 Excess glutathione and hyaluronidases helped to release capping molecules from the conjugate. 310 An elegant approach using nonspecific internalization based on differences between the metabolic activity of normal and cancer cells was reported. 311 Cancer cells exhibit increased glucose uptake with the enhanced uptake of synthesized glucose‐modified SiNPs. Celastrol used as an anticancer drug negatively affects the mitotic cycle, which is faster in cancer than in normal cells, thus making cancer cells more susceptible to the effect of the drug. Mesoporous SiNPs decorated with glucose via dendritic polymer linkers were found to induce the apoptosis of HeLa cells (50% apoptotic cells after 24 hr of incubation with 5.3 μM drug equivalent). On the other hand, a lower exposure time induced the production of heat shock proteins, suggesting that a subcytotoxic dosage (or exposure time) may even help to protect normal cells. 311 Figure 19 Schematic depiction of glutathione (GSH) induced drug release from hyaluronic acid (HA) capped pores of silica NPs. These pores are filled with disulfide‐bond tethered 6‐mercaptopurine (6‐MP) and, due to HA‐decoration, nanoparticles are captured and internalized via so called CD‐44 receptors—a common surface glycoprotein with significantly higher concentrations on the surface of certain cancer cells. Reprinted with permission from 309 . Copyright 2014 American Chemical Society. Carbohydrate modification can also be used for the selective and triggered drug release from mesoporous magnetic SiNPs. 312 Lectin‐like phenylboronic acid (PBA) displayed on their surface anchored dextran, which efficiently gated drug‐loaded pores in the body of magnetic SiNPs. Once glucose is present in sufficiently high concentration, it competitively binds to PBA, and the dextran "gatekeeper" layer is consequently released from the surface. Due to the much smaller size of glucose compared to dextran chains, drug molecules can easily migrate from the pores. This glucose‐dependent release has great potential, for example, in diabetes treatment, even though a selective cytotoxicity against HeLa cells in the presence of glucose was demonstrated in this study (60% decrease of viability compared to a 25% decrease upon treatment of the cells with unloaded and drug‐loaded NPs without dextran, respectively). 312 Employing the same gate‐keeping approach, Zhou et al. bound mannose units on the surface of mesoporous SiNPs via disulfide bonds. 313 These moieties tethered Con A lectin, a molecule large enough to prevent the migration of doxorubicin from the pores of mesoporous SiNPs. Selective release was triggered by the elevated concentration of glutathione in the cytoplasm of cancer cells, and the conjugate exhibited an IC 50 of 25 μg mL −1 against cancer cells, while a negligible viability change was observed when the conjugate was incubated with normal cells having a lower cytoplasmic glutathione concentration. 313 Chondroitin sulfate is another natural polysaccharide with the ability to promote the uptake of conjugates by selective interaction with CD44 receptors. Moreover, such a biopolymer has the ability to improve the biocompatibility and stability of NPs. All of these features of chondroitin sulfate were used to fabricate a mesoporous silica‐based vector loaded with a plasmid carrying the gene for the p53 protein. 314 A more than twofold increase in translated messenger RNA for p53 was observed in treated cancer cell lines, when the silica‐based vector was coated with chondroitin sulfate compared to bare mesoporous silica–plasmid conjugate. 314 Gene therapy induced expression of healthy p53 protein in human cancer cells inhibits tumor growth; thus, the above‐described vector represents another possible way for cancer treatment. 4. Polymer Scaffold Based Drug Carriers Liposomes prepared from a synthetic building block containing a hydrophobic tail conjugated with galactose as targeting units were capable of transporting doxorubicin easily into HepG2 cells displaying asialoglycoprotein 315 , 316 , 317 and other receptors. 44 A very similar system, only with incorporated disulfide bonds in the structure of micelles, was reported. 318 , 319 , 320 , 321 Such glycoconjugates were selectively internalized, and a drug was released after the glutathione‐mediated cleavage of S–S bonds. 318 , 319 , 320 , 321 Although these studies demonstrated the selective cytotoxicity toward targeted cancer cells in vitro, only Zhao et al. have reported the in vivo inhibition of tumor growth after the localized injection of galactosylated, doxorubicin‐loaded liposomes (tumor weight of approx. 50% compared to animals treated with an unconjugated drug). 316 The unfolding of synthetic 322 or natural 323 micelles triggered by UV irradiation for drug release was also reported, but such an approach was not tested for in vitro or in vivo cytotoxicity. In order to better control the form and chemical properties of nanocarriers, a tri‐block copolymer folded into different types of mannosylated micelles was synthesized. 324 Preliminary results suggested that there was some correlation between mannose receptor‐induced internalization and the shape of a micelle, but in vivo or in vitro studies investigating the inhibition of virus binding have yet to be conducted. 324 To discuss drug delivery from an application point of view, it is worth mentioning the use of a synthetic copolymer folded into micelles tested for the oral administration of sorafenib—the drug applied mainly for the treatment of kidney and liver cancers. 325 In vivo studies have demonstrated that galactosylated, drug‐loaded micelles efficiently and selectively administered sorafenib into the liver and secured its continuous release for several hours. 325 To improve the oral administration of darunavir, self‐assembled micelles made from synthetic copolymer poly(ethylene oxide)‐poly(propylene oxide) were glycosylated by microwave‐assisted ring opening reaction and loaded with the drug by a simple incubation of the two components in solution. 326 An approximately 500‐fold higher solubility of darunavir was observed upon its integration into micelles. This is a significant step in the development of less expensive anti‐HIV drugs, considering the fact that the low administration efficiency of present darunavir‐based delivery increases its cost. 326 Berberine hydrochloride is another drug with a reported anticancer effect with chemical properties not allowing oral administration. 327 Drug encapsulation into chitosan‐coated phospholipid micelles significantly improved pharmacokinetics and distribution parameters due to the increased biocompatibility of the conjugate. It should be noted, however, that only the biodistribution and drug release profile were assessed in the study, with results suggesting that chitosan‐coated liposomes are good vehicles for the oral administration of berberin. 327 Micelles prepared by the conjugation of α‐tocopherol (recognized as a potential anticancer drug) with pullulan were reported to release the drug 10‐hydroxycamptothecin at the low pH typically present in the cytoplasm of tumor cells. 328 In vitro cell assays revealed an approximately 60% decreased viability of targeted cells compared to a control experiment with another cell line. 328 A conjugate based on α‐tocopherol derivatives together with HA and cytostatic docetaxel in the micellar form was found to induce cytotoxicity and to decrease multidrug resistance via the α‐tocopherol‐mediated inhibition of the active drug transport from the cell. 329 The suppression of drug resistance on a molecular level was also achieved by the delivery of two recently synthesized inhibitors of efflux proteins (P‐glycoprotein and breast cancer resistance protein) loaded in poly‐lactic‐co‐glycolic‐acid. 330 A higher inhibition level of cancer cells incubated with conjugated compared to unconjugated inhibitors was demonstrated. 330 Interestingly, the same efflux mechanism was also suppressed by hydroxypropyl‐β‐cyclodextrin moieties grafted onto synthetic copolymer NPs, which enhance the effect of low‐soluble immunosuppressor tacrolimus by oral administration. 299 Atherosclerosis can be efficiently treated by the internalization of micelles prepared from a reconstituted high‐density lipoprotein (rHDL) loaded with levostatin mediated via Scavenger Receptor class B member 1 (SR‐BI) receptors accessible on the surface of cells present in a macrophage‐infiltrated atherosclerosis lesion. 331 Upon coating of the fabricated micelles with HA (Fig. 20 ), their administration became more efficient and precisely targeted in two steps: (i) because of their lower accumulation in the liver, this is achieved because HA‐decorated particles are not recognized by SR‐BI receptors and "scavenged" (recognition by SR‐BI receptors in liver cells leads to an increased level of phagocytosis). Thus, more particles are delivered into atherosclerotic veins. In the second stage, (ii) HA was readily recognized by CD44 receptors present in atherosclerotic plagues, where HA‐LT‐rHDL was accumulated. Next, due to the presence of hyaluronidases, HA was removed. The formed LT‐rHDL NPs were recognized by SR‐BI receptors of atherosclerosis lesion cells, which delivered them into desired "foam cells" where levostatin was released. 331 Figure 20 Scheme of fabrication and function of dual‐targeted reconstituted liposomes. Upper scheme represents preparation of micelles containing levostatin and coated with HA (HA‐LT‐rHDL). After injection in an atherosclerotic rabbit, HA coating allowed preferential distribution of HA‐LT‐rHDL into veins and their retention in an organism with decreased rate of phagocytosis in liver. Once HA‐LT‐rHDL reached a lesion site, it was recognized by CD44 receptors on the surface of endothelial cells which helped their transportation inside an atherosclerotic plague (the scheme in the lower part of the figure). In the plague, hyaluronidases removed HA from HA‐LT‐rHDL and the resulting LT‐rHDLs were preferentially recognized by SR‐BI receptors on surface of "foam cells" where lovastatin was intended to be released. Reprinted from 331 . Copyright 2014, with permission from Elsevier. The antiseptic drug NK007 was delivered successfully by oral application to regress murine colitis. 332 Polysaccharide components of micelles are responsible for efficient target delivery as could be concluded from the in vivo study. Nevertheless, a significant difference between the parameters recorded to evaluate treatment efficacy (i.e., colon length, stool consistency, etc.) was only observed between the control and treated groups but not between groups treated with unconjugated or NP‐conjugated NK007. 332 Amphotericin B, an antiparasitic drug effective, for example, against Leishmania donovani , induces several harmful side effects, with nephrotoxicity and hemolytic toxicity being the most severe. 333 Moreover, the drug is relatively expensive. Khan et al. conjugated this drug with mannosylated synthetic dendrimer to gain biocompatibility. As a result, no kidney damage or hemolysis was observed with the antiparasitic activity of the drug preserved. In fact, due to the mannose‐driven targeting of conjugates into macrophages, IC 50 (39 nM) decreased approximately fivefold compared to the unconjugated drug or its commercial analogue. 333 A supramolecular scaffold prepared from a polymer bearing PBA and a "complementary" glycopolymer was degraded under hyperglycaemic conditions. 334 The conjugate could be applied for the safe, comfortable, and glucose concentration‐sensitive distribution of insulin. 335 Zheng et al. also employed a PBA‐based conjugate with both PBA and lactobionic acid moieties displayed randomly on each copolymer chain. 334 Such molecules folded into a scaffold able to encapsulate insulin. PBA enhanced the mucoadhesion of the scaffolds, which could be then used for the nasal administration of insulin. Therapeutic effect was confirmed by in vivo experiments in diabetic mice with glucose levels decreased to approximately 40% 8 hr after the addition of therapeutic NPs, while such a decrease was observed within 0.5 hr when bare insulin was injected. 334 The single‐step ligation of HA with doxorubicin seems to be much simpler, forming conjugates that are selectively cytotoxic to cancer cells displaying CD44 receptors 336 similar to pullulan‐doxorubicin micelles selectively targeting cells with asialoglycoprotein receptors. 337 , 338 A similar weight reduction of induced tumor was observed in vivo after treatment with unconjugated and conjugated doxorubicin (approximately 0.4 g of tumor weight compared to 1.8 g developed in a control group), but the life span was longer and the animal weight was larger for NP‐treated mice compared to mice treated with unconjugated doxorubicin, suggesting the significant suppression of a negative side effect of a chemotherapy. 338 Plant glucomannan covalently conjugated with an inhibitor (bisphosphonate alendronate) of tumor‐associated macrophage was also prepared as a conjugate inducing selective, mannose receptor‐mediated macrophage apoptosis (a tumor weight of approx. 0.3 g was observed 14 days after the application of therapeutic NPs compared to 1.5 g developed in an untreated group). 339 Mannose receptors of macrophages were also targeted with ionic polymer‐based micelles containing the model protein to be delivered (BSA). 340 Upon coating of the micelles with β‐glucan from yeast cell walls, NPs were selectively internalized by macrophages. 340 So‐called "ivy NPs" are nature‐derived nanocarriers consisting of a heavily glycosylated protein core. 341 When loaded with doxorubicin, a targeted delivery and release of a drug into different cancer cell lines was observed. 341 The treatment of a solid tumor was equally efficient for the application of both the conjugate and unconjugated doxorubicin, but with a significantly lower weight loss of the animal observed after the application of the conjugate. 341 Quite a "minimalistic" approach was reported, relying on short polyethylene glycol chains decorated with several saccharidic units along their length and with a bioactive molecule on the chain end. 342 Such conjugates bearing two separated mannosyl units (the optimum distance between these units was found to be 5.6 nm) were internalized by macrophage cells via their mannose‐binding receptors. The authors declared a good biocompatibility of the carriers as demonstrated by the absence of lysosome storage disease‐like effect after their application. 342 On the other side, in vivo tests of applicability of such conjugates have yet to be performed. Similarly, Thomas et al. reported micelles prepared from bis‐ l ‐galactose lysine. 343 Galactose displayed on the surface of these NPs was enzymatically oxidized, and the aldehydes formed were used for the covalent crosslinking of more NPs together by di‐ or trivalent amine. The authors tested their delivery system only for the complexation of AuNPs, but selective delivery of drugs is anticipated. 343 A selective drug delivery into the central nervous system, passing the blood–brain barrier, is always very difficult. Recently, certain lectins displayed on NPs helped to enhance such crossing. 344 For example, polyethylene glycol based micelles decorated with lectin Solanum tuberosum were able to successfully deliver a basic fibroblast growth factor via the nasal route into the brain in order to improve the cognition of those affected by Alzheimer's disease. 344 Lactoferrin, an iron‐binding glycoprotein, was found to be helpful in delivering drug‐loaded micelles through the blood–brain barrier 345 , and the same protein was used to target asialoglycoprotein receptors on HepG2 cancer cells, as well. 346 Another way to pass the blood–brain barrier is using activated membrane transporters. Glucose transporter 1 was activated by p‐aminophenyl‐β‐ d ‐manno‐pyranoside with a subsequent transportation of anticancer drugs into the brain cells. 347 Kuo et al. prepared NPs based on solid lipid decorated with p‐aminophenyl‐β‐ d ‐manno‐pyranoside and folic acid and loaded with etoposide (a drug inhibiting the proliferation of malignant glioblastoma) that were efficiently distributed through the blood–brain barrier with efficient secondary targeting to glioblastoma cells secured by folic acid in in vitro experiments. 347 Nonviral gene transport is just another field where the aforementioned techniques of controlled targeted drug delivery have been employed. One of the materials used quite often for the preparation of nonviral vectors is PEI, which can be effectively complexed with the DNA or RNA of interest, but it possesses a certain level of toxicity and its targeting is not selective and thus suitable only for in vitro experiments. In order to apply PEI‐based conjugates for in vivo applications, several ways to improve the performance of such carriers were sought. For example, PEI was complexed with a biocompatible anionic glucose‐based glycopolymer, 348 GlcNAc, 349 chondroitin sulfate, 350 HA, 351 or depolymerized guar gum with available mannose units 352 with subsequent use in selective targeting. Alternatively, copolymers of the methacrylamide backbone decorated with glucose, 353 α‐ d ‐mannopyranosyl, 354 or trehalose 355 and cationic moieties could electrostatically complex plasmids with the subsequent selective transport of plasmid DNA into cancer cells 353 , 354 or siRNA into glioblastoma cells via glucose transporter‐1. 355 Furthermore, complexed siRNA was reported to retain its biological activity after freeze drying. 355 An electrostatic complexation of DNA with copolymer chains bearing a targeting GlcNAc moiety was also reported. 356 Targeted gene delivery was also achieved by DNA incorporation into PEGylated micelles decomposing at low pH and decorated with galactose 357 or mannose 358 to target asialoglycoprotein or mannose receptors, respectively. Noteworthy, a secondary target ligand was co‐entrapped with DNA to help penetrate through the nuclear membrane. 357 Mannose‐decorated cholesteryl‐based synthetic liposomes with virus‐like characteristics were reported to effectively transfect nonactivated dendritic cells by a plasmid‐containing gene for luciferase. 359 Mannosylated chitosan was also used for nonviral vector preparation 360 as well as mannosylated "bubble lipoplexes," that is, NPs with ultrasound‐triggered DNA release causing the transfection of M2 macrophages and their switch from tumor growth‐promoting into tumoricidal M1‐like macrophages. 361 , 362 In vivo experiments confirmed that this treatment could inhibit the growth of various tumors without adverse side effects. 361 , 362 A triblock copolymer consisting of hydrophobic and pH‐sensitive parts to assure the formation of micelles, polycation complexing oligonucleotides, and a mannose‐modified part was also applied to target M2 macrophages. 363 Synthetic vectors based on siRNA‐PEGylated cyclodextrin scaffolds with a peptide ligand as a targeting unit were also investigated. 364 In addition to the broadly used electrostatic interactions used for loading nucleotides into delivery NPs, Kim et al. employed hybridization with a specific DNA template with a condensation of the resulted DNA complex by a viral Mu peptide (Fig. 21 ). 365 The formed "DNA nanoballs" coated with HA were effectively internalized by cancer cells. In vivo experiments showed that antisense oligonucleotides were hybridized with their complementary mRNA in the cytoplasm, leading to a significant sensitization of the cells to doxorubicin treatment and to a significant decrease of tumor growth in mice. 365 Figure 21 RCA template and hybridization efficiency. (A) Secondary structure of scrambled rolling circle amplification (RCA) template. (B) Secondary structure of a dual antisense oligonucleotide (ASO) hybridizing RCA template for Dz‐13 and OGX‐427. (C) RCA products with poly ASO‐binding sequences were hybridized with two therapeutic ASOs, Dz‐13 and OGX‐427, to produce dual ASO‐hybridizing RCA products. (D) The hybridization efficiencies of the ASOs were tested for products of scrambled RCA templates and dual ASO‐hybridizing RCA templates using fluorescently labeled ASOs. Reprinted from 365 . Copyright 2015, with permission from Elsevier. A very simple approach relying only on the application of bare therapeutic oligonucleotides conjugated with a targeting ligand (e.g., a triantennary GalNAc 366 , 367 , 368 or glucose 369 ) was also tested. These agents were capable of blocking the adverse immunostimulation of an anti‐HIV oligonucleotide. This may be helpful in the fabrication of anti‐HIV drugs without adverse side effects. The conjugate can be easily prepared, but there is a concern regarding the stability of the carrier when exposed to endo‐ and exonucleases. B. Photothermal and Photodynamic Therapy Photothermal and photodynamic therapeutic methods rely on the targeted delivery of entities capable of strong energy absorption in the NIR spectrum. As a result, the NIR‐triggered increase of local temperature damages targeted (cancer) cells. Traditionally, metallic NPs have been used as photothermal agents, with AuNPs being the first choice. For example, HA‐covered gold nanostars very efficiently inhibit HeLa tumor growth in mice with almost complete tumor disappearance within 7 days after injection of the conjugate and laser irradiation. 370 The recent trend, however, is the use of a combination of simple photothermal therapy with other treatments, for example, chemotherapy. AuNRs wrapped in GO sheets conjugated with doxorubicin were reported to be fourfold‐stronger inducers of HeLa cell death compared to the same conjugates without doxorubicin used for sole photothermal treatment. 290 In addition to AuNPs, carbon mesoporous NPs were also employed simultaneously as carriers and photodynamic agents. 296 A synergic chemotherapeutic and photothermal effect was observed after loading their pores with doxorubicin and final coating with HA. The observed IC 50 against the MDA‐MB‐231 cell line decreased from 8.7 to 2.3 μM doxorubicin after simultaneous treatment with the drug‐loaded NPs and NIR laser. 296 A very simple but effective system was reported based on the electrostatic integration of HA and polyaniline applied as a targeting and photosensitizing component, respectively. 371 The use of these NPs with subsequent NIR laser treatment led to the complete degradation of induced tumors in mice approximately 8 days after treatment without any observable body weight loss. 371 However, the ternary design of multimodal therapeutic NPs was presented, containing poly(lactic acid) NPs coated with derivatized chitosan to increase the biocompatibility of the conjugate and IR 820 (a photosensitive dye). 372 Interestingly, two absorption peaks were found for IR 820 encapsulated in the polymer; thus, irradiation with two different wavelengths increased the photothermal effect with the decreased viability of the targeted breast cancer cell line by 70%. 372 In another treatment design called photodynamic therapy, photosensitizers were used to generate cytotoxic ROS upon the absorption of irradiation with a specific wavelength. Recently, AuNPs coated with HA (a targeting component) and a porphyrin‐based photosensitizer were reported to decrease the viability of cancer cells with CD44 receptors down to approximately 15% upon irradiation with UV light. 373 AuNPs can also exhibit a photosensitizing effect, but for the generation of cytotoxic radicals, X‐ray irradiation is required. In this sense, glucose‐decorated AuNPs were tested in vitro as selective sensitizers for radiotherapy, leading to increased cancer cell death by 20% compared to radiotherapy with unconjugated AuNPs. 374 Organic carriers for ROS‐generating photosensitizers were also reported. For example, porphyrin‐derived photosensitizer conjugated to glucose‐decorated poly(methacrylamide) polymer chains 375 or encapsulated into poly(lactide‐co‐glycolide)‐based NPs coated with HA. 376 In addition to porphyrin‐based photosensitizer, docetaxel was also co‐entrapped to induce a synergistic photochemotherapeutic effect with IC 50 of 8 ng mL −1 , while an IC 50 of 160 ng mL −1 was observed upon treatment with an unconjugated drug under the same conditions. 376 Fu et al. used chitosan‐coated SiNPs loaded with benzoyl peroxide as photosensitizers. 377 A cytotoxic test against human breast carcinoma cell line ZR75‐30 revealed that chitosan coating did not substantially affect cell viability (70 and 74% of retained viability after incubation with uncoated and coated NPs, respectively), even though the conjugation of benzoyl peroxide increased cytotoxicity (90 and 74% of viability retained after treatment with unconjugated vs. conjugated benzoyl peroxide, respectively). 377 GO was successfully applied as a carrier for the loading of photosensitizers, as reviewed recently. 378 Recently, HA was conjugated to GO nanosheets with the subsequent physisorption of a photosensitizer. 379 The efficient uptake of such modified GO NPs, with a size of 100 nm, was observed. The photosensitizer desorption led to enhanced NIR absorption and the generation of ROS in the cells. The authors reported IC 50 values of 1 and 0.1 μg mL −1 of photosensitizer equivalent, respectively, when targeted cells were treated with either free or NP‐conjugated photosensitizer. 379 A system combining both photodynamic and photothermal effects has also been developed. While an rGO/ZnO hybrid is responsible for ROS generation upon light irradiation, NIR laser irradiation induced local hyperthermia 322 because rGO effectively transferred heat to cells, with a local temperature increase upon exposure to NIR irradiation. 380 To secure targeted treatment, HA was conjugated with the rGO/ZnO complex. The function and structure of rGO/ZnO‐HA NPs are depicted in Figure 22 , where the viability of targeted cells is also depicted. From these plots, it is obvious that combined photodynamic and photothermal therapy decreased cell viability down to approximately 20%, much more than the application of only one type of treatment. 322 Figure 22 (A) Schematic illustration of sequential irradiation‐activated high‐performance apoptosis. (B) In vitro cell viability of MDA‐MB‐231 cells treated with rGO‐ZnO‐HA following PDT, PTT and combined PDT/PTT treatments. Reproduced from 322 , with permission by John Wiley & Sons. Targeted irradiation by neutron beam is yet another way to induce the cytotoxic effect of certain agents, such as 10 B ions using "boron neutron capture therapy." Radionuclide 10 B, after accumulating in the cells, captured supplied protons and underwent fission, generating heavily cytotoxic 7 Li and 4 He particles. 381 To achieve a sufficient amount of boron in cells (on average 10 11 10 B ions per cell), mesoporous SiNPs were used to carry 10 B coordinated to an organic ligand with high selectivity and an accumulation rate achieved by the decoration of the NPs with multivalent galactose units. The delivery efficiency was illustrated by having approximately 3 × 10 9 atoms in cells when treated with unconjugated boron ligand compared to 10 11 atoms when delivered by NPs. Preliminary results suggest that the NP‐delivered amount of 10 B atoms is high enough to efficiently perform boron neutron capture therapy. 381 C. Cell Imaging The principles applied to selective delivery can also be employed in cell imaging with appropriate carbohydrate‐binding receptors. Many studies have been published regarding the conjugation of glycan‐displaying NPs with dye molecules to facilitate detection in the UV‐VIS range, contrast enhancers for NMRI or computed tomography imaging of tissues or single cells accumulating nanocarriers. 382 1. UV‐VIS Probes The recent trend is to combine the therapeutic and imaging effects in so‐called theranostic (i.e., therapeutic + diagnostic) NPs loaded with both an imaging probe and a drug as discussed above. 44 , 381 , 383 The imaging function of the conjugate can also be combined with antiadhesive properties. 239 Noteworthy, NIR‐absorbing GO/doxorubicin/HA NPs for combined photochemotherapy could be applied to imaging even without any additional dye or label incorporated due to the fluorescent signal obtained from a carrier drug (e.g., doxorubicin). Such an approach was applied for the visualization of cells that internalized nanocarriers. 281 , 384 Dye‐loaded NPs releasing their cargo after internalization promoted by surface‐displayed carbohydrate moieties were reported. 385 The encapsulation of dyes into polymer NPs significantly increased the biocompatibility of imaging probes with the selective visualization of intracellular membranes of targeted cells (Fig. 23 ). 385 Self‐assembled NPs prepared from a natural polysaccharide levan were able to encapsulate indocyanine green. 386 The selective accumulation of the conjugate in tumor cells was mediated via overexpressed glucose receptors. 386 NPs formed from a copolymer containing fluorescent rhodamine and glucose were successfully internalized by cells with asialoglycoprotein receptors enabling their fluorescent imaging. 387 The chemical and enzymatic stability of the imaging NPs were very good, with a maximal release of only 2.5% of the loaded dye within 4 days in the presence of lipase and esterase. 387 Figure 23 Effective intracellular delivery of rhodamine B octadecyl ester in HDF cells mediated by (1:9 PGalSMA34 + PGMA51) PHPMA270 vesicles (prepared via thin film rehydration to ensure sterility and enable loading with a fluorescent dye). Cells were incubated with 1.0 mg/mL rhodamine B octadecyl ester loaded vesicles for 16 hr. (A) Confocal microscopy image of live HDF cells: note the intracellular staining of membranes (red) after exposure to the rhodamine‐loaded vesicles, cell nuclei are counter‐stained blue using Hoechst 33342. (B) HDF cells treated with the same vesicles containing no rhodamine dye (negative control). (C) Higher magnification image obtained for (A): effective intracellular delivery of rhodamine dye allows selective staining of the nuclear membrane (white arrows). Scale bar: 50 μm. Reprinted with permission from 385 . Copyright 2013 American Chemical Society. In addition to diagnosis, the dye labeling of NPs has also been used to evaluate the cell uptake rate. For example, the preparation of mannosylated micelles was optimized by the visualization of their uptake in the cells. 388 Further in vivo experiments confirmed their selective imaging capability. 388 Perylene bis(imide) derivative decorated with mannose was found to be fluorescent only after the degradation of the formed scaffold triggered by its ligation to a mannose‐specific lectin. 389 The fluorescent imaging of macrophage cells overexpressing such receptors has been reported (Fig. 24 ). 389 A novel NIR fluorophore activated upon isomerization at a low pH in lysosomes was used by Wu et al. 390 Biocompatible poly(styrene‐co‐maleic acid) NPs loaded with this probe specifically targeting surface‐displayed sialic acid were able to selectively tag and mark tumors in vivo. 390 Cell imaging using illumination outside the range of UV‐VIS and NIR has also been tested, for example, by Mäkilä et al., who radiolabeled ( 68 Ga) therapeutic siRNA oligonucleotides conjugated with multivalent galactose units to evaluate their in vivo distribution. 391 They observed that the valency of the displayed galactose was crucial to selective and efficient accumulation in liver, and the conjugate with seven glycan units exhibited the best performance, making them suitable for the further development of targeted genotherapy. 391 Figure 24 Confocal microscopic images of murine macrophage cells after incubation with water‐soluble glycocluster based on perylene bisimides PBI‐12‐Man (10 μg/mL, Man = mannose) without (a, excited at 559 nm; b, bright field; c, merging of photos a and b) or with mannose (18 mg/mL) inhibition (d, excited at 559 nm; e, bright field; f, merging of photos a and b) in PBS buffer at 37°C for 1 hr. As shown in a–c, the red fluorescence of PBI‐12‐Man was predominantly intracellular with punctate appearance, suggesting cell surface binding and endocytosis being the cell entry mechanism that results in vesicular (endosomal) localization. For the inhibition experiment, in the presence of α ‐d‐mannopyranoside (d–f), the remarkable decrease of the red fluorescence of PBI‐12‐Man indicated the selectively binding interactions of PBI‐12‐Man with the surface mannose receptor of the macrophage cells. Reprinted from 389 . Copyright 2014, with permission from Elsevier. Figure 25 Photoacoustic tumor imaging in mice with sialic acid‐decorated polymer nanoparticles (pNIR@P@SA). Nude mice bearing H22 subcutaneous tumors were intravenously injected with phosphate buffer saline (PBS; 100 mL) or pNIR@P@SA (40 mg kg −1 ). The mice were imaged 24 hr after vesicle injection. Control images were obtained from mice before intravenous injection of pNIR@P@SA or PBS. A significantly higher photoacoustic signal obviously resulted from tumor issue after incubation with pNIR@P@SA (lower right panel). Reproduced from 390 with permission of the Royal Society of Chemistry. Figure 26 CT number for tumor versus time. The blue line is the CT number of the control group, whereas the red line represents for the PEG Glc‐GNP injection group. Reprinted from 401 , with permission from Elsevier. Other recent studies reported the preparation of simple nanocarriers based on synthetic 392 , 393 or natural 394 , 395 polymer NPs conjugated with glycans and imaging probes with selective adhesion to specific cell surface receptors, enabling the visualization of the surfaces of these cells. A complex study was carried out to visualize the binding of radiolabeled mannose‐containing polymer (γ‐Tilmanocept, approved by the FDA as a lymphatic node imaging probe) to targeted cells. 396 The authors used a cyanine dye conjugated with Tilmanocept to specifically bind mannose receptors of macrophages, facilitating the delivery of a radiopharmaceutic drug. 396 Multivalent clusters of iminosugars conjugated with pyrene or boron‐dipyrromethene cores exhibited fluorescence spectra applicable in cell imaging. 397 As discussed earlier, multivalent iminosugars have been intensively tested as promising pharmacological chaperones. The conjugation of nanocarriers with photosensitizers applied in photodynamic therapy has also often been used for simultaneous therapy and imaging since the majority of molecules applied as photosensitizers exhibit fluorescence. For example, a porphyrin derivative conjugated with glycopolymer exhibited an emission peak at 633 nm, allowing the real‐time visualization of the NP uptake with confirmed targeted cytotoxic effect. 375 A similar effect was also observed with functionalized GO integrated into a pH‐sensitive HA nanogel serving as selective carrier for doxorubicin. 288 The conjugate based on GO allowed both photodynamic therapy and fluorescence imaging, 288 underlining the amazing versatility of graphene‐based nanomaterials. Deeper tissue penetration and better resolution are the main advantages of photoacoustic imaging. In this method, a probe with an absorption maximum at longer wavelengths is used. The absorption resulted in changes in the probe's properties (thermoelastic expansion); hence, after the activation by a deep‐penetrating laser pulse, the probe was transformed into a detectable ultrasound signal. 398 This imaging method was employed using sialic acid‐decorated polymer NPs loaded with a profluorophore, turning it into a photoacoustic probe by isomerization at a low lysosomal pH. 390 In the end, it became evident that a significantly higher photoacoustic signal was achieved in tumor vein imaging using this method—see Figure 25. 390 2. Metallic NPs in Selective Cell Imaging An inherent NIR‐induced fluorescence of small Au nanoclusters with the size of 2 nm was used in the preparation of selective imaging probes. Selectivity, stability, and biocompatibility were achieved by the conjugation of the nanoclusters with zwitterions and trivalent mannose ligands. 399 The integration of the latter component into the conjugate increased the internalization rate by 62 and 256% when targeted dendritic cells were incubated with 1 and 25 μg mL −1 of the NPs, respectively. 399 The absorption of energy by AuNPs from high‐frequency radiation has been employed to enhance the effectivity of radiotherapy (see Section 6.A.5) and as a contrast enhancer during computational tomography (CT) scans, as well. For example, in situ generated small AuNPs modified by dendrimers containing polyvalent lactobionic acid were reported to enter selectively hepatocarcinomic cells via their asialoglycoprotein receptors, allowing the high‐resolution computational tomography imaging of tumors. 400 The same method for selective tumor visualization was also enabled by glucose‐modified AuNPs mediated via glucose transporters overexpressed on the surface of cancer cells. 401 Tumors could be monitored for a week, as shown in Figure 26. 401 Glucose‐conjugated AuNPs bearing a ligand with 68 Ga as a tracer were biocompatible and excellent for in vivo imaging using positron emission spectroscopy. 402 Moreover, neuropeptides loaded on glucose‐ 68 Ga‐AuNPs enabled penetration through the blood–brain barrier. 402 Magnetic resonance is a broadly used imaging method in which the application of contrast agents significantly improves image quality, especially when these are in the form of NPs. In most experiments, iron oxide NPs decorated with stabilizing, targeting, or biocompatible materials have been tested. 403 As an example, versatile glycopeptide grafting to magnetic iron oxide NPs was reported but without in vivo testing. 404 A step further using sialic acid modified Fe 3 O 4 NPs for the selective in vivo MR imaging of β‐amyloids applicable for the early diagnosis of Alzheimer's disease was reported. 405 Mannose‐containing diblock copolymer grafted onto Fe 3 O 4 NPs enhanced the carbohydrate receptor‐driven uptake into lung cancer cells with potential application in the diagnosis and localization of this type of cancer. 406 HA‐grafted Fe 3 O 4 NPs employed in the selective MR imaging of tissues with CD44 receptor‐overexpressing cells, that is, tumors, or local inflammations were also described. 407 , 408 A similar system with core Fe 3 O 4 NPs having a thin golden shell covered with an HA layer enabling simultaneous MR imaging and photothermal therapy targeted toward cells with CD44 receptors has been reported. 370 In order to increase the in vivo stability of the Fe 3 O 4 core, its coating with a silica shell was described in several studies. 384 , 409 An enhanced imaging resolution of brain lesions in mice after simulated stroke can be observed in Figure 27 , showing that NPs with different glycosylation exhibited different biodistribution. 409 Sialic Le x present on NPs allowed their accumulation in ischemic, that is, damaged, brain tissue due to selectin overexpression; NPs decorated with Le x (without sialic acid present) could visualize such tissue less effectively, which was also confirmed by the software evaluation of dark and bright voxels in the images. 409 Nevertheless, good results were also obtained with NPs bearing hydroxyls groups on the NP surface instead of glycans. 409 In addition to the protective function of a silica shell, such a layer could provide a better environment for the covalent grafting of targeting ligands. The application of silica‐based nanocarriers for MR imaging has recently been reviewed by Caltagirone et al. 410 Figure 27 T2 maps and T2‐weighted images at 5 and 24 hr after MCAO in a representative animal tissue that received iron oxid‐silia nanoparticles decorated with (A) hydroxyl groups (HO@MNPs), (B) Lewis X without sialic acid (LX@MNPs), or (C) sialyl Lewis X (SX@MNPs). Note: the color scale bar corresponds to the T2 value (msec). Reprinted with permission from 409 . Copyright 2014 American Chemical Society. In addition to Fe 3 O 4 NPs, AuNPs were also applied in cell imaging. Small AuNPs of 1.8 nm in size grafted with lactobionic acid via PEG spacers were found to penetrate human hepatocellular carcinoma cells selectively via their overexpressed asialoglycoprotein receptors. 400 While T2‐weighted MRI was used for contrast enhancement applying iron oxide NPs, T1‐weighing was used for the same purpose using AuNPs loaded with a Gd 3+ complex. 411 The modification of AuNPs by a combination of different carbohydrates helped to achieve a high imaging selectivity of tumor cells in mice models. 412 AgNPs prepared by the reduction of Ag ions using saccharides present in sugar cane juice have ferromagnetic properties and could be applied for MRI. 411 3. Quantum Dots as Imaging Probes QDs as NPs composed of single nanocrystals of metal sulfides or other semiconductors have distinct optical properties resulting from their nano‐size. One of the most important features is their stable and adjustable fluorescence together with the overall stability inherent to all inorganic materials. Recent studies have reported that, after the encapsulation of QDs into a polymer shell with a final modification by carbohydrate moieties (for specific cell targeting), the resulted NPs still retain their fluorescence. The surface of QD‐modified NPs could be modified by a diverse range of functional groups with adjusted properties of NPs as reported by Schmidtke et al. 413 This study suggested applicability of such NPs in in vivo imaging, even though this has not been tested. 413 On the other side, the uptake and intracellular distribution of core‐shell CdSe‐ZnS QDs covalently grafted with different carbohydrates were assessed. 414 In addition to lactose, which enhanced the uptake of the QD‐glycan NPs into cancer cells, the composition of a mixed carbohydrate shell (i.e., lactose and mannose vs. lactose and maltotriose) significantly altered the internalization of NPs. 414 In another study, a secondary radiotracer ( 125 I) was used as the probe conjugated with CdSe/CdS QDs decorated with a sialic acid‐binding lectin. 415 These NPs were able to image breast cancer cells, as revealed by in vitro experiments. 415 Enhanced selectivity in the targeted imaging of muscle cells was observed for glucose‐decorated QDs with insulin‐induced uptake mediated via glucose receptors, as illustrated in Figure 28 , and such an approach could be applied for selective drug delivery as well. 416 Since the vast majority of inorganic QD particles exhibit cytotoxicity, Shinchi et al. focused on the development of cadmium‐free ZnS‐AgInS 2 NPs with lower cytotoxicity but still exhibiting good fluorescence properties and applicability in selective cell imaging after decoration with short targeting carbohydrates. 417 From the graphs shown in Figure 29 , it is obvious that all carbohydrate‐coated ZnS‐AgInS 2 QDs exhibited minimum cytotoxicity against HepG2 cells (panel A), unlike Cd‐based QDs (panel B). 417 Figure 28 (A) A molecular structure of glucose‐functionalized quantum dots (Glc‐QDs). (B) A schematic illustration of the cellular uptake of Glc‐QDs regulated by insulin and 2‐deoxyglucose (2‐DG) in C2C12 muscle cells. Reproduced from 416 with permission of the Royal Society of Chemistry. Figure 29 MTT assay for SFNPs. HepG2 cells were incubated with ZAIS/ZnS NPs (A) or CdTe/CdS QDs (B). The NP concentration was in the range of 5 to 50 μg mL −1 (left to right). Reprinted with permission from 417 . Copyright 2014 American Chemical Society. QD‐like fluorescence was assigned also to GO‐based NPs, 378 which were used, for example, for combined imaging and chemotherapy with an HA grafted as a targeting agent. 288 It is worth mentioning that the replacement of semiconductor‐based mostly cytotoxic QDs with less harmful graphene‐based NPs would be another step to develop novel, more efficient theranostic particles with low cytotoxicity. Silica QDs possessing inner fluorescence exhibited low cytotoxicity, having a surface that could be easily modified with glycans. 418 , 419 Furthermore, fluoride nanocrystals doped with rare earth element ions were also suggested as possible substitutes for "classical" Cd‐containing QDs—see, for example, 420 and references therein. 4. Toxicity of Nanomaterials It is beyond any doubt that the massive application and employment of nanomaterials have brought and should bring in the near future amazing achievements and technical possibilities—in fact, many authors are talking about the nanoage or nano‐revolution. Nevertheless, the other side of this "nano‐enthusiasm" is the persistent uncertainty about the possible toxicity of NPs. This issue is particularly delicate in medical applications, where NPs are supposed to be delivered directly into the human body and interact with cells at a molecular level. It should be noted that these concerns have been raised simultaneously with nanomaterial applications—see, for example, a 1982 study on the toxicity of a potential drug carrier. 421 In this study, one example of possible cytotoxicity mechanism of NPs was outlined, that is, the toxicity of monomers released from the intracellular degradation of NPs. Other known mechanisms, including unwanted aggregation of proteins, delivery of toxic molecules conjugated with NPs (including NPs from transportation and industrial emissions) and generation of highly cytotoxic ROS (the most toxic effect), were comprehensively reviewed. 422 , 423 , 424 , 425 The NPs' behavior in living organisms is prevalently determined by their physical and chemical properties. For example, decreased NP size increases their active surface and, consequently, the rate constant of any catalytic reaction taking place on the NP surface. Moreover, the size also determines the biodistribution mode in an organism—see, for example, different modes of interactions of small and large graphene sheets with cells (while the former tend to disrupt cell wall, 426 the latter more likely wrap it without any damage 427 ). Another crucial aspect is the nature of NP surface modification, including (i) surface charge, the adjustment of which can turn otherwise harmless NPs, for example, into cell disruptors (see references in 423 ) or other biological functions, 428 and (ii) the modification of NPs by biologically active moieties with examples, which can be found in sections of this review dealing with drug delivery, vaccine and other therapeutic effects, and cell imaging. Some toxic impacts can be estimated from the physical or chemical properties of the NPs acquired from a precise and accurate characterization of the NPs. This is usually accomplished since physical and chemical characterization and at least in vitro cytotoxicity tests are included in practically any study concerning the administration of NPs into the human body (drug carriers, imaging NPs, etc.). Nevertheless, the correlation between in vitro and model animal in vivo studies may significantly differ from the real impact on humans, nota bene when the subacute effect is considered. 429 Therefore, unless any negative health impacts are ruled out, appropriate precautions should be taken, and precise toxicity assessment must be performed in parallel with nanomedicine progress. A. Drug Delivery From a material point of view, diverse metallic NPs, carbon NPs, and SiNPs conjugated with carbohydrate moieties and drugs to be delivered have been reported in the past. A recent trend (since 2013), however, is the application of polymeric scaffolds or biopolymers, that is, either abundant polysaccharides based on glucose, chitosan, dextrans, or pullulans or protein‐based scaffolds for such purposes. 276 , 277 , 278 Saccharides were applied for the selective internalization of NPs by targeted cells due to receptors displayed on their surfaces with subsequent intracellular drug release. Saccharides were also used to substantially suppress adverse side effects of the treatment by the improved biocompatibility of otherwise hydrophobic, insoluble, or instable substances. After targeted delivery, drugs must be released from the conjugate to perform their therapeutic effect. Most nanocarriers have been prepared as pH‐sensitive conjugates, taking advantage of different pH values inside many tumor cells, where NP degradation occurs with subsequent drug release. Enzymatic nanocarrier degradation has also been described, as well as the degradation of the conjugate triggered by the presence of specific small molecules (glucose, glutathione, etc.). Irradiation‐sensitive nanocarriers have also been described where on‐demand and "remotely controlled" drug release occurred after near‐infrared (NIR) or UV irradiation. 279 In addition to a "model" drug (doxorubicin being one of the most frequently loaded cargoes in nanocarrier studies) or therapeutic proteins, carbohydrate‐based nanocarriers have also been used as nonviral vectors. 277 1. Carbon Nanomaterials Carbon NPs can be divided into four main groups regarding their dimensionality, that is, 0D fullerenes and spherical NPs, 1D nanotubes and nanofibers, 2D graphene sheets and 3D cryogels, aerogels, or foams. Recently, Li et al. compared of all types of carbon NPs for the simple encapsulation of an enzyme with its subsequent intracellular release. 280 While three‐dimensional graphene aerogel was suitable for the long‐term release of a cargo (3% in 24 hr), oxidized CNTs and GO provided fast cargo release (15–20% in 4 hr). 280 Without any doubts, a comparison of different types of NPs is needed for the development of ultraefficient drug carriers. Unfortunately, there is no recent comparative study focusing on the performance of different types of NPs modified with glycans applied in drug delivery. Many recent studies have explored the potential of graphene‐based NPs, that is, GO 281 , 282 , 283 or rGO. 284 , 285 Flakes of the former form of graphene derivative display many oxygen groups, which make GO hydrophilic and easily dispersible in aqueous solutions. 286 Furthermore, surface‐displayed oxygen moieties can be used for the covalent conjugation of various ligands. Most recent studies have investigated only slightly different variations of a basic scheme relying on GO nanosheets decorated with hyaluronic acid (HA) securing the selective binding of nanocarriers by CD44 receptors present on the surface of various kinds of tumor cells. Song et al. described the improved solubility of doxorubicin adsorbed on GO sheets and the "shielding" of a GO's surface charge by increasing its hydrophobicity. 281 In this study, a complex of GO‐doxorubicin with modified HA was reported to occur via H‐bonds between HA's amine and GO's epoxide group. 281 In another study, HA was covalently grafted onto the GO surface via adipic acid dihydrazide. 287 This GO–HA conjugate was treated with doxorubicin solution, and the drug was adsorbed onto a GO surface, most likely by hydrophobic and π–π stacking interactions. 287 HA could be attached to a GO surface via introduced sulfhydryl groups, forming disulfide bonds that were cleaved by glutathione inside targeted cells. 288 A similar approach was tested by Yang et al., who modified the GO surface with carboxymethylated chitosan. 289 Amine groups present in chitosan were used for covalent linkage to GO's hydroxyls and helped to sequentially attach HA with the subsequent loading of doxorubicin. 289 All of the authors declared the in vivo inhibition of tumor growth after injection of the prepared conjugates; however, the inhibition rate hardly exceeded 50% compared to the control. Thus, such relatively low efficiency of cancer growth inhibition complicates the further commercial exploration of doxorubicin‐based delivery systems. Interestingly, an almost 80% decrease in the size of epidermal tumors in mice was observed 28 days after the injection of the conjugate prepared from rGO (Fig. 18 ). 284 Most probably, higher hydrophobicity of rGO compared to GO allowed the adsorption of a higher amount of doxorubicin. Xu et al. combined the GO‐HA‐doxorubicin delivery system with a photothermal therapy by wrapping AuNPs in GO sheets with the subsequent loading of HA and doxorubicin. 290 Drug‐induced cytotoxicity based on a nanoconjugate was comparable to a treatment with the same concentration of unconjugated doxorubicin, but upon laser irradiation, the increased temperature of AuNPs decreased the cell viability of hepatoma cell line Huh7 from 44 to 18%. 290 Figure 18 Schematic depiction of reduced graphene oxide (black support) conjugated with cholesteryl (green), doxorubicin (red), and hyaluronic acid (blue) including hydrophobic and π–π stacking interactions. Reprinted from 284 . Copyright 2013, with permission from Elsevier. GO hydrogel modified with heparin‐mimicking sodium styrene sulfonate was perfectly biocompatible with blood. Such a nanoconjugate could have a broad range of therapeutic applications, including drug delivery. 291 Nevertheless, only a nonselective antimicrobial activity of particles loaded with gentamycin sulfate was tested. 291 CNTs offer similar surface properties as graphene‐based NPs, and their complexes with HA were also tested for doxorubicin delivery. However, when the drug was complexed with CNTs covalently grafted with PEI and HA, an in vitro study revealed a quite similar viability (approx. 50%) of model tumor cells treated with conjugated or unconjugated doxorubicin (24 hr, ∼4 μM doxorubicin). 292 More promising results were obtained with covalently grafted polyethylene glycol derivative to CNTs, crosslinked to a gel form by adding β‐cyclodextrin and treated with camptothecin (a hydrophobic cytotoxic drug). 293 An in vitro study showed the significant growth inhibition of a cancer cell line (∼80% inhibition after 7 days compared to cells treated with blank CNT‐polymer gel). The gelation process relayed on the natural ability of cyclodextrins to include certain molecules (PEG moieties, in this case) into their cavity via host–guest interaction. Such an absorption capacity also has great potential for hydrophobic drug loading. Furthermore, the feasibility of precise chemical modification makes cyclodextrins a great tool not only in smart drug development but generally in the molecular fabrication of diverse mechanisms. 293 These possibilities have been reviewed, for example, by Martínez et al. 294 In addition to carbon, boron nitride nanotubes have also been exploited for drug delivery. Such nanotubes with covalently grafted glycosylated chitosan exhibited low cytotoxicity; therefore, such NPs could be used as nonviral vectors. 295 A plasmid gene was more efficiently expressed in human carcinoma epithelial cells (A549) than in control cell lines even though these NPs did not exhibit any specific receptor ligand selectivity. 295 Enhanced plasmid expression could be assigned to more efficient endocytosis rather than to the presence of specific receptors in A549 cells. 295 An intriguing approach was reported by Zhou et al. by loading doxorubicin into pores of mesoporous carbon NPs. 296 These NPs were subsequently grafted by HA, which guided selective drug targeting. Furthermore, HA acted as a "gatekeeper" because it was specifically recognized and degraded by hyaluronidase‐1 after internalization and allowed drug release. In addition to enzymatic degradation, HA coating could also be removed by glutathione due to the nature of a covalent bond between HA and NPs, making drug release more efficient and controllable. The authors reported an IC 50 of 8.9 μM for doxorubicin carried by the nanocarrier. 296 2. Metal and Metal Oxide NP‐Glycan Conjugates for Drug Delivery AgNPs able to generate oxidative stress have strong antimicrobial properties. Their modification with glycans represents very simple carbohydrate‐guided therapy applicable in bactericidal surface fabrication (vide supra) and possibly also in cancer treatment. Kennedy et al. revealed that AgNPs modified with a monosaccharide galactose were internalized by model neural and cancer cells more efficiently than mannose and glucose conjugates, but interestingly, the former exhibited lower cell toxicity (approx. fourfold lower EC 50 against model cancer cells) compared to less‐internalized glycoconjugates, suggesting an important role of terminal saccharide in glycan moieties. 297 Similarly, Shahbazi et al. observed a significant level of cytotoxicity (approx. 80%) when the HT1080 sarcoma cell line was treated with iron oxide NPs (100 or 200 nm in diameter) coated with glucose, but a similar effect was observed with NPs coated by polyethylene glycol, suggesting the low targeting capability of such conjugates. 298 Apparently, AuNPs have not been considered preferential drug carriers, as discussed in a previous section, despite the fact that there are studies reporting the simple decoration of AuNPs with carbohydrate‐based targeting moieties (mannose modified 108 ) and even loaded with therapeutic agents. The latter was achieved by the conjugation of moieties capable of enhanced drug ligation to AuNPs, with cyclodextrins being very sound examples of such approach. Cyclodextrins (and their derivatives 299 ) can host hydrophobic, insoluble, drugs inside their structure 294 and can be easily conjugated to AuNPs. 300 Alternatively, AuNPs can be conjugated directly with cytotoxic ligands, for example, gold(I) triphenylphosphine. 301 On the other side, similar to the above‐mentioned carbon NP‐based nanocarriers, HA derivative coated on gold nanocages was used as a vehicle for the targeted delivery of doxorubicin into CD44 receptor‐expressing cells. 302 The intracellular hyaluronase‐1‐induced degradation of the conjugate resulted in drug release and selective cytotoxicity. In contrast to carbon NPs, gold nanocages could also be applied for thermodynamic therapy. This synergy of cancer cell treatment led to the complete elimination of a solid tumor in mice in 9 days by combined photo‐ and chemotherapy. 302 Terpen (borneol) stabilized selenium NPs decorated with glucose were also reported to deliver an anticancer drug selectively into Hep2 cells (a hepatocarcinomic cell line). 303 In addition to an efficient pH‐sensitive system of drug release from the conjugate, an intracellular, selenium NP‐directed production of reactive oxygen species (ROS) was also observed, making the fabricated glycoconjugate cytotoxic for drug resistant hepatocarcinomic cells. 303 Similarly, TiO 2 NPs coated with HA and loaded with cisplatin containing ligand were found to carry their payload selectively into an ovarian cell line (A2780) with pH‐induced drug release from these glycoconjugates. 304 Nevertheless, only an approximately twofold increased drug‐induced cytotoxicity against model cancer cells was observed upon treatment with conjugated versus unconjugated drug (a decrease to approx. 40 vs. 70% of the original viability, respectively, using 10 μM drug). 304 Polysaccharide carboxymethyl dextran was also used as a building block for the synthesis of magnetoliposomes loaded with doxorubicin. 305 Drug release was enhanced upon the exposure of the conjugates to a low‐frequency magnetic field in synergy with low pH induced structural changes leading to drug release. In vitro experiments confirmed the low cytotoxicity of the loaded magnetoliposomes to brain cells, while after incubation with 2.5 μg mL −1 of conjugated and free doxorubicin equivalents, the viability decreased to approx. 40 and 70% of original viability, respectively. 305 3. Silica‐Based Glycan‐Targeted Drug Carrier Systems In a rather "traditional" approach, small SiNPs were covalently modified with spacers; each one ended with trivalent glucose as a targeting unit. 306 The spacers provided a space for the inclusion of molecules of the anticancer drug paclitaxel. The drug delivery system was the same as described before, that is, whole conjugates were selectively internalized by cancer cell line HepG2, and the drug was released either by intracellular esterases or by low pH. Due to the good pharmacokinetics of the fabricated conjugate, a much higher toxicity (IC 50 = 0.7 μM paclitaxel equivalent) was induced than with a treatment by the unconjugated drug upon prolonged incubation (no viability change using 0.7 μM paclitaxel). 306 Polysaccharide moieties were also used to increase the biocompatibility of a mesoporous silica doxorubicin carrier. 307 Yu et al. adapted an already developed protocol to prepare HA‐coated mesoporous SiNPs transporting doxorubicin selectively into human colon cancer cells. 308 A decrease of targeted cell viability by approximately 50 and 20%, respectively, was observed upon incubation with 0.25 μM drug equivalent carried by HA‐coated NPs and uncoated NPs or free drug. 308 HA‐targeting mesoporous SiNPs with both a therapeutic agent and HA attached to NP's surface by disulfide bonds were investigated by Zhao et al. 309 Disulfide bonds were cleaved by glutathione, causing efficient drug release (Fig. 19 ). Cytotoxicity assessment showed that after 72 hr of treatment with the drug‐loaded conjugate, a cell line with CD44 receptors retained a viability of 39%, while the viability of a control cell line without CD44 was 49%, suggesting rather strong nonspecific internalization by the NIH 3T3 cell line. 309 The same group also prepared mesoporous SiNPs with pores loaded with doxorubicin and capped by HA attached to the surface by disulfide bonds. 310 Excess glutathione and hyaluronidases helped to release capping molecules from the conjugate. 310 An elegant approach using nonspecific internalization based on differences between the metabolic activity of normal and cancer cells was reported. 311 Cancer cells exhibit increased glucose uptake with the enhanced uptake of synthesized glucose‐modified SiNPs. Celastrol used as an anticancer drug negatively affects the mitotic cycle, which is faster in cancer than in normal cells, thus making cancer cells more susceptible to the effect of the drug. Mesoporous SiNPs decorated with glucose via dendritic polymer linkers were found to induce the apoptosis of HeLa cells (50% apoptotic cells after 24 hr of incubation with 5.3 μM drug equivalent). On the other hand, a lower exposure time induced the production of heat shock proteins, suggesting that a subcytotoxic dosage (or exposure time) may even help to protect normal cells. 311 Figure 19 Schematic depiction of glutathione (GSH) induced drug release from hyaluronic acid (HA) capped pores of silica NPs. These pores are filled with disulfide‐bond tethered 6‐mercaptopurine (6‐MP) and, due to HA‐decoration, nanoparticles are captured and internalized via so called CD‐44 receptors—a common surface glycoprotein with significantly higher concentrations on the surface of certain cancer cells. Reprinted with permission from 309 . Copyright 2014 American Chemical Society. Carbohydrate modification can also be used for the selective and triggered drug release from mesoporous magnetic SiNPs. 312 Lectin‐like phenylboronic acid (PBA) displayed on their surface anchored dextran, which efficiently gated drug‐loaded pores in the body of magnetic SiNPs. Once glucose is present in sufficiently high concentration, it competitively binds to PBA, and the dextran "gatekeeper" layer is consequently released from the surface. Due to the much smaller size of glucose compared to dextran chains, drug molecules can easily migrate from the pores. This glucose‐dependent release has great potential, for example, in diabetes treatment, even though a selective cytotoxicity against HeLa cells in the presence of glucose was demonstrated in this study (60% decrease of viability compared to a 25% decrease upon treatment of the cells with unloaded and drug‐loaded NPs without dextran, respectively). 312 Employing the same gate‐keeping approach, Zhou et al. bound mannose units on the surface of mesoporous SiNPs via disulfide bonds. 313 These moieties tethered Con A lectin, a molecule large enough to prevent the migration of doxorubicin from the pores of mesoporous SiNPs. Selective release was triggered by the elevated concentration of glutathione in the cytoplasm of cancer cells, and the conjugate exhibited an IC 50 of 25 μg mL −1 against cancer cells, while a negligible viability change was observed when the conjugate was incubated with normal cells having a lower cytoplasmic glutathione concentration. 313 Chondroitin sulfate is another natural polysaccharide with the ability to promote the uptake of conjugates by selective interaction with CD44 receptors. Moreover, such a biopolymer has the ability to improve the biocompatibility and stability of NPs. All of these features of chondroitin sulfate were used to fabricate a mesoporous silica‐based vector loaded with a plasmid carrying the gene for the p53 protein. 314 A more than twofold increase in translated messenger RNA for p53 was observed in treated cancer cell lines, when the silica‐based vector was coated with chondroitin sulfate compared to bare mesoporous silica–plasmid conjugate. 314 Gene therapy induced expression of healthy p53 protein in human cancer cells inhibits tumor growth; thus, the above‐described vector represents another possible way for cancer treatment. 4. Polymer Scaffold Based Drug Carriers Liposomes prepared from a synthetic building block containing a hydrophobic tail conjugated with galactose as targeting units were capable of transporting doxorubicin easily into HepG2 cells displaying asialoglycoprotein 315 , 316 , 317 and other receptors. 44 A very similar system, only with incorporated disulfide bonds in the structure of micelles, was reported. 318 , 319 , 320 , 321 Such glycoconjugates were selectively internalized, and a drug was released after the glutathione‐mediated cleavage of S–S bonds. 318 , 319 , 320 , 321 Although these studies demonstrated the selective cytotoxicity toward targeted cancer cells in vitro, only Zhao et al. have reported the in vivo inhibition of tumor growth after the localized injection of galactosylated, doxorubicin‐loaded liposomes (tumor weight of approx. 50% compared to animals treated with an unconjugated drug). 316 The unfolding of synthetic 322 or natural 323 micelles triggered by UV irradiation for drug release was also reported, but such an approach was not tested for in vitro or in vivo cytotoxicity. In order to better control the form and chemical properties of nanocarriers, a tri‐block copolymer folded into different types of mannosylated micelles was synthesized. 324 Preliminary results suggested that there was some correlation between mannose receptor‐induced internalization and the shape of a micelle, but in vivo or in vitro studies investigating the inhibition of virus binding have yet to be conducted. 324 To discuss drug delivery from an application point of view, it is worth mentioning the use of a synthetic copolymer folded into micelles tested for the oral administration of sorafenib—the drug applied mainly for the treatment of kidney and liver cancers. 325 In vivo studies have demonstrated that galactosylated, drug‐loaded micelles efficiently and selectively administered sorafenib into the liver and secured its continuous release for several hours. 325 To improve the oral administration of darunavir, self‐assembled micelles made from synthetic copolymer poly(ethylene oxide)‐poly(propylene oxide) were glycosylated by microwave‐assisted ring opening reaction and loaded with the drug by a simple incubation of the two components in solution. 326 An approximately 500‐fold higher solubility of darunavir was observed upon its integration into micelles. This is a significant step in the development of less expensive anti‐HIV drugs, considering the fact that the low administration efficiency of present darunavir‐based delivery increases its cost. 326 Berberine hydrochloride is another drug with a reported anticancer effect with chemical properties not allowing oral administration. 327 Drug encapsulation into chitosan‐coated phospholipid micelles significantly improved pharmacokinetics and distribution parameters due to the increased biocompatibility of the conjugate. It should be noted, however, that only the biodistribution and drug release profile were assessed in the study, with results suggesting that chitosan‐coated liposomes are good vehicles for the oral administration of berberin. 327 Micelles prepared by the conjugation of α‐tocopherol (recognized as a potential anticancer drug) with pullulan were reported to release the drug 10‐hydroxycamptothecin at the low pH typically present in the cytoplasm of tumor cells. 328 In vitro cell assays revealed an approximately 60% decreased viability of targeted cells compared to a control experiment with another cell line. 328 A conjugate based on α‐tocopherol derivatives together with HA and cytostatic docetaxel in the micellar form was found to induce cytotoxicity and to decrease multidrug resistance via the α‐tocopherol‐mediated inhibition of the active drug transport from the cell. 329 The suppression of drug resistance on a molecular level was also achieved by the delivery of two recently synthesized inhibitors of efflux proteins (P‐glycoprotein and breast cancer resistance protein) loaded in poly‐lactic‐co‐glycolic‐acid. 330 A higher inhibition level of cancer cells incubated with conjugated compared to unconjugated inhibitors was demonstrated. 330 Interestingly, the same efflux mechanism was also suppressed by hydroxypropyl‐β‐cyclodextrin moieties grafted onto synthetic copolymer NPs, which enhance the effect of low‐soluble immunosuppressor tacrolimus by oral administration. 299 Atherosclerosis can be efficiently treated by the internalization of micelles prepared from a reconstituted high‐density lipoprotein (rHDL) loaded with levostatin mediated via Scavenger Receptor class B member 1 (SR‐BI) receptors accessible on the surface of cells present in a macrophage‐infiltrated atherosclerosis lesion. 331 Upon coating of the fabricated micelles with HA (Fig. 20 ), their administration became more efficient and precisely targeted in two steps: (i) because of their lower accumulation in the liver, this is achieved because HA‐decorated particles are not recognized by SR‐BI receptors and "scavenged" (recognition by SR‐BI receptors in liver cells leads to an increased level of phagocytosis). Thus, more particles are delivered into atherosclerotic veins. In the second stage, (ii) HA was readily recognized by CD44 receptors present in atherosclerotic plagues, where HA‐LT‐rHDL was accumulated. Next, due to the presence of hyaluronidases, HA was removed. The formed LT‐rHDL NPs were recognized by SR‐BI receptors of atherosclerosis lesion cells, which delivered them into desired "foam cells" where levostatin was released. 331 Figure 20 Scheme of fabrication and function of dual‐targeted reconstituted liposomes. Upper scheme represents preparation of micelles containing levostatin and coated with HA (HA‐LT‐rHDL). After injection in an atherosclerotic rabbit, HA coating allowed preferential distribution of HA‐LT‐rHDL into veins and their retention in an organism with decreased rate of phagocytosis in liver. Once HA‐LT‐rHDL reached a lesion site, it was recognized by CD44 receptors on the surface of endothelial cells which helped their transportation inside an atherosclerotic plague (the scheme in the lower part of the figure). In the plague, hyaluronidases removed HA from HA‐LT‐rHDL and the resulting LT‐rHDLs were preferentially recognized by SR‐BI receptors on surface of "foam cells" where lovastatin was intended to be released. Reprinted from 331 . Copyright 2014, with permission from Elsevier. The antiseptic drug NK007 was delivered successfully by oral application to regress murine colitis. 332 Polysaccharide components of micelles are responsible for efficient target delivery as could be concluded from the in vivo study. Nevertheless, a significant difference between the parameters recorded to evaluate treatment efficacy (i.e., colon length, stool consistency, etc.) was only observed between the control and treated groups but not between groups treated with unconjugated or NP‐conjugated NK007. 332 Amphotericin B, an antiparasitic drug effective, for example, against Leishmania donovani , induces several harmful side effects, with nephrotoxicity and hemolytic toxicity being the most severe. 333 Moreover, the drug is relatively expensive. Khan et al. conjugated this drug with mannosylated synthetic dendrimer to gain biocompatibility. As a result, no kidney damage or hemolysis was observed with the antiparasitic activity of the drug preserved. In fact, due to the mannose‐driven targeting of conjugates into macrophages, IC 50 (39 nM) decreased approximately fivefold compared to the unconjugated drug or its commercial analogue. 333 A supramolecular scaffold prepared from a polymer bearing PBA and a "complementary" glycopolymer was degraded under hyperglycaemic conditions. 334 The conjugate could be applied for the safe, comfortable, and glucose concentration‐sensitive distribution of insulin. 335 Zheng et al. also employed a PBA‐based conjugate with both PBA and lactobionic acid moieties displayed randomly on each copolymer chain. 334 Such molecules folded into a scaffold able to encapsulate insulin. PBA enhanced the mucoadhesion of the scaffolds, which could be then used for the nasal administration of insulin. Therapeutic effect was confirmed by in vivo experiments in diabetic mice with glucose levels decreased to approximately 40% 8 hr after the addition of therapeutic NPs, while such a decrease was observed within 0.5 hr when bare insulin was injected. 334 The single‐step ligation of HA with doxorubicin seems to be much simpler, forming conjugates that are selectively cytotoxic to cancer cells displaying CD44 receptors 336 similar to pullulan‐doxorubicin micelles selectively targeting cells with asialoglycoprotein receptors. 337 , 338 A similar weight reduction of induced tumor was observed in vivo after treatment with unconjugated and conjugated doxorubicin (approximately 0.4 g of tumor weight compared to 1.8 g developed in a control group), but the life span was longer and the animal weight was larger for NP‐treated mice compared to mice treated with unconjugated doxorubicin, suggesting the significant suppression of a negative side effect of a chemotherapy. 338 Plant glucomannan covalently conjugated with an inhibitor (bisphosphonate alendronate) of tumor‐associated macrophage was also prepared as a conjugate inducing selective, mannose receptor‐mediated macrophage apoptosis (a tumor weight of approx. 0.3 g was observed 14 days after the application of therapeutic NPs compared to 1.5 g developed in an untreated group). 339 Mannose receptors of macrophages were also targeted with ionic polymer‐based micelles containing the model protein to be delivered (BSA). 340 Upon coating of the micelles with β‐glucan from yeast cell walls, NPs were selectively internalized by macrophages. 340 So‐called "ivy NPs" are nature‐derived nanocarriers consisting of a heavily glycosylated protein core. 341 When loaded with doxorubicin, a targeted delivery and release of a drug into different cancer cell lines was observed. 341 The treatment of a solid tumor was equally efficient for the application of both the conjugate and unconjugated doxorubicin, but with a significantly lower weight loss of the animal observed after the application of the conjugate. 341 Quite a "minimalistic" approach was reported, relying on short polyethylene glycol chains decorated with several saccharidic units along their length and with a bioactive molecule on the chain end. 342 Such conjugates bearing two separated mannosyl units (the optimum distance between these units was found to be 5.6 nm) were internalized by macrophage cells via their mannose‐binding receptors. The authors declared a good biocompatibility of the carriers as demonstrated by the absence of lysosome storage disease‐like effect after their application. 342 On the other side, in vivo tests of applicability of such conjugates have yet to be performed. Similarly, Thomas et al. reported micelles prepared from bis‐ l ‐galactose lysine. 343 Galactose displayed on the surface of these NPs was enzymatically oxidized, and the aldehydes formed were used for the covalent crosslinking of more NPs together by di‐ or trivalent amine. The authors tested their delivery system only for the complexation of AuNPs, but selective delivery of drugs is anticipated. 343 A selective drug delivery into the central nervous system, passing the blood–brain barrier, is always very difficult. Recently, certain lectins displayed on NPs helped to enhance such crossing. 344 For example, polyethylene glycol based micelles decorated with lectin Solanum tuberosum were able to successfully deliver a basic fibroblast growth factor via the nasal route into the brain in order to improve the cognition of those affected by Alzheimer's disease. 344 Lactoferrin, an iron‐binding glycoprotein, was found to be helpful in delivering drug‐loaded micelles through the blood–brain barrier 345 , and the same protein was used to target asialoglycoprotein receptors on HepG2 cancer cells, as well. 346 Another way to pass the blood–brain barrier is using activated membrane transporters. Glucose transporter 1 was activated by p‐aminophenyl‐β‐ d ‐manno‐pyranoside with a subsequent transportation of anticancer drugs into the brain cells. 347 Kuo et al. prepared NPs based on solid lipid decorated with p‐aminophenyl‐β‐ d ‐manno‐pyranoside and folic acid and loaded with etoposide (a drug inhibiting the proliferation of malignant glioblastoma) that were efficiently distributed through the blood–brain barrier with efficient secondary targeting to glioblastoma cells secured by folic acid in in vitro experiments. 347 Nonviral gene transport is just another field where the aforementioned techniques of controlled targeted drug delivery have been employed. One of the materials used quite often for the preparation of nonviral vectors is PEI, which can be effectively complexed with the DNA or RNA of interest, but it possesses a certain level of toxicity and its targeting is not selective and thus suitable only for in vitro experiments. In order to apply PEI‐based conjugates for in vivo applications, several ways to improve the performance of such carriers were sought. For example, PEI was complexed with a biocompatible anionic glucose‐based glycopolymer, 348 GlcNAc, 349 chondroitin sulfate, 350 HA, 351 or depolymerized guar gum with available mannose units 352 with subsequent use in selective targeting. Alternatively, copolymers of the methacrylamide backbone decorated with glucose, 353 α‐ d ‐mannopyranosyl, 354 or trehalose 355 and cationic moieties could electrostatically complex plasmids with the subsequent selective transport of plasmid DNA into cancer cells 353 , 354 or siRNA into glioblastoma cells via glucose transporter‐1. 355 Furthermore, complexed siRNA was reported to retain its biological activity after freeze drying. 355 An electrostatic complexation of DNA with copolymer chains bearing a targeting GlcNAc moiety was also reported. 356 Targeted gene delivery was also achieved by DNA incorporation into PEGylated micelles decomposing at low pH and decorated with galactose 357 or mannose 358 to target asialoglycoprotein or mannose receptors, respectively. Noteworthy, a secondary target ligand was co‐entrapped with DNA to help penetrate through the nuclear membrane. 357 Mannose‐decorated cholesteryl‐based synthetic liposomes with virus‐like characteristics were reported to effectively transfect nonactivated dendritic cells by a plasmid‐containing gene for luciferase. 359 Mannosylated chitosan was also used for nonviral vector preparation 360 as well as mannosylated "bubble lipoplexes," that is, NPs with ultrasound‐triggered DNA release causing the transfection of M2 macrophages and their switch from tumor growth‐promoting into tumoricidal M1‐like macrophages. 361 , 362 In vivo experiments confirmed that this treatment could inhibit the growth of various tumors without adverse side effects. 361 , 362 A triblock copolymer consisting of hydrophobic and pH‐sensitive parts to assure the formation of micelles, polycation complexing oligonucleotides, and a mannose‐modified part was also applied to target M2 macrophages. 363 Synthetic vectors based on siRNA‐PEGylated cyclodextrin scaffolds with a peptide ligand as a targeting unit were also investigated. 364 In addition to the broadly used electrostatic interactions used for loading nucleotides into delivery NPs, Kim et al. employed hybridization with a specific DNA template with a condensation of the resulted DNA complex by a viral Mu peptide (Fig. 21 ). 365 The formed "DNA nanoballs" coated with HA were effectively internalized by cancer cells. In vivo experiments showed that antisense oligonucleotides were hybridized with their complementary mRNA in the cytoplasm, leading to a significant sensitization of the cells to doxorubicin treatment and to a significant decrease of tumor growth in mice. 365 Figure 21 RCA template and hybridization efficiency. (A) Secondary structure of scrambled rolling circle amplification (RCA) template. (B) Secondary structure of a dual antisense oligonucleotide (ASO) hybridizing RCA template for Dz‐13 and OGX‐427. (C) RCA products with poly ASO‐binding sequences were hybridized with two therapeutic ASOs, Dz‐13 and OGX‐427, to produce dual ASO‐hybridizing RCA products. (D) The hybridization efficiencies of the ASOs were tested for products of scrambled RCA templates and dual ASO‐hybridizing RCA templates using fluorescently labeled ASOs. Reprinted from 365 . Copyright 2015, with permission from Elsevier. A very simple approach relying only on the application of bare therapeutic oligonucleotides conjugated with a targeting ligand (e.g., a triantennary GalNAc 366 , 367 , 368 or glucose 369 ) was also tested. These agents were capable of blocking the adverse immunostimulation of an anti‐HIV oligonucleotide. This may be helpful in the fabrication of anti‐HIV drugs without adverse side effects. The conjugate can be easily prepared, but there is a concern regarding the stability of the carrier when exposed to endo‐ and exonucleases. 1. Carbon Nanomaterials Carbon NPs can be divided into four main groups regarding their dimensionality, that is, 0D fullerenes and spherical NPs, 1D nanotubes and nanofibers, 2D graphene sheets and 3D cryogels, aerogels, or foams. Recently, Li et al. compared of all types of carbon NPs for the simple encapsulation of an enzyme with its subsequent intracellular release. 280 While three‐dimensional graphene aerogel was suitable for the long‐term release of a cargo (3% in 24 hr), oxidized CNTs and GO provided fast cargo release (15–20% in 4 hr). 280 Without any doubts, a comparison of different types of NPs is needed for the development of ultraefficient drug carriers. Unfortunately, there is no recent comparative study focusing on the performance of different types of NPs modified with glycans applied in drug delivery. Many recent studies have explored the potential of graphene‐based NPs, that is, GO 281 , 282 , 283 or rGO. 284 , 285 Flakes of the former form of graphene derivative display many oxygen groups, which make GO hydrophilic and easily dispersible in aqueous solutions. 286 Furthermore, surface‐displayed oxygen moieties can be used for the covalent conjugation of various ligands. Most recent studies have investigated only slightly different variations of a basic scheme relying on GO nanosheets decorated with hyaluronic acid (HA) securing the selective binding of nanocarriers by CD44 receptors present on the surface of various kinds of tumor cells. Song et al. described the improved solubility of doxorubicin adsorbed on GO sheets and the "shielding" of a GO's surface charge by increasing its hydrophobicity. 281 In this study, a complex of GO‐doxorubicin with modified HA was reported to occur via H‐bonds between HA's amine and GO's epoxide group. 281 In another study, HA was covalently grafted onto the GO surface via adipic acid dihydrazide. 287 This GO–HA conjugate was treated with doxorubicin solution, and the drug was adsorbed onto a GO surface, most likely by hydrophobic and π–π stacking interactions. 287 HA could be attached to a GO surface via introduced sulfhydryl groups, forming disulfide bonds that were cleaved by glutathione inside targeted cells. 288 A similar approach was tested by Yang et al., who modified the GO surface with carboxymethylated chitosan. 289 Amine groups present in chitosan were used for covalent linkage to GO's hydroxyls and helped to sequentially attach HA with the subsequent loading of doxorubicin. 289 All of the authors declared the in vivo inhibition of tumor growth after injection of the prepared conjugates; however, the inhibition rate hardly exceeded 50% compared to the control. Thus, such relatively low efficiency of cancer growth inhibition complicates the further commercial exploration of doxorubicin‐based delivery systems. Interestingly, an almost 80% decrease in the size of epidermal tumors in mice was observed 28 days after the injection of the conjugate prepared from rGO (Fig. 18 ). 284 Most probably, higher hydrophobicity of rGO compared to GO allowed the adsorption of a higher amount of doxorubicin. Xu et al. combined the GO‐HA‐doxorubicin delivery system with a photothermal therapy by wrapping AuNPs in GO sheets with the subsequent loading of HA and doxorubicin. 290 Drug‐induced cytotoxicity based on a nanoconjugate was comparable to a treatment with the same concentration of unconjugated doxorubicin, but upon laser irradiation, the increased temperature of AuNPs decreased the cell viability of hepatoma cell line Huh7 from 44 to 18%. 290 Figure 18 Schematic depiction of reduced graphene oxide (black support) conjugated with cholesteryl (green), doxorubicin (red), and hyaluronic acid (blue) including hydrophobic and π–π stacking interactions. Reprinted from 284 . Copyright 2013, with permission from Elsevier. GO hydrogel modified with heparin‐mimicking sodium styrene sulfonate was perfectly biocompatible with blood. Such a nanoconjugate could have a broad range of therapeutic applications, including drug delivery. 291 Nevertheless, only a nonselective antimicrobial activity of particles loaded with gentamycin sulfate was tested. 291 CNTs offer similar surface properties as graphene‐based NPs, and their complexes with HA were also tested for doxorubicin delivery. However, when the drug was complexed with CNTs covalently grafted with PEI and HA, an in vitro study revealed a quite similar viability (approx. 50%) of model tumor cells treated with conjugated or unconjugated doxorubicin (24 hr, ∼4 μM doxorubicin). 292 More promising results were obtained with covalently grafted polyethylene glycol derivative to CNTs, crosslinked to a gel form by adding β‐cyclodextrin and treated with camptothecin (a hydrophobic cytotoxic drug). 293 An in vitro study showed the significant growth inhibition of a cancer cell line (∼80% inhibition after 7 days compared to cells treated with blank CNT‐polymer gel). The gelation process relayed on the natural ability of cyclodextrins to include certain molecules (PEG moieties, in this case) into their cavity via host–guest interaction. Such an absorption capacity also has great potential for hydrophobic drug loading. Furthermore, the feasibility of precise chemical modification makes cyclodextrins a great tool not only in smart drug development but generally in the molecular fabrication of diverse mechanisms. 293 These possibilities have been reviewed, for example, by Martínez et al. 294 In addition to carbon, boron nitride nanotubes have also been exploited for drug delivery. Such nanotubes with covalently grafted glycosylated chitosan exhibited low cytotoxicity; therefore, such NPs could be used as nonviral vectors. 295 A plasmid gene was more efficiently expressed in human carcinoma epithelial cells (A549) than in control cell lines even though these NPs did not exhibit any specific receptor ligand selectivity. 295 Enhanced plasmid expression could be assigned to more efficient endocytosis rather than to the presence of specific receptors in A549 cells. 295 An intriguing approach was reported by Zhou et al. by loading doxorubicin into pores of mesoporous carbon NPs. 296 These NPs were subsequently grafted by HA, which guided selective drug targeting. Furthermore, HA acted as a "gatekeeper" because it was specifically recognized and degraded by hyaluronidase‐1 after internalization and allowed drug release. In addition to enzymatic degradation, HA coating could also be removed by glutathione due to the nature of a covalent bond between HA and NPs, making drug release more efficient and controllable. The authors reported an IC 50 of 8.9 μM for doxorubicin carried by the nanocarrier. 296 2. Metal and Metal Oxide NP‐Glycan Conjugates for Drug Delivery AgNPs able to generate oxidative stress have strong antimicrobial properties. Their modification with glycans represents very simple carbohydrate‐guided therapy applicable in bactericidal surface fabrication (vide supra) and possibly also in cancer treatment. Kennedy et al. revealed that AgNPs modified with a monosaccharide galactose were internalized by model neural and cancer cells more efficiently than mannose and glucose conjugates, but interestingly, the former exhibited lower cell toxicity (approx. fourfold lower EC 50 against model cancer cells) compared to less‐internalized glycoconjugates, suggesting an important role of terminal saccharide in glycan moieties. 297 Similarly, Shahbazi et al. observed a significant level of cytotoxicity (approx. 80%) when the HT1080 sarcoma cell line was treated with iron oxide NPs (100 or 200 nm in diameter) coated with glucose, but a similar effect was observed with NPs coated by polyethylene glycol, suggesting the low targeting capability of such conjugates. 298 Apparently, AuNPs have not been considered preferential drug carriers, as discussed in a previous section, despite the fact that there are studies reporting the simple decoration of AuNPs with carbohydrate‐based targeting moieties (mannose modified 108 ) and even loaded with therapeutic agents. The latter was achieved by the conjugation of moieties capable of enhanced drug ligation to AuNPs, with cyclodextrins being very sound examples of such approach. Cyclodextrins (and their derivatives 299 ) can host hydrophobic, insoluble, drugs inside their structure 294 and can be easily conjugated to AuNPs. 300 Alternatively, AuNPs can be conjugated directly with cytotoxic ligands, for example, gold(I) triphenylphosphine. 301 On the other side, similar to the above‐mentioned carbon NP‐based nanocarriers, HA derivative coated on gold nanocages was used as a vehicle for the targeted delivery of doxorubicin into CD44 receptor‐expressing cells. 302 The intracellular hyaluronase‐1‐induced degradation of the conjugate resulted in drug release and selective cytotoxicity. In contrast to carbon NPs, gold nanocages could also be applied for thermodynamic therapy. This synergy of cancer cell treatment led to the complete elimination of a solid tumor in mice in 9 days by combined photo‐ and chemotherapy. 302 Terpen (borneol) stabilized selenium NPs decorated with glucose were also reported to deliver an anticancer drug selectively into Hep2 cells (a hepatocarcinomic cell line). 303 In addition to an efficient pH‐sensitive system of drug release from the conjugate, an intracellular, selenium NP‐directed production of reactive oxygen species (ROS) was also observed, making the fabricated glycoconjugate cytotoxic for drug resistant hepatocarcinomic cells. 303 Similarly, TiO 2 NPs coated with HA and loaded with cisplatin containing ligand were found to carry their payload selectively into an ovarian cell line (A2780) with pH‐induced drug release from these glycoconjugates. 304 Nevertheless, only an approximately twofold increased drug‐induced cytotoxicity against model cancer cells was observed upon treatment with conjugated versus unconjugated drug (a decrease to approx. 40 vs. 70% of the original viability, respectively, using 10 μM drug). 304 Polysaccharide carboxymethyl dextran was also used as a building block for the synthesis of magnetoliposomes loaded with doxorubicin. 305 Drug release was enhanced upon the exposure of the conjugates to a low‐frequency magnetic field in synergy with low pH induced structural changes leading to drug release. In vitro experiments confirmed the low cytotoxicity of the loaded magnetoliposomes to brain cells, while after incubation with 2.5 μg mL −1 of conjugated and free doxorubicin equivalents, the viability decreased to approx. 40 and 70% of original viability, respectively. 305 3. Silica‐Based Glycan‐Targeted Drug Carrier Systems In a rather "traditional" approach, small SiNPs were covalently modified with spacers; each one ended with trivalent glucose as a targeting unit. 306 The spacers provided a space for the inclusion of molecules of the anticancer drug paclitaxel. The drug delivery system was the same as described before, that is, whole conjugates were selectively internalized by cancer cell line HepG2, and the drug was released either by intracellular esterases or by low pH. Due to the good pharmacokinetics of the fabricated conjugate, a much higher toxicity (IC 50 = 0.7 μM paclitaxel equivalent) was induced than with a treatment by the unconjugated drug upon prolonged incubation (no viability change using 0.7 μM paclitaxel). 306 Polysaccharide moieties were also used to increase the biocompatibility of a mesoporous silica doxorubicin carrier. 307 Yu et al. adapted an already developed protocol to prepare HA‐coated mesoporous SiNPs transporting doxorubicin selectively into human colon cancer cells. 308 A decrease of targeted cell viability by approximately 50 and 20%, respectively, was observed upon incubation with 0.25 μM drug equivalent carried by HA‐coated NPs and uncoated NPs or free drug. 308 HA‐targeting mesoporous SiNPs with both a therapeutic agent and HA attached to NP's surface by disulfide bonds were investigated by Zhao et al. 309 Disulfide bonds were cleaved by glutathione, causing efficient drug release (Fig. 19 ). Cytotoxicity assessment showed that after 72 hr of treatment with the drug‐loaded conjugate, a cell line with CD44 receptors retained a viability of 39%, while the viability of a control cell line without CD44 was 49%, suggesting rather strong nonspecific internalization by the NIH 3T3 cell line. 309 The same group also prepared mesoporous SiNPs with pores loaded with doxorubicin and capped by HA attached to the surface by disulfide bonds. 310 Excess glutathione and hyaluronidases helped to release capping molecules from the conjugate. 310 An elegant approach using nonspecific internalization based on differences between the metabolic activity of normal and cancer cells was reported. 311 Cancer cells exhibit increased glucose uptake with the enhanced uptake of synthesized glucose‐modified SiNPs. Celastrol used as an anticancer drug negatively affects the mitotic cycle, which is faster in cancer than in normal cells, thus making cancer cells more susceptible to the effect of the drug. Mesoporous SiNPs decorated with glucose via dendritic polymer linkers were found to induce the apoptosis of HeLa cells (50% apoptotic cells after 24 hr of incubation with 5.3 μM drug equivalent). On the other hand, a lower exposure time induced the production of heat shock proteins, suggesting that a subcytotoxic dosage (or exposure time) may even help to protect normal cells. 311 Figure 19 Schematic depiction of glutathione (GSH) induced drug release from hyaluronic acid (HA) capped pores of silica NPs. These pores are filled with disulfide‐bond tethered 6‐mercaptopurine (6‐MP) and, due to HA‐decoration, nanoparticles are captured and internalized via so called CD‐44 receptors—a common surface glycoprotein with significantly higher concentrations on the surface of certain cancer cells. Reprinted with permission from 309 . Copyright 2014 American Chemical Society. Carbohydrate modification can also be used for the selective and triggered drug release from mesoporous magnetic SiNPs. 312 Lectin‐like phenylboronic acid (PBA) displayed on their surface anchored dextran, which efficiently gated drug‐loaded pores in the body of magnetic SiNPs. Once glucose is present in sufficiently high concentration, it competitively binds to PBA, and the dextran "gatekeeper" layer is consequently released from the surface. Due to the much smaller size of glucose compared to dextran chains, drug molecules can easily migrate from the pores. This glucose‐dependent release has great potential, for example, in diabetes treatment, even though a selective cytotoxicity against HeLa cells in the presence of glucose was demonstrated in this study (60% decrease of viability compared to a 25% decrease upon treatment of the cells with unloaded and drug‐loaded NPs without dextran, respectively). 312 Employing the same gate‐keeping approach, Zhou et al. bound mannose units on the surface of mesoporous SiNPs via disulfide bonds. 313 These moieties tethered Con A lectin, a molecule large enough to prevent the migration of doxorubicin from the pores of mesoporous SiNPs. Selective release was triggered by the elevated concentration of glutathione in the cytoplasm of cancer cells, and the conjugate exhibited an IC 50 of 25 μg mL −1 against cancer cells, while a negligible viability change was observed when the conjugate was incubated with normal cells having a lower cytoplasmic glutathione concentration. 313 Chondroitin sulfate is another natural polysaccharide with the ability to promote the uptake of conjugates by selective interaction with CD44 receptors. Moreover, such a biopolymer has the ability to improve the biocompatibility and stability of NPs. All of these features of chondroitin sulfate were used to fabricate a mesoporous silica‐based vector loaded with a plasmid carrying the gene for the p53 protein. 314 A more than twofold increase in translated messenger RNA for p53 was observed in treated cancer cell lines, when the silica‐based vector was coated with chondroitin sulfate compared to bare mesoporous silica–plasmid conjugate. 314 Gene therapy induced expression of healthy p53 protein in human cancer cells inhibits tumor growth; thus, the above‐described vector represents another possible way for cancer treatment. 4. Polymer Scaffold Based Drug Carriers Liposomes prepared from a synthetic building block containing a hydrophobic tail conjugated with galactose as targeting units were capable of transporting doxorubicin easily into HepG2 cells displaying asialoglycoprotein 315 , 316 , 317 and other receptors. 44 A very similar system, only with incorporated disulfide bonds in the structure of micelles, was reported. 318 , 319 , 320 , 321 Such glycoconjugates were selectively internalized, and a drug was released after the glutathione‐mediated cleavage of S–S bonds. 318 , 319 , 320 , 321 Although these studies demonstrated the selective cytotoxicity toward targeted cancer cells in vitro, only Zhao et al. have reported the in vivo inhibition of tumor growth after the localized injection of galactosylated, doxorubicin‐loaded liposomes (tumor weight of approx. 50% compared to animals treated with an unconjugated drug). 316 The unfolding of synthetic 322 or natural 323 micelles triggered by UV irradiation for drug release was also reported, but such an approach was not tested for in vitro or in vivo cytotoxicity. In order to better control the form and chemical properties of nanocarriers, a tri‐block copolymer folded into different types of mannosylated micelles was synthesized. 324 Preliminary results suggested that there was some correlation between mannose receptor‐induced internalization and the shape of a micelle, but in vivo or in vitro studies investigating the inhibition of virus binding have yet to be conducted. 324 To discuss drug delivery from an application point of view, it is worth mentioning the use of a synthetic copolymer folded into micelles tested for the oral administration of sorafenib—the drug applied mainly for the treatment of kidney and liver cancers. 325 In vivo studies have demonstrated that galactosylated, drug‐loaded micelles efficiently and selectively administered sorafenib into the liver and secured its continuous release for several hours. 325 To improve the oral administration of darunavir, self‐assembled micelles made from synthetic copolymer poly(ethylene oxide)‐poly(propylene oxide) were glycosylated by microwave‐assisted ring opening reaction and loaded with the drug by a simple incubation of the two components in solution. 326 An approximately 500‐fold higher solubility of darunavir was observed upon its integration into micelles. This is a significant step in the development of less expensive anti‐HIV drugs, considering the fact that the low administration efficiency of present darunavir‐based delivery increases its cost. 326 Berberine hydrochloride is another drug with a reported anticancer effect with chemical properties not allowing oral administration. 327 Drug encapsulation into chitosan‐coated phospholipid micelles significantly improved pharmacokinetics and distribution parameters due to the increased biocompatibility of the conjugate. It should be noted, however, that only the biodistribution and drug release profile were assessed in the study, with results suggesting that chitosan‐coated liposomes are good vehicles for the oral administration of berberin. 327 Micelles prepared by the conjugation of α‐tocopherol (recognized as a potential anticancer drug) with pullulan were reported to release the drug 10‐hydroxycamptothecin at the low pH typically present in the cytoplasm of tumor cells. 328 In vitro cell assays revealed an approximately 60% decreased viability of targeted cells compared to a control experiment with another cell line. 328 A conjugate based on α‐tocopherol derivatives together with HA and cytostatic docetaxel in the micellar form was found to induce cytotoxicity and to decrease multidrug resistance via the α‐tocopherol‐mediated inhibition of the active drug transport from the cell. 329 The suppression of drug resistance on a molecular level was also achieved by the delivery of two recently synthesized inhibitors of efflux proteins (P‐glycoprotein and breast cancer resistance protein) loaded in poly‐lactic‐co‐glycolic‐acid. 330 A higher inhibition level of cancer cells incubated with conjugated compared to unconjugated inhibitors was demonstrated. 330 Interestingly, the same efflux mechanism was also suppressed by hydroxypropyl‐β‐cyclodextrin moieties grafted onto synthetic copolymer NPs, which enhance the effect of low‐soluble immunosuppressor tacrolimus by oral administration. 299 Atherosclerosis can be efficiently treated by the internalization of micelles prepared from a reconstituted high‐density lipoprotein (rHDL) loaded with levostatin mediated via Scavenger Receptor class B member 1 (SR‐BI) receptors accessible on the surface of cells present in a macrophage‐infiltrated atherosclerosis lesion. 331 Upon coating of the fabricated micelles with HA (Fig. 20 ), their administration became more efficient and precisely targeted in two steps: (i) because of their lower accumulation in the liver, this is achieved because HA‐decorated particles are not recognized by SR‐BI receptors and "scavenged" (recognition by SR‐BI receptors in liver cells leads to an increased level of phagocytosis). Thus, more particles are delivered into atherosclerotic veins. In the second stage, (ii) HA was readily recognized by CD44 receptors present in atherosclerotic plagues, where HA‐LT‐rHDL was accumulated. Next, due to the presence of hyaluronidases, HA was removed. The formed LT‐rHDL NPs were recognized by SR‐BI receptors of atherosclerosis lesion cells, which delivered them into desired "foam cells" where levostatin was released. 331 Figure 20 Scheme of fabrication and function of dual‐targeted reconstituted liposomes. Upper scheme represents preparation of micelles containing levostatin and coated with HA (HA‐LT‐rHDL). After injection in an atherosclerotic rabbit, HA coating allowed preferential distribution of HA‐LT‐rHDL into veins and their retention in an organism with decreased rate of phagocytosis in liver. Once HA‐LT‐rHDL reached a lesion site, it was recognized by CD44 receptors on the surface of endothelial cells which helped their transportation inside an atherosclerotic plague (the scheme in the lower part of the figure). In the plague, hyaluronidases removed HA from HA‐LT‐rHDL and the resulting LT‐rHDLs were preferentially recognized by SR‐BI receptors on surface of "foam cells" where lovastatin was intended to be released. Reprinted from 331 . Copyright 2014, with permission from Elsevier. The antiseptic drug NK007 was delivered successfully by oral application to regress murine colitis. 332 Polysaccharide components of micelles are responsible for efficient target delivery as could be concluded from the in vivo study. Nevertheless, a significant difference between the parameters recorded to evaluate treatment efficacy (i.e., colon length, stool consistency, etc.) was only observed between the control and treated groups but not between groups treated with unconjugated or NP‐conjugated NK007. 332 Amphotericin B, an antiparasitic drug effective, for example, against Leishmania donovani , induces several harmful side effects, with nephrotoxicity and hemolytic toxicity being the most severe. 333 Moreover, the drug is relatively expensive. Khan et al. conjugated this drug with mannosylated synthetic dendrimer to gain biocompatibility. As a result, no kidney damage or hemolysis was observed with the antiparasitic activity of the drug preserved. In fact, due to the mannose‐driven targeting of conjugates into macrophages, IC 50 (39 nM) decreased approximately fivefold compared to the unconjugated drug or its commercial analogue. 333 A supramolecular scaffold prepared from a polymer bearing PBA and a "complementary" glycopolymer was degraded under hyperglycaemic conditions. 334 The conjugate could be applied for the safe, comfortable, and glucose concentration‐sensitive distribution of insulin. 335 Zheng et al. also employed a PBA‐based conjugate with both PBA and lactobionic acid moieties displayed randomly on each copolymer chain. 334 Such molecules folded into a scaffold able to encapsulate insulin. PBA enhanced the mucoadhesion of the scaffolds, which could be then used for the nasal administration of insulin. Therapeutic effect was confirmed by in vivo experiments in diabetic mice with glucose levels decreased to approximately 40% 8 hr after the addition of therapeutic NPs, while such a decrease was observed within 0.5 hr when bare insulin was injected. 334 The single‐step ligation of HA with doxorubicin seems to be much simpler, forming conjugates that are selectively cytotoxic to cancer cells displaying CD44 receptors 336 similar to pullulan‐doxorubicin micelles selectively targeting cells with asialoglycoprotein receptors. 337 , 338 A similar weight reduction of induced tumor was observed in vivo after treatment with unconjugated and conjugated doxorubicin (approximately 0.4 g of tumor weight compared to 1.8 g developed in a control group), but the life span was longer and the animal weight was larger for NP‐treated mice compared to mice treated with unconjugated doxorubicin, suggesting the significant suppression of a negative side effect of a chemotherapy. 338 Plant glucomannan covalently conjugated with an inhibitor (bisphosphonate alendronate) of tumor‐associated macrophage was also prepared as a conjugate inducing selective, mannose receptor‐mediated macrophage apoptosis (a tumor weight of approx. 0.3 g was observed 14 days after the application of therapeutic NPs compared to 1.5 g developed in an untreated group). 339 Mannose receptors of macrophages were also targeted with ionic polymer‐based micelles containing the model protein to be delivered (BSA). 340 Upon coating of the micelles with β‐glucan from yeast cell walls, NPs were selectively internalized by macrophages. 340 So‐called "ivy NPs" are nature‐derived nanocarriers consisting of a heavily glycosylated protein core. 341 When loaded with doxorubicin, a targeted delivery and release of a drug into different cancer cell lines was observed. 341 The treatment of a solid tumor was equally efficient for the application of both the conjugate and unconjugated doxorubicin, but with a significantly lower weight loss of the animal observed after the application of the conjugate. 341 Quite a "minimalistic" approach was reported, relying on short polyethylene glycol chains decorated with several saccharidic units along their length and with a bioactive molecule on the chain end. 342 Such conjugates bearing two separated mannosyl units (the optimum distance between these units was found to be 5.6 nm) were internalized by macrophage cells via their mannose‐binding receptors. The authors declared a good biocompatibility of the carriers as demonstrated by the absence of lysosome storage disease‐like effect after their application. 342 On the other side, in vivo tests of applicability of such conjugates have yet to be performed. Similarly, Thomas et al. reported micelles prepared from bis‐ l ‐galactose lysine. 343 Galactose displayed on the surface of these NPs was enzymatically oxidized, and the aldehydes formed were used for the covalent crosslinking of more NPs together by di‐ or trivalent amine. The authors tested their delivery system only for the complexation of AuNPs, but selective delivery of drugs is anticipated. 343 A selective drug delivery into the central nervous system, passing the blood–brain barrier, is always very difficult. Recently, certain lectins displayed on NPs helped to enhance such crossing. 344 For example, polyethylene glycol based micelles decorated with lectin Solanum tuberosum were able to successfully deliver a basic fibroblast growth factor via the nasal route into the brain in order to improve the cognition of those affected by Alzheimer's disease. 344 Lactoferrin, an iron‐binding glycoprotein, was found to be helpful in delivering drug‐loaded micelles through the blood–brain barrier 345 , and the same protein was used to target asialoglycoprotein receptors on HepG2 cancer cells, as well. 346 Another way to pass the blood–brain barrier is using activated membrane transporters. Glucose transporter 1 was activated by p‐aminophenyl‐β‐ d ‐manno‐pyranoside with a subsequent transportation of anticancer drugs into the brain cells. 347 Kuo et al. prepared NPs based on solid lipid decorated with p‐aminophenyl‐β‐ d ‐manno‐pyranoside and folic acid and loaded with etoposide (a drug inhibiting the proliferation of malignant glioblastoma) that were efficiently distributed through the blood–brain barrier with efficient secondary targeting to glioblastoma cells secured by folic acid in in vitro experiments. 347 Nonviral gene transport is just another field where the aforementioned techniques of controlled targeted drug delivery have been employed. One of the materials used quite often for the preparation of nonviral vectors is PEI, which can be effectively complexed with the DNA or RNA of interest, but it possesses a certain level of toxicity and its targeting is not selective and thus suitable only for in vitro experiments. In order to apply PEI‐based conjugates for in vivo applications, several ways to improve the performance of such carriers were sought. For example, PEI was complexed with a biocompatible anionic glucose‐based glycopolymer, 348 GlcNAc, 349 chondroitin sulfate, 350 HA, 351 or depolymerized guar gum with available mannose units 352 with subsequent use in selective targeting. Alternatively, copolymers of the methacrylamide backbone decorated with glucose, 353 α‐ d ‐mannopyranosyl, 354 or trehalose 355 and cationic moieties could electrostatically complex plasmids with the subsequent selective transport of plasmid DNA into cancer cells 353 , 354 or siRNA into glioblastoma cells via glucose transporter‐1. 355 Furthermore, complexed siRNA was reported to retain its biological activity after freeze drying. 355 An electrostatic complexation of DNA with copolymer chains bearing a targeting GlcNAc moiety was also reported. 356 Targeted gene delivery was also achieved by DNA incorporation into PEGylated micelles decomposing at low pH and decorated with galactose 357 or mannose 358 to target asialoglycoprotein or mannose receptors, respectively. Noteworthy, a secondary target ligand was co‐entrapped with DNA to help penetrate through the nuclear membrane. 357 Mannose‐decorated cholesteryl‐based synthetic liposomes with virus‐like characteristics were reported to effectively transfect nonactivated dendritic cells by a plasmid‐containing gene for luciferase. 359 Mannosylated chitosan was also used for nonviral vector preparation 360 as well as mannosylated "bubble lipoplexes," that is, NPs with ultrasound‐triggered DNA release causing the transfection of M2 macrophages and their switch from tumor growth‐promoting into tumoricidal M1‐like macrophages. 361 , 362 In vivo experiments confirmed that this treatment could inhibit the growth of various tumors without adverse side effects. 361 , 362 A triblock copolymer consisting of hydrophobic and pH‐sensitive parts to assure the formation of micelles, polycation complexing oligonucleotides, and a mannose‐modified part was also applied to target M2 macrophages. 363 Synthetic vectors based on siRNA‐PEGylated cyclodextrin scaffolds with a peptide ligand as a targeting unit were also investigated. 364 In addition to the broadly used electrostatic interactions used for loading nucleotides into delivery NPs, Kim et al. employed hybridization with a specific DNA template with a condensation of the resulted DNA complex by a viral Mu peptide (Fig. 21 ). 365 The formed "DNA nanoballs" coated with HA were effectively internalized by cancer cells. In vivo experiments showed that antisense oligonucleotides were hybridized with their complementary mRNA in the cytoplasm, leading to a significant sensitization of the cells to doxorubicin treatment and to a significant decrease of tumor growth in mice. 365 Figure 21 RCA template and hybridization efficiency. (A) Secondary structure of scrambled rolling circle amplification (RCA) template. (B) Secondary structure of a dual antisense oligonucleotide (ASO) hybridizing RCA template for Dz‐13 and OGX‐427. (C) RCA products with poly ASO‐binding sequences were hybridized with two therapeutic ASOs, Dz‐13 and OGX‐427, to produce dual ASO‐hybridizing RCA products. (D) The hybridization efficiencies of the ASOs were tested for products of scrambled RCA templates and dual ASO‐hybridizing RCA templates using fluorescently labeled ASOs. Reprinted from 365 . Copyright 2015, with permission from Elsevier. A very simple approach relying only on the application of bare therapeutic oligonucleotides conjugated with a targeting ligand (e.g., a triantennary GalNAc 366 , 367 , 368 or glucose 369 ) was also tested. These agents were capable of blocking the adverse immunostimulation of an anti‐HIV oligonucleotide. This may be helpful in the fabrication of anti‐HIV drugs without adverse side effects. The conjugate can be easily prepared, but there is a concern regarding the stability of the carrier when exposed to endo‐ and exonucleases. B. Photothermal and Photodynamic Therapy Photothermal and photodynamic therapeutic methods rely on the targeted delivery of entities capable of strong energy absorption in the NIR spectrum. As a result, the NIR‐triggered increase of local temperature damages targeted (cancer) cells. Traditionally, metallic NPs have been used as photothermal agents, with AuNPs being the first choice. For example, HA‐covered gold nanostars very efficiently inhibit HeLa tumor growth in mice with almost complete tumor disappearance within 7 days after injection of the conjugate and laser irradiation. 370 The recent trend, however, is the use of a combination of simple photothermal therapy with other treatments, for example, chemotherapy. AuNRs wrapped in GO sheets conjugated with doxorubicin were reported to be fourfold‐stronger inducers of HeLa cell death compared to the same conjugates without doxorubicin used for sole photothermal treatment. 290 In addition to AuNPs, carbon mesoporous NPs were also employed simultaneously as carriers and photodynamic agents. 296 A synergic chemotherapeutic and photothermal effect was observed after loading their pores with doxorubicin and final coating with HA. The observed IC 50 against the MDA‐MB‐231 cell line decreased from 8.7 to 2.3 μM doxorubicin after simultaneous treatment with the drug‐loaded NPs and NIR laser. 296 A very simple but effective system was reported based on the electrostatic integration of HA and polyaniline applied as a targeting and photosensitizing component, respectively. 371 The use of these NPs with subsequent NIR laser treatment led to the complete degradation of induced tumors in mice approximately 8 days after treatment without any observable body weight loss. 371 However, the ternary design of multimodal therapeutic NPs was presented, containing poly(lactic acid) NPs coated with derivatized chitosan to increase the biocompatibility of the conjugate and IR 820 (a photosensitive dye). 372 Interestingly, two absorption peaks were found for IR 820 encapsulated in the polymer; thus, irradiation with two different wavelengths increased the photothermal effect with the decreased viability of the targeted breast cancer cell line by 70%. 372 In another treatment design called photodynamic therapy, photosensitizers were used to generate cytotoxic ROS upon the absorption of irradiation with a specific wavelength. Recently, AuNPs coated with HA (a targeting component) and a porphyrin‐based photosensitizer were reported to decrease the viability of cancer cells with CD44 receptors down to approximately 15% upon irradiation with UV light. 373 AuNPs can also exhibit a photosensitizing effect, but for the generation of cytotoxic radicals, X‐ray irradiation is required. In this sense, glucose‐decorated AuNPs were tested in vitro as selective sensitizers for radiotherapy, leading to increased cancer cell death by 20% compared to radiotherapy with unconjugated AuNPs. 374 Organic carriers for ROS‐generating photosensitizers were also reported. For example, porphyrin‐derived photosensitizer conjugated to glucose‐decorated poly(methacrylamide) polymer chains 375 or encapsulated into poly(lactide‐co‐glycolide)‐based NPs coated with HA. 376 In addition to porphyrin‐based photosensitizer, docetaxel was also co‐entrapped to induce a synergistic photochemotherapeutic effect with IC 50 of 8 ng mL −1 , while an IC 50 of 160 ng mL −1 was observed upon treatment with an unconjugated drug under the same conditions. 376 Fu et al. used chitosan‐coated SiNPs loaded with benzoyl peroxide as photosensitizers. 377 A cytotoxic test against human breast carcinoma cell line ZR75‐30 revealed that chitosan coating did not substantially affect cell viability (70 and 74% of retained viability after incubation with uncoated and coated NPs, respectively), even though the conjugation of benzoyl peroxide increased cytotoxicity (90 and 74% of viability retained after treatment with unconjugated vs. conjugated benzoyl peroxide, respectively). 377 GO was successfully applied as a carrier for the loading of photosensitizers, as reviewed recently. 378 Recently, HA was conjugated to GO nanosheets with the subsequent physisorption of a photosensitizer. 379 The efficient uptake of such modified GO NPs, with a size of 100 nm, was observed. The photosensitizer desorption led to enhanced NIR absorption and the generation of ROS in the cells. The authors reported IC 50 values of 1 and 0.1 μg mL −1 of photosensitizer equivalent, respectively, when targeted cells were treated with either free or NP‐conjugated photosensitizer. 379 A system combining both photodynamic and photothermal effects has also been developed. While an rGO/ZnO hybrid is responsible for ROS generation upon light irradiation, NIR laser irradiation induced local hyperthermia 322 because rGO effectively transferred heat to cells, with a local temperature increase upon exposure to NIR irradiation. 380 To secure targeted treatment, HA was conjugated with the rGO/ZnO complex. The function and structure of rGO/ZnO‐HA NPs are depicted in Figure 22 , where the viability of targeted cells is also depicted. From these plots, it is obvious that combined photodynamic and photothermal therapy decreased cell viability down to approximately 20%, much more than the application of only one type of treatment. 322 Figure 22 (A) Schematic illustration of sequential irradiation‐activated high‐performance apoptosis. (B) In vitro cell viability of MDA‐MB‐231 cells treated with rGO‐ZnO‐HA following PDT, PTT and combined PDT/PTT treatments. Reproduced from 322 , with permission by John Wiley & Sons. Targeted irradiation by neutron beam is yet another way to induce the cytotoxic effect of certain agents, such as 10 B ions using "boron neutron capture therapy." Radionuclide 10 B, after accumulating in the cells, captured supplied protons and underwent fission, generating heavily cytotoxic 7 Li and 4 He particles. 381 To achieve a sufficient amount of boron in cells (on average 10 11 10 B ions per cell), mesoporous SiNPs were used to carry 10 B coordinated to an organic ligand with high selectivity and an accumulation rate achieved by the decoration of the NPs with multivalent galactose units. The delivery efficiency was illustrated by having approximately 3 × 10 9 atoms in cells when treated with unconjugated boron ligand compared to 10 11 atoms when delivered by NPs. Preliminary results suggest that the NP‐delivered amount of 10 B atoms is high enough to efficiently perform boron neutron capture therapy. 381 C. Cell Imaging The principles applied to selective delivery can also be employed in cell imaging with appropriate carbohydrate‐binding receptors. Many studies have been published regarding the conjugation of glycan‐displaying NPs with dye molecules to facilitate detection in the UV‐VIS range, contrast enhancers for NMRI or computed tomography imaging of tissues or single cells accumulating nanocarriers. 382 1. UV‐VIS Probes The recent trend is to combine the therapeutic and imaging effects in so‐called theranostic (i.e., therapeutic + diagnostic) NPs loaded with both an imaging probe and a drug as discussed above. 44 , 381 , 383 The imaging function of the conjugate can also be combined with antiadhesive properties. 239 Noteworthy, NIR‐absorbing GO/doxorubicin/HA NPs for combined photochemotherapy could be applied to imaging even without any additional dye or label incorporated due to the fluorescent signal obtained from a carrier drug (e.g., doxorubicin). Such an approach was applied for the visualization of cells that internalized nanocarriers. 281 , 384 Dye‐loaded NPs releasing their cargo after internalization promoted by surface‐displayed carbohydrate moieties were reported. 385 The encapsulation of dyes into polymer NPs significantly increased the biocompatibility of imaging probes with the selective visualization of intracellular membranes of targeted cells (Fig. 23 ). 385 Self‐assembled NPs prepared from a natural polysaccharide levan were able to encapsulate indocyanine green. 386 The selective accumulation of the conjugate in tumor cells was mediated via overexpressed glucose receptors. 386 NPs formed from a copolymer containing fluorescent rhodamine and glucose were successfully internalized by cells with asialoglycoprotein receptors enabling their fluorescent imaging. 387 The chemical and enzymatic stability of the imaging NPs were very good, with a maximal release of only 2.5% of the loaded dye within 4 days in the presence of lipase and esterase. 387 Figure 23 Effective intracellular delivery of rhodamine B octadecyl ester in HDF cells mediated by (1:9 PGalSMA34 + PGMA51) PHPMA270 vesicles (prepared via thin film rehydration to ensure sterility and enable loading with a fluorescent dye). Cells were incubated with 1.0 mg/mL rhodamine B octadecyl ester loaded vesicles for 16 hr. (A) Confocal microscopy image of live HDF cells: note the intracellular staining of membranes (red) after exposure to the rhodamine‐loaded vesicles, cell nuclei are counter‐stained blue using Hoechst 33342. (B) HDF cells treated with the same vesicles containing no rhodamine dye (negative control). (C) Higher magnification image obtained for (A): effective intracellular delivery of rhodamine dye allows selective staining of the nuclear membrane (white arrows). Scale bar: 50 μm. Reprinted with permission from 385 . Copyright 2013 American Chemical Society. In addition to diagnosis, the dye labeling of NPs has also been used to evaluate the cell uptake rate. For example, the preparation of mannosylated micelles was optimized by the visualization of their uptake in the cells. 388 Further in vivo experiments confirmed their selective imaging capability. 388 Perylene bis(imide) derivative decorated with mannose was found to be fluorescent only after the degradation of the formed scaffold triggered by its ligation to a mannose‐specific lectin. 389 The fluorescent imaging of macrophage cells overexpressing such receptors has been reported (Fig. 24 ). 389 A novel NIR fluorophore activated upon isomerization at a low pH in lysosomes was used by Wu et al. 390 Biocompatible poly(styrene‐co‐maleic acid) NPs loaded with this probe specifically targeting surface‐displayed sialic acid were able to selectively tag and mark tumors in vivo. 390 Cell imaging using illumination outside the range of UV‐VIS and NIR has also been tested, for example, by Mäkilä et al., who radiolabeled ( 68 Ga) therapeutic siRNA oligonucleotides conjugated with multivalent galactose units to evaluate their in vivo distribution. 391 They observed that the valency of the displayed galactose was crucial to selective and efficient accumulation in liver, and the conjugate with seven glycan units exhibited the best performance, making them suitable for the further development of targeted genotherapy. 391 Figure 24 Confocal microscopic images of murine macrophage cells after incubation with water‐soluble glycocluster based on perylene bisimides PBI‐12‐Man (10 μg/mL, Man = mannose) without (a, excited at 559 nm; b, bright field; c, merging of photos a and b) or with mannose (18 mg/mL) inhibition (d, excited at 559 nm; e, bright field; f, merging of photos a and b) in PBS buffer at 37°C for 1 hr. As shown in a–c, the red fluorescence of PBI‐12‐Man was predominantly intracellular with punctate appearance, suggesting cell surface binding and endocytosis being the cell entry mechanism that results in vesicular (endosomal) localization. For the inhibition experiment, in the presence of α ‐d‐mannopyranoside (d–f), the remarkable decrease of the red fluorescence of PBI‐12‐Man indicated the selectively binding interactions of PBI‐12‐Man with the surface mannose receptor of the macrophage cells. Reprinted from 389 . Copyright 2014, with permission from Elsevier. Figure 25 Photoacoustic tumor imaging in mice with sialic acid‐decorated polymer nanoparticles (pNIR@P@SA). Nude mice bearing H22 subcutaneous tumors were intravenously injected with phosphate buffer saline (PBS; 100 mL) or pNIR@P@SA (40 mg kg −1 ). The mice were imaged 24 hr after vesicle injection. Control images were obtained from mice before intravenous injection of pNIR@P@SA or PBS. A significantly higher photoacoustic signal obviously resulted from tumor issue after incubation with pNIR@P@SA (lower right panel). Reproduced from 390 with permission of the Royal Society of Chemistry. Figure 26 CT number for tumor versus time. The blue line is the CT number of the control group, whereas the red line represents for the PEG Glc‐GNP injection group. Reprinted from 401 , with permission from Elsevier. Other recent studies reported the preparation of simple nanocarriers based on synthetic 392 , 393 or natural 394 , 395 polymer NPs conjugated with glycans and imaging probes with selective adhesion to specific cell surface receptors, enabling the visualization of the surfaces of these cells. A complex study was carried out to visualize the binding of radiolabeled mannose‐containing polymer (γ‐Tilmanocept, approved by the FDA as a lymphatic node imaging probe) to targeted cells. 396 The authors used a cyanine dye conjugated with Tilmanocept to specifically bind mannose receptors of macrophages, facilitating the delivery of a radiopharmaceutic drug. 396 Multivalent clusters of iminosugars conjugated with pyrene or boron‐dipyrromethene cores exhibited fluorescence spectra applicable in cell imaging. 397 As discussed earlier, multivalent iminosugars have been intensively tested as promising pharmacological chaperones. The conjugation of nanocarriers with photosensitizers applied in photodynamic therapy has also often been used for simultaneous therapy and imaging since the majority of molecules applied as photosensitizers exhibit fluorescence. For example, a porphyrin derivative conjugated with glycopolymer exhibited an emission peak at 633 nm, allowing the real‐time visualization of the NP uptake with confirmed targeted cytotoxic effect. 375 A similar effect was also observed with functionalized GO integrated into a pH‐sensitive HA nanogel serving as selective carrier for doxorubicin. 288 The conjugate based on GO allowed both photodynamic therapy and fluorescence imaging, 288 underlining the amazing versatility of graphene‐based nanomaterials. Deeper tissue penetration and better resolution are the main advantages of photoacoustic imaging. In this method, a probe with an absorption maximum at longer wavelengths is used. The absorption resulted in changes in the probe's properties (thermoelastic expansion); hence, after the activation by a deep‐penetrating laser pulse, the probe was transformed into a detectable ultrasound signal. 398 This imaging method was employed using sialic acid‐decorated polymer NPs loaded with a profluorophore, turning it into a photoacoustic probe by isomerization at a low lysosomal pH. 390 In the end, it became evident that a significantly higher photoacoustic signal was achieved in tumor vein imaging using this method—see Figure 25. 390 2. Metallic NPs in Selective Cell Imaging An inherent NIR‐induced fluorescence of small Au nanoclusters with the size of 2 nm was used in the preparation of selective imaging probes. Selectivity, stability, and biocompatibility were achieved by the conjugation of the nanoclusters with zwitterions and trivalent mannose ligands. 399 The integration of the latter component into the conjugate increased the internalization rate by 62 and 256% when targeted dendritic cells were incubated with 1 and 25 μg mL −1 of the NPs, respectively. 399 The absorption of energy by AuNPs from high‐frequency radiation has been employed to enhance the effectivity of radiotherapy (see Section 6.A.5) and as a contrast enhancer during computational tomography (CT) scans, as well. For example, in situ generated small AuNPs modified by dendrimers containing polyvalent lactobionic acid were reported to enter selectively hepatocarcinomic cells via their asialoglycoprotein receptors, allowing the high‐resolution computational tomography imaging of tumors. 400 The same method for selective tumor visualization was also enabled by glucose‐modified AuNPs mediated via glucose transporters overexpressed on the surface of cancer cells. 401 Tumors could be monitored for a week, as shown in Figure 26. 401 Glucose‐conjugated AuNPs bearing a ligand with 68 Ga as a tracer were biocompatible and excellent for in vivo imaging using positron emission spectroscopy. 402 Moreover, neuropeptides loaded on glucose‐ 68 Ga‐AuNPs enabled penetration through the blood–brain barrier. 402 Magnetic resonance is a broadly used imaging method in which the application of contrast agents significantly improves image quality, especially when these are in the form of NPs. In most experiments, iron oxide NPs decorated with stabilizing, targeting, or biocompatible materials have been tested. 403 As an example, versatile glycopeptide grafting to magnetic iron oxide NPs was reported but without in vivo testing. 404 A step further using sialic acid modified Fe 3 O 4 NPs for the selective in vivo MR imaging of β‐amyloids applicable for the early diagnosis of Alzheimer's disease was reported. 405 Mannose‐containing diblock copolymer grafted onto Fe 3 O 4 NPs enhanced the carbohydrate receptor‐driven uptake into lung cancer cells with potential application in the diagnosis and localization of this type of cancer. 406 HA‐grafted Fe 3 O 4 NPs employed in the selective MR imaging of tissues with CD44 receptor‐overexpressing cells, that is, tumors, or local inflammations were also described. 407 , 408 A similar system with core Fe 3 O 4 NPs having a thin golden shell covered with an HA layer enabling simultaneous MR imaging and photothermal therapy targeted toward cells with CD44 receptors has been reported. 370 In order to increase the in vivo stability of the Fe 3 O 4 core, its coating with a silica shell was described in several studies. 384 , 409 An enhanced imaging resolution of brain lesions in mice after simulated stroke can be observed in Figure 27 , showing that NPs with different glycosylation exhibited different biodistribution. 409 Sialic Le x present on NPs allowed their accumulation in ischemic, that is, damaged, brain tissue due to selectin overexpression; NPs decorated with Le x (without sialic acid present) could visualize such tissue less effectively, which was also confirmed by the software evaluation of dark and bright voxels in the images. 409 Nevertheless, good results were also obtained with NPs bearing hydroxyls groups on the NP surface instead of glycans. 409 In addition to the protective function of a silica shell, such a layer could provide a better environment for the covalent grafting of targeting ligands. The application of silica‐based nanocarriers for MR imaging has recently been reviewed by Caltagirone et al. 410 Figure 27 T2 maps and T2‐weighted images at 5 and 24 hr after MCAO in a representative animal tissue that received iron oxid‐silia nanoparticles decorated with (A) hydroxyl groups (HO@MNPs), (B) Lewis X without sialic acid (LX@MNPs), or (C) sialyl Lewis X (SX@MNPs). Note: the color scale bar corresponds to the T2 value (msec). Reprinted with permission from 409 . Copyright 2014 American Chemical Society. In addition to Fe 3 O 4 NPs, AuNPs were also applied in cell imaging. Small AuNPs of 1.8 nm in size grafted with lactobionic acid via PEG spacers were found to penetrate human hepatocellular carcinoma cells selectively via their overexpressed asialoglycoprotein receptors. 400 While T2‐weighted MRI was used for contrast enhancement applying iron oxide NPs, T1‐weighing was used for the same purpose using AuNPs loaded with a Gd 3+ complex. 411 The modification of AuNPs by a combination of different carbohydrates helped to achieve a high imaging selectivity of tumor cells in mice models. 412 AgNPs prepared by the reduction of Ag ions using saccharides present in sugar cane juice have ferromagnetic properties and could be applied for MRI. 411 3. Quantum Dots as Imaging Probes QDs as NPs composed of single nanocrystals of metal sulfides or other semiconductors have distinct optical properties resulting from their nano‐size. One of the most important features is their stable and adjustable fluorescence together with the overall stability inherent to all inorganic materials. Recent studies have reported that, after the encapsulation of QDs into a polymer shell with a final modification by carbohydrate moieties (for specific cell targeting), the resulted NPs still retain their fluorescence. The surface of QD‐modified NPs could be modified by a diverse range of functional groups with adjusted properties of NPs as reported by Schmidtke et al. 413 This study suggested applicability of such NPs in in vivo imaging, even though this has not been tested. 413 On the other side, the uptake and intracellular distribution of core‐shell CdSe‐ZnS QDs covalently grafted with different carbohydrates were assessed. 414 In addition to lactose, which enhanced the uptake of the QD‐glycan NPs into cancer cells, the composition of a mixed carbohydrate shell (i.e., lactose and mannose vs. lactose and maltotriose) significantly altered the internalization of NPs. 414 In another study, a secondary radiotracer ( 125 I) was used as the probe conjugated with CdSe/CdS QDs decorated with a sialic acid‐binding lectin. 415 These NPs were able to image breast cancer cells, as revealed by in vitro experiments. 415 Enhanced selectivity in the targeted imaging of muscle cells was observed for glucose‐decorated QDs with insulin‐induced uptake mediated via glucose receptors, as illustrated in Figure 28 , and such an approach could be applied for selective drug delivery as well. 416 Since the vast majority of inorganic QD particles exhibit cytotoxicity, Shinchi et al. focused on the development of cadmium‐free ZnS‐AgInS 2 NPs with lower cytotoxicity but still exhibiting good fluorescence properties and applicability in selective cell imaging after decoration with short targeting carbohydrates. 417 From the graphs shown in Figure 29 , it is obvious that all carbohydrate‐coated ZnS‐AgInS 2 QDs exhibited minimum cytotoxicity against HepG2 cells (panel A), unlike Cd‐based QDs (panel B). 417 Figure 28 (A) A molecular structure of glucose‐functionalized quantum dots (Glc‐QDs). (B) A schematic illustration of the cellular uptake of Glc‐QDs regulated by insulin and 2‐deoxyglucose (2‐DG) in C2C12 muscle cells. Reproduced from 416 with permission of the Royal Society of Chemistry. Figure 29 MTT assay for SFNPs. HepG2 cells were incubated with ZAIS/ZnS NPs (A) or CdTe/CdS QDs (B). The NP concentration was in the range of 5 to 50 μg mL −1 (left to right). Reprinted with permission from 417 . Copyright 2014 American Chemical Society. QD‐like fluorescence was assigned also to GO‐based NPs, 378 which were used, for example, for combined imaging and chemotherapy with an HA grafted as a targeting agent. 288 It is worth mentioning that the replacement of semiconductor‐based mostly cytotoxic QDs with less harmful graphene‐based NPs would be another step to develop novel, more efficient theranostic particles with low cytotoxicity. Silica QDs possessing inner fluorescence exhibited low cytotoxicity, having a surface that could be easily modified with glycans. 418 , 419 Furthermore, fluoride nanocrystals doped with rare earth element ions were also suggested as possible substitutes for "classical" Cd‐containing QDs—see, for example, 420 and references therein. 4. Toxicity of Nanomaterials It is beyond any doubt that the massive application and employment of nanomaterials have brought and should bring in the near future amazing achievements and technical possibilities—in fact, many authors are talking about the nanoage or nano‐revolution. Nevertheless, the other side of this "nano‐enthusiasm" is the persistent uncertainty about the possible toxicity of NPs. This issue is particularly delicate in medical applications, where NPs are supposed to be delivered directly into the human body and interact with cells at a molecular level. It should be noted that these concerns have been raised simultaneously with nanomaterial applications—see, for example, a 1982 study on the toxicity of a potential drug carrier. 421 In this study, one example of possible cytotoxicity mechanism of NPs was outlined, that is, the toxicity of monomers released from the intracellular degradation of NPs. Other known mechanisms, including unwanted aggregation of proteins, delivery of toxic molecules conjugated with NPs (including NPs from transportation and industrial emissions) and generation of highly cytotoxic ROS (the most toxic effect), were comprehensively reviewed. 422 , 423 , 424 , 425 The NPs' behavior in living organisms is prevalently determined by their physical and chemical properties. For example, decreased NP size increases their active surface and, consequently, the rate constant of any catalytic reaction taking place on the NP surface. Moreover, the size also determines the biodistribution mode in an organism—see, for example, different modes of interactions of small and large graphene sheets with cells (while the former tend to disrupt cell wall, 426 the latter more likely wrap it without any damage 427 ). Another crucial aspect is the nature of NP surface modification, including (i) surface charge, the adjustment of which can turn otherwise harmless NPs, for example, into cell disruptors (see references in 423 ) or other biological functions, 428 and (ii) the modification of NPs by biologically active moieties with examples, which can be found in sections of this review dealing with drug delivery, vaccine and other therapeutic effects, and cell imaging. Some toxic impacts can be estimated from the physical or chemical properties of the NPs acquired from a precise and accurate characterization of the NPs. This is usually accomplished since physical and chemical characterization and at least in vitro cytotoxicity tests are included in practically any study concerning the administration of NPs into the human body (drug carriers, imaging NPs, etc.). Nevertheless, the correlation between in vitro and model animal in vivo studies may significantly differ from the real impact on humans, nota bene when the subacute effect is considered. 429 Therefore, unless any negative health impacts are ruled out, appropriate precautions should be taken, and precise toxicity assessment must be performed in parallel with nanomedicine progress. 1. UV‐VIS Probes The recent trend is to combine the therapeutic and imaging effects in so‐called theranostic (i.e., therapeutic + diagnostic) NPs loaded with both an imaging probe and a drug as discussed above. 44 , 381 , 383 The imaging function of the conjugate can also be combined with antiadhesive properties. 239 Noteworthy, NIR‐absorbing GO/doxorubicin/HA NPs for combined photochemotherapy could be applied to imaging even without any additional dye or label incorporated due to the fluorescent signal obtained from a carrier drug (e.g., doxorubicin). Such an approach was applied for the visualization of cells that internalized nanocarriers. 281 , 384 Dye‐loaded NPs releasing their cargo after internalization promoted by surface‐displayed carbohydrate moieties were reported. 385 The encapsulation of dyes into polymer NPs significantly increased the biocompatibility of imaging probes with the selective visualization of intracellular membranes of targeted cells (Fig. 23 ). 385 Self‐assembled NPs prepared from a natural polysaccharide levan were able to encapsulate indocyanine green. 386 The selective accumulation of the conjugate in tumor cells was mediated via overexpressed glucose receptors. 386 NPs formed from a copolymer containing fluorescent rhodamine and glucose were successfully internalized by cells with asialoglycoprotein receptors enabling their fluorescent imaging. 387 The chemical and enzymatic stability of the imaging NPs were very good, with a maximal release of only 2.5% of the loaded dye within 4 days in the presence of lipase and esterase. 387 Figure 23 Effective intracellular delivery of rhodamine B octadecyl ester in HDF cells mediated by (1:9 PGalSMA34 + PGMA51) PHPMA270 vesicles (prepared via thin film rehydration to ensure sterility and enable loading with a fluorescent dye). Cells were incubated with 1.0 mg/mL rhodamine B octadecyl ester loaded vesicles for 16 hr. (A) Confocal microscopy image of live HDF cells: note the intracellular staining of membranes (red) after exposure to the rhodamine‐loaded vesicles, cell nuclei are counter‐stained blue using Hoechst 33342. (B) HDF cells treated with the same vesicles containing no rhodamine dye (negative control). (C) Higher magnification image obtained for (A): effective intracellular delivery of rhodamine dye allows selective staining of the nuclear membrane (white arrows). Scale bar: 50 μm. Reprinted with permission from 385 . Copyright 2013 American Chemical Society. In addition to diagnosis, the dye labeling of NPs has also been used to evaluate the cell uptake rate. For example, the preparation of mannosylated micelles was optimized by the visualization of their uptake in the cells. 388 Further in vivo experiments confirmed their selective imaging capability. 388 Perylene bis(imide) derivative decorated with mannose was found to be fluorescent only after the degradation of the formed scaffold triggered by its ligation to a mannose‐specific lectin. 389 The fluorescent imaging of macrophage cells overexpressing such receptors has been reported (Fig. 24 ). 389 A novel NIR fluorophore activated upon isomerization at a low pH in lysosomes was used by Wu et al. 390 Biocompatible poly(styrene‐co‐maleic acid) NPs loaded with this probe specifically targeting surface‐displayed sialic acid were able to selectively tag and mark tumors in vivo. 390 Cell imaging using illumination outside the range of UV‐VIS and NIR has also been tested, for example, by Mäkilä et al., who radiolabeled ( 68 Ga) therapeutic siRNA oligonucleotides conjugated with multivalent galactose units to evaluate their in vivo distribution. 391 They observed that the valency of the displayed galactose was crucial to selective and efficient accumulation in liver, and the conjugate with seven glycan units exhibited the best performance, making them suitable for the further development of targeted genotherapy. 391 Figure 24 Confocal microscopic images of murine macrophage cells after incubation with water‐soluble glycocluster based on perylene bisimides PBI‐12‐Man (10 μg/mL, Man = mannose) without (a, excited at 559 nm; b, bright field; c, merging of photos a and b) or with mannose (18 mg/mL) inhibition (d, excited at 559 nm; e, bright field; f, merging of photos a and b) in PBS buffer at 37°C for 1 hr. As shown in a–c, the red fluorescence of PBI‐12‐Man was predominantly intracellular with punctate appearance, suggesting cell surface binding and endocytosis being the cell entry mechanism that results in vesicular (endosomal) localization. For the inhibition experiment, in the presence of α ‐d‐mannopyranoside (d–f), the remarkable decrease of the red fluorescence of PBI‐12‐Man indicated the selectively binding interactions of PBI‐12‐Man with the surface mannose receptor of the macrophage cells. Reprinted from 389 . Copyright 2014, with permission from Elsevier. Figure 25 Photoacoustic tumor imaging in mice with sialic acid‐decorated polymer nanoparticles (pNIR@P@SA). Nude mice bearing H22 subcutaneous tumors were intravenously injected with phosphate buffer saline (PBS; 100 mL) or pNIR@P@SA (40 mg kg −1 ). The mice were imaged 24 hr after vesicle injection. Control images were obtained from mice before intravenous injection of pNIR@P@SA or PBS. A significantly higher photoacoustic signal obviously resulted from tumor issue after incubation with pNIR@P@SA (lower right panel). Reproduced from 390 with permission of the Royal Society of Chemistry. Figure 26 CT number for tumor versus time. The blue line is the CT number of the control group, whereas the red line represents for the PEG Glc‐GNP injection group. Reprinted from 401 , with permission from Elsevier. Other recent studies reported the preparation of simple nanocarriers based on synthetic 392 , 393 or natural 394 , 395 polymer NPs conjugated with glycans and imaging probes with selective adhesion to specific cell surface receptors, enabling the visualization of the surfaces of these cells. A complex study was carried out to visualize the binding of radiolabeled mannose‐containing polymer (γ‐Tilmanocept, approved by the FDA as a lymphatic node imaging probe) to targeted cells. 396 The authors used a cyanine dye conjugated with Tilmanocept to specifically bind mannose receptors of macrophages, facilitating the delivery of a radiopharmaceutic drug. 396 Multivalent clusters of iminosugars conjugated with pyrene or boron‐dipyrromethene cores exhibited fluorescence spectra applicable in cell imaging. 397 As discussed earlier, multivalent iminosugars have been intensively tested as promising pharmacological chaperones. The conjugation of nanocarriers with photosensitizers applied in photodynamic therapy has also often been used for simultaneous therapy and imaging since the majority of molecules applied as photosensitizers exhibit fluorescence. For example, a porphyrin derivative conjugated with glycopolymer exhibited an emission peak at 633 nm, allowing the real‐time visualization of the NP uptake with confirmed targeted cytotoxic effect. 375 A similar effect was also observed with functionalized GO integrated into a pH‐sensitive HA nanogel serving as selective carrier for doxorubicin. 288 The conjugate based on GO allowed both photodynamic therapy and fluorescence imaging, 288 underlining the amazing versatility of graphene‐based nanomaterials. Deeper tissue penetration and better resolution are the main advantages of photoacoustic imaging. In this method, a probe with an absorption maximum at longer wavelengths is used. The absorption resulted in changes in the probe's properties (thermoelastic expansion); hence, after the activation by a deep‐penetrating laser pulse, the probe was transformed into a detectable ultrasound signal. 398 This imaging method was employed using sialic acid‐decorated polymer NPs loaded with a profluorophore, turning it into a photoacoustic probe by isomerization at a low lysosomal pH. 390 In the end, it became evident that a significantly higher photoacoustic signal was achieved in tumor vein imaging using this method—see Figure 25. 390 2. Metallic NPs in Selective Cell Imaging An inherent NIR‐induced fluorescence of small Au nanoclusters with the size of 2 nm was used in the preparation of selective imaging probes. Selectivity, stability, and biocompatibility were achieved by the conjugation of the nanoclusters with zwitterions and trivalent mannose ligands. 399 The integration of the latter component into the conjugate increased the internalization rate by 62 and 256% when targeted dendritic cells were incubated with 1 and 25 μg mL −1 of the NPs, respectively. 399 The absorption of energy by AuNPs from high‐frequency radiation has been employed to enhance the effectivity of radiotherapy (see Section 6.A.5) and as a contrast enhancer during computational tomography (CT) scans, as well. For example, in situ generated small AuNPs modified by dendrimers containing polyvalent lactobionic acid were reported to enter selectively hepatocarcinomic cells via their asialoglycoprotein receptors, allowing the high‐resolution computational tomography imaging of tumors. 400 The same method for selective tumor visualization was also enabled by glucose‐modified AuNPs mediated via glucose transporters overexpressed on the surface of cancer cells. 401 Tumors could be monitored for a week, as shown in Figure 26. 401 Glucose‐conjugated AuNPs bearing a ligand with 68 Ga as a tracer were biocompatible and excellent for in vivo imaging using positron emission spectroscopy. 402 Moreover, neuropeptides loaded on glucose‐ 68 Ga‐AuNPs enabled penetration through the blood–brain barrier. 402 Magnetic resonance is a broadly used imaging method in which the application of contrast agents significantly improves image quality, especially when these are in the form of NPs. In most experiments, iron oxide NPs decorated with stabilizing, targeting, or biocompatible materials have been tested. 403 As an example, versatile glycopeptide grafting to magnetic iron oxide NPs was reported but without in vivo testing. 404 A step further using sialic acid modified Fe 3 O 4 NPs for the selective in vivo MR imaging of β‐amyloids applicable for the early diagnosis of Alzheimer's disease was reported. 405 Mannose‐containing diblock copolymer grafted onto Fe 3 O 4 NPs enhanced the carbohydrate receptor‐driven uptake into lung cancer cells with potential application in the diagnosis and localization of this type of cancer. 406 HA‐grafted Fe 3 O 4 NPs employed in the selective MR imaging of tissues with CD44 receptor‐overexpressing cells, that is, tumors, or local inflammations were also described. 407 , 408 A similar system with core Fe 3 O 4 NPs having a thin golden shell covered with an HA layer enabling simultaneous MR imaging and photothermal therapy targeted toward cells with CD44 receptors has been reported. 370 In order to increase the in vivo stability of the Fe 3 O 4 core, its coating with a silica shell was described in several studies. 384 , 409 An enhanced imaging resolution of brain lesions in mice after simulated stroke can be observed in Figure 27 , showing that NPs with different glycosylation exhibited different biodistribution. 409 Sialic Le x present on NPs allowed their accumulation in ischemic, that is, damaged, brain tissue due to selectin overexpression; NPs decorated with Le x (without sialic acid present) could visualize such tissue less effectively, which was also confirmed by the software evaluation of dark and bright voxels in the images. 409 Nevertheless, good results were also obtained with NPs bearing hydroxyls groups on the NP surface instead of glycans. 409 In addition to the protective function of a silica shell, such a layer could provide a better environment for the covalent grafting of targeting ligands. The application of silica‐based nanocarriers for MR imaging has recently been reviewed by Caltagirone et al. 410 Figure 27 T2 maps and T2‐weighted images at 5 and 24 hr after MCAO in a representative animal tissue that received iron oxid‐silia nanoparticles decorated with (A) hydroxyl groups (HO@MNPs), (B) Lewis X without sialic acid (LX@MNPs), or (C) sialyl Lewis X (SX@MNPs). Note: the color scale bar corresponds to the T2 value (msec). Reprinted with permission from 409 . Copyright 2014 American Chemical Society. In addition to Fe 3 O 4 NPs, AuNPs were also applied in cell imaging. Small AuNPs of 1.8 nm in size grafted with lactobionic acid via PEG spacers were found to penetrate human hepatocellular carcinoma cells selectively via their overexpressed asialoglycoprotein receptors. 400 While T2‐weighted MRI was used for contrast enhancement applying iron oxide NPs, T1‐weighing was used for the same purpose using AuNPs loaded with a Gd 3+ complex. 411 The modification of AuNPs by a combination of different carbohydrates helped to achieve a high imaging selectivity of tumor cells in mice models. 412 AgNPs prepared by the reduction of Ag ions using saccharides present in sugar cane juice have ferromagnetic properties and could be applied for MRI. 411 3. Quantum Dots as Imaging Probes QDs as NPs composed of single nanocrystals of metal sulfides or other semiconductors have distinct optical properties resulting from their nano‐size. One of the most important features is their stable and adjustable fluorescence together with the overall stability inherent to all inorganic materials. Recent studies have reported that, after the encapsulation of QDs into a polymer shell with a final modification by carbohydrate moieties (for specific cell targeting), the resulted NPs still retain their fluorescence. The surface of QD‐modified NPs could be modified by a diverse range of functional groups with adjusted properties of NPs as reported by Schmidtke et al. 413 This study suggested applicability of such NPs in in vivo imaging, even though this has not been tested. 413 On the other side, the uptake and intracellular distribution of core‐shell CdSe‐ZnS QDs covalently grafted with different carbohydrates were assessed. 414 In addition to lactose, which enhanced the uptake of the QD‐glycan NPs into cancer cells, the composition of a mixed carbohydrate shell (i.e., lactose and mannose vs. lactose and maltotriose) significantly altered the internalization of NPs. 414 In another study, a secondary radiotracer ( 125 I) was used as the probe conjugated with CdSe/CdS QDs decorated with a sialic acid‐binding lectin. 415 These NPs were able to image breast cancer cells, as revealed by in vitro experiments. 415 Enhanced selectivity in the targeted imaging of muscle cells was observed for glucose‐decorated QDs with insulin‐induced uptake mediated via glucose receptors, as illustrated in Figure 28 , and such an approach could be applied for selective drug delivery as well. 416 Since the vast majority of inorganic QD particles exhibit cytotoxicity, Shinchi et al. focused on the development of cadmium‐free ZnS‐AgInS 2 NPs with lower cytotoxicity but still exhibiting good fluorescence properties and applicability in selective cell imaging after decoration with short targeting carbohydrates. 417 From the graphs shown in Figure 29 , it is obvious that all carbohydrate‐coated ZnS‐AgInS 2 QDs exhibited minimum cytotoxicity against HepG2 cells (panel A), unlike Cd‐based QDs (panel B). 417 Figure 28 (A) A molecular structure of glucose‐functionalized quantum dots (Glc‐QDs). (B) A schematic illustration of the cellular uptake of Glc‐QDs regulated by insulin and 2‐deoxyglucose (2‐DG) in C2C12 muscle cells. Reproduced from 416 with permission of the Royal Society of Chemistry. Figure 29 MTT assay for SFNPs. HepG2 cells were incubated with ZAIS/ZnS NPs (A) or CdTe/CdS QDs (B). The NP concentration was in the range of 5 to 50 μg mL −1 (left to right). Reprinted with permission from 417 . Copyright 2014 American Chemical Society. QD‐like fluorescence was assigned also to GO‐based NPs, 378 which were used, for example, for combined imaging and chemotherapy with an HA grafted as a targeting agent. 288 It is worth mentioning that the replacement of semiconductor‐based mostly cytotoxic QDs with less harmful graphene‐based NPs would be another step to develop novel, more efficient theranostic particles with low cytotoxicity. Silica QDs possessing inner fluorescence exhibited low cytotoxicity, having a surface that could be easily modified with glycans. 418 , 419 Furthermore, fluoride nanocrystals doped with rare earth element ions were also suggested as possible substitutes for "classical" Cd‐containing QDs—see, for example, 420 and references therein. 4. Toxicity of Nanomaterials It is beyond any doubt that the massive application and employment of nanomaterials have brought and should bring in the near future amazing achievements and technical possibilities—in fact, many authors are talking about the nanoage or nano‐revolution. Nevertheless, the other side of this "nano‐enthusiasm" is the persistent uncertainty about the possible toxicity of NPs. This issue is particularly delicate in medical applications, where NPs are supposed to be delivered directly into the human body and interact with cells at a molecular level. It should be noted that these concerns have been raised simultaneously with nanomaterial applications—see, for example, a 1982 study on the toxicity of a potential drug carrier. 421 In this study, one example of possible cytotoxicity mechanism of NPs was outlined, that is, the toxicity of monomers released from the intracellular degradation of NPs. Other known mechanisms, including unwanted aggregation of proteins, delivery of toxic molecules conjugated with NPs (including NPs from transportation and industrial emissions) and generation of highly cytotoxic ROS (the most toxic effect), were comprehensively reviewed. 422 , 423 , 424 , 425 The NPs' behavior in living organisms is prevalently determined by their physical and chemical properties. For example, decreased NP size increases their active surface and, consequently, the rate constant of any catalytic reaction taking place on the NP surface. Moreover, the size also determines the biodistribution mode in an organism—see, for example, different modes of interactions of small and large graphene sheets with cells (while the former tend to disrupt cell wall, 426 the latter more likely wrap it without any damage 427 ). Another crucial aspect is the nature of NP surface modification, including (i) surface charge, the adjustment of which can turn otherwise harmless NPs, for example, into cell disruptors (see references in 423 ) or other biological functions, 428 and (ii) the modification of NPs by biologically active moieties with examples, which can be found in sections of this review dealing with drug delivery, vaccine and other therapeutic effects, and cell imaging. Some toxic impacts can be estimated from the physical or chemical properties of the NPs acquired from a precise and accurate characterization of the NPs. This is usually accomplished since physical and chemical characterization and at least in vitro cytotoxicity tests are included in practically any study concerning the administration of NPs into the human body (drug carriers, imaging NPs, etc.). Nevertheless, the correlation between in vitro and model animal in vivo studies may significantly differ from the real impact on humans, nota bene when the subacute effect is considered. 429 Therefore, unless any negative health impacts are ruled out, appropriate precautions should be taken, and precise toxicity assessment must be performed in parallel with nanomedicine progress. 7. GLYCAN ENRICHMENT AND SEPARATION NPs by definition are any objects with at least one dimension below 100 nm exhibiting unique physicochemical properties, which can be effectively applied to enhance the performance of glycan analysis by MS or to increase the efficiency of glycan enrichment. This is especially true due to the high surface‐to‐volume ratio, which can help to address problems with mass transfer; high surface area also means that the high density of active functional groups could be attached to NPs to increase separation efficiency or aid in MS analysis. 430 NPs can be synthesized with controlled size and morphology, and it is also possible to prepare hybrid NPs with unique characteristics (e.g., MNPs covered by a gold film suitable for the formation of a SAM on their surface, etc.). 63 SiNPs are chemically stable over a wide pH, a feature that is important for column packing material. Nanoporous silica materials, with a pore size spanning range from 2 to 150 nm with various modifications, are quite often applied as packing material for the separation of glycans. Other NPs, including metal oxides, CNTs, fullerene, and graphene, or GO can be used as packing/enrichment material either in their pristine or modified state. Recently, imprinted polymers have become popular because they are regarded as affinity matrices with tailored properties for the separation of a particular analyte with lectin‐like properties. 430 A. Glycan Release Glycans attached to the protein backbone can be analyzed either in their intact form (i.e., glycoproteins), after protein digestion by endopeptidases or endoproteinases (i.e., glycopeptides) or as pure glycans released from the protein or peptide backbone using the enzyme PNGase F or PNGase A (Fig. 30 ). The release of glycans from O ‐glycans can be done not using enzymes but rather chemical reaction, that is, β‐elimination or the use of hydrazine is needed. When glycopeptides are produced from glycoproteins by a tryptic digest, glycopeptides are only a minor component of the tryptic digest, and their enrichment is important for high‐performance MS analysis. Peptides isolated from glycopeptides can be further applied for protein fingerprinting. 431 , 432 Figure 30 A graphical summary of different approaches, which can be applied for glycoprotein analysis including glycomics (analysis of released glycans), proteomics (identification of peptides after removal of glycans) and glycoproteomics (characterization of intact glycoproteins in order to indentify a site‐specific connectivity between glycans and proteins). B. Mass Spectrometry MALDI‐time‐of‐flight‐MS (MALDI‐TOF‐MS) is a gentle technique allowing the analysis of large biomolecules in both positive and negative ionization modes. This is why this technique is the most commonly applied in glycan, glycopeptide, or glycoprotein analysis, even though the MALDI‐TOF process itself is not yet fully understood. 433 On the other hand MALDI‐TOF and electrospray ionization (ESI) might not distinguish glycan isomers when operated in a conventional mode. Various on‐line separation techniques (i.e., using porous graphitized carbon packed columns), which can be coupled to ESI MS, can effectively resolve this problem. 434 Alternatively, tandem MS, that is, MS/MS or MS n , can resolve this issue as well. 435 MALDI‐TOF compared to ESI exhibits higher sensitivity for the determination of glycans and efficiently ionizes molecules with a higher molecular weight. At the same time, MALDI‐TOF is more tolerant to contaminants. 435 MALDI‐TOF when compared to ESI introduces higher energy to the sample, resulting in the fragmentation of labile functional groups, including sulfates, phosphates, and sialic acids. 435 Since sialic acids are quite labile molecules with significant involvement in many biological processes, an analysis of sialic acids requires the development of special analytical strategies. 436 , 437 , 438 , 439 Moreover, a proper matrix must be chosen to effectively generate ions of large molecules for MALDI‐TOF analysis, and as for ESI, the sensitivity of detection decreases with the increasing molecular weight of glycans. 435 C. Glycan Enrichment There are few main reasons why glycans have to be enriched prior to MS analysis: (i) the glycan level in the studies sample can be well below detection limit of an MS instrument; (ii) glycopeptides or glycans must be removed from peptides due to their lower abundance compared to peptides; and (iii) the ionization of glycans or glycopeptides in the presence of peptides is frequently less efficient. 63 Glycans, glycopeptides, or glycoproteins can be enriched in numerous ways, including boronate affinity, hydrazide chemistry, lectin affinity, hydrophilic interactions, and other approaches (Fig. 31 ). 440 Figure 31 Various ways glycans can be enriched using hydrophobic, hydrophilic, or ionic interactions (upper left part) or biorecognition interactions involving glycan binding proteins (lectins, lower left part). Alternatively for glycan capture/release particles modified by boronate functional groups or hydrazine can be applied, as well. 1. Boronate‐Based Enrichment Boronate‐modified materials are applied for glycan enrichment due to the strong interactions between boronate functional groups and cis ‐diols present in glycans under basic conditions, that is, at pH 8–9. 441 The main prerequisite for glycan binding to boronate‐modified material is the choice of a proper pH, which should be above the p K a of the boronate moiety of a boronate‐terminated molecule (e.g., p K a = 8.9 for frequently applied PBA). Sialic acids are exceptions to this rule since the effective binding of SA to boronate‐modified material can occur at a pH lower than the p K a of boronate functional group. Recent studies however suggest that besides proper choice of pH, other factors come into play during the interaction of glycans with boronate‐modified material, such as steric factors, and the nature of buffer components. 442 Decreased pH can be effectively applied to release captured glycans for further analysis. Relatively high pH needed for effective glycan enrichment based on boronate affinity can be a problem for quite complex biomolecules. Such glycopeptides or glycoproteins can become unstable at a high pH. To counter this, various strategies have been applied to lower the pH needed for such binding, including the proper choice of substituents (electron withdrawing groups) in proximity to boronate functional groups. 443 Alternatively, benzoboroxozole derivatives can be applied to lower the pH required for glycan binding via boronate affinity. 444 Such approaches are also useful due to the widening of the applicable pH window for glycan enrichment. 2. Hydrazide‐Based Enrichment Hydrazide‐modified materials can be efficiently applied for glycan enrichment, but glycans must be oxidized first to form aldehydes from cis‐diol groups. Such aldehydes react with hydrazide‐forming hydrazone bonds, which are stable under basic conditions, and glycan can be released from enrichment material by the application of acidic pH, similar to glycan release from boronate affinity materials. Alternatively, PNGase F could be applied for the release of glycopeptides. 445 , 446 3. Oxime‐Based Enrichment Oxime‐modified surfaces can be a good alternative to the hydrazide‐based enrichment of glycans since the process of enrichment can be considerably shorter (1 hr) compared to hydrazide‐based enrichment (12–16 hr), with high enrichment sensitivity with LOD in the low femtomole range and with the selectivity of glycopeptide enrichment even in the presence of a 100‐fold excess of nonglycopeptides. Glycopeptides must be oxidized in order to interact with oxime functionalities, and overnight incubation with PNGase F was required for glycan release for subsequent MALDI‐TOF‐MS analysis. 447 4. Hydrophilic Interactions A hydrophilic interaction‐based method for glycan enrichment involves the application of a polar stationary phase and a nonpolar mobile phase for separation. Analytes partition between a water‐rich layer on the surface of the stationary phase and the mobile phase. Glycans as hydrophilic molecules will be attached to the column/material, while other more hydrophobic molecules (including peptides) will be dissolved in the mobile phase and washed away. Hydrophilic interaction based on zwitterionic functional groups has become popular in recent years due to the ability of zwitterions to form a stable water layer on the surface of the separation phase, which is effective for glycan partitioning. 430 5. Lectin‐Based Enrichment A lectin‐based approach for glycan enrichment is based on the natural bioaffinity interaction between glycans and glycan‐binding proteins (lectins) triggered by the interaction of glycans, especially with aromatic stacking amino acids and amino acids providing hydrogen bonds. 448 The application of lectins for glycan enrichment is advantageous compared to boronate, hydrophilic interaction or hydrazide‐modified materials due to the higher degree of specificity for some glycan structures, but selectivity and binding affinity are weaker compared to antibody‐based biorecognition. Moreover, lectins can distinguish minor changes in glycan structures. A good example is SNA, recognizing sialic acid bound to galactose via an α‐2,6 bond, while Maackia amurensis agglutinin binds to α‐2,3‐linked sialic acid. Another example is the arm‐specific recognition of bisecting glycans (i.e., glycans with two arms). 449 Thus, not all glycans present in the sample will be enriched using lectins, and a lectin‐affinity enrichment strategy can be applied to fractionate glycans into several groups. Finally, glycans can be released from lectins by the application of free carbohydrates or by exposure to an acidic eluting buffer. 6. Other Ways for Enrichment There are other options for glycan enrichment, such as reverse‐phase mode, electrostatic mode (for glycans containing negatively charged sialic acids), electrostatic repulsion hydrophilic interactions, and affinity capture based on protein A, which is applicable for the selective enrichment of antibodies with the subsequent glycoprofiling of glycans present especially in the Fc fragment of IgGs. 21 , 430 D. Nanoporous Materials Porous materials are, according to IUPAC definition, considered nanoporous only if the pore size is below 1 μm, but earlier terminology defining materials as macroporous (above 50 nm), mesoporous (2–50 nm), and microporous (below 2 nm) is frequently used in the literature. 450 In the forthcoming paragraphs we will stick to the latter division of nanoporous materials into three categories. The other important parameters of porous materials are distribution of pore size, pore shape (cylindrical, slit shaped, funnel shaped, etc.), and degree of pore interconnections. The porosity of the given material can be characterized by the total pore volume and a specific surface area. Such porous materials are quite frequently applied as packing material for the separation columns for the enrichment/separation of glycoproteins, glycopeptides, or glycans and have been reviewed in the past. 430 , 446 , 451 , 452 The size and modification of pores play a prominent role in interactions with ligands due to the confinement effect. 453 For example, when glycoprotein (ribonuclease B) interacted with a mesoporous material, the enhancement factor (i.e., K D of the interaction with mesoporous silica compared to non‐porous silica) increased from 45 to 900, when, in addition boronate affinity, electrostatic interactions were also involved. Upon the decreased pore size of mesoporous silica from 2.6 to 2.1 nm, a further increase in the enhancement factor to 2100 was observed. 453 This indicates that for effective ligand capture, the careful design of surface chemistry and the morphology of porous materials are important. 454 An interesting approach for how to increase the surface area of composites (i.e., from 69 to 93 m 2 g −1 ) while decreasing the pore size (from 7.6 to 5.0 nm) by mixing GO sheets into composites was demonstrated by He et al. 455 1. Silica‐Based Nanoporous Materials Porous silica material with an ordered/controlled size and shape of large pores (1 μm) was prepared by Yan et al. 456 using polystyrene colloidal crystals as a hard template, with mesopores of 4.6 nm using a surfactant as a soft template and applying tetraethylorthosilicate (TEOS) as a silica source. Moreover, macropores were interconnected via pores of 50 nm in diameter (Fig. 32 ). The specific surface area of the material was 255 m 2 g −1 with a pore volume of 0.46 cm 3 g −1 . Such hierarchically ordered material was then modified with 4‐vinylphenylboronic acid to form polyboronic acid‐modified material, which was successfully applied for the effective enrichment of glycopeptides prepared by the tryptic digest of a model glycoprotein ‐ HRP. The method using glycan enrichment (10 min) revealed the presence of 12 glycans by MALDI‐TOF MS, while only four glycans were obtained without any glycan enrichment. The glycan of HRP could be detected with an LOD of 1 ng μL −1 (1 fmol) even in the presence of BSA, which was in 50:1 excess compared to HRP. 456 Figure 32 SEM image of hierarchically ordered macro/mesoporous silica material before (left) and after (right) boronate modification. Reproduced from 456 with permission by John Wiley & Sons. Mechanically stable silica microparticles of 1.6 μm in diameter with macropores ranging in size from 50 to 150 nm were applied to immobilize two lectins, Con A and Aleuria aurantia lectin, by Mann et al. 457 Such particles with interconnected pores and with a high surface area of 200 m 2 g −1 accommodated a high density of immobilized lectins (16 mg g −1 ). When lectin‐modified particles were applied as a support for loading a chromatography column, up to 75 μg of glycoproteins could be bound to the column (50 × 1 mm) with immobilized Con A. Finally, the columns were used to enrich glycoproteins from just 1 μL of blood serum after removing IgG and albumin. Subsequently, the glycomic and glycoproteomic profiling of bound glycoproteins was performed using MALDI TOF/TOF MS. 457 A capillary with boronate‐modified mesoporous walls was used for the selective enrichment of neutral or acidic glycans via pH manipulation by Lu et al. 458 When the binding solution had a pH lower than the p K a of boronate by one pH unit or more, the boronate‐modified capillary column captured sialylated glycoproteins. In contrast, when the pH of the binding buffer was higher than the p K a value of boronate by one pH unit or more, glycoproteins containing SA were electrostatically repelled from interaction with the monolith, and glycoproteins with neutral glycans were preferentially captured. Furthermore, the ionic strength of the binding buffer played an important role in the capture of sialylated glycoproteins. 458 A boronate‐modified affinity column was also helpful in the identification of DNA aptamers binding to glycoproteins using the systematic evolution of ligands by exponential enrichment as proposed by Nie et al. 459 The boronate column served as a matrix for binding glycoprotein (HRP), with the complete process performed on the column; the authors identified seven DNA aptamers binding HRP with an affinity constant down to 10 nM. 459 DNA aptamers selected specifically against glycoproteins can be applied for the selective capture of some glycoproteins in the future since their K D (10 −8 M) can be lower than that of lectins. 459 2. Carbon‐Based Nanoporous Materials TEOS with a surfactant as a structure‐directing agent was applied by Qin et al. 460 to prepare SiNPs with a uniform size of 70 nm. The particles with a surface area of 901 m 2 g −1 with a pore size of 2.8 nm were treated with sulfuric acid and subsequently carbonized with the appearance of sp 2 ‐hybridized aromatic carbon structures, which were effective in glycan enrichment. Thirteen N ‐glycans released from just 5 ng of ovalbumin (i.e., LOD of 100 fmol) by PNGase F could be detected. Finally, the particles were applied for the enrichment of N ‐glycans from 50 nL of human serum and analyzed by MALDI‐TOF‐MS. Eight new N ‐glycans that were not observed in the MS spectrum when standard carbon adsorbent was applied were observed, including low‐abundant fucose and high‐mannose glycans. 460 Mesoporous silica composites were prepared by Sun et al. 461 using a graphene layer as a support. In this case, graphene was oxidized by nitric acid to deposit oxygen‐rich functionalities, which were used in the subsequent step for the formation of a mesoporous silica layer by the deposition of TEOS in the presence of a surfactant as a structure‐directing agent. Finally, the whole nanocomposite was carbonized to transform the surfactant into a carbon layer. The resulting material exhibited a uniform pore size of 2.8 nm with protein‐excluding properties with a surface area of 372 m 2 g −1 . MALDI‐TOF‐MS showed the presence of 25 N ‐linked glycans released from ovalbumin when glycan enrichment with graphene‐based nanocomposite was applied, while enrichment with active carbon material was less effective in terms of the number of glycans detected and the signal intensity obtained. MALDI‐TOF‐MS analysis without any enrichment did not show any sign of glycans. 461 Commercially available mesoporous graphitic carbon with a pore size of 25 or 30 nm was applied by Zhao et al. as a packing material for fully automatable two‐dimensional liquid chromatography for high‐throughput proteomic and glycomic analysis of various lysates/samples. 462 With this fully automated approach, it was possible to identify up to 2678 proteins, 11,984 unique peptides in neurons or up to 130 N ‐glycoproteins, 705 N ‐glycans, and 254 glycosylation sites in macaques plasma ( Macaca fascicularis ) with a total analysis time of 19 hr using ESI‐MS/MS. According to authors, the technology offered unattended, robust, and scalable analyses of samples on a submicrogram scale when sample injection was performed once a day, and the system was able to operate continuously for up to 14 days without experiencing any major problems. 462 A porous silica column was applied for the covalent immobilization of PNGase F by Jmeian et al., and such a column was used for the continuous (online) release of neutral and acidic glycans from glycoproteins with subsequent continuous glycan enrichment in a column loaded with porous graphitic carbon and with a subsequent analysis of glycans. 463 The capillary was coupled to a C 8 trap and a porous graphitic carbon HPLC‐chip and finally interfaced to perform liquid chromatography‐MS and liquid chromatography‐MS/MS analyses. The main advantage of the system is the short time (6 min) needed to release glycans compared to the usual overnight incubation at 37°C. Moreover, glycans could be effectively analyzed from just 100 fmol of protein or 0.1 μL of human blood serum. 463 The time needed for online glycan removal is shorter compared to that needed for the microwave‐assisted release of glycans in the presence of PNGase F (30 min at 37°C). 464 GO was due to the large surface area (theoretical relative surface area of 2630 m 2 g −1 ) effectively applied by Ren et al. as a support for the covalent immobilization of a high amount of PNGase F (688 μg g −1 of GO, which is threefold higher compared to SiNPs). Highly efficient N ‐glycan release from a protein backbone of two model glycoproteins (ribonuclease B and asialofetuin) can be completed within 2 min, and the enzyme can be reused. The stability of the enzyme complex is remarkable, that is, up to 8 weeks, when stored at 4°C, and as low as 2 μg μL −1 of the enzyme complex was needed to analyze the plasma extracts. 465 3. Other Nanoporous Materials Hydrophilic mesoporous phenol‐based material with added PEI, prepared by Jin et al. with a template‐free method, exhibited a large surface area (548 m 2 g −1 ) with 13 nm 466 mesopores. When the amount of PEI changed, it was possible to change both the porosities and surface area of the material. Glycopeptides could be detected with an enrichment factor (MS intensity ratio after and before enrichment) of 62 with a recovery index of 80.4%. A tryptic digest of IgG could be detected down to 5 fmol with only a moderate improvement compared to commercial hydrophilic interaction chromatography beads (40 fmol). 466 A capillary modified with a polymeric material with an introduced –SH group having a surface area of 30.3 m 2 g −1 was applied for the further selective modification of the surface with AuNPs (13 nm) by Wu et al. 467 The gold surface was further modified by two thiols terminated in boronate and amine functional groups making an intramolecular B‐N coordination for the enhanced specificity of glycoprotein capture with a binding capacity of 0.39 mg g −1 . The capillary was applied for the selective enrichment of a glycoprotein over a protein (1:1000 ratio) with a recovery of 85% and finally applied for the glycoproteomic profiling of just 9 μg of human plasma with 160 glycoproteins identified by MS. 467 A boronate‐modified affinity monolithic column with a surface area of 13.7 m 2 g −1 and with bimodal pores containing mesopores (3.9 nm) and macropores (1.4 μm) was prepared by Liu et al. using ring‐opening polymerization to mimic the function of protein A to bind antibodies. 468 In this case, mesopores could be accommodated by the glycan of IgG via the interaction of glycans with boronate, while the protein backbone of IgG could not enter the mesopores (Fig. 33 ). The binding capacity of such material for IgG binding was comparable (23.7 mg g −1 ) to a number of protein A mimics. The monolith was a quite stable and cost‐effective alternative to using protein A with the price of 40 USD per 1 g. Moreover, the IgG attached to the monolith still exhibited its binding affinity toward antigens. 468 Figure 33 A scheme showing a specific recognition of IgG by the monolith exhibiting protein A‐like binding with presence of bimodal pores (macropores and mesopores) modified by boronate functional groups. Reproduced from 468 with permission of the Royal Society of Chemistry. E. Imprinted Polymers MIP can be effectively applied for the selective capture and release of analytes because the analyte molecule is imprinted, and after the formation of a thin layer of polymer around the template (analyte), the analyte is removed from the MIP. Thus, such MIP is then able to selectively recognize the analyte over other molecules. Such an approach was applied to imprint HRP as a model glycoprotein on the inner wall of the capillary tube modified by boronate functionality as described by Lin et al. (Fig. 34 ). 469 Then, a polydopamine film was formed around HRP with the final removal of HRP. Such an MIP was able to recognize HRP with a bound amount of 4.6 mg g −1 over other glycoproteins, while the boronate‐modified capillary also effectively captured other glycoproteins. 469 Figure 34 (A) Preparation of vinylphenylboronic acid‐based molecularly imprinted monolith with polydopamine coating and (B) its recognition mechanism toward glycoproteins. In the figure the following four major steps are shown: (1) preparation of the VPBA‐based polymeric skeletons; (2) reversible immobilization of glycoprotein via boronate affinity interaction; (3) self‐polymerization of DA on the surface of the glycoprotein‐immobilized boronate affinity monolithic skeletons; (4) the formation of glycoprotein‐imprinted monolith after removal of template that is complementary in shape, size, and functionality with respect to the template. Reprinted from 469 . Copyright (2013), with permission from Elsevier. An interesting imprinting approach for the selective binding of a glycoprotein was recently applied by Mendes et al. (Fig. 35 ). 470 Gold surface was modified by an acrylamide‐alkyne cysteine derivative to make a thin film terminated in alkyne and acrylamide functionalities. Then, a glycoprotein was incubated with (3‐acryl‐amidophenyl)boronic acid when boronate functional groups interacted with a glycan part of the glycoprotein (PSA), and the other functional groups were applied for the attachment of PSA on the surface via acrylamide copolymerization. Finally, the surface was blocked by the addition of OEG‐N 3 molecules via click chemistry. Because the boronate–glycan interaction can be disrupted by low pH, a simple pH lowering was applied to remove the PSA template from the surface. Such an approach exhibited a 30‐fold higher affinity for its target compared to that of other (glyco)proteins. 470 Even though this imprinting strategy was applied for SPR sensing, there is a possibility of using this approach for selective PSA enrichment from a complex sample. Figure 35 Experimental design for the formation of surface‐restricted click‐imprinted binding sites for glycoproteins. Disulfide dimer (DFC) SAMs were prepared by immersing clean gold substrates in 0.1 mM methanolic solutions of DFC for 24 h (step 1). In step 2, BA receptor units are introduced via (3‐acrylamidophenyl)boronic acid (AM‐BA) that is incubated for 30 min at an optimized pH (8.5) with a template target glycoprotein. Multiple boronate esters are formed reversibly between the AM‐BAs and the carbohydrate structures of the glycoprotein template. The pre‐assembled glycoprotein–AM‐BA complex is then grafted on the DFC SAM via acrylamide co‐polymerization, affording the creation of spatially arranged sets of BAs on the surface that are specific for the target glycoprotein (step 3). In order to provide complimentary allosteric specificity, a mould or imprint is created around the glycoprotein template at the surface by so‐called click chemistry functionalization of the alkynes of the DFC on the SAM by reacting azide‐terminated heptaethylene glycol (Az‐OEG) moieties with the terminal alkynes on the DFC SAM via a copper‐catalysed alkyne–azide cycloaddition (CuCAAC) reaction (step 4). The glycoprotein targets are removed by washing under acidic conditions (step 5). Reproduced from 470 with permission of the Royal Society of Chemistry. In their study, Liu et al. focused on finding the relationship between the thickness of the imprinted polymer and the size of the imprinted glycoprotein regarding the binding efficiency. 471 Three glycoproteins (ribonuclease B, HRP, and GOx) were imprinted within a polydopamine film deposited over magnetic particles modified by boronate functional groups. The thickness of the polydopamine‐deposited film (5.7, 10.2, 16.3, and 25.3 nm) was controlled by polymerization time (1, 3, 6, and 15 hr). The higher the size of the glycoprotein (15–80 kDa), the thicker the film (5.7–16.3 nm) needed to be to achieve maximal binding capacity (16.7–19.5 mg g −1 ). 471 Polymers with imprinted glycans isolated from two model glycoproteins—ribonuclease B and transferrin—were prepared by Bie et al. on 100 nm boronate‐modified MNPs. 472 Glycans released from a glycoprotein were attached to boronate‐modified MNPs, and glycans were imprinted by forming a thin (2.5 nm) silane layer. Finally, the glycan template was removed, and such NPs were applied for the enrichment of glycoproteins. The imprinting ratio (i.e., the ratio of glycoprotein attached to imprinted vs. non‐imprinted NPs) was 8.4 for ribonuclease B and 21.8 for transferrin, with a maximal sorption capacity of 1.4 mg g −1 and a K D of 25 μM, suggesting only moderate glycan enrichment. The transferrin glycan‐imprinted polymer was applied for the enrichment of transferrin from a human serum sample. 472 F. Magnetic NPs (MNPs) Glycan enrichment can be effectively performed using MNPs with different glycan‐recognizing functional groups present on the surface of such NPs. Core‐shell NPs with an MNP (Fe 3 O 4 ‐based) core and silica shell having uniform 2.2 nm mesopores (formed by the removal of a surfactant from the shell layer) were prepared by Zheng et al. 473 Such NPs were further activated, and glucose was covalently linked to it by click chemistry (Fig. 36 ) to provide hydrophilic interactions. Such particles exhibited a large surface area (324 m 2 g −1 ), high ability to enrich glycopeptides (250 mg g −1 ), high sensitivity (50 fmol), short incubation time (5 min), and high recovery (94.6%) using ESI MS analysis. When 0.25 μL of human serum without enrichment was analyzed, only 12 N ‐glycans were found, while 42 N ‐glycans were identified after the application of the MNP‐based enrichment step in the ESI MS spectra. 473 Figure 36 (A) Schematic representation of the synthesis of hydrophilic MMNs and (B) the selective enrichment process for glycopeptides using hydrophilic MMNs. Reprinted from 473 . Copyright 2014, with permission from Elsevier. MNPs were modified by Cao et al. using a polymer to which hydrazide was attached, forming a 3D matrix for glycan enrichment with a threefold higher loading of hydrazide compared to single‐layer NPs. 474 The enrichment of glycopeptides was effective even in the presence of a 100‐fold excess of peptides with a recovery index of 78% and with a binding capacity for glycopeptides of 25 μg mg −1 . Glycopeptides enriched from mouse liver tissue were analyzed by MALDI‐TOF‐MS. 474 In order to enhance the hydrophilicity of the MNPs, Chen et al. covered such particles by a thin layer of SiO 2 to which zwitterionic molecules were attached by precipitative polymerization. 475 Such particles possessing a highly hydrophilic surface were effective in glycan capture, when by MALDI‐TOF‐MS, glycans from just 0.1 fmol of IgG could be detected with a high binding capacity of 100 mg g −1 and high enrichment recovery of 74% with a rapid magnetic separation. Finally, the particles were applied for the glycoprofiling of 65 μg of proteins extracted from mouse liver. 475 MNPs with co‐precipitated ethylenediaminetetraacetic acid having a size of 15 nm were modified by Dong et al. using Cu(II) with the subsequent high loading of Con A lectin (up to 28 wt%). 476 Such particles were applied for the enrichment of a model glycoprotein, namely ovalbumin, with a binding capacity of 72 mg g −1 , while a high affinity with K D of 38 nM was observed. Glycoproteins were successfully enriched, even in the presence of a non‐glycoprotein in molar excess of 600:1 with a fast magnetic separation within 15 sec. 476 MNPs with a size of 15–20 nm during synthesis formed larger aggregates with a size of 140 nm and finally were modified by chitosan using the one‐pot method as described by Fang et al. 477 Such particles were able to detect glycans from as low as 8 fmol of a tryptic digest of IgG using MALDI‐TOF‐MS with a binding capacity of 17.5 mg g −1 . Finally, 45 μL of tryptic digest from HeLa cells was successfully glycoprofiled. 477 Magnetic mesoporous (pores with 3.8 nm) particles with a surface area of 211 m 2 g −1 and with a total pore volume of 0.38 cm 3 g −1 were applied by Deng et al. for the enrichment of as low as 10 fmol of a model glycoprotein in 300‐fold excess of protein (BSA), and finally, glycans enriched from human serum were determined by MALDI‐TOF‐MS. 478 Liu et al. applied MNPs ( d = 100 nm) patterned by a PAMAM dendrimer further modified by boronate functional groups for model glycoprotein (HRP) enrichment. 479 Boronate present on a dendrimer modified MNPs had a 3–4 orders of magnitude higher affinity constant for glycoproteins compared to single boronate, while glycoprotein enrichment was possible even in a 1 million‐fold excess of competing monosaccharide with an efficiency of 42–88%. HRP could be adsorbed with density of 21 mg g −1 of support with LOD of 180 amol with extraction equilibria reached within 1 min (with desorption equilibria of 5 min). The reusability of glycoprotein enrichment was highly reproducible, with an RSD of 7.5% for five consecutive enrichment steps. Finally, the approach was also applied for the analysis of glycoproteins in human saliva. 479 Deng et al. applied MNPs (100 nm) modified by poly(styrene‐co‐vinylbenzene‐boronic acid) for glycopeptide enrichment. 480 Such particles could detect glycopeptides down to 125 fmol from a tryptic digest of HRP, even in 120‐fold excess of peptides. Glycopeptides enriched from human serum were analyzed using MALDI‐TOF‐MS. 480 Lu et al. developed an interesting strategy to increase the selectivity of glycan enrichment using two types of nanomaterials. 481 The first nanomaterials were MNPs (70 nm) modified by boronate functionalities for glycopeptide capture, and the second were poly(methyl methacrylate) beads (200 nm) for selective peptide capture. Model glycoprotein (HRP) was detected down to 220 fmol, and enrichment could be performed in 100‐fold excess of a nonglycoprotein with a maximal binding capacity for glycopeptides of 150 mg g −1 . MALDI‐TOF‐MS revealed 90% recovery using this synergetic enrichment, and as little as 1 μL of human serum was sufficient for analysis. 481 MNPs could also be successfully used for the prefractionation of glycoconjugates (mainly glycopeptides and glycoproteins). Although the traditional method is hydrazide chemistry‐based solid‐phase extraction, solid‐phase extraction through reductive amination by amine‐functionalized MNPs had also been developed, 482 shortening the extraction time to 4 hr and improving the LOD by 2 orders of magnitude. Magnetic Fe 3 O 4 NPs were functionalized by 3‐aminopropyltriethoxysilane and subsequently incubated with glycoconjugate sample, which was converted into aldehydes by sodium periodate oxidation prior to incubation. While nonspecifically adsorbed proteins could be easily washed away, glycopeptides/glycoproteins remained immobilized on the surface. Using a specific PNGase F enzyme, an excellent isolation performance and identification of glycosylation sites using nano‐LC‐MS/MS analysis were achieved. 482 The highest binding capacity toward glycoproteins was achieved by boronate‐modified magnetic particles (265 nm) using ovalbumin as a model glycoprotein of 778 483 or 882 mg g −1 using magnetic particles with size of 500 nm. 484 An elegant method to increase the low affinity of glycan binding by three different lectins was addressed by the modification of lectins with boronate linkers—called boronic acid decorated lectins (Fig. 37 )—was suggested by Lu et al. 485 When such hybrid biomaterial was immobilized on MNPs, a 2‐ to 60‐fold increase in the detection sensitivity for glycoproteins was observed due to the increased affinity from 2.2‐fold ( Aleuria aurantia lectin) to 5.6‐fold (Con A) for particular glycans. Glycoproteins could be detected with an LOD of 33 fmol using MALDI‐TOF‐MS. Finally, the enrichment step was utilized for the glycoproteomic analysis of a tryptic digest of HeLa cells. 485 Figure 37 A schematic illustration of dual binding of a BAD‐lectin (BAD = boronic acid decorated) to a glycoprotein. (i) A BAD‐lectin. (ii) A glycoprotein captured by the lectin via noncovalent glycan‐specific recognition. (iii) A glycoprotein captured by both lectin and BA; the latter mediates the formation of a boronate ester generating a stable covalent lectin‐glycoprotein complex. (iv) A glycoprotein captured by a BA ligand alone. Reprinted with permission from 485 . Copyright 2013 American Chemical Society. G. Gold NPs (AuNPs) An interesting approach to simplify glycan enrichment based on the formation of a polymeric monolith within a pipette tip was presented by Alwael et al. 486 In order to enhance the overall surface area of the monolith, 20 nm AuNPs were attached to the polymer, and the gold surface was further modified by thiols for the covalent immobilization of Erythrina cristagalli lectin (Fig. 38 ). Such a lectin‐modified monolith could selectively enrich galactosylated glycans based on the lectin´s preferential affinity with a high recovery of 95%. Finally, the device was applied in combination with reversed‐phase capillary HPLC for the analysis of E. coli lysate. 486 Figure 38 SEM images of a porous polymer monolith formed within a polypropylene pipette tip with different magnification (on left and in the middle). Field emission scanning electron microscopy images of a porous polymer monolith agglomerated with covalently attached 20 nm AuNPs with 60,000× magnification. Reproduced from 486 with permission of the Royal Society of Chemistry. A quite interesting approach to increase the sensitivity of glycopeptide/glycoprotein detection using LDI‐TOF (MALDI‐TOF‐MS, but without a need to use a matrix) was proposed by Liu et al. 487 Boronate‐modified magnetic particles with a size of 412 nm were applied to capture glycopeptides from a tryptic digestion of a model protein (HRP) or intact HRP. In the subsequent step, magnetic particles with captured glycopeptides/glycoprotein were incubated with activated AuNPs (13 nm) to covalently capture glycopeptides/glycoprotein. Unbound AuNPs were removed from the system, and glycopeptides/glycoproteins were released from boronate magnetic particles by exposure to acidic pH. Finally, AuNPs with captured glycopeptides/glycoproteins were measured using MS. AuNPs containing 64,000 Au atoms each could enhance the MS analysis of glycopeptides when applied as MS tags even without the need to use a matrix since MS could detect AuNPs with an LOD of 0.03 amol and HRP with an LOD of 45 fM. Since detection is based on the analysis of the Au 2 + ion rather than the analysis of glycopeptides/glycoproteins, the method can be applied not for the identification of glycoproteins but rather for obtaining information with high sensitivity when glycoproteins are present in a particular sample. 487 Another interesting approach for the matrix‐free analysis of various low‐molecular‐weight compounds (but not glycans) was proposed by Razunguzwa et al. using an array of silicon pillars with a diameter of 150 nm, a spacing of 337 nm, and a length of 1.2 μm, which can detect analytes down to femtomole without any enrichment. 488 Hydrazide‐functionalized ultrasmall AuNPs with a size of 1.2 nm were applied by Tran et al. for the very selective capture of periodate‐oxidized glycopeptides, when as much as 97% of all of the peptides captured from rat kidney tissue were glycopeptides. 489 This highly selective capture of glycopeptides was possible due to the extremely high density of hydrazide on AuNPs, that is, 630 nmol mg −1 , which is a 79‐fold higher density compared to hydrazide density on magnetic particles of 200–500 nm. 489 H. Silica NPs (SiNPs) Three different strategies to deposit zwitterionic brushes with a thickness of 5 nm on the surface of SiNPs (90 nm) were compared by Huang et al. using the tryptic digest of IgG as a model glycoprotein. 490 The results showed that the most efficient strategy for glycan enrichment was the application of NPs with the polymer grafted using reversible addition–fragmentation chain transfer with a low LOD of 10 fmol and a high recovery index of 88%. The approach was also successfully applied to analyze the tryptic digest of mouse liver with 303 unique glycosylation sites and with an enrichment efficiency of 70% using MALDI‐TOF‐MS, which was much higher compared to the other two approaches tested in the paper and other commercially available approaches. 490 I. Carbon NPs 1. Graphene and Graphene Oxide (GO) Graphene as a 2D crystalline material consisting of a single layer of carbon atoms with unique properties has attracted considerable attraction since its discovery in 2004. 6 Graphene, having a high surface area, could be quite effectively applied for the enrichment of glycans. Zhang et al. recently published an approach with the application of GO noncovalently modified by pyrenebutyric acid via π–π stacking interaction between graphene and a pyrene moiety. 491 In the subsequent step, the –COOH group was activated by SOCl 2 to prepare pyrenebutyryl chloride‐derivatized GO, which was applied for selective glycan binding (Fig. 39 ). A simple visual monitoring of glycan enrichment could be conducted due to GO crosslinking by attached glycans, leading to aggregation. Finally, glycans were released from the GO surface in an acidic environment by sonication at 60°C, supporting the hydrolysis of an ester bond. The enrichment protocol allowed the detection of the main glycan from just 10 ng (i.e., 0.2 pmol) of fetuin using MALDI‐TOF‐MS, and the method was applied to analyze glycans in cancer cells. Interestingly, glycoprofiling using graphene‐based enrichment method was successfully applied even when glycoprotein was highly diluted in a protein sample not having glycans at ratio of 1:100. 491 Figure 39 Graphene oxide modified by 1‐pyrenebutyric acid with subsequent activation of –COOH group by SOCl 2 for effective glycan enrichment. Reprinted with permission from 491 . Copyright 2013 American Chemical Society. GO chemically modified with pyrenebutyric acid hydrazide was applied by Bai et al. for glycan enrichment. 492 In addition to the model protein fetuin, human plasma was applied in the study with a final MALDI‐TOF‐MS analysis of the captured glycans. Glycan (maltoheptaose) recovery in a model sample spiked with BSA was 100%, but human serum spiked with maltoheptaose showed a recovery index of 73%. 492 2. Nanodiamonds Boronate‐modified nanodiamonds could also be applied for glycan enrichment, as shown by Xu et al. 493 In this particular case a tryptic digest of HRP as a model glycoprotein was performed, and 50‐fold enhanced MALDI‐MS spectra were obtained after glycan enrichment compared to a tryptic HRP digest without any enrichment, with HRP being detectable at an LOD of 0.5 nM in 100 μL (i.e., 0.5 fmol). The recovery index for glycopeptide analysis was 72%, and the purification efficiency for an analysis of glycopeptides in the mouse liver fraction was 69% with 24 newly identified glycosylation sites compared to databases. 493 Loh et al. clearly showed that the modification of nanodiamonds by boronate functionalities must be performed to have a linker between the surface of nanodiamonds and boronate functional groups in order to prevent the nonspecific binding of proteins to the hydrophobic nanodiamond surface, which would negatively influence enrichment specificity. 494 Nanodiamond particles (DNPs) were effectively applied by Wu et al. to enhance the sensitivity of glycan analysis by MALDI‐TOF‐MS. 495 In this case DNPs were used to transfer energy from matrix to glycans within a trilayer (i.e., by making a sandwich: matrix‐DNPs‐sample). The most important functions of DNPs were to mediate heat transfer between matrix‐ and glycan‐containing samples, avoiding direct heat exchange between these two components, and to prepare the highly homogeneous morphology of the spot on the MALDI target (Fig. 40 ) compared to two other methods not involving DNPs. Such particles have a low extinction coefficient not absorbing laser energy in the near‐UV region, and since DNPs are inert, they do not compete for charges with the analyte. When a model analyte dextran was analyzed by MALDI‐TOF, a 79‐ or 7‐fold increase in the detection sensitivity was achieved compared to dried‐droplet (a mixture of a matrix and a sample) or thin‐layer (layer of a sample deposited over a matrix layer) methods, respectively. The size of such particles (50–500 nm) did not influence sensitivity enhancement. 495 Figure 40 Morphology of trilayer and other samples. (A) Schematic of the configuration of matrix, DNP, and analyte in trilayer samples, (B) image of a dried‐droplet sample, (C) image of a thin‐layer sample, and (D) image of a trilayer sample containing 3 μg DNPs. The scale bars represent 1 mm. Reprinted with permission from 495 . Copyright 2013 American Chemical Society. 3. Single‐Walled CNTs Strano´s group modified SWCNTs with derivatives of PBAs containing –COOH, ‐NO 2 , and –NH 2 functional groups with ortho‐ , meta‐ , and para‐ substitutions generating 144 distinct corona phases on the surface of SWCNTs and some of them exhibited remarkable binding affinity to certain monosaccharides, while others were not bound. 496 , 497 Even though the method was applied for sensing purposes (i.e., monitoring of quenching of intrinsic fluorescence of SWCNTs 496 , 497 ), this can be applied for the enrichment of glycans. 498 J. Hybrid NPs A polydopamine film was deposited by Bi et al. using self‐polymerization on an rGO‐Fe 3 O 4 ‐modified surface and applied for the subsequent deposition of AuNPs. 499 Thiolated mannose finally formed SAM on AuNPs, and such a nanocomposite was employed for glycan enrichment. HRP as a model glycoprotein could be detected by MALDI‐TOF from a tryptic digest of the protein down to a concentration of 0.1 ng uL −1 (i.e., 40 ng or 1 pmol). 499 Hu et al . modified silica bubbles (30 um) with AuNPs (20 nm), which were further functionalized with boronate‐terminated thiol. 500 Glycopeptides prepared by a tryptic digest from two glycoproteins (HRP and IgG) could be enriched with a binding capacity of 60 mg g −1 , and as much as 10 ng (∼200 fmol) of protein was needed for glycoprofiling with a rather low enhancement (approximately tenfold) of MALDI‐TOF signal compared to the signal obtained without enrichment. 500 Ju et al. applied hybrid NPs (magnetic CNTs) prepared on CNTs with d = 40–60 nm as a scaffold by the in situ formation of MNPs ( d = 10–15 nm) from Fe 3+ ions. 501 Finally, hybrid NPs were patterned by the boronate functional group. Model glycoprotein HRP could be enriched with a sorption capacity of 346 mg g −1 , while the sorption of nonglycoprotein HSA was quite low (52 mg g −1 ). The LOD for the detection of HRP by MALDI‐TOF‐MS using an enrichment step was 1 pmol, while with commercially available boronate‐modified agarose gel, it was not possible to detect 10 pmol of the glycoprotein, and hybrid NPs could enrich glycopeptides in the presence of nonglycosylated peptides, while the latter being in 50‐fold excess. 501 Zou et al. applied a quite sophisticated strategy for glycan enrichment. 502 GO with deposited MNPs were covered by a silica shell of a final thickness of 50–60 nm. In the next step, the PAMAM dendrimer was grafted to the surface with subsequent modification by AuNPs on which thiolated maltose was anchored. Such a maltose‐modified hydrophilic composite with a surface area of 57.8 m 2 g −1 and with a maltose density of 2.4 μmol m −2 could enrich glycopeptides with an LOD of 0.5 fmol. This approach was finally applied for the analysis of as low as 50 μg of mouse liver tryptic digest. 502 For comparison, a commercially available meta‐aminophenylboronic acid modified agarose could enrich glycopeptides with an LOD of 30 nmol. 503 Yang et al. prepared a nanocomposite with graphene as a support to accommodate MNPs (100 nm) and a phenolic‐formaldehyde resin (condensation of formaldehyde and hydroquinone) with a final modification by aminophenylboronic acid for glycopeptide enrichment. 504 The graphene‐based composite with a thickness of 10 nm and a surface area of 76.3 m 2 g −1 could detect glycopeptides down to 1 fmol using MALDI‐TOF‐MS, even when the concentration of peptides was in excess of 100:1 compared to glycopeptides. The composite was applied to analyze human serum, when as little as 1 μL of a sample was sufficient. 504 K. Other Interesting Approaches Herein are described interesting strategies of effective glycan enrichment not necessarily based on the application of NPs. The approach developed by Jiao et al. is based on the use of hydrazinonicotinic acid as a matrix for the MALDI‐TOF‐MS analysis of glycans, where the matrix, besides adsorbing laser energy, also contains a hydrazine moiety for selective interaction with glycans. 505 Thus, no glycan enrichment is needed since laser energy is mainly adsorbed by glycans even in the presence of proteins/peptides. The detection limit for a glycan is down to 1 amol, which is 5 orders of magnitude lower amount compared to 2,5‐dihydroxybenzoic acid used as a matrix. The other advantages of using this novel matrix are the higher homogeneity of glycan spots and better salt tolerance compared to the traditional matrix. The approach was finally applied for the analysis of a human serum. 505 The online enzyme digestion of glycoproteins within a microbore hollow fiber reactor (ID = 200 um) was suggested by Kim et al. using trypsin to digest glycoproteins into peptides and glycopeptides within 30 min, which were then isolated from peptides using lectins. 506 Finally, glycans were released from glycopeptides by the application of PNGase F, and the method was applied for the analysis of human urinary samples (PC patients). 506 An improved assay protocol from the same group was applied for the analysis of sera from patients with liver cancer. 507 Time needed for tryptic digestion was comparable to that needed for the microwave‐assisted release of glycans in the presence of trypsin (10 min at 50°C). 464 Direct selective glycan enrichment was performed by Li et al. using a hydrophobic fluorinated carbon tag with an –NH 2 terminal group for specific coupling to the reducing end of glycans. 508 Such modified glycans could be ionized more efficiently by one order of magnitude compared to unmodified glycans using MALDI‐TOF‐MS. Alternatively, hydrophobized glycans could be selectively isolated from a mixture with other biomolecules using fluorous solid extraction. 508 Liu et al. developed a matrix‐free strategy for MS using a lithium‐rich metal oxide composite with a particle size of 200–300 nm applied to analyze low‐molecular‐weight analytes, including several oligosaccharides (not glycans). 509 Ruman et al. applied AgNPs with a size of 100 nm for the matrix‐free MS of various low‐molecular‐weight analytes, but not glycans, with a remarkable sensitivity of detection down to 33 amol for ribose. 510 Moreover, the application of AgNPs as a matrix in the MS analysis of low‐molecular‐weight molecules was recently reviewed by Sekula et al. 511 Yang et al. developed an on‐plate glycopeptide enrichment procedure using gold‐coated silicon wafer modified by SAM with a final modification by boronate functional groups taking 24 hr to complete. 512 This approach could effectively pre‐concentrate glycopeptides released from three model glycoproteins with an LOD of 1 fmol, which increased 93‐ to 248‐fold compared to a procedure without enrichment. The plate could, however, be reused only three times because its capability to enrich glycopeptides then decreased sharply. 512 Lu et al. deposited 900 μm gold spots on a hydrophobic silica wafer and the gold layer was patterned by 4‐mercaptophenylboronic acid for specific glycoproteins/glycopeptides enrichment. 513 The LOD for glycopeptides released from a model glycoprotein HRP was 230 fmol, which is 1 order of magnitude lower compared to MALDI‐TOF‐MS using a standard stainless steel plate. 513 Magnetic particles with d = 220 nm were used in the study performed by Xiong et al. 514 The particles were covered by silane terminated in the –NH 2 group, thus forming a positively charged surface. A layer‐by‐layer method was then applied to cover modified particles by the alternate deposition of negatively charged HA and positively charged chitosan. The best glycan enrichment strategy was obtained with particles having ten layers of both polysaccharides. Glycan enrichment with a binding capacity of 200 mg g −1 was achieved by hydrophilic interaction with glycans and polysaccharide‐modified MNPs. Three glycopeptides from a tryptic digest of IgG could be detected from just 0.2 fmol of IgG with a recovery as high as 69%. The most abundant peak in MALDI‐TOF‐MS was enhanced 111‐fold compared to the procedure without any glycan enrichment. Finally, the approach was applied for analysis of the glycoproteome from just 20 μg of a mouse liver protein sample with the identification of 605 unique N ‐glycosylation sites in 616 distinct glycopeptides. 514 Another approach using particles larger than 100 nm described by Ma et al. is worth mentioning. 515 This ligand‐free approach is based on the application of hybrid particles having a magnetic core with 22 nm AgNPs deposited on the surface. Selective glycan enrichment by such hybrid particles was achieved by the reversible interaction between glycans and the Ag surface, and glycopeptides could be effectively separated from a mixture having a molar ratio of glycopeptides to nonglycopeptides of 1:100 within 1 min. Finally, the method was applied for MALDI‐TOF‐MS glycan analysis from just 1 μL of rat serum. 515 A homogeneous system for glycan enrichment based on the pH‐responsive polymer poly‐(acrylic acid‐co‐methyl acrylate) was proven by Bai et al. to be much more effective (96.2% of glycoproteins captured within 1 hr) compared to a solid‐phase glycan enrichment (90% captured in 8 hr). 516 The polymer with a hydrodynamic size of 30 nm was soluble at pH 6.0 but became insoluble at pH 2.0. When the hydrazide‐modified polymer was incubated with oxidized glycopeptides/glycoproteins, their separation could be performed by switching the pH to 2.0 with subsequent centrifugation. From a mixture of glycoproteins with proteins at a ratio of 1:100, low levels of glycoproteins (1 fmol) can be effectively captured. Using polymer‐based glycan enrichment MALDI‐TOF‐MS, the signal intensity increased 29‐fold and the S/N ratio increased 325‐fold. The glycan enrichment strategy was finally applied for the analysis of mouse brain samples. 516 Zhang et al. applied a hydrophilic support‐based amino‐functionalized metal‐organic framework containing 25 nm NPs with a specific surface are of 2187 m 2 g −1 for glycopeptide enrichment. 517 A tryptic digest of IgG as a model glycoprotein could be detected after enrichment using MALDI‐TOF‐MS with an LOD of 20 fmol, and an analysis of human serum was performed from a volume as small as 10 μL. 517 A. Glycan Release Glycans attached to the protein backbone can be analyzed either in their intact form (i.e., glycoproteins), after protein digestion by endopeptidases or endoproteinases (i.e., glycopeptides) or as pure glycans released from the protein or peptide backbone using the enzyme PNGase F or PNGase A (Fig. 30 ). The release of glycans from O ‐glycans can be done not using enzymes but rather chemical reaction, that is, β‐elimination or the use of hydrazine is needed. When glycopeptides are produced from glycoproteins by a tryptic digest, glycopeptides are only a minor component of the tryptic digest, and their enrichment is important for high‐performance MS analysis. Peptides isolated from glycopeptides can be further applied for protein fingerprinting. 431 , 432 Figure 30 A graphical summary of different approaches, which can be applied for glycoprotein analysis including glycomics (analysis of released glycans), proteomics (identification of peptides after removal of glycans) and glycoproteomics (characterization of intact glycoproteins in order to indentify a site‐specific connectivity between glycans and proteins). B. Mass Spectrometry MALDI‐time‐of‐flight‐MS (MALDI‐TOF‐MS) is a gentle technique allowing the analysis of large biomolecules in both positive and negative ionization modes. This is why this technique is the most commonly applied in glycan, glycopeptide, or glycoprotein analysis, even though the MALDI‐TOF process itself is not yet fully understood. 433 On the other hand MALDI‐TOF and electrospray ionization (ESI) might not distinguish glycan isomers when operated in a conventional mode. Various on‐line separation techniques (i.e., using porous graphitized carbon packed columns), which can be coupled to ESI MS, can effectively resolve this problem. 434 Alternatively, tandem MS, that is, MS/MS or MS n , can resolve this issue as well. 435 MALDI‐TOF compared to ESI exhibits higher sensitivity for the determination of glycans and efficiently ionizes molecules with a higher molecular weight. At the same time, MALDI‐TOF is more tolerant to contaminants. 435 MALDI‐TOF when compared to ESI introduces higher energy to the sample, resulting in the fragmentation of labile functional groups, including sulfates, phosphates, and sialic acids. 435 Since sialic acids are quite labile molecules with significant involvement in many biological processes, an analysis of sialic acids requires the development of special analytical strategies. 436 , 437 , 438 , 439 Moreover, a proper matrix must be chosen to effectively generate ions of large molecules for MALDI‐TOF analysis, and as for ESI, the sensitivity of detection decreases with the increasing molecular weight of glycans. 435 C. Glycan Enrichment There are few main reasons why glycans have to be enriched prior to MS analysis: (i) the glycan level in the studies sample can be well below detection limit of an MS instrument; (ii) glycopeptides or glycans must be removed from peptides due to their lower abundance compared to peptides; and (iii) the ionization of glycans or glycopeptides in the presence of peptides is frequently less efficient. 63 Glycans, glycopeptides, or glycoproteins can be enriched in numerous ways, including boronate affinity, hydrazide chemistry, lectin affinity, hydrophilic interactions, and other approaches (Fig. 31 ). 440 Figure 31 Various ways glycans can be enriched using hydrophobic, hydrophilic, or ionic interactions (upper left part) or biorecognition interactions involving glycan binding proteins (lectins, lower left part). Alternatively for glycan capture/release particles modified by boronate functional groups or hydrazine can be applied, as well. 1. Boronate‐Based Enrichment Boronate‐modified materials are applied for glycan enrichment due to the strong interactions between boronate functional groups and cis ‐diols present in glycans under basic conditions, that is, at pH 8–9. 441 The main prerequisite for glycan binding to boronate‐modified material is the choice of a proper pH, which should be above the p K a of the boronate moiety of a boronate‐terminated molecule (e.g., p K a = 8.9 for frequently applied PBA). Sialic acids are exceptions to this rule since the effective binding of SA to boronate‐modified material can occur at a pH lower than the p K a of boronate functional group. Recent studies however suggest that besides proper choice of pH, other factors come into play during the interaction of glycans with boronate‐modified material, such as steric factors, and the nature of buffer components. 442 Decreased pH can be effectively applied to release captured glycans for further analysis. Relatively high pH needed for effective glycan enrichment based on boronate affinity can be a problem for quite complex biomolecules. Such glycopeptides or glycoproteins can become unstable at a high pH. To counter this, various strategies have been applied to lower the pH needed for such binding, including the proper choice of substituents (electron withdrawing groups) in proximity to boronate functional groups. 443 Alternatively, benzoboroxozole derivatives can be applied to lower the pH required for glycan binding via boronate affinity. 444 Such approaches are also useful due to the widening of the applicable pH window for glycan enrichment. 2. Hydrazide‐Based Enrichment Hydrazide‐modified materials can be efficiently applied for glycan enrichment, but glycans must be oxidized first to form aldehydes from cis‐diol groups. Such aldehydes react with hydrazide‐forming hydrazone bonds, which are stable under basic conditions, and glycan can be released from enrichment material by the application of acidic pH, similar to glycan release from boronate affinity materials. Alternatively, PNGase F could be applied for the release of glycopeptides. 445 , 446 3. Oxime‐Based Enrichment Oxime‐modified surfaces can be a good alternative to the hydrazide‐based enrichment of glycans since the process of enrichment can be considerably shorter (1 hr) compared to hydrazide‐based enrichment (12–16 hr), with high enrichment sensitivity with LOD in the low femtomole range and with the selectivity of glycopeptide enrichment even in the presence of a 100‐fold excess of nonglycopeptides. Glycopeptides must be oxidized in order to interact with oxime functionalities, and overnight incubation with PNGase F was required for glycan release for subsequent MALDI‐TOF‐MS analysis. 447 4. Hydrophilic Interactions A hydrophilic interaction‐based method for glycan enrichment involves the application of a polar stationary phase and a nonpolar mobile phase for separation. Analytes partition between a water‐rich layer on the surface of the stationary phase and the mobile phase. Glycans as hydrophilic molecules will be attached to the column/material, while other more hydrophobic molecules (including peptides) will be dissolved in the mobile phase and washed away. Hydrophilic interaction based on zwitterionic functional groups has become popular in recent years due to the ability of zwitterions to form a stable water layer on the surface of the separation phase, which is effective for glycan partitioning. 430 5. Lectin‐Based Enrichment A lectin‐based approach for glycan enrichment is based on the natural bioaffinity interaction between glycans and glycan‐binding proteins (lectins) triggered by the interaction of glycans, especially with aromatic stacking amino acids and amino acids providing hydrogen bonds. 448 The application of lectins for glycan enrichment is advantageous compared to boronate, hydrophilic interaction or hydrazide‐modified materials due to the higher degree of specificity for some glycan structures, but selectivity and binding affinity are weaker compared to antibody‐based biorecognition. Moreover, lectins can distinguish minor changes in glycan structures. A good example is SNA, recognizing sialic acid bound to galactose via an α‐2,6 bond, while Maackia amurensis agglutinin binds to α‐2,3‐linked sialic acid. Another example is the arm‐specific recognition of bisecting glycans (i.e., glycans with two arms). 449 Thus, not all glycans present in the sample will be enriched using lectins, and a lectin‐affinity enrichment strategy can be applied to fractionate glycans into several groups. Finally, glycans can be released from lectins by the application of free carbohydrates or by exposure to an acidic eluting buffer. 6. Other Ways for Enrichment There are other options for glycan enrichment, such as reverse‐phase mode, electrostatic mode (for glycans containing negatively charged sialic acids), electrostatic repulsion hydrophilic interactions, and affinity capture based on protein A, which is applicable for the selective enrichment of antibodies with the subsequent glycoprofiling of glycans present especially in the Fc fragment of IgGs. 21 , 430 1. Boronate‐Based Enrichment Boronate‐modified materials are applied for glycan enrichment due to the strong interactions between boronate functional groups and cis ‐diols present in glycans under basic conditions, that is, at pH 8–9. 441 The main prerequisite for glycan binding to boronate‐modified material is the choice of a proper pH, which should be above the p K a of the boronate moiety of a boronate‐terminated molecule (e.g., p K a = 8.9 for frequently applied PBA). Sialic acids are exceptions to this rule since the effective binding of SA to boronate‐modified material can occur at a pH lower than the p K a of boronate functional group. Recent studies however suggest that besides proper choice of pH, other factors come into play during the interaction of glycans with boronate‐modified material, such as steric factors, and the nature of buffer components. 442 Decreased pH can be effectively applied to release captured glycans for further analysis. Relatively high pH needed for effective glycan enrichment based on boronate affinity can be a problem for quite complex biomolecules. Such glycopeptides or glycoproteins can become unstable at a high pH. To counter this, various strategies have been applied to lower the pH needed for such binding, including the proper choice of substituents (electron withdrawing groups) in proximity to boronate functional groups. 443 Alternatively, benzoboroxozole derivatives can be applied to lower the pH required for glycan binding via boronate affinity. 444 Such approaches are also useful due to the widening of the applicable pH window for glycan enrichment. 2. Hydrazide‐Based Enrichment Hydrazide‐modified materials can be efficiently applied for glycan enrichment, but glycans must be oxidized first to form aldehydes from cis‐diol groups. Such aldehydes react with hydrazide‐forming hydrazone bonds, which are stable under basic conditions, and glycan can be released from enrichment material by the application of acidic pH, similar to glycan release from boronate affinity materials. Alternatively, PNGase F could be applied for the release of glycopeptides. 445 , 446 3. Oxime‐Based Enrichment Oxime‐modified surfaces can be a good alternative to the hydrazide‐based enrichment of glycans since the process of enrichment can be considerably shorter (1 hr) compared to hydrazide‐based enrichment (12–16 hr), with high enrichment sensitivity with LOD in the low femtomole range and with the selectivity of glycopeptide enrichment even in the presence of a 100‐fold excess of nonglycopeptides. Glycopeptides must be oxidized in order to interact with oxime functionalities, and overnight incubation with PNGase F was required for glycan release for subsequent MALDI‐TOF‐MS analysis. 447 4. Hydrophilic Interactions A hydrophilic interaction‐based method for glycan enrichment involves the application of a polar stationary phase and a nonpolar mobile phase for separation. Analytes partition between a water‐rich layer on the surface of the stationary phase and the mobile phase. Glycans as hydrophilic molecules will be attached to the column/material, while other more hydrophobic molecules (including peptides) will be dissolved in the mobile phase and washed away. Hydrophilic interaction based on zwitterionic functional groups has become popular in recent years due to the ability of zwitterions to form a stable water layer on the surface of the separation phase, which is effective for glycan partitioning. 430 5. Lectin‐Based Enrichment A lectin‐based approach for glycan enrichment is based on the natural bioaffinity interaction between glycans and glycan‐binding proteins (lectins) triggered by the interaction of glycans, especially with aromatic stacking amino acids and amino acids providing hydrogen bonds. 448 The application of lectins for glycan enrichment is advantageous compared to boronate, hydrophilic interaction or hydrazide‐modified materials due to the higher degree of specificity for some glycan structures, but selectivity and binding affinity are weaker compared to antibody‐based biorecognition. Moreover, lectins can distinguish minor changes in glycan structures. A good example is SNA, recognizing sialic acid bound to galactose via an α‐2,6 bond, while Maackia amurensis agglutinin binds to α‐2,3‐linked sialic acid. Another example is the arm‐specific recognition of bisecting glycans (i.e., glycans with two arms). 449 Thus, not all glycans present in the sample will be enriched using lectins, and a lectin‐affinity enrichment strategy can be applied to fractionate glycans into several groups. Finally, glycans can be released from lectins by the application of free carbohydrates or by exposure to an acidic eluting buffer. 6. Other Ways for Enrichment There are other options for glycan enrichment, such as reverse‐phase mode, electrostatic mode (for glycans containing negatively charged sialic acids), electrostatic repulsion hydrophilic interactions, and affinity capture based on protein A, which is applicable for the selective enrichment of antibodies with the subsequent glycoprofiling of glycans present especially in the Fc fragment of IgGs. 21 , 430 D. Nanoporous Materials Porous materials are, according to IUPAC definition, considered nanoporous only if the pore size is below 1 μm, but earlier terminology defining materials as macroporous (above 50 nm), mesoporous (2–50 nm), and microporous (below 2 nm) is frequently used in the literature. 450 In the forthcoming paragraphs we will stick to the latter division of nanoporous materials into three categories. The other important parameters of porous materials are distribution of pore size, pore shape (cylindrical, slit shaped, funnel shaped, etc.), and degree of pore interconnections. The porosity of the given material can be characterized by the total pore volume and a specific surface area. Such porous materials are quite frequently applied as packing material for the separation columns for the enrichment/separation of glycoproteins, glycopeptides, or glycans and have been reviewed in the past. 430 , 446 , 451 , 452 The size and modification of pores play a prominent role in interactions with ligands due to the confinement effect. 453 For example, when glycoprotein (ribonuclease B) interacted with a mesoporous material, the enhancement factor (i.e., K D of the interaction with mesoporous silica compared to non‐porous silica) increased from 45 to 900, when, in addition boronate affinity, electrostatic interactions were also involved. Upon the decreased pore size of mesoporous silica from 2.6 to 2.1 nm, a further increase in the enhancement factor to 2100 was observed. 453 This indicates that for effective ligand capture, the careful design of surface chemistry and the morphology of porous materials are important. 454 An interesting approach for how to increase the surface area of composites (i.e., from 69 to 93 m 2 g −1 ) while decreasing the pore size (from 7.6 to 5.0 nm) by mixing GO sheets into composites was demonstrated by He et al. 455 1. Silica‐Based Nanoporous Materials Porous silica material with an ordered/controlled size and shape of large pores (1 μm) was prepared by Yan et al. 456 using polystyrene colloidal crystals as a hard template, with mesopores of 4.6 nm using a surfactant as a soft template and applying tetraethylorthosilicate (TEOS) as a silica source. Moreover, macropores were interconnected via pores of 50 nm in diameter (Fig. 32 ). The specific surface area of the material was 255 m 2 g −1 with a pore volume of 0.46 cm 3 g −1 . Such hierarchically ordered material was then modified with 4‐vinylphenylboronic acid to form polyboronic acid‐modified material, which was successfully applied for the effective enrichment of glycopeptides prepared by the tryptic digest of a model glycoprotein ‐ HRP. The method using glycan enrichment (10 min) revealed the presence of 12 glycans by MALDI‐TOF MS, while only four glycans were obtained without any glycan enrichment. The glycan of HRP could be detected with an LOD of 1 ng μL −1 (1 fmol) even in the presence of BSA, which was in 50:1 excess compared to HRP. 456 Figure 32 SEM image of hierarchically ordered macro/mesoporous silica material before (left) and after (right) boronate modification. Reproduced from 456 with permission by John Wiley & Sons. Mechanically stable silica microparticles of 1.6 μm in diameter with macropores ranging in size from 50 to 150 nm were applied to immobilize two lectins, Con A and Aleuria aurantia lectin, by Mann et al. 457 Such particles with interconnected pores and with a high surface area of 200 m 2 g −1 accommodated a high density of immobilized lectins (16 mg g −1 ). When lectin‐modified particles were applied as a support for loading a chromatography column, up to 75 μg of glycoproteins could be bound to the column (50 × 1 mm) with immobilized Con A. Finally, the columns were used to enrich glycoproteins from just 1 μL of blood serum after removing IgG and albumin. Subsequently, the glycomic and glycoproteomic profiling of bound glycoproteins was performed using MALDI TOF/TOF MS. 457 A capillary with boronate‐modified mesoporous walls was used for the selective enrichment of neutral or acidic glycans via pH manipulation by Lu et al. 458 When the binding solution had a pH lower than the p K a of boronate by one pH unit or more, the boronate‐modified capillary column captured sialylated glycoproteins. In contrast, when the pH of the binding buffer was higher than the p K a value of boronate by one pH unit or more, glycoproteins containing SA were electrostatically repelled from interaction with the monolith, and glycoproteins with neutral glycans were preferentially captured. Furthermore, the ionic strength of the binding buffer played an important role in the capture of sialylated glycoproteins. 458 A boronate‐modified affinity column was also helpful in the identification of DNA aptamers binding to glycoproteins using the systematic evolution of ligands by exponential enrichment as proposed by Nie et al. 459 The boronate column served as a matrix for binding glycoprotein (HRP), with the complete process performed on the column; the authors identified seven DNA aptamers binding HRP with an affinity constant down to 10 nM. 459 DNA aptamers selected specifically against glycoproteins can be applied for the selective capture of some glycoproteins in the future since their K D (10 −8 M) can be lower than that of lectins. 459 2. Carbon‐Based Nanoporous Materials TEOS with a surfactant as a structure‐directing agent was applied by Qin et al. 460 to prepare SiNPs with a uniform size of 70 nm. The particles with a surface area of 901 m 2 g −1 with a pore size of 2.8 nm were treated with sulfuric acid and subsequently carbonized with the appearance of sp 2 ‐hybridized aromatic carbon structures, which were effective in glycan enrichment. Thirteen N ‐glycans released from just 5 ng of ovalbumin (i.e., LOD of 100 fmol) by PNGase F could be detected. Finally, the particles were applied for the enrichment of N ‐glycans from 50 nL of human serum and analyzed by MALDI‐TOF‐MS. Eight new N ‐glycans that were not observed in the MS spectrum when standard carbon adsorbent was applied were observed, including low‐abundant fucose and high‐mannose glycans. 460 Mesoporous silica composites were prepared by Sun et al. 461 using a graphene layer as a support. In this case, graphene was oxidized by nitric acid to deposit oxygen‐rich functionalities, which were used in the subsequent step for the formation of a mesoporous silica layer by the deposition of TEOS in the presence of a surfactant as a structure‐directing agent. Finally, the whole nanocomposite was carbonized to transform the surfactant into a carbon layer. The resulting material exhibited a uniform pore size of 2.8 nm with protein‐excluding properties with a surface area of 372 m 2 g −1 . MALDI‐TOF‐MS showed the presence of 25 N ‐linked glycans released from ovalbumin when glycan enrichment with graphene‐based nanocomposite was applied, while enrichment with active carbon material was less effective in terms of the number of glycans detected and the signal intensity obtained. MALDI‐TOF‐MS analysis without any enrichment did not show any sign of glycans. 461 Commercially available mesoporous graphitic carbon with a pore size of 25 or 30 nm was applied by Zhao et al. as a packing material for fully automatable two‐dimensional liquid chromatography for high‐throughput proteomic and glycomic analysis of various lysates/samples. 462 With this fully automated approach, it was possible to identify up to 2678 proteins, 11,984 unique peptides in neurons or up to 130 N ‐glycoproteins, 705 N ‐glycans, and 254 glycosylation sites in macaques plasma ( Macaca fascicularis ) with a total analysis time of 19 hr using ESI‐MS/MS. According to authors, the technology offered unattended, robust, and scalable analyses of samples on a submicrogram scale when sample injection was performed once a day, and the system was able to operate continuously for up to 14 days without experiencing any major problems. 462 A porous silica column was applied for the covalent immobilization of PNGase F by Jmeian et al., and such a column was used for the continuous (online) release of neutral and acidic glycans from glycoproteins with subsequent continuous glycan enrichment in a column loaded with porous graphitic carbon and with a subsequent analysis of glycans. 463 The capillary was coupled to a C 8 trap and a porous graphitic carbon HPLC‐chip and finally interfaced to perform liquid chromatography‐MS and liquid chromatography‐MS/MS analyses. The main advantage of the system is the short time (6 min) needed to release glycans compared to the usual overnight incubation at 37°C. Moreover, glycans could be effectively analyzed from just 100 fmol of protein or 0.1 μL of human blood serum. 463 The time needed for online glycan removal is shorter compared to that needed for the microwave‐assisted release of glycans in the presence of PNGase F (30 min at 37°C). 464 GO was due to the large surface area (theoretical relative surface area of 2630 m 2 g −1 ) effectively applied by Ren et al. as a support for the covalent immobilization of a high amount of PNGase F (688 μg g −1 of GO, which is threefold higher compared to SiNPs). Highly efficient N ‐glycan release from a protein backbone of two model glycoproteins (ribonuclease B and asialofetuin) can be completed within 2 min, and the enzyme can be reused. The stability of the enzyme complex is remarkable, that is, up to 8 weeks, when stored at 4°C, and as low as 2 μg μL −1 of the enzyme complex was needed to analyze the plasma extracts. 465 3. Other Nanoporous Materials Hydrophilic mesoporous phenol‐based material with added PEI, prepared by Jin et al. with a template‐free method, exhibited a large surface area (548 m 2 g −1 ) with 13 nm 466 mesopores. When the amount of PEI changed, it was possible to change both the porosities and surface area of the material. Glycopeptides could be detected with an enrichment factor (MS intensity ratio after and before enrichment) of 62 with a recovery index of 80.4%. A tryptic digest of IgG could be detected down to 5 fmol with only a moderate improvement compared to commercial hydrophilic interaction chromatography beads (40 fmol). 466 A capillary modified with a polymeric material with an introduced –SH group having a surface area of 30.3 m 2 g −1 was applied for the further selective modification of the surface with AuNPs (13 nm) by Wu et al. 467 The gold surface was further modified by two thiols terminated in boronate and amine functional groups making an intramolecular B‐N coordination for the enhanced specificity of glycoprotein capture with a binding capacity of 0.39 mg g −1 . The capillary was applied for the selective enrichment of a glycoprotein over a protein (1:1000 ratio) with a recovery of 85% and finally applied for the glycoproteomic profiling of just 9 μg of human plasma with 160 glycoproteins identified by MS. 467 A boronate‐modified affinity monolithic column with a surface area of 13.7 m 2 g −1 and with bimodal pores containing mesopores (3.9 nm) and macropores (1.4 μm) was prepared by Liu et al. using ring‐opening polymerization to mimic the function of protein A to bind antibodies. 468 In this case, mesopores could be accommodated by the glycan of IgG via the interaction of glycans with boronate, while the protein backbone of IgG could not enter the mesopores (Fig. 33 ). The binding capacity of such material for IgG binding was comparable (23.7 mg g −1 ) to a number of protein A mimics. The monolith was a quite stable and cost‐effective alternative to using protein A with the price of 40 USD per 1 g. Moreover, the IgG attached to the monolith still exhibited its binding affinity toward antigens. 468 Figure 33 A scheme showing a specific recognition of IgG by the monolith exhibiting protein A‐like binding with presence of bimodal pores (macropores and mesopores) modified by boronate functional groups. Reproduced from 468 with permission of the Royal Society of Chemistry. 1. Silica‐Based Nanoporous Materials Porous silica material with an ordered/controlled size and shape of large pores (1 μm) was prepared by Yan et al. 456 using polystyrene colloidal crystals as a hard template, with mesopores of 4.6 nm using a surfactant as a soft template and applying tetraethylorthosilicate (TEOS) as a silica source. Moreover, macropores were interconnected via pores of 50 nm in diameter (Fig. 32 ). The specific surface area of the material was 255 m 2 g −1 with a pore volume of 0.46 cm 3 g −1 . Such hierarchically ordered material was then modified with 4‐vinylphenylboronic acid to form polyboronic acid‐modified material, which was successfully applied for the effective enrichment of glycopeptides prepared by the tryptic digest of a model glycoprotein ‐ HRP. The method using glycan enrichment (10 min) revealed the presence of 12 glycans by MALDI‐TOF MS, while only four glycans were obtained without any glycan enrichment. The glycan of HRP could be detected with an LOD of 1 ng μL −1 (1 fmol) even in the presence of BSA, which was in 50:1 excess compared to HRP. 456 Figure 32 SEM image of hierarchically ordered macro/mesoporous silica material before (left) and after (right) boronate modification. Reproduced from 456 with permission by John Wiley & Sons. Mechanically stable silica microparticles of 1.6 μm in diameter with macropores ranging in size from 50 to 150 nm were applied to immobilize two lectins, Con A and Aleuria aurantia lectin, by Mann et al. 457 Such particles with interconnected pores and with a high surface area of 200 m 2 g −1 accommodated a high density of immobilized lectins (16 mg g −1 ). When lectin‐modified particles were applied as a support for loading a chromatography column, up to 75 μg of glycoproteins could be bound to the column (50 × 1 mm) with immobilized Con A. Finally, the columns were used to enrich glycoproteins from just 1 μL of blood serum after removing IgG and albumin. Subsequently, the glycomic and glycoproteomic profiling of bound glycoproteins was performed using MALDI TOF/TOF MS. 457 A capillary with boronate‐modified mesoporous walls was used for the selective enrichment of neutral or acidic glycans via pH manipulation by Lu et al. 458 When the binding solution had a pH lower than the p K a of boronate by one pH unit or more, the boronate‐modified capillary column captured sialylated glycoproteins. In contrast, when the pH of the binding buffer was higher than the p K a value of boronate by one pH unit or more, glycoproteins containing SA were electrostatically repelled from interaction with the monolith, and glycoproteins with neutral glycans were preferentially captured. Furthermore, the ionic strength of the binding buffer played an important role in the capture of sialylated glycoproteins. 458 A boronate‐modified affinity column was also helpful in the identification of DNA aptamers binding to glycoproteins using the systematic evolution of ligands by exponential enrichment as proposed by Nie et al. 459 The boronate column served as a matrix for binding glycoprotein (HRP), with the complete process performed on the column; the authors identified seven DNA aptamers binding HRP with an affinity constant down to 10 nM. 459 DNA aptamers selected specifically against glycoproteins can be applied for the selective capture of some glycoproteins in the future since their K D (10 −8 M) can be lower than that of lectins. 459 2. Carbon‐Based Nanoporous Materials TEOS with a surfactant as a structure‐directing agent was applied by Qin et al. 460 to prepare SiNPs with a uniform size of 70 nm. The particles with a surface area of 901 m 2 g −1 with a pore size of 2.8 nm were treated with sulfuric acid and subsequently carbonized with the appearance of sp 2 ‐hybridized aromatic carbon structures, which were effective in glycan enrichment. Thirteen N ‐glycans released from just 5 ng of ovalbumin (i.e., LOD of 100 fmol) by PNGase F could be detected. Finally, the particles were applied for the enrichment of N ‐glycans from 50 nL of human serum and analyzed by MALDI‐TOF‐MS. Eight new N ‐glycans that were not observed in the MS spectrum when standard carbon adsorbent was applied were observed, including low‐abundant fucose and high‐mannose glycans. 460 Mesoporous silica composites were prepared by Sun et al. 461 using a graphene layer as a support. In this case, graphene was oxidized by nitric acid to deposit oxygen‐rich functionalities, which were used in the subsequent step for the formation of a mesoporous silica layer by the deposition of TEOS in the presence of a surfactant as a structure‐directing agent. Finally, the whole nanocomposite was carbonized to transform the surfactant into a carbon layer. The resulting material exhibited a uniform pore size of 2.8 nm with protein‐excluding properties with a surface area of 372 m 2 g −1 . MALDI‐TOF‐MS showed the presence of 25 N ‐linked glycans released from ovalbumin when glycan enrichment with graphene‐based nanocomposite was applied, while enrichment with active carbon material was less effective in terms of the number of glycans detected and the signal intensity obtained. MALDI‐TOF‐MS analysis without any enrichment did not show any sign of glycans. 461 Commercially available mesoporous graphitic carbon with a pore size of 25 or 30 nm was applied by Zhao et al. as a packing material for fully automatable two‐dimensional liquid chromatography for high‐throughput proteomic and glycomic analysis of various lysates/samples. 462 With this fully automated approach, it was possible to identify up to 2678 proteins, 11,984 unique peptides in neurons or up to 130 N ‐glycoproteins, 705 N ‐glycans, and 254 glycosylation sites in macaques plasma ( Macaca fascicularis ) with a total analysis time of 19 hr using ESI‐MS/MS. According to authors, the technology offered unattended, robust, and scalable analyses of samples on a submicrogram scale when sample injection was performed once a day, and the system was able to operate continuously for up to 14 days without experiencing any major problems. 462 A porous silica column was applied for the covalent immobilization of PNGase F by Jmeian et al., and such a column was used for the continuous (online) release of neutral and acidic glycans from glycoproteins with subsequent continuous glycan enrichment in a column loaded with porous graphitic carbon and with a subsequent analysis of glycans. 463 The capillary was coupled to a C 8 trap and a porous graphitic carbon HPLC‐chip and finally interfaced to perform liquid chromatography‐MS and liquid chromatography‐MS/MS analyses. The main advantage of the system is the short time (6 min) needed to release glycans compared to the usual overnight incubation at 37°C. Moreover, glycans could be effectively analyzed from just 100 fmol of protein or 0.1 μL of human blood serum. 463 The time needed for online glycan removal is shorter compared to that needed for the microwave‐assisted release of glycans in the presence of PNGase F (30 min at 37°C). 464 GO was due to the large surface area (theoretical relative surface area of 2630 m 2 g −1 ) effectively applied by Ren et al. as a support for the covalent immobilization of a high amount of PNGase F (688 μg g −1 of GO, which is threefold higher compared to SiNPs). Highly efficient N ‐glycan release from a protein backbone of two model glycoproteins (ribonuclease B and asialofetuin) can be completed within 2 min, and the enzyme can be reused. The stability of the enzyme complex is remarkable, that is, up to 8 weeks, when stored at 4°C, and as low as 2 μg μL −1 of the enzyme complex was needed to analyze the plasma extracts. 465 3. Other Nanoporous Materials Hydrophilic mesoporous phenol‐based material with added PEI, prepared by Jin et al. with a template‐free method, exhibited a large surface area (548 m 2 g −1 ) with 13 nm 466 mesopores. When the amount of PEI changed, it was possible to change both the porosities and surface area of the material. Glycopeptides could be detected with an enrichment factor (MS intensity ratio after and before enrichment) of 62 with a recovery index of 80.4%. A tryptic digest of IgG could be detected down to 5 fmol with only a moderate improvement compared to commercial hydrophilic interaction chromatography beads (40 fmol). 466 A capillary modified with a polymeric material with an introduced –SH group having a surface area of 30.3 m 2 g −1 was applied for the further selective modification of the surface with AuNPs (13 nm) by Wu et al. 467 The gold surface was further modified by two thiols terminated in boronate and amine functional groups making an intramolecular B‐N coordination for the enhanced specificity of glycoprotein capture with a binding capacity of 0.39 mg g −1 . The capillary was applied for the selective enrichment of a glycoprotein over a protein (1:1000 ratio) with a recovery of 85% and finally applied for the glycoproteomic profiling of just 9 μg of human plasma with 160 glycoproteins identified by MS. 467 A boronate‐modified affinity monolithic column with a surface area of 13.7 m 2 g −1 and with bimodal pores containing mesopores (3.9 nm) and macropores (1.4 μm) was prepared by Liu et al. using ring‐opening polymerization to mimic the function of protein A to bind antibodies. 468 In this case, mesopores could be accommodated by the glycan of IgG via the interaction of glycans with boronate, while the protein backbone of IgG could not enter the mesopores (Fig. 33 ). The binding capacity of such material for IgG binding was comparable (23.7 mg g −1 ) to a number of protein A mimics. The monolith was a quite stable and cost‐effective alternative to using protein A with the price of 40 USD per 1 g. Moreover, the IgG attached to the monolith still exhibited its binding affinity toward antigens. 468 Figure 33 A scheme showing a specific recognition of IgG by the monolith exhibiting protein A‐like binding with presence of bimodal pores (macropores and mesopores) modified by boronate functional groups. Reproduced from 468 with permission of the Royal Society of Chemistry. E. Imprinted Polymers MIP can be effectively applied for the selective capture and release of analytes because the analyte molecule is imprinted, and after the formation of a thin layer of polymer around the template (analyte), the analyte is removed from the MIP. Thus, such MIP is then able to selectively recognize the analyte over other molecules. Such an approach was applied to imprint HRP as a model glycoprotein on the inner wall of the capillary tube modified by boronate functionality as described by Lin et al. (Fig. 34 ). 469 Then, a polydopamine film was formed around HRP with the final removal of HRP. Such an MIP was able to recognize HRP with a bound amount of 4.6 mg g −1 over other glycoproteins, while the boronate‐modified capillary also effectively captured other glycoproteins. 469 Figure 34 (A) Preparation of vinylphenylboronic acid‐based molecularly imprinted monolith with polydopamine coating and (B) its recognition mechanism toward glycoproteins. In the figure the following four major steps are shown: (1) preparation of the VPBA‐based polymeric skeletons; (2) reversible immobilization of glycoprotein via boronate affinity interaction; (3) self‐polymerization of DA on the surface of the glycoprotein‐immobilized boronate affinity monolithic skeletons; (4) the formation of glycoprotein‐imprinted monolith after removal of template that is complementary in shape, size, and functionality with respect to the template. Reprinted from 469 . Copyright (2013), with permission from Elsevier. An interesting imprinting approach for the selective binding of a glycoprotein was recently applied by Mendes et al. (Fig. 35 ). 470 Gold surface was modified by an acrylamide‐alkyne cysteine derivative to make a thin film terminated in alkyne and acrylamide functionalities. Then, a glycoprotein was incubated with (3‐acryl‐amidophenyl)boronic acid when boronate functional groups interacted with a glycan part of the glycoprotein (PSA), and the other functional groups were applied for the attachment of PSA on the surface via acrylamide copolymerization. Finally, the surface was blocked by the addition of OEG‐N 3 molecules via click chemistry. Because the boronate–glycan interaction can be disrupted by low pH, a simple pH lowering was applied to remove the PSA template from the surface. Such an approach exhibited a 30‐fold higher affinity for its target compared to that of other (glyco)proteins. 470 Even though this imprinting strategy was applied for SPR sensing, there is a possibility of using this approach for selective PSA enrichment from a complex sample. Figure 35 Experimental design for the formation of surface‐restricted click‐imprinted binding sites for glycoproteins. Disulfide dimer (DFC) SAMs were prepared by immersing clean gold substrates in 0.1 mM methanolic solutions of DFC for 24 h (step 1). In step 2, BA receptor units are introduced via (3‐acrylamidophenyl)boronic acid (AM‐BA) that is incubated for 30 min at an optimized pH (8.5) with a template target glycoprotein. Multiple boronate esters are formed reversibly between the AM‐BAs and the carbohydrate structures of the glycoprotein template. The pre‐assembled glycoprotein–AM‐BA complex is then grafted on the DFC SAM via acrylamide co‐polymerization, affording the creation of spatially arranged sets of BAs on the surface that are specific for the target glycoprotein (step 3). In order to provide complimentary allosteric specificity, a mould or imprint is created around the glycoprotein template at the surface by so‐called click chemistry functionalization of the alkynes of the DFC on the SAM by reacting azide‐terminated heptaethylene glycol (Az‐OEG) moieties with the terminal alkynes on the DFC SAM via a copper‐catalysed alkyne–azide cycloaddition (CuCAAC) reaction (step 4). The glycoprotein targets are removed by washing under acidic conditions (step 5). Reproduced from 470 with permission of the Royal Society of Chemistry. In their study, Liu et al. focused on finding the relationship between the thickness of the imprinted polymer and the size of the imprinted glycoprotein regarding the binding efficiency. 471 Three glycoproteins (ribonuclease B, HRP, and GOx) were imprinted within a polydopamine film deposited over magnetic particles modified by boronate functional groups. The thickness of the polydopamine‐deposited film (5.7, 10.2, 16.3, and 25.3 nm) was controlled by polymerization time (1, 3, 6, and 15 hr). The higher the size of the glycoprotein (15–80 kDa), the thicker the film (5.7–16.3 nm) needed to be to achieve maximal binding capacity (16.7–19.5 mg g −1 ). 471 Polymers with imprinted glycans isolated from two model glycoproteins—ribonuclease B and transferrin—were prepared by Bie et al. on 100 nm boronate‐modified MNPs. 472 Glycans released from a glycoprotein were attached to boronate‐modified MNPs, and glycans were imprinted by forming a thin (2.5 nm) silane layer. Finally, the glycan template was removed, and such NPs were applied for the enrichment of glycoproteins. The imprinting ratio (i.e., the ratio of glycoprotein attached to imprinted vs. non‐imprinted NPs) was 8.4 for ribonuclease B and 21.8 for transferrin, with a maximal sorption capacity of 1.4 mg g −1 and a K D of 25 μM, suggesting only moderate glycan enrichment. The transferrin glycan‐imprinted polymer was applied for the enrichment of transferrin from a human serum sample. 472 F. Magnetic NPs (MNPs) Glycan enrichment can be effectively performed using MNPs with different glycan‐recognizing functional groups present on the surface of such NPs. Core‐shell NPs with an MNP (Fe 3 O 4 ‐based) core and silica shell having uniform 2.2 nm mesopores (formed by the removal of a surfactant from the shell layer) were prepared by Zheng et al. 473 Such NPs were further activated, and glucose was covalently linked to it by click chemistry (Fig. 36 ) to provide hydrophilic interactions. Such particles exhibited a large surface area (324 m 2 g −1 ), high ability to enrich glycopeptides (250 mg g −1 ), high sensitivity (50 fmol), short incubation time (5 min), and high recovery (94.6%) using ESI MS analysis. When 0.25 μL of human serum without enrichment was analyzed, only 12 N ‐glycans were found, while 42 N ‐glycans were identified after the application of the MNP‐based enrichment step in the ESI MS spectra. 473 Figure 36 (A) Schematic representation of the synthesis of hydrophilic MMNs and (B) the selective enrichment process for glycopeptides using hydrophilic MMNs. Reprinted from 473 . Copyright 2014, with permission from Elsevier. MNPs were modified by Cao et al. using a polymer to which hydrazide was attached, forming a 3D matrix for glycan enrichment with a threefold higher loading of hydrazide compared to single‐layer NPs. 474 The enrichment of glycopeptides was effective even in the presence of a 100‐fold excess of peptides with a recovery index of 78% and with a binding capacity for glycopeptides of 25 μg mg −1 . Glycopeptides enriched from mouse liver tissue were analyzed by MALDI‐TOF‐MS. 474 In order to enhance the hydrophilicity of the MNPs, Chen et al. covered such particles by a thin layer of SiO 2 to which zwitterionic molecules were attached by precipitative polymerization. 475 Such particles possessing a highly hydrophilic surface were effective in glycan capture, when by MALDI‐TOF‐MS, glycans from just 0.1 fmol of IgG could be detected with a high binding capacity of 100 mg g −1 and high enrichment recovery of 74% with a rapid magnetic separation. Finally, the particles were applied for the glycoprofiling of 65 μg of proteins extracted from mouse liver. 475 MNPs with co‐precipitated ethylenediaminetetraacetic acid having a size of 15 nm were modified by Dong et al. using Cu(II) with the subsequent high loading of Con A lectin (up to 28 wt%). 476 Such particles were applied for the enrichment of a model glycoprotein, namely ovalbumin, with a binding capacity of 72 mg g −1 , while a high affinity with K D of 38 nM was observed. Glycoproteins were successfully enriched, even in the presence of a non‐glycoprotein in molar excess of 600:1 with a fast magnetic separation within 15 sec. 476 MNPs with a size of 15–20 nm during synthesis formed larger aggregates with a size of 140 nm and finally were modified by chitosan using the one‐pot method as described by Fang et al. 477 Such particles were able to detect glycans from as low as 8 fmol of a tryptic digest of IgG using MALDI‐TOF‐MS with a binding capacity of 17.5 mg g −1 . Finally, 45 μL of tryptic digest from HeLa cells was successfully glycoprofiled. 477 Magnetic mesoporous (pores with 3.8 nm) particles with a surface area of 211 m 2 g −1 and with a total pore volume of 0.38 cm 3 g −1 were applied by Deng et al. for the enrichment of as low as 10 fmol of a model glycoprotein in 300‐fold excess of protein (BSA), and finally, glycans enriched from human serum were determined by MALDI‐TOF‐MS. 478 Liu et al. applied MNPs ( d = 100 nm) patterned by a PAMAM dendrimer further modified by boronate functional groups for model glycoprotein (HRP) enrichment. 479 Boronate present on a dendrimer modified MNPs had a 3–4 orders of magnitude higher affinity constant for glycoproteins compared to single boronate, while glycoprotein enrichment was possible even in a 1 million‐fold excess of competing monosaccharide with an efficiency of 42–88%. HRP could be adsorbed with density of 21 mg g −1 of support with LOD of 180 amol with extraction equilibria reached within 1 min (with desorption equilibria of 5 min). The reusability of glycoprotein enrichment was highly reproducible, with an RSD of 7.5% for five consecutive enrichment steps. Finally, the approach was also applied for the analysis of glycoproteins in human saliva. 479 Deng et al. applied MNPs (100 nm) modified by poly(styrene‐co‐vinylbenzene‐boronic acid) for glycopeptide enrichment. 480 Such particles could detect glycopeptides down to 125 fmol from a tryptic digest of HRP, even in 120‐fold excess of peptides. Glycopeptides enriched from human serum were analyzed using MALDI‐TOF‐MS. 480 Lu et al. developed an interesting strategy to increase the selectivity of glycan enrichment using two types of nanomaterials. 481 The first nanomaterials were MNPs (70 nm) modified by boronate functionalities for glycopeptide capture, and the second were poly(methyl methacrylate) beads (200 nm) for selective peptide capture. Model glycoprotein (HRP) was detected down to 220 fmol, and enrichment could be performed in 100‐fold excess of a nonglycoprotein with a maximal binding capacity for glycopeptides of 150 mg g −1 . MALDI‐TOF‐MS revealed 90% recovery using this synergetic enrichment, and as little as 1 μL of human serum was sufficient for analysis. 481 MNPs could also be successfully used for the prefractionation of glycoconjugates (mainly glycopeptides and glycoproteins). Although the traditional method is hydrazide chemistry‐based solid‐phase extraction, solid‐phase extraction through reductive amination by amine‐functionalized MNPs had also been developed, 482 shortening the extraction time to 4 hr and improving the LOD by 2 orders of magnitude. Magnetic Fe 3 O 4 NPs were functionalized by 3‐aminopropyltriethoxysilane and subsequently incubated with glycoconjugate sample, which was converted into aldehydes by sodium periodate oxidation prior to incubation. While nonspecifically adsorbed proteins could be easily washed away, glycopeptides/glycoproteins remained immobilized on the surface. Using a specific PNGase F enzyme, an excellent isolation performance and identification of glycosylation sites using nano‐LC‐MS/MS analysis were achieved. 482 The highest binding capacity toward glycoproteins was achieved by boronate‐modified magnetic particles (265 nm) using ovalbumin as a model glycoprotein of 778 483 or 882 mg g −1 using magnetic particles with size of 500 nm. 484 An elegant method to increase the low affinity of glycan binding by three different lectins was addressed by the modification of lectins with boronate linkers—called boronic acid decorated lectins (Fig. 37 )—was suggested by Lu et al. 485 When such hybrid biomaterial was immobilized on MNPs, a 2‐ to 60‐fold increase in the detection sensitivity for glycoproteins was observed due to the increased affinity from 2.2‐fold ( Aleuria aurantia lectin) to 5.6‐fold (Con A) for particular glycans. Glycoproteins could be detected with an LOD of 33 fmol using MALDI‐TOF‐MS. Finally, the enrichment step was utilized for the glycoproteomic analysis of a tryptic digest of HeLa cells. 485 Figure 37 A schematic illustration of dual binding of a BAD‐lectin (BAD = boronic acid decorated) to a glycoprotein. (i) A BAD‐lectin. (ii) A glycoprotein captured by the lectin via noncovalent glycan‐specific recognition. (iii) A glycoprotein captured by both lectin and BA; the latter mediates the formation of a boronate ester generating a stable covalent lectin‐glycoprotein complex. (iv) A glycoprotein captured by a BA ligand alone. Reprinted with permission from 485 . Copyright 2013 American Chemical Society. G. Gold NPs (AuNPs) An interesting approach to simplify glycan enrichment based on the formation of a polymeric monolith within a pipette tip was presented by Alwael et al. 486 In order to enhance the overall surface area of the monolith, 20 nm AuNPs were attached to the polymer, and the gold surface was further modified by thiols for the covalent immobilization of Erythrina cristagalli lectin (Fig. 38 ). Such a lectin‐modified monolith could selectively enrich galactosylated glycans based on the lectin´s preferential affinity with a high recovery of 95%. Finally, the device was applied in combination with reversed‐phase capillary HPLC for the analysis of E. coli lysate. 486 Figure 38 SEM images of a porous polymer monolith formed within a polypropylene pipette tip with different magnification (on left and in the middle). Field emission scanning electron microscopy images of a porous polymer monolith agglomerated with covalently attached 20 nm AuNPs with 60,000× magnification. Reproduced from 486 with permission of the Royal Society of Chemistry. A quite interesting approach to increase the sensitivity of glycopeptide/glycoprotein detection using LDI‐TOF (MALDI‐TOF‐MS, but without a need to use a matrix) was proposed by Liu et al. 487 Boronate‐modified magnetic particles with a size of 412 nm were applied to capture glycopeptides from a tryptic digestion of a model protein (HRP) or intact HRP. In the subsequent step, magnetic particles with captured glycopeptides/glycoprotein were incubated with activated AuNPs (13 nm) to covalently capture glycopeptides/glycoprotein. Unbound AuNPs were removed from the system, and glycopeptides/glycoproteins were released from boronate magnetic particles by exposure to acidic pH. Finally, AuNPs with captured glycopeptides/glycoproteins were measured using MS. AuNPs containing 64,000 Au atoms each could enhance the MS analysis of glycopeptides when applied as MS tags even without the need to use a matrix since MS could detect AuNPs with an LOD of 0.03 amol and HRP with an LOD of 45 fM. Since detection is based on the analysis of the Au 2 + ion rather than the analysis of glycopeptides/glycoproteins, the method can be applied not for the identification of glycoproteins but rather for obtaining information with high sensitivity when glycoproteins are present in a particular sample. 487 Another interesting approach for the matrix‐free analysis of various low‐molecular‐weight compounds (but not glycans) was proposed by Razunguzwa et al. using an array of silicon pillars with a diameter of 150 nm, a spacing of 337 nm, and a length of 1.2 μm, which can detect analytes down to femtomole without any enrichment. 488 Hydrazide‐functionalized ultrasmall AuNPs with a size of 1.2 nm were applied by Tran et al. for the very selective capture of periodate‐oxidized glycopeptides, when as much as 97% of all of the peptides captured from rat kidney tissue were glycopeptides. 489 This highly selective capture of glycopeptides was possible due to the extremely high density of hydrazide on AuNPs, that is, 630 nmol mg −1 , which is a 79‐fold higher density compared to hydrazide density on magnetic particles of 200–500 nm. 489 H. Silica NPs (SiNPs) Three different strategies to deposit zwitterionic brushes with a thickness of 5 nm on the surface of SiNPs (90 nm) were compared by Huang et al. using the tryptic digest of IgG as a model glycoprotein. 490 The results showed that the most efficient strategy for glycan enrichment was the application of NPs with the polymer grafted using reversible addition–fragmentation chain transfer with a low LOD of 10 fmol and a high recovery index of 88%. The approach was also successfully applied to analyze the tryptic digest of mouse liver with 303 unique glycosylation sites and with an enrichment efficiency of 70% using MALDI‐TOF‐MS, which was much higher compared to the other two approaches tested in the paper and other commercially available approaches. 490 I. Carbon NPs 1. Graphene and Graphene Oxide (GO) Graphene as a 2D crystalline material consisting of a single layer of carbon atoms with unique properties has attracted considerable attraction since its discovery in 2004. 6 Graphene, having a high surface area, could be quite effectively applied for the enrichment of glycans. Zhang et al. recently published an approach with the application of GO noncovalently modified by pyrenebutyric acid via π–π stacking interaction between graphene and a pyrene moiety. 491 In the subsequent step, the –COOH group was activated by SOCl 2 to prepare pyrenebutyryl chloride‐derivatized GO, which was applied for selective glycan binding (Fig. 39 ). A simple visual monitoring of glycan enrichment could be conducted due to GO crosslinking by attached glycans, leading to aggregation. Finally, glycans were released from the GO surface in an acidic environment by sonication at 60°C, supporting the hydrolysis of an ester bond. The enrichment protocol allowed the detection of the main glycan from just 10 ng (i.e., 0.2 pmol) of fetuin using MALDI‐TOF‐MS, and the method was applied to analyze glycans in cancer cells. Interestingly, glycoprofiling using graphene‐based enrichment method was successfully applied even when glycoprotein was highly diluted in a protein sample not having glycans at ratio of 1:100. 491 Figure 39 Graphene oxide modified by 1‐pyrenebutyric acid with subsequent activation of –COOH group by SOCl 2 for effective glycan enrichment. Reprinted with permission from 491 . Copyright 2013 American Chemical Society. GO chemically modified with pyrenebutyric acid hydrazide was applied by Bai et al. for glycan enrichment. 492 In addition to the model protein fetuin, human plasma was applied in the study with a final MALDI‐TOF‐MS analysis of the captured glycans. Glycan (maltoheptaose) recovery in a model sample spiked with BSA was 100%, but human serum spiked with maltoheptaose showed a recovery index of 73%. 492 2. Nanodiamonds Boronate‐modified nanodiamonds could also be applied for glycan enrichment, as shown by Xu et al. 493 In this particular case a tryptic digest of HRP as a model glycoprotein was performed, and 50‐fold enhanced MALDI‐MS spectra were obtained after glycan enrichment compared to a tryptic HRP digest without any enrichment, with HRP being detectable at an LOD of 0.5 nM in 100 μL (i.e., 0.5 fmol). The recovery index for glycopeptide analysis was 72%, and the purification efficiency for an analysis of glycopeptides in the mouse liver fraction was 69% with 24 newly identified glycosylation sites compared to databases. 493 Loh et al. clearly showed that the modification of nanodiamonds by boronate functionalities must be performed to have a linker between the surface of nanodiamonds and boronate functional groups in order to prevent the nonspecific binding of proteins to the hydrophobic nanodiamond surface, which would negatively influence enrichment specificity. 494 Nanodiamond particles (DNPs) were effectively applied by Wu et al. to enhance the sensitivity of glycan analysis by MALDI‐TOF‐MS. 495 In this case DNPs were used to transfer energy from matrix to glycans within a trilayer (i.e., by making a sandwich: matrix‐DNPs‐sample). The most important functions of DNPs were to mediate heat transfer between matrix‐ and glycan‐containing samples, avoiding direct heat exchange between these two components, and to prepare the highly homogeneous morphology of the spot on the MALDI target (Fig. 40 ) compared to two other methods not involving DNPs. Such particles have a low extinction coefficient not absorbing laser energy in the near‐UV region, and since DNPs are inert, they do not compete for charges with the analyte. When a model analyte dextran was analyzed by MALDI‐TOF, a 79‐ or 7‐fold increase in the detection sensitivity was achieved compared to dried‐droplet (a mixture of a matrix and a sample) or thin‐layer (layer of a sample deposited over a matrix layer) methods, respectively. The size of such particles (50–500 nm) did not influence sensitivity enhancement. 495 Figure 40 Morphology of trilayer and other samples. (A) Schematic of the configuration of matrix, DNP, and analyte in trilayer samples, (B) image of a dried‐droplet sample, (C) image of a thin‐layer sample, and (D) image of a trilayer sample containing 3 μg DNPs. The scale bars represent 1 mm. Reprinted with permission from 495 . Copyright 2013 American Chemical Society. 3. Single‐Walled CNTs Strano´s group modified SWCNTs with derivatives of PBAs containing –COOH, ‐NO 2 , and –NH 2 functional groups with ortho‐ , meta‐ , and para‐ substitutions generating 144 distinct corona phases on the surface of SWCNTs and some of them exhibited remarkable binding affinity to certain monosaccharides, while others were not bound. 496 , 497 Even though the method was applied for sensing purposes (i.e., monitoring of quenching of intrinsic fluorescence of SWCNTs 496 , 497 ), this can be applied for the enrichment of glycans. 498 1. Graphene and Graphene Oxide (GO) Graphene as a 2D crystalline material consisting of a single layer of carbon atoms with unique properties has attracted considerable attraction since its discovery in 2004. 6 Graphene, having a high surface area, could be quite effectively applied for the enrichment of glycans. Zhang et al. recently published an approach with the application of GO noncovalently modified by pyrenebutyric acid via π–π stacking interaction between graphene and a pyrene moiety. 491 In the subsequent step, the –COOH group was activated by SOCl 2 to prepare pyrenebutyryl chloride‐derivatized GO, which was applied for selective glycan binding (Fig. 39 ). A simple visual monitoring of glycan enrichment could be conducted due to GO crosslinking by attached glycans, leading to aggregation. Finally, glycans were released from the GO surface in an acidic environment by sonication at 60°C, supporting the hydrolysis of an ester bond. The enrichment protocol allowed the detection of the main glycan from just 10 ng (i.e., 0.2 pmol) of fetuin using MALDI‐TOF‐MS, and the method was applied to analyze glycans in cancer cells. Interestingly, glycoprofiling using graphene‐based enrichment method was successfully applied even when glycoprotein was highly diluted in a protein sample not having glycans at ratio of 1:100. 491 Figure 39 Graphene oxide modified by 1‐pyrenebutyric acid with subsequent activation of –COOH group by SOCl 2 for effective glycan enrichment. Reprinted with permission from 491 . Copyright 2013 American Chemical Society. GO chemically modified with pyrenebutyric acid hydrazide was applied by Bai et al. for glycan enrichment. 492 In addition to the model protein fetuin, human plasma was applied in the study with a final MALDI‐TOF‐MS analysis of the captured glycans. Glycan (maltoheptaose) recovery in a model sample spiked with BSA was 100%, but human serum spiked with maltoheptaose showed a recovery index of 73%. 492 2. Nanodiamonds Boronate‐modified nanodiamonds could also be applied for glycan enrichment, as shown by Xu et al. 493 In this particular case a tryptic digest of HRP as a model glycoprotein was performed, and 50‐fold enhanced MALDI‐MS spectra were obtained after glycan enrichment compared to a tryptic HRP digest without any enrichment, with HRP being detectable at an LOD of 0.5 nM in 100 μL (i.e., 0.5 fmol). The recovery index for glycopeptide analysis was 72%, and the purification efficiency for an analysis of glycopeptides in the mouse liver fraction was 69% with 24 newly identified glycosylation sites compared to databases. 493 Loh et al. clearly showed that the modification of nanodiamonds by boronate functionalities must be performed to have a linker between the surface of nanodiamonds and boronate functional groups in order to prevent the nonspecific binding of proteins to the hydrophobic nanodiamond surface, which would negatively influence enrichment specificity. 494 Nanodiamond particles (DNPs) were effectively applied by Wu et al. to enhance the sensitivity of glycan analysis by MALDI‐TOF‐MS. 495 In this case DNPs were used to transfer energy from matrix to glycans within a trilayer (i.e., by making a sandwich: matrix‐DNPs‐sample). The most important functions of DNPs were to mediate heat transfer between matrix‐ and glycan‐containing samples, avoiding direct heat exchange between these two components, and to prepare the highly homogeneous morphology of the spot on the MALDI target (Fig. 40 ) compared to two other methods not involving DNPs. Such particles have a low extinction coefficient not absorbing laser energy in the near‐UV region, and since DNPs are inert, they do not compete for charges with the analyte. When a model analyte dextran was analyzed by MALDI‐TOF, a 79‐ or 7‐fold increase in the detection sensitivity was achieved compared to dried‐droplet (a mixture of a matrix and a sample) or thin‐layer (layer of a sample deposited over a matrix layer) methods, respectively. The size of such particles (50–500 nm) did not influence sensitivity enhancement. 495 Figure 40 Morphology of trilayer and other samples. (A) Schematic of the configuration of matrix, DNP, and analyte in trilayer samples, (B) image of a dried‐droplet sample, (C) image of a thin‐layer sample, and (D) image of a trilayer sample containing 3 μg DNPs. The scale bars represent 1 mm. Reprinted with permission from 495 . Copyright 2013 American Chemical Society. 3. Single‐Walled CNTs Strano´s group modified SWCNTs with derivatives of PBAs containing –COOH, ‐NO 2 , and –NH 2 functional groups with ortho‐ , meta‐ , and para‐ substitutions generating 144 distinct corona phases on the surface of SWCNTs and some of them exhibited remarkable binding affinity to certain monosaccharides, while others were not bound. 496 , 497 Even though the method was applied for sensing purposes (i.e., monitoring of quenching of intrinsic fluorescence of SWCNTs 496 , 497 ), this can be applied for the enrichment of glycans. 498 J. Hybrid NPs A polydopamine film was deposited by Bi et al. using self‐polymerization on an rGO‐Fe 3 O 4 ‐modified surface and applied for the subsequent deposition of AuNPs. 499 Thiolated mannose finally formed SAM on AuNPs, and such a nanocomposite was employed for glycan enrichment. HRP as a model glycoprotein could be detected by MALDI‐TOF from a tryptic digest of the protein down to a concentration of 0.1 ng uL −1 (i.e., 40 ng or 1 pmol). 499 Hu et al . modified silica bubbles (30 um) with AuNPs (20 nm), which were further functionalized with boronate‐terminated thiol. 500 Glycopeptides prepared by a tryptic digest from two glycoproteins (HRP and IgG) could be enriched with a binding capacity of 60 mg g −1 , and as much as 10 ng (∼200 fmol) of protein was needed for glycoprofiling with a rather low enhancement (approximately tenfold) of MALDI‐TOF signal compared to the signal obtained without enrichment. 500 Ju et al. applied hybrid NPs (magnetic CNTs) prepared on CNTs with d = 40–60 nm as a scaffold by the in situ formation of MNPs ( d = 10–15 nm) from Fe 3+ ions. 501 Finally, hybrid NPs were patterned by the boronate functional group. Model glycoprotein HRP could be enriched with a sorption capacity of 346 mg g −1 , while the sorption of nonglycoprotein HSA was quite low (52 mg g −1 ). The LOD for the detection of HRP by MALDI‐TOF‐MS using an enrichment step was 1 pmol, while with commercially available boronate‐modified agarose gel, it was not possible to detect 10 pmol of the glycoprotein, and hybrid NPs could enrich glycopeptides in the presence of nonglycosylated peptides, while the latter being in 50‐fold excess. 501 Zou et al. applied a quite sophisticated strategy for glycan enrichment. 502 GO with deposited MNPs were covered by a silica shell of a final thickness of 50–60 nm. In the next step, the PAMAM dendrimer was grafted to the surface with subsequent modification by AuNPs on which thiolated maltose was anchored. Such a maltose‐modified hydrophilic composite with a surface area of 57.8 m 2 g −1 and with a maltose density of 2.4 μmol m −2 could enrich glycopeptides with an LOD of 0.5 fmol. This approach was finally applied for the analysis of as low as 50 μg of mouse liver tryptic digest. 502 For comparison, a commercially available meta‐aminophenylboronic acid modified agarose could enrich glycopeptides with an LOD of 30 nmol. 503 Yang et al. prepared a nanocomposite with graphene as a support to accommodate MNPs (100 nm) and a phenolic‐formaldehyde resin (condensation of formaldehyde and hydroquinone) with a final modification by aminophenylboronic acid for glycopeptide enrichment. 504 The graphene‐based composite with a thickness of 10 nm and a surface area of 76.3 m 2 g −1 could detect glycopeptides down to 1 fmol using MALDI‐TOF‐MS, even when the concentration of peptides was in excess of 100:1 compared to glycopeptides. The composite was applied to analyze human serum, when as little as 1 μL of a sample was sufficient. 504 K. Other Interesting Approaches Herein are described interesting strategies of effective glycan enrichment not necessarily based on the application of NPs. The approach developed by Jiao et al. is based on the use of hydrazinonicotinic acid as a matrix for the MALDI‐TOF‐MS analysis of glycans, where the matrix, besides adsorbing laser energy, also contains a hydrazine moiety for selective interaction with glycans. 505 Thus, no glycan enrichment is needed since laser energy is mainly adsorbed by glycans even in the presence of proteins/peptides. The detection limit for a glycan is down to 1 amol, which is 5 orders of magnitude lower amount compared to 2,5‐dihydroxybenzoic acid used as a matrix. The other advantages of using this novel matrix are the higher homogeneity of glycan spots and better salt tolerance compared to the traditional matrix. The approach was finally applied for the analysis of a human serum. 505 The online enzyme digestion of glycoproteins within a microbore hollow fiber reactor (ID = 200 um) was suggested by Kim et al. using trypsin to digest glycoproteins into peptides and glycopeptides within 30 min, which were then isolated from peptides using lectins. 506 Finally, glycans were released from glycopeptides by the application of PNGase F, and the method was applied for the analysis of human urinary samples (PC patients). 506 An improved assay protocol from the same group was applied for the analysis of sera from patients with liver cancer. 507 Time needed for tryptic digestion was comparable to that needed for the microwave‐assisted release of glycans in the presence of trypsin (10 min at 50°C). 464 Direct selective glycan enrichment was performed by Li et al. using a hydrophobic fluorinated carbon tag with an –NH 2 terminal group for specific coupling to the reducing end of glycans. 508 Such modified glycans could be ionized more efficiently by one order of magnitude compared to unmodified glycans using MALDI‐TOF‐MS. Alternatively, hydrophobized glycans could be selectively isolated from a mixture with other biomolecules using fluorous solid extraction. 508 Liu et al. developed a matrix‐free strategy for MS using a lithium‐rich metal oxide composite with a particle size of 200–300 nm applied to analyze low‐molecular‐weight analytes, including several oligosaccharides (not glycans). 509 Ruman et al. applied AgNPs with a size of 100 nm for the matrix‐free MS of various low‐molecular‐weight analytes, but not glycans, with a remarkable sensitivity of detection down to 33 amol for ribose. 510 Moreover, the application of AgNPs as a matrix in the MS analysis of low‐molecular‐weight molecules was recently reviewed by Sekula et al. 511 Yang et al. developed an on‐plate glycopeptide enrichment procedure using gold‐coated silicon wafer modified by SAM with a final modification by boronate functional groups taking 24 hr to complete. 512 This approach could effectively pre‐concentrate glycopeptides released from three model glycoproteins with an LOD of 1 fmol, which increased 93‐ to 248‐fold compared to a procedure without enrichment. The plate could, however, be reused only three times because its capability to enrich glycopeptides then decreased sharply. 512 Lu et al. deposited 900 μm gold spots on a hydrophobic silica wafer and the gold layer was patterned by 4‐mercaptophenylboronic acid for specific glycoproteins/glycopeptides enrichment. 513 The LOD for glycopeptides released from a model glycoprotein HRP was 230 fmol, which is 1 order of magnitude lower compared to MALDI‐TOF‐MS using a standard stainless steel plate. 513 Magnetic particles with d = 220 nm were used in the study performed by Xiong et al. 514 The particles were covered by silane terminated in the –NH 2 group, thus forming a positively charged surface. A layer‐by‐layer method was then applied to cover modified particles by the alternate deposition of negatively charged HA and positively charged chitosan. The best glycan enrichment strategy was obtained with particles having ten layers of both polysaccharides. Glycan enrichment with a binding capacity of 200 mg g −1 was achieved by hydrophilic interaction with glycans and polysaccharide‐modified MNPs. Three glycopeptides from a tryptic digest of IgG could be detected from just 0.2 fmol of IgG with a recovery as high as 69%. The most abundant peak in MALDI‐TOF‐MS was enhanced 111‐fold compared to the procedure without any glycan enrichment. Finally, the approach was applied for analysis of the glycoproteome from just 20 μg of a mouse liver protein sample with the identification of 605 unique N ‐glycosylation sites in 616 distinct glycopeptides. 514 Another approach using particles larger than 100 nm described by Ma et al. is worth mentioning. 515 This ligand‐free approach is based on the application of hybrid particles having a magnetic core with 22 nm AgNPs deposited on the surface. Selective glycan enrichment by such hybrid particles was achieved by the reversible interaction between glycans and the Ag surface, and glycopeptides could be effectively separated from a mixture having a molar ratio of glycopeptides to nonglycopeptides of 1:100 within 1 min. Finally, the method was applied for MALDI‐TOF‐MS glycan analysis from just 1 μL of rat serum. 515 A homogeneous system for glycan enrichment based on the pH‐responsive polymer poly‐(acrylic acid‐co‐methyl acrylate) was proven by Bai et al. to be much more effective (96.2% of glycoproteins captured within 1 hr) compared to a solid‐phase glycan enrichment (90% captured in 8 hr). 516 The polymer with a hydrodynamic size of 30 nm was soluble at pH 6.0 but became insoluble at pH 2.0. When the hydrazide‐modified polymer was incubated with oxidized glycopeptides/glycoproteins, their separation could be performed by switching the pH to 2.0 with subsequent centrifugation. From a mixture of glycoproteins with proteins at a ratio of 1:100, low levels of glycoproteins (1 fmol) can be effectively captured. Using polymer‐based glycan enrichment MALDI‐TOF‐MS, the signal intensity increased 29‐fold and the S/N ratio increased 325‐fold. The glycan enrichment strategy was finally applied for the analysis of mouse brain samples. 516 Zhang et al. applied a hydrophilic support‐based amino‐functionalized metal‐organic framework containing 25 nm NPs with a specific surface are of 2187 m 2 g −1 for glycopeptide enrichment. 517 A tryptic digest of IgG as a model glycoprotein could be detected after enrichment using MALDI‐TOF‐MS with an LOD of 20 fmol, and an analysis of human serum was performed from a volume as small as 10 μL. 517 8. ON‐PLATE ANALYSIS OF GLYCOSYLTRANSFERASE ACTIVITIES The identification of novel glycosyltransferases is important for the effective production of glycans since enzymatic production does not require tedious protection/deprotection steps, and even complex glycans can be effectively produced either enzymatically or by using a chemoenzymatic approach. 151 Thus, new ways for effective and label‐free screening approaches are vital in this field. One of the most promising approaches besides offering the label‐free characterization of enzymatic activities also providing a high‐throughput analysis is to perform enzymatic reactions directly on the MALDI targets as pioneered by Mrksich´s group 518 , 519 and later extended also by Flitsch´s 520 , 521 and Reichardt´s 522 groups. The most promising material for such targets is gold because SAMs can form on gold through a modification of thiols. When two thiols are mixed during the formation of SAM, a mixed SAM can be produced with an adjusted density of functional groups on the surface. 151 , 523 Mrksich´s group applied this strategy to screen 57,120 reactions (14,280 combinations of enzymes, immobilized acceptor substrates, and donor substrates in four different buffer systems) on a plate having 384 spots with 24 oligosaccharide acceptors. 524 In the assays, 85 glycosyltransferases (including 76 bacterial enzymes that had not been previously characterized) in their crude form produced using an in vitro expression system were tested. The results using SAMs for MALDI‐MS showed 44 new glycosylation products, the enzymatic activities of four previously uncharacterized enzymes and the unknown donor preference for two known galactosyltransferases. These assays are quick, with the screening of 24 disaccharides taking 4 hr. Although various reaction conditions were tested (pH, type of a buffer, concentration of donors, time of analysis, presence of various divalent metal ions), the density of functional thiol (thiolated glycan, maleimide terminated thiols or azide terminated thiols) within a mixed SAM did not vary, with a ratio between functional and diluting thiols (OEG thiol) fixed at 1:4. The only drawback of this approach was a need to use NMR to identify the type of glycosidic linkage formed. 524 This drawback was recently addressed by Both et al. using travelling wave ion mobility spectrometry MS to identify diastereomeric glycoconjugates (Fig. 41 ) with several asymmetric centers differing only in their configuration at a single position. 525 Moreover, GlcNAc and GalNAc oxonium ions generated by collision‐induced dissociation could be identified by ion mobility with an application for screening native glycopeptides built from different epimeric carbohydrates. 525 The most recent study from Seeberger´s group showed that IM‐MS can unambiguously identify carbohydrate stereoisomers and linkage‐isomers even in a mixture and that the minor isomer could be detected in 1000‐fold excess of the other isomer. 526 Kolarich´s group recently focused on the optimization of collision‐induced dissociation energy parameters for improved fragmentation of both glycans and peptides. 527 Figure 41 Examples of common epimeric glycoconjugates and families of enzymes involved in their biosynthesis. (a) The pp‐α‐GnT family of enzymes mediate the transfer of an α ‐linked GalNAc residue to the hydroxyl groups of serine and threonine residues in proteins. This modification is common in higher eukaryotes (including humans) and represents the first step in the biosynthesis of mucin‐type O ‐glycans. (b) In unicellular eukaryotes, mucin‐type O ‐glycans have been found initiated with an α‐linked GlcNAc residue generated by a pp‐a‐GnT family of enzymes. (c) β‐linked GlcNAc is a common post‐translational modification (PTM) observed in many organisms and is especially prevalent among multicellular eukaryotes. The attachment of this residue is mediated by O ‐GlcNAc transferase (OGT) β‐linked GlcNAc is also transferred to the notch epidermal growth factor repeats by extracellular OGT. The differentiation of epimeric glycopeptides by mass spectrometry has not been reported previously, (d) Travelling wave ion mobility spectrometry mass spectrometry (TWIMS) arrival time distribution showing the discrimination of epimeric glycopeptides 19 and 20 and the distinction of HexNAc oxonium ions generated following collision induced dissociation (CID). CID of glycopeptides 19 and 20 before ion mobility separation results in the formation of epimeric oxonium ions that are distinguishable by TWIMS‐MS. Vertical dashed lines represent drift time as identified by full width half height. Reprinted by permission from Macmillan Publishers Ltd: Nature Chemistry, 525 copyright 2014. A very interesting approach for the measurement of enzyme activities using a nanostructured MALDI surface when the matrix is not needed for laser desorption/ionization was proposed by Northen et al. 528 The nanostructured surface was prepared as described earlier by the same group with 10 nm nanostructured pores. 529 The enzyme substrate for glycan processing enzymes (i.e., glycosyltransferase or glycosidase) with a fluorous tag was attached to a nanohole covered by fluorous‐containing silane molecules via a soft immobilization process allowing the efficient desorption/ionization by a laser. 528 Moreover, the substrate contained easily ionizable arginine and a five‐carbon linker to avoid steric hindrance for the binding. Such an interface offered an extremely high S/N ratio of 20 for 200 amol of a substrate, and the method could detect 500 fg of the enzyme (∼5 amol). The method was applied to detect β‐1,4‐galactosidase activity in the lysate of E. coli and thermophilic microbial community and to analyze enzymatic inhibitors, as well. 528 Recently, such an approach was extended by the same group using oxime derivatization for the rapid kinetic characterization of glycosyl hydrolases from Clostridium thermocellum , with a possibility to identify novel hydrolases applicable for the effective production of biofuels from plant biomass. 530 Reichert´s group then built upon Wong´s approach using a hydrophobic SAM deposited on a gold surface, which was then applied for the reversible insertion/extraction of glycans with a lipid tag (Fig. 42 ). 522 Such an approach was successfully applied for the analysis of activities of glycosyltransferases and glycosidases using MALDI‐TOF‐MS detection. Although the action of glycosyltransferases could be detected after 10 min of incubation, 60 hr was needed for the full conversion. For the analysis of the activity of glycosidases, 1.5 hr of incubation resulted in complete hydrolysis. In addition, the binding of lectins to glycan‐modified surfaces was performed. 522 Later, the same group applied a slightly modified immobilization strategy (i.e., modification of the indium tin oxide [ITO] surface by hydrophobic silane, incubation with hydrophobic hydrocarbon terminated in active ester) for the covalent immobilization of either amine‐terminated glycans or lectins. 531 Such a surface with arrayed ligands was then utilized to (i) study the enzyme activity of eight recombinant glycosyltransferases, (ii) determine the specificity of a fucosyltransferase, (iii) profile glycoproteins bound to immobilized lectins, and (iv) identify lectins bound to immobilized glycans with an on‐chip tryptic digestion and in situ peptide sequencing. A total of 54 glycan spots could be produced in 3 min, and 30 fmol of glycan could be detected by MS. 531 Lipid and fluorous tags, if long enough, could resist their removal upon extensive washing. 55 This approach for glycan immobilization resembles the insertion of glycans/glycoproteins within a cell membrane, offering free movement within a hydrophobic SAM. 104 , 532 Reichardt´s group also proved that ITO with controlled grain size (50–100 nm), roughness (1 nm), and film thickness (20–40 nm) affected by a sputtering could be applied for surface‐assisted laser desorption/ionization MS without the need to use a matrix, 533 and such an approach has the potential to be applied in glycan enrichment due to the hydrophilic nature of the unmodified ITO surface. Another matrix‐free approach for laser desorption/ionization MS from the same group involved the use of nanostructured weathered steel (i.e., nanostructured rust layer formed on a polished steel plate with a roughness of 0.8 μm by its incubation at 37°C with 90% humidity for 2 months) hydrophobized by octadecyl trichlorosilane for sensitive glycan detection down to 5 fmol. 534 Six different nanomaterials (AuNPs, Pt nanosponges, MNPs, TiO 2 NPs, Se NPs, CdTe QDs) as a matrix were recently applied by Wang et al. in surface‐assisted laser desorption/ionization MS of peptides/proteins, with some of them being detected with an LOD down to 6 fmol, but glycans (glycopeptides) were not investigated in the study. 535 Figure 42 Formation of glycan‐functionalized surfaces for MALDI‐TOF‐MS analysis. Reproduced with permission from 522 by John Wiley & Sons. 9. MICROENGINE/MICROROCKET‐BASED ACTIVE GLYCOPROFILING Nanotechnology has helped to develop novel devices that can be applied for the active glycoprofiling of various samples. Such devices are propelled by different means and, when modified by glycan binding agents (boronate functional groups or lectins), can selectively pick, transport, and release a glycan cargo (or cells) on demand. A polycarbonate membrane with 2 μm pores was applied by Joseph Wang et al. 536 as a template for the preparation of microengines of tubular shape with a length of 6 μm and a diameter of 2 μm. The microengine was constructed using a sequential deposition of polyaniline and platinum within the pores of the membrane. Finally, the device was patterned by a thin adhesive layer of Ti (10 nm), 26 nm of Ni (a magnetic layer), and 12 nm of Au film (for the further patterning of the device). The Au surface was patterned by a mixed SAM composed of 11‐mercaptoundecanoic acid and 6‐mercaptohexanol for the covalent attachment of lectins (Con A and Ulex Europaeus agglutinin). The device could then be controlled magnetically and was propelled by the formation of oxygen bubbles generated by the catalytic decomposition of H 2 O 2 (a fuel) by platinum (Fig. 43 ). The device selectively picked up E. coli cells even in the presence of yeast cells and transported them, and cargo release was triggered by a change of pH. The device could travel at a speed of 33 μm s −1 , which is sufficient to remove bacterial cells fixed on a glass slide. 536 In the next report from the same group, the overall concept of active glycoprofiling by a microengine/microrocket 10 μm long with ID = 800 nm and OD = 1000 nm was simplified. 537 The device was in this case constructed from polyaminophenylboronic acid applied instead of polyaniline, which was used to selectively capture glycans. The body of the device consisted of Pt and Pt‐Ni layers within microtubules, which simplified the overall design. The device was applied for the selective capture of yeast cells via a glycan‐boronate interaction, and the cargo was released by the addition of a carbohydrate and could travel at a speed of 20 μm s −1 using H 2 O 2 as a fuel. 537 Figure 43 Lectin‐modified microengines for bacteria isolation. Schemes depicting (A) the selective pick‐up, transport, and release of the target bacteria by a Con A‐modified microengine, and (B) surface chemistry involved on the microengines functionalization with the lectin receptor. Upon encountering the cells, the Con A‐functionalized microengines recognize the E. coli cell walls by O ‐antigen structure binding‐allowing for selective pick‐up and transport. Inset (in Scheme A, top left side): a SEM image of a portion of a Con A‐modified microengine loaded with an E. coli cell. Scheme A, right side: Release of the capture bacteria by navigation in a 10 mM glycine solution, pH 2.5. Scheme B: Steps involved in the microengines gold surface functionalization. (1) Self‐assembling of 11‐mercaptoundecanoic acid/6‐mercaptohexanol binary monolayer; (2) activation of the carboxylic terminal groups of the MUA to amine‐reactive esters by the EDC and NHS coupling agents; and (3) reaction of NHS ester groups with the primary amines of the Con A to yield stable amide bonds. Reprinted with permission from 536 . Copyright 2012 American Chemical Society. Since the application of H 2 O 2 as a fuel can negatively impact biomolecules/cells, other options for propulsion have been sought. Ultrasound was employed by the same group to propel a device 250 nm in diameter and 1.8 μm in length (a three‐segment Au/Ni/Au device). 538 Most likely, the combination of the small size of the device with an optimized design and ultrasound propulsion resulted in the high speed of 255 μm s −1 at which the device could travel while it was guided magnetically. A Con A‐modified device was applied for the selective capture of E. coli cells, and when antibodies were immobilized on the device, other bacterial cells could be selectively captured. 538 The highest speed of 6 m s −1 for such micromotors was achieved by Joseph Wang´s group using acoustic droplet vaporization, and such microengines have strength to penetrate tissues. 539 Furthermore, the surface roughness of the devices propelled by ultrasound can be tuned by the deposition of an Au‐Ag alloy during device fabrication with a subsequent dealloying of the silver component forming pores of 4–8 nm for the enhanced sorption or loading of biorecognition elements. 540 The application of activated carbon for the production of highly sorptive spherical micromotors was described. 340 In addition to bubble‐based and ultrasound propulsion, other ways to propel microengines are available, including electrical, magnetic, and self‐electrophoretic propulsion, 541 but other propulsion principles, including light‐driven, enzyme‐catalysed reaction‐driven, and via intact motile cells, are also possible. 542 Microengines could be made smaller for the construction of tubular devices for membrane templated synthesis with pores as small as 5 nm, 543 and spherical nanomotors less than 100 nm in size have already been prepared. 544 Such small micromotors have been successful in blood/serum 545 , 546 and even inside living cells, 542 , 547 , 548 with the potential to work as miniature mixers, 549 , 550 and such devices could also be used for active glycoprofiling/enrichment in blood/serum or other clinically relevant samples. Moreover, in order to avoid an immune response toward artificial micromotors, red blood cells were recently turned into functional micromotors with the ability to operate in blood/serum. 551 , 552 The most recent study suggests that microcannons acoustically firing cargoes with an average speed of 1.05 ± 0.26 m s −1 could penetrate tissues with higher effectivity than microengines. 553 10. CONCLUSIONS Since quite a few different fields are covered in this review, conclusions and perspectives will be addressed for every field covered separately in the following sections. A. Nanoglycosensing The section dedicated to nanomaterial‐based bioanalytical methods and biosensors described in detail some of the most commonly used nanomaterials in such applications, including metal NPs, carbon nanomaterials, polymer nanostructures, and QDs. Great attention was paid to the combination of two or more (nano)materials, called nanohybrids and nanocomposites. Although some properties of the nanomaterials are of great importance in the field of nanobiotechnology and nanobiosensing (high conductivity, large surface area, and unique shape allowing multivalent binding, optical properties, etc.), 554 , 555 few real applications (i.e., analysis of real samples) have been described in the current literature. 556 Most of the presented papers dealt only with commercially available substances or model cell lines, although there is a growing number of papers trying to dynamically evaluate the number of membrane‐bound glycans on living cells for clinical diagnostics. 557 When used in a real clinical laboratory, such cells have to be isolated from patient tissues (taken during biopsy); therefore, less invasive methods to isolate the samples are welcome. The use of human sera or even whole blood is limited, mostly because of a small level of a particular biomarker present in body fluids. 558 , 559 , 560 , 561 Using a combination of immobilized antibodies (able to specifically bind a biomarker from the real sample) together with a lectin able to "decipher" the biomarker´s glycocode might be a solution in some cases. 562 B. Carbohydrate‐Based Vaccines and Therapeutics To summarize the therapeutic applications of carbohydrates integrated with nanomaterials or carbohydrates synthetized with a controlled density on various supports at the nanoscale, three main application directions can be distinguished: (i) employment of carbohydrate‐containing immunogens, especially viral and bacterial glycans responsible for cell recognition and consequent immune response, and tumor‐associated antigens 188 ; (ii) development of therapeutics based on the selective blocking of cell receptors via competitive lectin‐carbohydrate binding 563 ; and (iii) employment of carbohydrates as building blocks to develop NPs with therapeutic effects of molecules other than carbohydrate‐based moieties. The most recent progress in the development of carbohydrate nano‐vaccines is represented by the development of a method of chemoenzymatic synthesis allowing the tailored conjugation of glycans to carrier proteins leading to the increased immunoefficiency of the prepared NPs. 188 Another promising achievement was to conjugate tumor‐associated antigens to NPs even though non‐carbohydrate antigens seem to provide more significant results in this field. 564 , 565 The therapeutic effect outlined in (ii) has recently led to the development of NP‐glycan conjugates with higher affinity to specific receptors than bacterial or viral glycans. In the majority of studies, such NPs were employed as antibacterial and antiviral drugs because of their ability to selectively inhibit the first phase of infection. Recent studies have suggested that a combination of multivalent glycans and the selectivity of lectin‐glycan binding could be a step further in the development of completely new class of antibiotics. 566 It should be noted that, regardless of studies employing metallic and carbon NPs, (bio)polymers and synthetic micelles are the most preferable carriers for the same reasons as in the case of glycan nano‐vaccines. C. Cell Targeting While the nanomaterials used for delivery of carbohydrate‐based therapeutic agents (antigens or non‐immunogens) were mostly AuNPs, (bio)polymers, and dendrimers, there was much higher diversity of nanomaterials tested or developed as nanocarriers for drug delivery of drugs other than glycans, targeted by conjugated glycan moieties. Very intensive research is currently underway to explore targeting abilities, biocompatibility, and other advantages of HA, which has been used in many studies as an NP‐coating, securing selective internalization by tumor cells via their surface displayed CD44 receptors. 567 In a typical approach, HA has been used for the decoration of gold, 568 , 569 carbon, 288 or silica 310 NPs or synthetic NPs 570 conjugated with doxorubicin (a model cytostatic drug). Many studies have reported the inhibition of tumor growth after treatment by such NPs, but not at a significant rate. Nevertheless, efficient targeting is a crucial step in avoiding the severe adverse side effects of conventional cancer chemotherapy. In addition to glycan‐driven targeting, recent studies have also focused on the selective release of the delivered drug after internalization. Here, again, HA is an intriguing material because it could be degraded by intracellular hyaluronidases, liberating the drug from HA‐based NPs. 571 For many nanocarriers, the final drug release was secured by a significantly lower pH in lysosomes, which the NP could enter only after surface receptor‐driven internalization. 310 , 568 Another recently studied way to keep the nanocarriers with their cargo intact until they reach the target is the introduction of disulfide bonds into the NP structure. This bond could be readily reduced by glutathione, which is typically present at an elevated concentration in the cytoplasm of cancer cells. 310 Interesting results have also been achieved by the development of nanocarriers capable of passing through the blood brain barrier, for example, by the activation of a glucose transporter or by the conjugation of NPs with glycoproteins or lectins. This is important for the selective chemotherapy of brain tumors. In addition to drug delivery, carbohydrate‐targeted NPs were also promising as nonviral vectors, which has opened new possibilities in gene therapy for cancer. In addition to chemo‐ and geno‐therapies, targeting delivery has also been recently used for so‐called photodynamic and photothermal therapies. While in the former case, photosensitizers are delivered into cells that generate cytotoxic ROS, 375 , 571 , 572 the latter method relies on the delivery of NPs causing local overheating by irradiation after their internalization into selected cells. 573 In both cases, the desired action is typically triggered by NIR irradiation, which is harmless to healthy tissues without internalized photosensitizers. Another advantage of these methods is that it is not necessary to release the cargo from the carriers; the only aim is to get into the desired cells as many nanocarrier particles as possible, while a minimal amount should be in healthy tissue, as indicated by recent progress in carbohydrate‐based targeting. D. Cell Imaging Upon the replacement of the therapeutic agents discussed in the previous section with imaging probes, efficient diagnostic tools can be obtained not only for the visualization of tissue with an unusual glycopattern (most often tumors 390 or inflammation sites 574 ) but also for the very precise detection and assessment of, for example, surface cell receptors, lectin‐binding properties, and other features necessary for the further development of a glycan‐based targeted therapy. While such assessments are typically performed in vitro, NIR or UV‐VIS imaging probes can be used. Organic dyes coupled to glycan‐based nanocarriers are typically used for imaging in this range, and recently, so‐called theranostic NPs have been developed where imaging probes together with therapeutic agents are incorporated (see for example 575 ); thus, drug delivery can be monitored and confirmed. It should be noted that some organic therapeutic agents could also simultaneously function as imaging probes (e.g., doxorubicin or most photosensitizers), making theranostic NPs less expensive and easier to fabricate. For efficient in vivo imaging, CT and MRI have been routinely used for many years. Recent studies have focused on the targeted delivery of contrast enhancers, allowing the efficient visualization of selected tissues. While selectively distributed AuNPs have typically been used for CT imaging, 400 iron oxide containing NPs have been tested extensively as MRI imaging probes. 576 E. Glycan Enrichment and Separation It can be concluded that the application of nanomaterials or mesoporous materials with various functionalities present on the surface can significantly decrease the time needed for glycan enrichment (i.e., from a typical time of 160 577 to 1 515 min). Moreover, when using hybrid MNPs, the separation time can be as short as 15 sec by the simple application of an external magnet, 476 thus rendering the use of additional equipment (centrifuge) with a much longer separation time (few minutes) unnecessary. Nanomaterial‐based glycan enrichment could be applied for a minute amount of a starting material. Glycans can be enriched with commercially available hydrophilic beads with an LOD of 40 fmol, 466 and the most sensitive enrichment procedure involving MNPs with thin SiO 2 layer and zwitterion modification could detect as low as 100 amol 475 of glycan. Thus, nanomaterials can help to enrich glycans with an LOD 3–9 orders of magnitude lower compared to commercially available supports, but some approaches can be quite sophisticated, involving five different components. The high effectivity of glycan enrichment also means that an ultralow amount of often precious samples is required, that is, 50 nL of human serum with an application of a carbonized mesoporous silica composite. 460 Glycan enrichment can be effective in the presence of nonglycoprotein or nonglycopeptides being in excess over glycoproteins/glycopeptides with a ratio of 1000:1 467 and also in a 1 million‐fold excess of a competing monosaccharide 479 due to high affinity toward glycans. Not only glycan enrichment can be made more effective using nanomaterial‐based approaches, but nanomaterial can also be applied as a support for the immobilization of trypsin or PNGase to perform an enzymatic reaction on a surface. The advantage of using immobilized enzymes is their reuse, since, for example, the stability of PNGase F covalently immobilized on GO was up to 8 weeks when stored at 4°C, and as little as 2 μg uL −1 of the enzyme complex was needed to analyze plasma extracts. 465 Moreover, the reaction was completed within 2 min, while usually overnight incubation was needed with a soluble enzyme. 465 When PNGase F was covalently immobilized on a porous silica column, the reaction was completed within 6 min. 463 Time needed for on‐line glycan removal was even shorter compared to that needed for the microwave‐assisted release of glycans in the presence of PNGase F (30 min at 37°C). 464 Immobilized trypsin was also applied for on‐line protein digestion, 506 which could be used for continuous protein digestion suitable for automatic glycan‐detecting systems. Nanomaterials can have additional beneficial effects on glycan analysis. For example, the addition of DNPs to the MALDI target in a layer‐by‐layer approach improved spot morphology and enhanced the sensitivity of MALDI‐TOF‐MS detection 79‐ or 7‐fold compared to the dried‐droplet (a mixture of matrix and sample) or thin‐layer (layer of sample deposited over a matrix layer) method, respectively. 495 The main principle behind this is the mediation of heat transfer between matrix and glycans by controlled architecture. 495 Considering the efficiency of energy absorption from the laser, the most effective way was the use of hydrazinonicotinic acid as both the matrix and glycan‐binding agent. 505 Thus, glycan enrichment directly on a plate was applied with an impressive LOD for glycan MALDI‐TOF‐MS analysis down to 1 amol, which is 5 orders of magnitude lower than that of 2,5‐dihydroxybenzoic acid, which is traditionally applied as the matrix. 505 MIP with a controlled thickness over imprinted analyte at the nanometer scale could offer a quite high imprinting ratio (i.e., ratio of glycoprotein attached to imprinted vs. nonimprinted NPs) of 22 472 and a quite high selectivity for its analyte of 30 over other glycoproteins, 470 which is useful for selecting a particular glycoprotein even from a quite complex sample. A. Nanoglycosensing The section dedicated to nanomaterial‐based bioanalytical methods and biosensors described in detail some of the most commonly used nanomaterials in such applications, including metal NPs, carbon nanomaterials, polymer nanostructures, and QDs. Great attention was paid to the combination of two or more (nano)materials, called nanohybrids and nanocomposites. Although some properties of the nanomaterials are of great importance in the field of nanobiotechnology and nanobiosensing (high conductivity, large surface area, and unique shape allowing multivalent binding, optical properties, etc.), 554 , 555 few real applications (i.e., analysis of real samples) have been described in the current literature. 556 Most of the presented papers dealt only with commercially available substances or model cell lines, although there is a growing number of papers trying to dynamically evaluate the number of membrane‐bound glycans on living cells for clinical diagnostics. 557 When used in a real clinical laboratory, such cells have to be isolated from patient tissues (taken during biopsy); therefore, less invasive methods to isolate the samples are welcome. The use of human sera or even whole blood is limited, mostly because of a small level of a particular biomarker present in body fluids. 558 , 559 , 560 , 561 Using a combination of immobilized antibodies (able to specifically bind a biomarker from the real sample) together with a lectin able to "decipher" the biomarker´s glycocode might be a solution in some cases. 562 B. Carbohydrate‐Based Vaccines and Therapeutics To summarize the therapeutic applications of carbohydrates integrated with nanomaterials or carbohydrates synthetized with a controlled density on various supports at the nanoscale, three main application directions can be distinguished: (i) employment of carbohydrate‐containing immunogens, especially viral and bacterial glycans responsible for cell recognition and consequent immune response, and tumor‐associated antigens 188 ; (ii) development of therapeutics based on the selective blocking of cell receptors via competitive lectin‐carbohydrate binding 563 ; and (iii) employment of carbohydrates as building blocks to develop NPs with therapeutic effects of molecules other than carbohydrate‐based moieties. The most recent progress in the development of carbohydrate nano‐vaccines is represented by the development of a method of chemoenzymatic synthesis allowing the tailored conjugation of glycans to carrier proteins leading to the increased immunoefficiency of the prepared NPs. 188 Another promising achievement was to conjugate tumor‐associated antigens to NPs even though non‐carbohydrate antigens seem to provide more significant results in this field. 564 , 565 The therapeutic effect outlined in (ii) has recently led to the development of NP‐glycan conjugates with higher affinity to specific receptors than bacterial or viral glycans. In the majority of studies, such NPs were employed as antibacterial and antiviral drugs because of their ability to selectively inhibit the first phase of infection. Recent studies have suggested that a combination of multivalent glycans and the selectivity of lectin‐glycan binding could be a step further in the development of completely new class of antibiotics. 566 It should be noted that, regardless of studies employing metallic and carbon NPs, (bio)polymers and synthetic micelles are the most preferable carriers for the same reasons as in the case of glycan nano‐vaccines. C. Cell Targeting While the nanomaterials used for delivery of carbohydrate‐based therapeutic agents (antigens or non‐immunogens) were mostly AuNPs, (bio)polymers, and dendrimers, there was much higher diversity of nanomaterials tested or developed as nanocarriers for drug delivery of drugs other than glycans, targeted by conjugated glycan moieties. Very intensive research is currently underway to explore targeting abilities, biocompatibility, and other advantages of HA, which has been used in many studies as an NP‐coating, securing selective internalization by tumor cells via their surface displayed CD44 receptors. 567 In a typical approach, HA has been used for the decoration of gold, 568 , 569 carbon, 288 or silica 310 NPs or synthetic NPs 570 conjugated with doxorubicin (a model cytostatic drug). Many studies have reported the inhibition of tumor growth after treatment by such NPs, but not at a significant rate. Nevertheless, efficient targeting is a crucial step in avoiding the severe adverse side effects of conventional cancer chemotherapy. In addition to glycan‐driven targeting, recent studies have also focused on the selective release of the delivered drug after internalization. Here, again, HA is an intriguing material because it could be degraded by intracellular hyaluronidases, liberating the drug from HA‐based NPs. 571 For many nanocarriers, the final drug release was secured by a significantly lower pH in lysosomes, which the NP could enter only after surface receptor‐driven internalization. 310 , 568 Another recently studied way to keep the nanocarriers with their cargo intact until they reach the target is the introduction of disulfide bonds into the NP structure. This bond could be readily reduced by glutathione, which is typically present at an elevated concentration in the cytoplasm of cancer cells. 310 Interesting results have also been achieved by the development of nanocarriers capable of passing through the blood brain barrier, for example, by the activation of a glucose transporter or by the conjugation of NPs with glycoproteins or lectins. This is important for the selective chemotherapy of brain tumors. In addition to drug delivery, carbohydrate‐targeted NPs were also promising as nonviral vectors, which has opened new possibilities in gene therapy for cancer. In addition to chemo‐ and geno‐therapies, targeting delivery has also been recently used for so‐called photodynamic and photothermal therapies. While in the former case, photosensitizers are delivered into cells that generate cytotoxic ROS, 375 , 571 , 572 the latter method relies on the delivery of NPs causing local overheating by irradiation after their internalization into selected cells. 573 In both cases, the desired action is typically triggered by NIR irradiation, which is harmless to healthy tissues without internalized photosensitizers. Another advantage of these methods is that it is not necessary to release the cargo from the carriers; the only aim is to get into the desired cells as many nanocarrier particles as possible, while a minimal amount should be in healthy tissue, as indicated by recent progress in carbohydrate‐based targeting. D. Cell Imaging Upon the replacement of the therapeutic agents discussed in the previous section with imaging probes, efficient diagnostic tools can be obtained not only for the visualization of tissue with an unusual glycopattern (most often tumors 390 or inflammation sites 574 ) but also for the very precise detection and assessment of, for example, surface cell receptors, lectin‐binding properties, and other features necessary for the further development of a glycan‐based targeted therapy. While such assessments are typically performed in vitro, NIR or UV‐VIS imaging probes can be used. Organic dyes coupled to glycan‐based nanocarriers are typically used for imaging in this range, and recently, so‐called theranostic NPs have been developed where imaging probes together with therapeutic agents are incorporated (see for example 575 ); thus, drug delivery can be monitored and confirmed. It should be noted that some organic therapeutic agents could also simultaneously function as imaging probes (e.g., doxorubicin or most photosensitizers), making theranostic NPs less expensive and easier to fabricate. For efficient in vivo imaging, CT and MRI have been routinely used for many years. Recent studies have focused on the targeted delivery of contrast enhancers, allowing the efficient visualization of selected tissues. While selectively distributed AuNPs have typically been used for CT imaging, 400 iron oxide containing NPs have been tested extensively as MRI imaging probes. 576 E. Glycan Enrichment and Separation It can be concluded that the application of nanomaterials or mesoporous materials with various functionalities present on the surface can significantly decrease the time needed for glycan enrichment (i.e., from a typical time of 160 577 to 1 515 min). Moreover, when using hybrid MNPs, the separation time can be as short as 15 sec by the simple application of an external magnet, 476 thus rendering the use of additional equipment (centrifuge) with a much longer separation time (few minutes) unnecessary. Nanomaterial‐based glycan enrichment could be applied for a minute amount of a starting material. Glycans can be enriched with commercially available hydrophilic beads with an LOD of 40 fmol, 466 and the most sensitive enrichment procedure involving MNPs with thin SiO 2 layer and zwitterion modification could detect as low as 100 amol 475 of glycan. Thus, nanomaterials can help to enrich glycans with an LOD 3–9 orders of magnitude lower compared to commercially available supports, but some approaches can be quite sophisticated, involving five different components. The high effectivity of glycan enrichment also means that an ultralow amount of often precious samples is required, that is, 50 nL of human serum with an application of a carbonized mesoporous silica composite. 460 Glycan enrichment can be effective in the presence of nonglycoprotein or nonglycopeptides being in excess over glycoproteins/glycopeptides with a ratio of 1000:1 467 and also in a 1 million‐fold excess of a competing monosaccharide 479 due to high affinity toward glycans. Not only glycan enrichment can be made more effective using nanomaterial‐based approaches, but nanomaterial can also be applied as a support for the immobilization of trypsin or PNGase to perform an enzymatic reaction on a surface. The advantage of using immobilized enzymes is their reuse, since, for example, the stability of PNGase F covalently immobilized on GO was up to 8 weeks when stored at 4°C, and as little as 2 μg uL −1 of the enzyme complex was needed to analyze plasma extracts. 465 Moreover, the reaction was completed within 2 min, while usually overnight incubation was needed with a soluble enzyme. 465 When PNGase F was covalently immobilized on a porous silica column, the reaction was completed within 6 min. 463 Time needed for on‐line glycan removal was even shorter compared to that needed for the microwave‐assisted release of glycans in the presence of PNGase F (30 min at 37°C). 464 Immobilized trypsin was also applied for on‐line protein digestion, 506 which could be used for continuous protein digestion suitable for automatic glycan‐detecting systems. Nanomaterials can have additional beneficial effects on glycan analysis. For example, the addition of DNPs to the MALDI target in a layer‐by‐layer approach improved spot morphology and enhanced the sensitivity of MALDI‐TOF‐MS detection 79‐ or 7‐fold compared to the dried‐droplet (a mixture of matrix and sample) or thin‐layer (layer of sample deposited over a matrix layer) method, respectively. 495 The main principle behind this is the mediation of heat transfer between matrix and glycans by controlled architecture. 495 Considering the efficiency of energy absorption from the laser, the most effective way was the use of hydrazinonicotinic acid as both the matrix and glycan‐binding agent. 505 Thus, glycan enrichment directly on a plate was applied with an impressive LOD for glycan MALDI‐TOF‐MS analysis down to 1 amol, which is 5 orders of magnitude lower than that of 2,5‐dihydroxybenzoic acid, which is traditionally applied as the matrix. 505 MIP with a controlled thickness over imprinted analyte at the nanometer scale could offer a quite high imprinting ratio (i.e., ratio of glycoprotein attached to imprinted vs. nonimprinted NPs) of 22 472 and a quite high selectivity for its analyte of 30 over other glycoproteins, 470 which is useful for selecting a particular glycoprotein even from a quite complex sample. 11. PERSPECTIVES A. Nanoglycosensing Moreover, the next step for researchers dealing with lectin–glycan interactions should be to overcome lectin "promiscuity" by tuning their specificity (e.g., using recombinant technology to generate mutant proteins 578 ) or to prepare more specific aptamers (DNA or peptide) with better binding properties for recognising glycans, which could be immobilized on surfaces with controlled orientation/precision because of their smaller size. 579 , 580 Molecular imprinting technology for the differentiation of two glycan moieties with the same composition but different glycosidic bond(s), as in the case of 3´‐ and 6´‐sialyllactose mentioned in the Section A. .A, should also be investigated. 581 , 582 Most of the presented methods involve expensive equipment (SPR) or methods with a complicated theoretical background (EIS), which could be interpreted only by a skilled person. Cheap protocols (for instance, those using common microplates) with possible naked‐eye detection are welcome. In addition, some of the methods, even well understood, still cannot be used outside of the laboratory and thus integrated into commercially mass produced point‐of‐care devices. Such barriers to commercialization are significant problems in the case of impedimetric biosensors because of the lower stability and reliability of these biosensors compared to other types. 583 Sensor regeneration and flow systems are an issue to focus on, even though they are not as important as the need for detection in an array format for highly parallel analysis. An array assay format in combination with more sensitive platforms compared to generally employed fluorescent techniques (ideally, electrochemical methods compatible with miniaturization and nanomaterial‐based modifications) still needs to be developed for a wider use. B. Carbohydrate‐Based Vaccines and Therapeutics From a different perspective, there are barely any studies exploring the use of nanomaterials other than AuNPs and synthetic micelles or biopolymer scaffolds. In this regard, reported iron oxide NPs conjugated with glycans 226 can be considered a way for the future preparation of immunogenic nanoconjugates based on nanomaterials being cheaper than AuNPs. Natural polysaccharides are also known to increase the stability of vaccine NPs. More importantly, the immunogenicity of these materials and their use as excellent adjuvants has to be tested. From this point of view, chitosan and its derivatives are polysaccharides that are frequently applied for such purposes. 202 , 584 Also very important is the recent progress in the controlled synthesis 585 and evaluation of iminosugars 586 conjugated with NPs, that is, pharmacological chaperones responsible for partial functional restoration of glycosidases, whose misfolding results in lysosomal storage diseases. It seems that in this particular field, the synthesis of dendrimers or scaffolded iminosugar carriers requires a meticulously defined conformal structure, conjugate size, valency, and spacers between a carrier and an iminosugar moiety. 586 C. Cell Targeting The targeting, delivery, and release mechanisms are very similar to those employed in targeted drug delivery. Importantly, even though genotherapy itself may not provide the highest efficiency, it is possible to increase sensitivity of tumor cells to another treatment (e.g., to cytostatics) by the application of specific short nucleotides, thus allowing such a combined therapy to become more efficient. 351 Furthermore, a reliable and specific gene delivery would probably be necessary for further progress in personalized medicine. Finally, it should be noted that the most rapid tumor growth inhibition has been observed typically when a combination of methods was used; for example, chemophototherapy or chemothermal therapy. 288 In the further development of carbohydrate‐based nano‐therapeutics, such synergy of effects would probably be the most promising path to follow. D. Cell Imaging Considering that AuNPs are excellent thermodynamic therapy agents, we see another way for the development of multifunctional NPs without any additional fabrication steps or additional components used. From this perspective, graphene and its derivatives were found to have even more promising attributes—they could be used as drug (and glycan‐based targeting moieties) carriers, thermodynamic therapy agents, and fluorophores. 288 , 587 Furthermore, graphene QDs, that is, nanometer‐sized graphene sheets, can replace conventional semiconductors and usually cytotoxic (unless they are made biocompatible, for example, by carbohydrate coating) QDs. 588 It is anticipated that graphene derivatives will be investigated more intensively in years to come to obtain cheap, easy‐to‐fabricate and efficient theranostic NPs. E. Glycan Enrichment and Separation Finally, we can say that even though nanomaterials or nanoporous materials exhibit remarkable efficiency in glycan profiling due to their robustness, the reproducibility of the glycan enrichment process together with their mechanical/chemical stability has to be more rigorously compared, ideally in studies involving multiple institutions, as recently performed for the glycoprofiling of N ‐glycans of IgG 589 , 590 , 591 or O ‐ and N ‐glycans from cultured cell lines, 592 to identify strong and weak attributes for each glycan enrichment technique. The other class of affinity‐based molecules, which can be applied in the future to selectively detect glycoproteins/glycopeptides or for glycan enrichment, are DNA/RNA aptamers designed to recognize glycoproteins since their K D for glycans can be lower compared to lectins. 459 There is however hope for the more frequent application of modified lectins in future glycan enrichment, when lectin modified with boronate will be applied for the more sensitive binding of glycans. 485 Moreover, various types of microengines/microrockets can also be applied for active glycan enrichment from quite complex samples. 12. ABBREVIATIONS AgNPs silver nanoparticles AuNPs gold nanoparticles AuNRs gold nanorods BSA bovine serum albumin CEA carcinoembryonic antigen cfu colony‐forming unit CNTs carbon nanotubes Con A Concanavalin A CT computational tomography CuAAC Cu I catalyzed Huisgen azide‐alkyne cycloaddition CVD chemical vapor deposition DC‐SIGN Dendritic Cell‐Specific Intercellular adhesion molecule‐3‐Grabbing Non‐integrin DNPs nanodiamond particles DPV differential pulse voltammetry ECL electrochemiluminiscence EIS electrochemical impedance spectroscopy ESI electrospray ionization FimH type 1 fimbrial adhesin FITC fluorescein isothiocyanate GCE glassy carbon electrode GlcNAc N ‐acetylglucosamine GO graphene oxide GOx glucose oxidase HA hyaluronic acid HRP horseradish peroxidase IgG immunoglobulin G ITO indium tin oxide LacNAc N ‐acetyl‐ d ‐lactosamine Le a Lewis a antigen Le x Lewis x antigen Le y Lewis y antigen LOD limit of detection LPS lipopolysaccharide LSPR localized surface plasmon resonance MALDI matrix‐assisted laser desorption/ionization MALDI‐TOF‐MS matrix‐assisted laser desorption/ionization time‐of‐flight mass spectrometry MIP molecularly imprinted polymers MNPs magnetic nanoparticles MS mass spectrometry MUC1 immunogenic synthetic mucin MWCNTs multiwalled carbon nanotubes NIR near infrared NMR nuclear magnetic resonance NMRi nuclear magnetic resonance imaging NPs nanoparticles OEG oligoethylene glycol PAMAM poly(amidoamine) PBA phenylboronic acid PC prostate cancer PdNPs palladium nanoparticles PEG polyethyleneglycol PEI poly(ethylenimine) PNGase A peptide‐ N ‐glycosidase A PNGase F peptide‐ N ‐glycosidase F PSA prostate‐specific antigen PtNPs platinum nanoparticles QCM quartz crystal microbalance QDs quantum dots rHDL reconstituted high‐density lipoprotein ROS reactive oxygen species SAMs self‐assembled monolayers SERS surface enhanced Raman scattering SiNPs silica NPs siRNA small interfering RNA sLe a sialyl Lewis a antigen sLe x sialyl Lewis x antigen SNA Sambucus nigra agglutinin SPR surface plasmon resonance SR‐BI Scavenger Receptor class B member 1 SWV square‐wave voltammetry RCA 120 Ricinus communis agglutinin with M w = 120 kDa rGO reduced graphene oxide SWCNTs single‐walled carbon nanotubes TEOS tetraethylorthosilicate TF ag Thomas Friedrich antigen WGA wheat germ agglutinin A. Nanoglycosensing Moreover, the next step for researchers dealing with lectin–glycan interactions should be to overcome lectin "promiscuity" by tuning their specificity (e.g., using recombinant technology to generate mutant proteins 578 ) or to prepare more specific aptamers (DNA or peptide) with better binding properties for recognising glycans, which could be immobilized on surfaces with controlled orientation/precision because of their smaller size. 579 , 580 Molecular imprinting technology for the differentiation of two glycan moieties with the same composition but different glycosidic bond(s), as in the case of 3´‐ and 6´‐sialyllactose mentioned in the Section A. .A, should also be investigated. 581 , 582 Most of the presented methods involve expensive equipment (SPR) or methods with a complicated theoretical background (EIS), which could be interpreted only by a skilled person. Cheap protocols (for instance, those using common microplates) with possible naked‐eye detection are welcome. In addition, some of the methods, even well understood, still cannot be used outside of the laboratory and thus integrated into commercially mass produced point‐of‐care devices. Such barriers to commercialization are significant problems in the case of impedimetric biosensors because of the lower stability and reliability of these biosensors compared to other types. 583 Sensor regeneration and flow systems are an issue to focus on, even though they are not as important as the need for detection in an array format for highly parallel analysis. An array assay format in combination with more sensitive platforms compared to generally employed fluorescent techniques (ideally, electrochemical methods compatible with miniaturization and nanomaterial‐based modifications) still needs to be developed for a wider use. B. Carbohydrate‐Based Vaccines and Therapeutics From a different perspective, there are barely any studies exploring the use of nanomaterials other than AuNPs and synthetic micelles or biopolymer scaffolds. In this regard, reported iron oxide NPs conjugated with glycans 226 can be considered a way for the future preparation of immunogenic nanoconjugates based on nanomaterials being cheaper than AuNPs. Natural polysaccharides are also known to increase the stability of vaccine NPs. More importantly, the immunogenicity of these materials and their use as excellent adjuvants has to be tested. From this point of view, chitosan and its derivatives are polysaccharides that are frequently applied for such purposes. 202 , 584 Also very important is the recent progress in the controlled synthesis 585 and evaluation of iminosugars 586 conjugated with NPs, that is, pharmacological chaperones responsible for partial functional restoration of glycosidases, whose misfolding results in lysosomal storage diseases. It seems that in this particular field, the synthesis of dendrimers or scaffolded iminosugar carriers requires a meticulously defined conformal structure, conjugate size, valency, and spacers between a carrier and an iminosugar moiety. 586 C. Cell Targeting The targeting, delivery, and release mechanisms are very similar to those employed in targeted drug delivery. Importantly, even though genotherapy itself may not provide the highest efficiency, it is possible to increase sensitivity of tumor cells to another treatment (e.g., to cytostatics) by the application of specific short nucleotides, thus allowing such a combined therapy to become more efficient. 351 Furthermore, a reliable and specific gene delivery would probably be necessary for further progress in personalized medicine. Finally, it should be noted that the most rapid tumor growth inhibition has been observed typically when a combination of methods was used; for example, chemophototherapy or chemothermal therapy. 288 In the further development of carbohydrate‐based nano‐therapeutics, such synergy of effects would probably be the most promising path to follow. D. Cell Imaging Considering that AuNPs are excellent thermodynamic therapy agents, we see another way for the development of multifunctional NPs without any additional fabrication steps or additional components used. From this perspective, graphene and its derivatives were found to have even more promising attributes—they could be used as drug (and glycan‐based targeting moieties) carriers, thermodynamic therapy agents, and fluorophores. 288 , 587 Furthermore, graphene QDs, that is, nanometer‐sized graphene sheets, can replace conventional semiconductors and usually cytotoxic (unless they are made biocompatible, for example, by carbohydrate coating) QDs. 588 It is anticipated that graphene derivatives will be investigated more intensively in years to come to obtain cheap, easy‐to‐fabricate and efficient theranostic NPs. E. Glycan Enrichment and Separation Finally, we can say that even though nanomaterials or nanoporous materials exhibit remarkable efficiency in glycan profiling due to their robustness, the reproducibility of the glycan enrichment process together with their mechanical/chemical stability has to be more rigorously compared, ideally in studies involving multiple institutions, as recently performed for the glycoprofiling of N ‐glycans of IgG 589 , 590 , 591 or O ‐ and N ‐glycans from cultured cell lines, 592 to identify strong and weak attributes for each glycan enrichment technique. The other class of affinity‐based molecules, which can be applied in the future to selectively detect glycoproteins/glycopeptides or for glycan enrichment, are DNA/RNA aptamers designed to recognize glycoproteins since their K D for glycans can be lower compared to lectins. 459 There is however hope for the more frequent application of modified lectins in future glycan enrichment, when lectin modified with boronate will be applied for the more sensitive binding of glycans. 485 Moreover, various types of microengines/microrockets can also be applied for active glycan enrichment from quite complex samples. 12. ABBREVIATIONS AgNPs silver nanoparticles AuNPs gold nanoparticles AuNRs gold nanorods BSA bovine serum albumin CEA carcinoembryonic antigen cfu colony‐forming unit CNTs carbon nanotubes Con A Concanavalin A CT computational tomography CuAAC Cu I catalyzed Huisgen azide‐alkyne cycloaddition CVD chemical vapor deposition DC‐SIGN Dendritic Cell‐Specific Intercellular adhesion molecule‐3‐Grabbing Non‐integrin DNPs nanodiamond particles DPV differential pulse voltammetry ECL electrochemiluminiscence EIS electrochemical impedance spectroscopy ESI electrospray ionization FimH type 1 fimbrial adhesin FITC fluorescein isothiocyanate GCE glassy carbon electrode GlcNAc N ‐acetylglucosamine GO graphene oxide GOx glucose oxidase HA hyaluronic acid HRP horseradish peroxidase IgG immunoglobulin G ITO indium tin oxide LacNAc N ‐acetyl‐ d ‐lactosamine Le a Lewis a antigen Le x Lewis x antigen Le y Lewis y antigen LOD limit of detection LPS lipopolysaccharide LSPR localized surface plasmon resonance MALDI matrix‐assisted laser desorption/ionization MALDI‐TOF‐MS matrix‐assisted laser desorption/ionization time‐of‐flight mass spectrometry MIP molecularly imprinted polymers MNPs magnetic nanoparticles MS mass spectrometry MUC1 immunogenic synthetic mucin MWCNTs multiwalled carbon nanotubes NIR near infrared NMR nuclear magnetic resonance NMRi nuclear magnetic resonance imaging NPs nanoparticles OEG oligoethylene glycol PAMAM poly(amidoamine) PBA phenylboronic acid PC prostate cancer PdNPs palladium nanoparticles PEG polyethyleneglycol PEI poly(ethylenimine) PNGase A peptide‐ N ‐glycosidase A PNGase F peptide‐ N ‐glycosidase F PSA prostate‐specific antigen PtNPs platinum nanoparticles QCM quartz crystal microbalance QDs quantum dots rHDL reconstituted high‐density lipoprotein ROS reactive oxygen species SAMs self‐assembled monolayers SERS surface enhanced Raman scattering SiNPs silica NPs siRNA small interfering RNA sLe a sialyl Lewis a antigen sLe x sialyl Lewis x antigen SNA Sambucus nigra agglutinin SPR surface plasmon resonance SR‐BI Scavenger Receptor class B member 1 SWV square‐wave voltammetry RCA 120 Ricinus communis agglutinin with M w = 120 kDa rGO reduced graphene oxide SWCNTs single‐walled carbon nanotubes TEOS tetraethylorthosilicate TF ag Thomas Friedrich antigen WGA wheat germ agglutinin

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7777151/

Prognosis of Patients with Sepsis and Non-Hepatic Hyperammonemia: A Cohort Study

Background Hyperammonemia has been reported in some critically ill patients with sepsis who do not have hepatic failure. A significant proportion of patients with non-hepatic hyperammonemia have underlying sepsis, but the association between non-hepatic hyperammonemia and prognosis is unclear. Material/Methods Information about patients with sepsis and non-hepatic hyperammonemia was retrieved from the Medical Information Mart for Intensive Care-III database. Survival rates were analyzed using the Kaplan-Meier method. Multivariate logistic regression models were employed to identify prognostic factors. Receiver operating characteristic (ROC) curve analysis was used to measure the predictive ability of ammonia in terms of patient mortality. Results A total of 265 patients with sepsis were enrolled in this study. Compared with the non-hyperammonemia group, the patients with hyperammonemia had significantly higher rates of hospital (59.8% vs. 43.0%, P =0.007), 30-day (47.7% vs. 34.8%, P =0.036), 90-day (61.7% vs. 43.7%, P =0.004), and 1-year mortality (67.3% vs. 49.4%, P =0.004). In the survival analysis, hyperammonemia was associated with these outcomes. Serum ammonia level was an independent predictor of hospital mortality. The area under the ROC curve for the ammonia levels had poor discriminative capacity. The hyperammonemia group also had significantly lower Glasgow Coma Scale scores ( P =0.020) and higher incidences of delirium (15.9% vs. 8.2%, P =0.034) and encephalopathy (37.4% vs. 19.6%, P =0.001). Intestinal infection and urinary tract infection with organisms such as Escherichia coli may be risk factors for hyperammonemia in patients who have sepsis. Conclusions Higher ammonia levels are associated with poorer prognosis in patients with sepsis. Ammonia also may be associated with sepsis-associated encephalopathy. Therefore, we recommend that serum ammonia levels be measured in patients who are suspected of having sepsis. Background Hyperammonemia has been reported in some critically ill patients with sepsis who do not have hepatic failure. A significant proportion of patients with non-hepatic hyperammonemia have underlying sepsis, but the association between non-hepatic hyperammonemia and prognosis is unclear. Material/Methods Information about patients with sepsis and non-hepatic hyperammonemia was retrieved from the Medical Information Mart for Intensive Care-III database. Survival rates were analyzed using the Kaplan-Meier method. Multivariate logistic regression models were employed to identify prognostic factors. Receiver operating characteristic (ROC) curve analysis was used to measure the predictive ability of ammonia in terms of patient mortality. Results A total of 265 patients with sepsis were enrolled in this study. Compared with the non-hyperammonemia group, the patients with hyperammonemia had significantly higher rates of hospital (59.8% vs. 43.0%, P =0.007), 30-day (47.7% vs. 34.8%, P =0.036), 90-day (61.7% vs. 43.7%, P =0.004), and 1-year mortality (67.3% vs. 49.4%, P =0.004). In the survival analysis, hyperammonemia was associated with these outcomes. Serum ammonia level was an independent predictor of hospital mortality. The area under the ROC curve for the ammonia levels had poor discriminative capacity. The hyperammonemia group also had significantly lower Glasgow Coma Scale scores ( P =0.020) and higher incidences of delirium (15.9% vs. 8.2%, P =0.034) and encephalopathy (37.4% vs. 19.6%, P =0.001). Intestinal infection and urinary tract infection with organisms such as Escherichia coli may be risk factors for hyperammonemia in patients who have sepsis. Conclusions Higher ammonia levels are associated with poorer prognosis in patients with sepsis. Ammonia also may be associated with sepsis-associated encephalopathy. Therefore, we recommend that serum ammonia levels be measured in patients who are suspected of having sepsis. Background Sepsis is a serious medical condition responsible for approximately 19.77% of all deaths worldwide [ 1 , 2 ]. The mortality is a result of the systemic inflammation and end-organ dysfunction associated with these infections [ 3 ]. The rate of mortality in patients diagnosed with sepsis is 30%, and 50% in individuals with severe sepsis. In patients in whom the disease progresses to septic shock, the mortality rate can rise to as high as 80%. As an individual's infection worsens, the risk of mortality gradually increases [ 4 ]. Sepsis-associated encephalopathy (SAE) can be found in up to 70% of patients with severe sepsis and it is a common neurological complication [ 5 ], with a mortality rate of up to 70% [ 6 ]. Ammonia is a major factor in the pathogenesis of hepatic encephalopathy and it crosses the blood-brain barrier readily, resulting in significant neurotoxicity [ 7 ]. Disorders of ammonia metabolism can lead to hyperammonemia, which usually is a consequence of hepatic failure. Hyperammonemia also can occur in critically ill patients who do not have hepatic disease [ 8 ], including individuals with sepsis, gastrointestinal bleeding, kidney failure, elevations in sodium, and exposure to valproate [ 8 , 9 ]. In recent reports, serum ammonia has been suggested as a possible predictor of 28-day mortality and hospital stay in patients with sepsis. While elevation in ammonia level has been reported as a novel biomarker for sepsis [ 10 , 11 ], its roles in long-term prognosis and as a risk factor for non-hepatic hyperammonemia in patients with sepsis are unclear. The relationship between serum ammonia and the development of sepsis and its prognosis in patients with the condition remains under-explored. The aim of this study was to determine the significance of elevated serum ammonia levels to both the short- and long-term prognosis of patients with sepsis. We also explored risk factors for non-hepatic hyperammonemia in sepsis and the association between non-hepatic hyperammonemia and SAE. Material and Methods Database This was a retrospective study based on information recorded in the publicly available Medical Information Mart for Intensive Care (MIMIC-III) database between 2001 and 2012. Use of the database was approved by the Massachusetts Institute of Technology (Cambridge, Massachusetts, U.S.A.) and the Institutional Review Board of Beth Israel Deaconess Medical Center (Boston, Massachusetts, U.S.A.). Individual patient consent was not required because the study was a retrospective review of publicly available, anonymized data and the analysis did not affect the care of individual patients. The raw data were extracted using structure query language (SQL) with Navicat and further processed with R software. Patient population Inclusion criteria for the study were as follows: (1) a diagnosis of sepsis, severe sepsis, or septic shock according to International Classification of Diseases, Ninth Revision (ICD-9) codes; (2) age ≥18 and ≤89 years; (3) admission for >24 hours in the intensive care unit (ICU); and documentation of blood ammonia levels. A blood ammonia level >35 μmol/L was defined as hyperammonemia in the MIMIC-III database. Exclusion criteria for the study were as follows: (1) a history of acute or chronic liver disease, including hepatitis, hepatic cirrhosis, hepatic encephalopathy, hepatorenal syndrome, hepatic injury, and other chronic liver disease, according to ICD-9 diagnosis codes on patient discharge ( Supplementary Table 1 ); and (2) no documentation of vital signs or ICD-9 diagnostic codes. Data extraction R statistical software (R Foundation for Statistical Computing, Vienna, Austria) was used to collect data on baseline characteristics information such as age, sex, and vital signs and laboratory parameters during the first 24 hours of ICU admission. The maximum value for ammonia during each patient's ICU stay also was retrieved. Infection type ( Supplementary Table 2 ), microbiology type ( Supplementary Table 3 ), and patient comorbidities ( Supplementary Table 4 ) were determined according to the primary ICD-9 codes, as documented in each patient's discharge summary. We retrieved the SQL scripts from the GitHub website ( https://github.com/MIT-LCP/mimic-code/tree/master/concepts/severity-scores ) and used them to calculate the severity scores. Simplified Acute Physiology Score (SAPSII), Sequential Organ Failure Assessment (SOFA) score, and Glasgow Coma Scale (GCS) ratings also were recorded during the first 24 hours of each patient's ICU stay. Outcomes of patient conditions such as delirium, encephalopathy, mechanical ventilation, renal replacement therapy ( Supplementary Table 5 ), and survival status were recorded. Relevant information was obtained about patients who were diagnosed with "sepsis," "severe sepsis," and "septic shock" on discharge, according to ICD-9 codes ( Supplementary Table 6 ). Patients were assigned to the hyperammonemia and non-hyperammonemia groups based on serum ammonia levels. They were also divided into conscious (GCS=15), sub-coma (GCS 9–14), and deep coma groups (GCS 3–8) based on GCS scores. Statistical analysis The statistical analysis compared the hyperammonemia and non-hyperammonemia groups. Data distribution was tested using the Shapiro-Wilk test. Continuous variables were expressed as means with standard deviation for normal distributed data, and for non-normally distributed data, medians (interquartile range [IQR]) were expressed. Categorical variables were represented as frequencies with percentage and compared using a chi-square test. Variables with missing data are relatively common in the MIMIC-III database and we replaced them with median values (Supplementary Material 1). A non-parametric test (Mann-Whitney U or Kruskal-Wallis) was used for comparisons between the baseline characteristics and outcomes in the hyperammonemia and non-hyperammonemia groups and the relationship between serum ammonia and consciousness. Kaplan-Meier curves were analyzed using log-rank tests for comparison of hospital mortality between the hyperammonemia and non-hepatic hyperammonemia groups. A Cox regression model was used to screen for variables associated with hospital mortality in survivors versus non-survivors. A 2-tailed P 24 hours in the intensive care unit (ICU); and documentation of blood ammonia levels. A blood ammonia level >35 μmol/L was defined as hyperammonemia in the MIMIC-III database. Exclusion criteria for the study were as follows: (1) a history of acute or chronic liver disease, including hepatitis, hepatic cirrhosis, hepatic encephalopathy, hepatorenal syndrome, hepatic injury, and other chronic liver disease, according to ICD-9 diagnosis codes on patient discharge ( Supplementary Table 1 ); and (2) no documentation of vital signs or ICD-9 diagnostic codes. Data extraction R statistical software (R Foundation for Statistical Computing, Vienna, Austria) was used to collect data on baseline characteristics information such as age, sex, and vital signs and laboratory parameters during the first 24 hours of ICU admission. The maximum value for ammonia during each patient's ICU stay also was retrieved. Infection type ( Supplementary Table 2 ), microbiology type ( Supplementary Table 3 ), and patient comorbidities ( Supplementary Table 4 ) were determined according to the primary ICD-9 codes, as documented in each patient's discharge summary. We retrieved the SQL scripts from the GitHub website ( https://github.com/MIT-LCP/mimic-code/tree/master/concepts/severity-scores ) and used them to calculate the severity scores. Simplified Acute Physiology Score (SAPSII), Sequential Organ Failure Assessment (SOFA) score, and Glasgow Coma Scale (GCS) ratings also were recorded during the first 24 hours of each patient's ICU stay. Outcomes of patient conditions such as delirium, encephalopathy, mechanical ventilation, renal replacement therapy ( Supplementary Table 5 ), and survival status were recorded. Relevant information was obtained about patients who were diagnosed with "sepsis," "severe sepsis," and "septic shock" on discharge, according to ICD-9 codes ( Supplementary Table 6 ). Patients were assigned to the hyperammonemia and non-hyperammonemia groups based on serum ammonia levels. They were also divided into conscious (GCS=15), sub-coma (GCS 9–14), and deep coma groups (GCS 3–8) based on GCS scores. Statistical analysis The statistical analysis compared the hyperammonemia and non-hyperammonemia groups. Data distribution was tested using the Shapiro-Wilk test. Continuous variables were expressed as means with standard deviation for normal distributed data, and for non-normally distributed data, medians (interquartile range [IQR]) were expressed. Categorical variables were represented as frequencies with percentage and compared using a chi-square test. Variables with missing data are relatively common in the MIMIC-III database and we replaced them with median values (Supplementary Material 1). A non-parametric test (Mann-Whitney U or Kruskal-Wallis) was used for comparisons between the baseline characteristics and outcomes in the hyperammonemia and non-hyperammonemia groups and the relationship between serum ammonia and consciousness. Kaplan-Meier curves were analyzed using log-rank tests for comparison of hospital mortality between the hyperammonemia and non-hepatic hyperammonemia groups. A Cox regression model was used to screen for variables associated with hospital mortality in survivors versus non-survivors. A 2-tailed P <0.05 was considered statistically significant. All statistical analyses were performed with R software (version 3.4.3). Results Baseline patient characteristics The patient inclusion flowchart is shown in Figure 1 . A total of 2159 patients were tested for blood ammonia according to information in the MIMIC-III database. Using the inclusion and exclusion criteria, 1051 patients were identified for further screening. Of those patients, 265 were diagnosed with "sepsis," "severe sepsis," or "septic shock" on discharge, according to ICD-9 codes, and were enrolled in the study. The incidence of non-hepatic hyperammonemia was 40.4% with a 67.3% rate of 1-year mortality. Information on the patients' baseline characteristics, vital signs, laboratory parameters, infection type, microbiology type, and comorbid diseases is summarized in Table 1 . There were 107 patients in the hyperammonemia group and 158 patients in the non-hyperammonemia group. Patients in the hyperammonemia group had significantly more intestinal infections (23.4% vs. 13.3%, P =0.034) and urinary tract infections (UTIs) (45.8% vs. 24.7%, P <0.001) than patients in the non-hyperammonemia group. Patients with hyperammonemia were more likely to be infected with Escherichia coli (42.1% vs. 22.8%, P =0.001). Patients in the hyperammonemia group had lower GCS scores than patients in the non-hyperammonemia group ( P =0.020). No correlation was found between ammonia levels and respiratory infection, gastrointestinal bleeding, heart failure, kidney failure, or infection in other tissues by E. coli . In addition, there were no significant differences in SAPSII or SOFA scores between the 2 groups. Patient outcomes Table 2 shows the outcomes in the hyperammonemia and non-hyperammonemia groups. As illustrated, a greater proportion of patients in the hyperammonemia group were diagnosed with delirium (15.9% vs. 8.2%, P =0.034) and encephalopathy (37.4% vs. 19.6%, P =0.001). Patients with hyperammonemia also had higher rates of short- and long-term mortality (in-hospital, 59.8% vs. 43.0%; 30-day, 47.7% vs. 34.8%; 90-day, 61.7% vs. 43.7%; 1-year, 67.3% vs. 49.4%). Ammonia was an independent prognostic predictor in patients with sepsis Patients in the hyperammonemia group had worse survival rates (in-hospital, 90-day, and 1-year mortality) ( Figure 2 ). Furthermore, univariate and multivariate Cox analysis was performed of baseline variables (age and sex) and results of laboratory tests (alanine aminotransferase, aspartate aminotransferase, creatinine, blood urea nitrogen, hemoglobin, platelet count, partial thromboplastin time, international normalized ratio, prothrombin time, white blood cell count, and ammonia). The factors significantly correlated with survival were adjusted for in the multivariate analysis. The analysis revealed that ammonia remained an independent prognostic factor in patients with sepsis. ( P <0.01 or P <0.05) ( Table 3 ). Receiver operating characteristic curves of ammonia indices for predicting mortality To further confirm the reliability of ammonia, we plotted the area under the receiver operating characteristic (ROC) curve for 90-day and 1-year survival, and in-hospital mortality. The discriminative ability of ammonia levels based on the ROC curve analysis was 0.625 for in-hospital mortality, 0.620 for 90-day survival, and 0.624 for 1-year survival ( Figure 3 ). Relationship between serum ammonia and consciousness Patients were divided into conscious (n=109), sub-coma (n=112), and deep coma groups (n=44) based on GCS score. As shown in Figure 4 , patients with lower GCS scores had higher serum ammonia levels. The serum ammonia levels were highest in the deep coma group, compared with the other 2 groups ( P <0.001), and they were significantly higher in the sub-coma group than in the conscious group ( P <0.001) ( Figure 4 ). Baseline patient characteristics The patient inclusion flowchart is shown in Figure 1 . A total of 2159 patients were tested for blood ammonia according to information in the MIMIC-III database. Using the inclusion and exclusion criteria, 1051 patients were identified for further screening. Of those patients, 265 were diagnosed with "sepsis," "severe sepsis," or "septic shock" on discharge, according to ICD-9 codes, and were enrolled in the study. The incidence of non-hepatic hyperammonemia was 40.4% with a 67.3% rate of 1-year mortality. Information on the patients' baseline characteristics, vital signs, laboratory parameters, infection type, microbiology type, and comorbid diseases is summarized in Table 1 . There were 107 patients in the hyperammonemia group and 158 patients in the non-hyperammonemia group. Patients in the hyperammonemia group had significantly more intestinal infections (23.4% vs. 13.3%, P =0.034) and urinary tract infections (UTIs) (45.8% vs. 24.7%, P <0.001) than patients in the non-hyperammonemia group. Patients with hyperammonemia were more likely to be infected with Escherichia coli (42.1% vs. 22.8%, P =0.001). Patients in the hyperammonemia group had lower GCS scores than patients in the non-hyperammonemia group ( P =0.020). No correlation was found between ammonia levels and respiratory infection, gastrointestinal bleeding, heart failure, kidney failure, or infection in other tissues by E. coli . In addition, there were no significant differences in SAPSII or SOFA scores between the 2 groups. Patient outcomes Table 2 shows the outcomes in the hyperammonemia and non-hyperammonemia groups. As illustrated, a greater proportion of patients in the hyperammonemia group were diagnosed with delirium (15.9% vs. 8.2%, P =0.034) and encephalopathy (37.4% vs. 19.6%, P =0.001). Patients with hyperammonemia also had higher rates of short- and long-term mortality (in-hospital, 59.8% vs. 43.0%; 30-day, 47.7% vs. 34.8%; 90-day, 61.7% vs. 43.7%; 1-year, 67.3% vs. 49.4%). Ammonia was an independent prognostic predictor in patients with sepsis Patients in the hyperammonemia group had worse survival rates (in-hospital, 90-day, and 1-year mortality) ( Figure 2 ). Furthermore, univariate and multivariate Cox analysis was performed of baseline variables (age and sex) and results of laboratory tests (alanine aminotransferase, aspartate aminotransferase, creatinine, blood urea nitrogen, hemoglobin, platelet count, partial thromboplastin time, international normalized ratio, prothrombin time, white blood cell count, and ammonia). The factors significantly correlated with survival were adjusted for in the multivariate analysis. The analysis revealed that ammonia remained an independent prognostic factor in patients with sepsis. ( P <0.01 or P <0.05) ( Table 3 ). Receiver operating characteristic curves of ammonia indices for predicting mortality To further confirm the reliability of ammonia, we plotted the area under the receiver operating characteristic (ROC) curve for 90-day and 1-year survival, and in-hospital mortality. The discriminative ability of ammonia levels based on the ROC curve analysis was 0.625 for in-hospital mortality, 0.620 for 90-day survival, and 0.624 for 1-year survival ( Figure 3 ). Relationship between serum ammonia and consciousness Patients were divided into conscious (n=109), sub-coma (n=112), and deep coma groups (n=44) based on GCS score. As shown in Figure 4 , patients with lower GCS scores had higher serum ammonia levels. The serum ammonia levels were highest in the deep coma group, compared with the other 2 groups ( P <0.001), and they were significantly higher in the sub-coma group than in the conscious group ( P <0.001) ( Figure 4 ). Discussion Our study demonstrated that the incidence of non-hepatic hyperammonemia is 40.4% in patients with sepsis and the incidence of sepsis with encephalopathy in patients with non-hepatic hyperammonemia is 37.4%. Serum ammonia level may be a predictor of mortality in patients with sepsis who do not have hepatic disease. In addition, we found that intestinal infection, UTI, and infections in other tissues caused by E. coli were risk factors for non-hepatic hyperammonemia in patients with sepsis. We also found that the rate of hospital mortality in patients with sepsis who had non-hepatic hyperammonemia was 59.8%, which was significantly higher than in patients with sepsis who had normal serum ammonia levels (46.4%) [ 1 ]. A higher serum ammonia level may be a risk factor for mortality. Our results are consistent with the findings of Zhao et al., which showed that in patients with sepsis, an increased serum ammonia level on admission to the emergency department was correlated with an increased rate of mortality at 28 days. Our study explored mortality levels for up to 1 year, and we found that serum ammonia is an independent risk factor for long-term prognosis in patients with sepsis. In a case series, McEwan et al. suggested that higher serum ammonia levels are related to adverse clinical outcomes, which correlates with our findings. However, Zhao et al. showed that serum ammonia levels had a robust ability to predict the 28-day mortality rate in patients with sepsis, with an area under the ROC curve of 0.813, which is in contrast to our findings. That discrepancy may be attributable to differences in basic patient characteristics between the 2 studies. It suggests that serum ammonia level may be a new prognostic marker for patients with sepsis. An interesting finding in our study is that non-hepatic hyperammonemia may be associated with an increased risk of SAE [ 12 ]. SAE is mainly characterized by symptoms of delirium with changes in a patient's consciousness, and it also can lead to coma [ 13 ]. Our study demonstrated that patients with hyperammonemia had lower GCS scores. In the absence of previous cerebrovascular and encephalopathic brain disease, SAE is more likely to occur as the serum ammonia level increases. SAE is a diffuse brain dysfunction that occurs secondary to sepsis in the body without overt infection of the central nervous system. Its pathogenesis is multifaceted and is attributed to a combination of astrocyte swelling, an increase in glutamine synthesis, and a disproportionate ratio of aromatic amino acids to branched chain amino acids [ 14 – 16 ]. Based on our study results, we hypothesize that non-hepatic hyperammonemia may be associated with SAE. Unfortunately, in our study, some primary brain diseases (such as cerebral hemorrhage and cerebral infarction) and some secondary brain diseases (such as metabolic encephalopathy and pulmonary encephalopathy) were not excluded. The association between non-hepatic hyperammonemia and SAE needs to be validated in future well-designed experimental trials. Intestinal infection, UTI, and infection of other tissues by E. coli may be risk factors for non-hepatic hyperammonemia in patients with sepsis. Our results showed that the incidence of intestinal infections in the hyperammonemia group was 23.4% higher than in the non-hyperammonemia group. This is consistent with research by Wang et al., which found that in patients with infection-induced hepatic encephalopathy, levels of plasma ammonia were significantly higher in association with intestinal tract infection compared with other sites of infection. Their results, along with our findings, support the notion that intestinal infection is related to hyperammonemia [ 17 ]. A possible explanation for the link between intestinal infection and non-hepatic hyperammonemia is intestinal flora. Colonic bacteria have been known to produce ammonia from amino acid deamination or via urease, the hydrolysis of urea into carbon dioxide and ammonia [ 18 ]. When the body develops sepsis, the composition of intestinal microbes changes, due to factors such as antibiotic usage, systemic inflammation, and intestinal leakage [ 19 ]. In the patient's feces, the composition of the microbial components changes rapidly, the microbial diversity is largely lost, and the proportion of anaerobic bacteria significantly reduced and of Enterobacteriaceae increased [ 20 ]. Ammonia production is increased by converting nitrate to nitrite, and subsequently to ammonia [ 21 ]. Our results are consistent with previous studies, in which an increase in ammonia was associated with higher rates of infection by Enterobacteriaceae [ 3 , 13 , 22 ]. Therefore, serum ammonia should be measured when risk factors are present, such as intestinal infection or infection by E. coli . Our study showed that UTI is significantly associated with non-hepatic hyperammonemia in patients with sepsis, which is in line with the literature [ 23 – 25 ]. The possible explanation for the link between non-hepatic hyperammonemia and UTI is urease-producing bacteria and distal renal tubular acidosis [ 26 ]. With the entry of urea into the urinary tract, urease-producing bacteria form "ammonia," which results in alkalinization of the urine. The pH of the urine, when relatively high compared with that of the blood, enhances the diffusion of "ammonia" into the bloodstream [ 27 , 28 ]. Another plausible explanation for the linkage between hyperammonemia and UTI is distal renal tubular acidosis. Severe UTIs occasionally are accompanied by altered distal renal tubular function, which results in reduced bicarbonates, and in turn, leads to increased renal "ammonia" production [ 29 ]. The last explanation could be urinary retention associated with a neurogenic bladder. As the pressure in the bladder increases, the area of the bladder expands and promotes drainage of more ammonia directly into the inferior vena cava via the internal iliac veins [ 30 ]. Therefore, in patients with UTIs, serum ammonia levels should be closely monitored and timely measures taken to reduce them. Several limitations of the present study must be acknowledged. First, the result suggests a link between higher serum ammonia levels and lower GCS scores. Because of the nature of the retrospective analysis, the onset times of coma were not always available or documented, and some patients with primary and secondary encephalopathy in this study were not excluded. Therefore, whether there is a causal relationship between ammonia and SAE cannot be determined based on our results. Second, due to the limitations of the database, information was missing on some clinical variables, such as bilirubin, albumin, and intravenous nutrition. Inclusion of those data may have led to a more comprehensive understanding of the role of other biomarkers in sepsis with non-hepatic hyperammonemia. Third, our cohort study used ICD-9 diagnostic codes for sepsis, severe sepsis, and septic shock, but the concept of severe sepsis was eliminated in Sepsis 3.0, which may have led to bias in our research results. Conclusions Non-hepatic hyperammonemia is associated with mortality in patients with sepsis. The present study was essentially a pilot that requires validation. We recommend that serum ammonia levels be measured in patients who have risk factors, such as intestinal infection, UTI, and E. coli infection. Infection caused by E. coli is a potential biomarker for sepsis in patients who have non-hepatic hyperammonemia. Our study also demonstrated a correlation between non-hepatic hyperammonemia and an increased risk of SAE. Supplementary Information Supplementary Table 1 Exclusion of patients with acute and chronic liver disease from the MIMIC-III database according to International Classification of Diseases, Ninth Revision codes. ICD9-code Description 700 Hepatitis A with coma 0701 Viral hepatitis A without mention of hepatic coma 07020 Viral hepatitis B with hepatic coma, acute or unspecified, without mention of hepatitis delta 07021 Viral hepatitis B with hepatic coma, acute or unspecified, with hepatitis delta 07022 Chronic viral hepatitis B with hepatic coma without hepatitis delta 07023 Chronic viral hepatitis B with hepatic coma with hepatitis delta 07030 Viral hepatitis B without mention of hepatic coma, acute or unspecified, without mention of hepatitis 07031 Viral hepatitis B without mention of hepatic coma, acute or unspecified, with hepatitis delta 07032 Chronic viral hepatitis B without mention of hepatic coma without mention of hepatitis delta 07033 Chronic viral hepatitis B without mention of hepatic coma with hepatitis delta 07041 Acute hepatitis C with hepatic coma 07042 Hepatitis delta without mention of active hepatitis B disease with hepatic coma 07043 Hepatitis E with hepatic coma 07044 Chronic hepatitis C with hepatic coma 07049 Other specified viral hepatitis with hepatic coma 07051 Acute hepatitis C without mention of hepatic coma 07052 Hepatitis delta without mention of active hepatitis B disease or hepatic coma 07053 Hepatitis E without mention of hepatic coma 07054 Chronic hepatitis C without mention of hepatic coma 07059 Other specified viral hepatitis without mention of hepatic coma 0706 Unspecified viral hepatitis with hepatic coma 07070 Unspecified viral hepatitis C without hepatic coma 07071 Unspecified viral hepatitis C with hepatic coma 0709 Unspecified viral hepatitis without mention of hepatic coma 5712 Alcoholic cirrhosis of liver 5713 Alcoholic liver damage, unspecified 57140 Chronic hepatitis, unspecified 57141 Chronic persistent hepatitis 57142 Autoimmune hepatitis 57149 Other chronic hepatitis 5715 Cirrhosis of liver without mention of alcohol 5716 Biliary cirrhosis 5718 Other chronic nonalcoholic liver disease 5719 Unspecified chronic liver disease without mention of alcohol 5722 Hepatic encephalopathy 5724 Hepatorenal syndrome 5728 Other sequelae of chronic liver disease 5738 Other specified disorders of liver 5735 Hepatopulmonary syndrome 5734 Hepatic infarction 5733 Hepatitis, unspecified 5732 Hepatitis in other infectious diseases classified elsewhere 5731 Hepatitis in viral diseases classified elsewhere 5730 Chronic passive congestion of liver V0260 Viral hepatitis carrier, unspecified V0261 Hepatitis B carrier V0262 Hepatitis C carrier V0269 Other viral hepatitis carrier 86400 Injury to liver without mention of open wound into cavity, unspecified injury 86401 Injury to liver without mention of open wound into cavity, hematoma and contusion 86402 Injury to liver without mention of open wound into cavity, laceration, minor 86403 Injury to liver without mention of open wound into cavity, laceration, moderate 86404 Injury to liver without mention of open wound into cavity, laceration, major 86405 Injury to liver without mention of open wound into cavity laceration, unspecified 86409 Other injury to liver without mention of open wound into cavity 86410 Injury to liver with open wound into cavity, unspecified injury 4560 Esophageal varices with bleeding 4561 Esophageal varices without mention of bleeding 45620 Esophageal varices in diseases classified elsewhere, with bleeding 45621 Esophageal varices in diseases classified elsewhere, without mention of bleeding Supplementary Table 2 Type of Microbiology type and org_itemid. org_itemid Description Microbiology type 80026 Pseudomonas aeruginosa 80155 Staphylococcus , coagulase negative 80223 Probable enterococcus 80023 Staph aureus coag + 80155 Staphylococcus , coagulase negative 80280 Viridans streptococci 80081 Gram positive bacteria 80075 Yeast 80004 Klebsiella pneumoniae 80060 Albicans 80254 Candida albicans , presumptive identification 80058 Gram negative rod(s) 80260 Positive for pneumocystis carinii 80002 Escherichia coli 80053 Enterococcus sp. 80293 Positive for methicillin resistant staph aureus 80168 Enterococcus faecium 80139 Clostridium difficile 80112 Bacteroides fragilis group 80087 Stenotrophomonas (xanthomonas) maltophilia 80066 Aspergillus fumigatus Supplementary Table 3 Type of infection and International Classification of Diseases, Ninth Revision codes. ICD9 code Description Intestinal infection 845 Intestinal infection due to Clostridium difficile 847 Intestinal infection due to other gram-negative bacteria 88 Intestinal infection due to other organism, not elsewhere classified 90 Infectious colitis, enteritis, and gastroenteritis 93 Diarrhea of presumed infectious origin 56081 Intestinal or peritoneal adhesions with obstruction (postoperative) (postinfection) 56982 Ulceration of intestine 56983 Perforation of intestine Urinary tract infection 5990 Urinary tract infection lung infection 322 Salmonella pneumonia 1160 Tuberculous pneumonia [any form], unspecified 1161 Tuberculous pneumonia [any form], bacteriological or histological examination not done 1162 Tuberculous pneumonia [any form], bacteriological or histological examination unknown (at present) 1163 Tuberculous pneumonia [any form], tubercle bacilli found (in sputum) by microscopy 1164 Tuberculous pneumonia [any form], tubercle bacilli not found (in sputum) by microscopy, but found by bacterial culture 1165 Tuberculous pneumonia [any form], tubercle bacilli not found by bacteriological examination, but tuberculosis confirmed histologically 413 Klebsiella pneumoniae 551 Postmeasles pneumonia 382 Pneumococcal septicemia [ Streptococcus pneumoniae septicemia] 11505 Histoplasm caps pneumon 11515 Infection by Histoplasma duboisii , pneumonia 11595 Histoplasmosis, unspecified, pneumonia 730 Ornithosis with pneumonia 48249 Other Staphylococcus pneumonia 48281 Pneumonia due to anaerobes 48282 Pneumonia due to Escherichia coli 48283 Pneumonia due to other gram-negative bacteria 4800 Pneumonia due to adenovirus 4801 Pneumonia due to respiratory syncytial virus 4802 Pneumonia due to parainfluenza virus 4803 Pneumonia due to SARS-associated coronavirus 4808 Pneumonia due to other virus not elsewhere classified 4809 Viral pneumonia, unspecified 481 Pneumococcal pneumonia [ Streptococcus pneumoniae pneumonia] 4820 Pneumonia due to Klebsiella pneumoniae 4821 Pneumonia due to Pseudomonas 4822 Pneumonia due to Hemophilus influenzae [ H. influenzae ] 48230 Pneumonia due to Streptococcus , unspecified 48231 Pneumonia due to Streptococcus , group A 48232 Pneumonia due to Streptococcus , group B 48239 Pneumonia due to other Streptococcus 48240 Pneumonia due to Staphylococcus , unspecified 48241 Methicillin susceptible pneumonia due to Staphylococcus aureus 48242 Methicillin resistant pneumonia due to Staphylococcus aureus 48284 Pneumonia due to Legionnaires' disease 48289 Pneumonia due to other specified bacteria 4829 Bacterial pneumonia NOS Bacterial pneumonia, unspecified 4830 Pneumonia due to mycoplasma pneumoniae 4831 Pneumonia due to chlamydia 4838 Pneumonia due to other specified organism 4841 Pneumonia in cytomegalic inclusion disease 4843 Pneumonia in whoop cough 4845 Pneumonia in anthrax 4846 Pneum in aspergillosis 4847 Pneumonia in other systemic mycoses 4848 Pneumonia in other infectious diseases classified elsewhere 485 Bronchopneumonia, organism unspecified 486 Pneumonia, organism unspecified 4870 Influenza with pneumonia 4871 Influenza with other respiratory manifestations 4878 Influenza with other manifestations 48801 Influenza due to identified avian influenza virus with pneumonia 48802 Influenza due to identified avian influenza virus with other respiratory manifestations 48809 Influenza due to identified avian influenza virus with other manifestations 48811 Influenza due to identified 2009 H1N1 influenza virus with pneumonia 48812 Influenza due to identified 2009 H1N1 influenza virus with other respiratory manifestations 48819 Influenza due to identified 2009 H1N1 influenza virus with other manifestations 48881 Influenza due to identified novel influenza A virus with pneumonia 48882 Influenza due to identified novel influenza A virus with other respiratory manifestations 48889 Influenza due to identified novel influenza A virus with other manifestations 51630 Idiopathic interstitial pneumonia, not otherwise specified 51635 Idiopathic lymphoid interstitial pneumonia 51636 Cryptogenic organizing pneumonia 51637 Desquamative interstitial pneumonia 5171 Rheumatic pneumonia 7700 Congenital pneumonia V066 Need for prophylactic vaccination and inoculation against streptococcus pneumoniae [pneumococcus] and influenza 99731 Ventilator associated pneumonia 99732 Postprocedural aspiration pneumonia V0382 Other specified vaccinations against Streptococcus pneumoniae [pneumococcus] 1166 Tuberculous pneumonia [any form], tubercle bacilli not found by bacteriological or histological examination, but tuberculosis confirmed by other methods [inoculation of animals] 3453 Grand mal status Supplementary Table 4 Type of disease and International Classification of Diseases, Ninth Revision codes. Disease ICD9-Code Description Gastrointestinal bleeding 5789 Hemorrhage of gastrointestinal tract, unspecified 5780 Hematemesis 5781 Blood in stool 5693 Hemorrhage of rectum and anus 4560 Esophageal varices with bleeding 45620 Esophageal varices in diseases classified elsewhere, with bleeding 53100 Acute gastric ulcer with hemorrhage, without mention of obstruction 53101 Acute gastric ulcer with hemorrhage, with obstruction 53120 Acute gastric ulcer with hemorrhage and perforation, without mention of obstruction 53121 Acute gastric ulcer with hemorrhage and perforation, with obstruction 53300 Acute peptic ulcer of unspecified site with hemorrhage, without mention of obstruction 53320 Acute peptic ulcer of unspecified site with hemorrhage and perforation, without mention of obstruction 53321 Acute peptic ulcer of unspecified site with hemorrhage and perforation, with obstruction 53200 Acute duodenal ulcer with hemorrhage, without mention of obstruction 53201 Acute duodenal ulcer with hemorrhage, with obstruction 53220 Acute duodenal ulcer with hemorrhage and perforation, without mention of obstruction 53221 Acute duodenal ulcer with hemorrhage and perforation, with obstruction 53400 Acute gastrojejunal ulcer with hemorrhage, without mention of obstruction 53401 Acute gastrojejunal ulcer, with hemorrhage, with obstruction 53420 Acute gastrojejunal ulcer with hemorrhage and perforation, without mention of obstruction 53421 Acute gastrojejunal ulcer with hemorrhage and perforation, with obstruction 53501 Acute gastritis, with hemorrhage Heart failure 4280 Congestive heart failure, unspecified 4281 Left heart failure 42830 Diastolic heart failure, unspecified 42831 Acute diastolic heart failure 42832 Chronic diastolic heart failure 42833 Acute on chronic diastolic heart failure 42840 Combined systolic and diastolic heart failure, unspecified 42841 Acute combined systolic and diastolic heart failure 42842 Chronic combined systolic and diastolic heart failure 42843 Acute on chronic combined systolic and diastolic heart failure 39891 Rheumatic heart failure (congestive) Kidney failure 5845 Acute kidney failure with lesion of tubular necrosis 5846 Acute kidney failure with lesion of renal cortical necrosis 5848 Acute kidney failure with other specified pathological lesion in kidney 5849 Acute kidney failure, unspecified 5852 Chronic kidney disease, Stage II (mild) 5853 Chronic kidney disease, Stage III (moderate) 5854 Chronic kidney disease, Stage IV (severe) 5855 Chronic kidney disease, Stage V 5856 End stage renal disease Supplementary Table 5 Outcome of patients in the hyperammonemia and non-hyperammonemia groups and International Classification of Diseases, Ninth Revision codes. ICD9-code Description Delirium 29041 Vascular dementia, with delirium 29043 Vascular dementia, with depressed mood 29281 Drug-induced delirium 2910 Alcohol withdrawal delirium 2930 Delirium due to conditions classified elsewhere Encephalopathy 4372 Hypertensive encephalopathy 34982 Toxic encephalopathy 34831 Metabolic encephalopathy 34830 Encephalopathy, unspecified 34839 Other encephalopathy Supplemrentary Table 6 Definition of sepsis based on International Classification of Diseases, Ninth Revision codes. ICD9-code Description 99591 Sepsis 99592 Severe sepsis 78552 Septic shock ICD-9 – International Classification of Diseases, Ninth Revision.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6313897/

Fast Tracks and Roadblocks for Zika Vaccines

In early 2014, a relatively obscure virus, the Zika virus, made headlines worldwide following an increase in the number of congenital malformations. Since then, research on Zika virus, treatment and vaccines have progressed swiftly with various drugs being repurposed and vaccines heading into clinical trials. Nonetheless, the need for a vaccine is crucial in order to eradicate this re-emerging arthropod-borne virus which remained silent since its first discovery in 1947. In this review, we focused on how the inconspicuous virus managed to spread, the key immunological factors required for a vaccine and the various vaccine platforms that are currently being studied. 1. History of Zika Virus Zika virus (ZIKV) was first isolated within the Zika forest in Uganda back in 1947 when a group of scientists working on yellow fever unexpectedly discovered an unknown new virus from a sentinel rhesus macaque [ 1 ]. It was then characterized and classified as an arthropod-borne virus belonging to the Flaviviridae family and Flavivirus genus. Years later, the first confirmed human ZIKV infection case came in 1952 when live virus was isolated from a patient in Eastern Nigeria [ 2 ]. Since then, it was confirmed that ZIKV is a human viral pathogen. The first recorded human outbreak of ZIKV originated on the Island of Yap in 2007 whereby an estimated 73% of the population were infected. Prior to this, there were only 14 reported human ZIKV infections [ 3 ]. Subsequently, in 2008, Foy et al., found evidence that human-to-human transmission of ZIKV was plausible through sexual intercourse [ 4 ]. Unfortunately, the start of 2012 ushered in a wave of various mosquito-borne diseases which included Zika, Chikungunya and Dengue virus in the Pacific islands [ 5 , 6 ]. A routine check of blood donations in French Polynesia found samples which were positive for ZIKV when tested using PCR, triggering a closer surveillance of the spread of ZIKV [ 7 ]. In November 2013, the first suspected case of perinatal transmission of ZIKV was reported in the French Polynesia where a newborn had symptoms of mild pruritic rash at the time of delivery with the mother displaying symptoms of ZIKV infection two weeks prior delivery [ 8 ]. Following the recent surge of ZIKV cases in 2013 and 2014, an increase in congenital cerebral malformations, brainstem dysfunction and microcephaly among fetuses and newborns were observed, however the cause was unknown [ 9 ]. It was only when the Brazilian Government reported a correlation between ZIKV virus infections with congenital malformations in November 2015 that triggered health authorities to keep a vigilant watch for this inconspicuous disease [ 10 ]. By February 2016, a public health emergency was declared following proof of the association between ZIKV, microcephaly and other neurologic disorders such as the Guillain-Barre syndrome. 2. Treatment Due to the massive outbreak of ZIKV infection in 2015 without a known cure, various groups have been working on repurposing FDA-approved drugs for treatment [ 11 , 12 , 13 ]. Repurposing drugs would save a lot of time spent on finding and testing new molecules as human pharmacological data could be found in the pharmacopoeia; however more studies are needed to address issues such as the concentration of drugs needed to kill 90% of the virus and the drug exposure levels in the relevant organs or tissues. There are currently no FDA approved drugs under investigation in clinical trials solely for the treatment of ZIKV infection [ 14 ]. Hence despite efforts to repurpose known drugs for treatment, there is still an urgent need for a vaccine against ZIKV due to its implications for a pregnant mother and child. 3. Challenges for a Zika Vaccine There are many challenges when designing a vaccine targeting ZIKV. Firstly, there is currently no perfect small animal model for pre-clinical testing. Wild-type (WT) mice are poorly susceptible to Zika infection and develop low and short-lived viremia without apparent pathologies [ 15 , 16 , 17 , 18 ]. This is due to the strong type I interferon response during the first days of infection [ 19 , 20 ]. Mice deficient in molecules of this pathway have been shown to sustain high viremia and display a large array of pathologies [ 16 ]. WT models includes the use of antibodies against the interferon alpha/beta receptor have also been used extensively as a model [ 16 ]. Thus, so far, non-human primates (NHP) which is one of the natural reservoirs of the virus are the best pre-clinical models as discussed in depth by Lee & Ng [ 19 ]. The cost and maintenance of testing vaccines on NHP is extremely high and this poses a hurdle in vaccine development. As ZIKV is linked to congenital malformations, one of the toughest hurdles to overcome is the need for a vaccine which is safe for pregnant women and their offspring. Another key challenge in designing a vaccine against ZIKV is the possible risk of antibody-dependent enhancement of heterologous flavivirus infection such as dengue due to the sequence and antigenic similarity between them. This has clearly been showcased in the clinical trials of a dengue vaccine developed by Sanofi [ 21 ]. 4. Structure of Zika Virus A clear understanding of structure-based functional analyses of the virus is needed in order to design a vaccine against ZIKV. ZIKV has a structure akin to other flaviviruses, an icosahedral symmetry with a radius of 220 à [ 22 ] ( Figure 1 ). The full genome of ZIKV contains 11,000 bases which is enclosed by a lipid membrane encoding three structural proteins i.e. capsid (C), pre-membrane (prM), envelope (E) and seven non-structural (NS) proteins i.e. NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5 [ 23 , 24 ]. The icosahedral shape of the virus is attributed to the C protein which comprises the viral capsid whilst the outer transmembrane is decorated with E and prM/M proteins [ 22 ], which is the most commonly used target in vaccine designs which is further discussed in this review. In a resting state, the immature virus contains 60 trimeric E:M heterodimers, however, these particles lose their spiky surface following maturation process and cleavage of prM [ 23 ]. Consequently, the fusion loop is exposed, allowing the particles to fuse and become infectious in the low pH environment in the endosomes. 5. Immunity Induced by ZIKV Vaccines Vaccination is the most effective strategy for combating infectious diseases and has saved millions of lives. A successful vaccine should be able to induce efficient immune responses and generate immunological memory to provide long lasting protection against the particular pathogen [ 25 ]. In this section, we discuss a brief overview of how anti-Zika vaccines should trigger an immune response leading to memory. Most of the vaccines against ZIKV today focus on the induction of long-lived neutralizing antibody (nAb) responses. This was based on the fact that prophylactic and curative treatments with monoclonal antibodies against Zika have been shown to protect mice or NHP against ZIKV challenges [ 26 , 27 ] Thus Zika vaccines should induce long-lasting memory B cells. B-cells internalize, and process viruses or vaccine molecules. These cells present viral antigen-derived peptides on their MHC II. After maximum activation and amplification, most B cells will die, and a fraction of B cells will be developed into B cell memory [ 25 , 28 , 29 ]. It has to be noted that a ZIKV vaccine will also need to induce good Zika-specific helper T cell responses, since activated T-helper recognizing MHC II/antigen complex found on B-cells promotes B cell multiplication, antibody maturation and secretion [ 25 ]. 6. Vaccine Formulations 6.1. Live Attenuated Virus (LAV) Vaccine Live attenuated virus vaccine is arguably one of the favored vaccination strategies due to their previous success with the yellow fever virus vaccine, YF-17D, in the 1930s. A single dose of YF-17D vaccine is able to induce high titers of nAb which confers protection on at least 95% of recipients [ 30 , 31 ]. This strategy has been employed with many other diseases such as polio, measles and mumps [ 18 ]. Moreover, the production of attenuated vaccine is cost effective and fairly simple in comparison to other vaccine strategies. Kwek et al., isolated a small-plaque ZIKV variant by passaging French Polynesian ZIKV isolate in Vero cells and in C6/36 cells prior to infection on BHK-21 cells [ 32 ]. These few passages resulted in a small proportion of genetic variants, of which smaller plaques were selected. The authors picked the smallest variant of ZIKV, encoded DN-2, which had a substitution of alanine to guanine at position 948. This corresponded to a methionine to valine change in the amino acid sequence of the membrane gene. Nonetheless, DN-2 was able to infect monocyte-derived APCs, human umbilical vein endothelial cells, human embryonic stem cell-derived endothelial progenitor cells and HuH-7cells at significantly lower levels in comparison to other ZIKV strains. Subsequently, A129 mice which were inoculated with DN-2 were able to survive a 10 4 PFU challenge without any weight loss or detectable viremia in the peripheral circulation. However, in a subsequent study, 40% of mice had low levels of DN-2 RNA in the testes 15 days post-challenge. The authors also looked at attenuated maternal-fetal transmission in A129 dams following inoculation and challenge. Fetuses of DN-2 infected dams did not show any pathological differences from the uninfected dams however, one of the fetuses did succumb to infection. Another issue concerning the use of LAV in this study is the stability of the virus. The authors noted that DN-2 is stable up to 7 passages, which questions the stability of the LAV that may have resulted in genetic variants in the following passages. Shan et al., manipulated the complementary deoxyribonucleic acid (cDNA) clone of the ZIKV Cambodian strain FSS13025 with various deletions to mutate the local RNA structure of the viral 3'UTR which was consequently transfected into Vero cells [ 33 , 34 , 35 ]. Immunogenicity and efficacy were then tested on AG129 mice where the authors showed that the nAb titers of mice immunized with mutated virus were comparable to those induced by the wildtype. In addition, mice inoculated with the mutated virus showed a delay in the peak viremia which was a 100-fold lower than the wild type (WT) [ 34 ]. Subsequently, the mice were challenged intraperitoneally with the Puerto Rican strain (PRVABC59 ZIKV). Interestingly, the authors found that mice immunized with mutated virus, produced four-fold higher levels of IFN-γ than those immunized with the wild type. Remarkably, none of the A129 mice which received the mutated virus showed any detectable viremia in the periphery circulation or organs following challenge. The results facilitated in narrowing down the groups of mutated viruses for their potential vaccine candidate i.e. virus with 10 nucleotide deletions (del-10). Subsequently, the authors evaluated the neurovirulence of del-10 through intracranial immunization of CD-1 pups with lower virus challenging doses. A dose as low as 10 IFU of WT virus resulted in 25% mortality whereas all the pups given del-10 survived. Finally, the authors exposed Aedes aegypti mosquitoes to artificial blood meals containing either WT virus or del-10. Following 7 days incubation, 56% of the engorged mosquitoes were infected with the WT virus but none with del-10 mutants [ 34 ]. In a follow up study, Shan et al. also tested the efficacy of del-10 to prevent vertical transmission in C57BL/6 female mice, following anti-IFNAR treatment [ 35 ]. At day 28 post immunization, the authors noted that all immunized mice developed high titers of nAb. Subsequently, the mice were mated and monitored for vaginal plugs. Pregnant mice were challenged with mouse adapted ZIKV African strain Dakar 41519 following anti-IFNAR antibody treatment. Mice vaccinated with a single dose of del-10, had a significant reduction in viremia and in viral loads in tissues. As expected, the levels of nAb titers were inversely correlated to the levels of infectious virus. The authors subsequently assessed if del-10 was able to prevent testicular damage in male mice. Three weeks-old A129 male mice were vaccinated with del-10 and subsequently challenged with PRVABC59. No viremia was detected in mice immunized with del-10 unlike the sham vaccinated mice. In addition, there was no reduction in total sperm or motile sperm counts in comparison to healthy uninfected mice. Results with vaccinated male mice were in total contrast with the sham vaccinated mice, which saw a significant reduction in both total and motile sperm counts. In addition, the pathology of the testes also differed whereby mice vaccinated with del-10 did not show any reduction in size or weight unlike the sham treated mice. Interestingly, these results were not reproducible in older male mice as the sham vaccinated mice did not show any weight or size reduction. The others then tested their vaccine candidate on non-human primates. Primarily, the authors evaluated the level of attenuation of their vaccine and found that only one of the four rhesus macaques showed detectable viremia. Subsequently, the authors noted that the levels of nAb titers elicited in macaques vaccinated with del-10, were comparatively lower than those inoculated with the WT ZIKV. However, both groups of macaques did not display any viremia following challenge at day 56. This was in contrast to PBS vaccinated group where high levels of viremia were detected. Neutralizing antibody titers increased between 10–100 folds in mice vaccinated with del-10 however, no significant increase was seen in the group given WT ZIKV. As del-10 was unable to confer sterilizing immunity in macaques, the authors tested another candidate, annotated as del-20 (containing a 20 nucleotide deletion) as a second LAV candidate. The same sets of experiments were performed to test the safety and immunogenicity of the vaccine candidate, del-20. Male A129 mice were immunized with del-20 and by day 6, all mice had exhibited lower virus titers in comparison to WT ZIKV inoculated mice. This was as expected as del-20 was previously shown to be less attenuated than del-10 as it was less sensitive to type I interferon inhibition. Similar to the results of del-10, when young (3-week-old) male A129 mice were challenged with PRVABC59 following immunization, the mice did not show any loss of sperm counts, zero detectable viremia in testis or any changes in testis pathology. Interestingly, when neonate CD-1 mice were subjected to 10 4 FFU of del-20 intracranially, there was a mortality rate of 29% whereas all neonates survived when given 10 3 FFU of del-20. Aedes aegypti mosquitoes were then fed to a blood meal containing WT ZIKV and del-20. 50% of mosquitoes fed with WT were infected with ZIKV whereas none showed any viral RNA when fed with del-20. Finally, the authors proceeded with testing del-20 on macaques. Two of the three macaques that were vaccinated with del-20 subcutaneously exhibited low levels of viremia in various organs. Nonetheless, these NHPs soon developed high nAb titers by day 10 following vaccination. Subsequently, the three macaques were challenged with PRVABC59 on day 56 and displayed no viremia which the authors concluded that del-20 induces sterilizing immunity. In conclusion, a live attenuated vaccine undoubtedly has its advantages of able to elicit strong immune responses with a single dose, however, there are certain drawbacks to using live attenuated vaccines as well. These include its limited use in immunocompromised or pregnant patients due to the risk of adverse effects. Hence there is still a need for other avenues in order to provide complementary options for controlling ZIKV infections. 6.2. Messenger RNA (mRNA) Vaccines The use of mRNA vaccines is a relatively new trend that has gained popularity after it first made headlines in the 1990's. As the minimal genetic construct, mRNA is containing only the elements required for expression of the specific encoded protein region. In addition, mRNA is incapable of interacting with the genome but acts only as a transient carrier of information. Other advantages for its use as vaccine platforms include its safety profile; as it is non-immunogenic, there is no risk of potential infection unlike LAV [ 36 ]. 2017 saw an influx of new mRNA vaccines coming to play in targeting the disease, however, one of the disadvantages of utilizing mRNA in an approach to vaccine design is its rapid degradation by ribonucleases [ 37 ]. This setback however, could be easily tackled with the use of delivery systems as discussed in the examples given below. Early in 2017, Pardi et al. designed a mRNA vaccine against ZIKV utilizing the prM-E glycoproteins of ZIKV H/PF/2013 with a modified nucleoside 1-methypseudourine which abrogate innate sensing and increases mRNA translation in vivo [ 38 ]. mRNA was subsequently packaged into lipid nanoparticles composed of ionizable cationic lipids, phosphatidylcholine, cholesterol and polyethylene glycol. The authors initially tested the efficacy of their vaccine candidate on C57BL/6 mice which were given ZIKV-prM-E mRNA LNP intradermally. By week 8, vaccinated mice were able to elicit high titers of IgG and nAb. Subsequently, similar studies were repeated in BALB/c mice whereby mice showed peak antibody production by week 8. The authors proceeded with a challenge study on the immunized BALB/c mice. Mice were challenged with PRVABC59 on either week 2 or week 20 following immunization. Mice immunized with mRNA LNP did not exhibit any detectable viremia in the blood whereas 8 out of 9 showed presence of viral RNA in the sham group. Consequently, the authors assessed the efficacy of their vaccine candidate on rhesus macaques. Animals were immunized i.d. with varying doses of the vaccine. Interestingly, there were no significant differences in E-protein specific IgG and nAb titers induced by the varying amounts of vaccine given to the various groups of macaques. A challenge with the PRVABC59 virus was conducted 5 weeks following immunization. Negative control animals which did not receive the vaccine were susceptible to infection by day 3 post infection. In contrast, all vaccinated animals, regardless of the amount of vaccine received, were protected against ZIKV infection with one exception. A macaque that received the highest dose of vaccine had low transient viral RNA detected on day 3 however, it resolved by day 5. Unsurprisingly, this macaque exhibited one of the lowest nAb titers amongst the group of vaccinated animals. Further studies need to be conducted with bigger sample size as a single blip in this study, had implications in determining correlates of protection. In 2017, Richner et al., engineered a lipid nanoparticle encapsulating modified mRNA encoding ZIKV prM/M and E genes from the Asian strain together with the signal sequence of human IgE (IgE sig -prM-E LNP) [ 39 ]. The vaccine candidate, IgE sig -prM-E LNP, was initially assessed for its efficacy in AG129 mice. Mice were injected intramuscularly (i.m.) with either a single or double dose of the vaccine. The groups receiving two doses were able to elicit high titers of sera nAb. Subsequently, these mice were challenged with ZIKV P6-740 (Malaysian strain) on day 42. Unsurprisingly, mice which received two doses of the vaccine all survived with the exception of a single animal. Subsequently, the authors tested the vaccine on immunocompetent C57BL/6 mice. Eight week old male mice were given two doses of the vaccine prior to receiving anti-ifnar1 antibody and challenged. All mice survived without detection of viremia post infection. As the fusion loop (FL) in DII of the E protein is immunodominant in humans, there is a possible risk that vaccination with ZIKV full protein could elicit cross-reactive antibodies which may enhance other flavivirus infection through ADE. Hence, the authors generated mutations in the mRNA vaccine (IgE sig -prM-E FL LNP) and a series of Japanese encephalitis virus (JEV) leader sequence replacing the IgE in mRNA LNPs i.e. JEV sig -prM-E-FL LNP, to eliminate the antibody reactivity of FL-specific antibody. BALB/c mice vaccinated with JEV sig -prM-E LNP showed the highest level of protection (one mouse had viral RNA detected in the brain) following challenge with ZIKV Dakar 41,519 after anti-Ifnar1 treatment. Unfortunately, vaccine candidates with FL mutations were unable to confer full protection as viral RNA was detected in some animals albeit at much lower levels in comparison to the placebo LNPs. Similar to other studies, the authors found that levels of nAb were inversely correlated to the levels of ZIKV RNA in the sera and tissues. In order to evaluate whether the mutated vaccines were able to induce cross-reactive antibodies, sera of vaccinated mice were incubated with dengue virus serotype 1 (DENV-1) RVP prior overlaying on Fcγ receptor IIA-expressing K562 cells. As expected, sera from mice immunized with the FL mutation displayed a lower peak sera enhancement titer (PET) i.e., a reduction in cross-reactivity. Subsequently, authors passively transferred sera from vaccinated mice to AG129 mice 1 day prior to challenge with DENV-2. Unsurprisingly, at least 80% of mice that had sera from the FL mutations survived with lower morbidity clinical scores in comparison to those mice given the sera from full IgE sig -prM-E LNP. 6.3. DNA Vaccines DNA vaccines are of the earliest vaccine platforms to be proposed for human clinical trials following the ZIKV outbreak (clinicaltrials.gov). The use of genetically engineered DNA plasmids encoding various antigens to induce both humoral and cellular responses has been explored against various infectious diseases caused by parasites [ 40 , 41 ], bacteria [ 17 , 42 ] and other viruses [ 43 , 44 ]. Here, we explore the recent advances of DNA vaccines against ZIKV infection focusing on those that have reached studies on NHPs and human clinical trials. Larocca et al. designed their DNA vaccine based on the pre-membrane and envelope (prM-Env) of the Brazil strain BeH815744, scrapping the first 93 amino acids of prM [ 45 ]. The authors also tested a few other vaccine candidates, carrying various mutations in the plasmid DNA prM-Env i.e. lacking prM and/or lacking the transmembrane region or the full stem of Env, in parallel with their initial candidate. The immunogenicity of the various DNA plasmids was assessed using BALB/c mice. Mice given prM-Env vaccine intravascularly was able to elicit higher Env-specific IgG titers in comparison to all other groups. In addition, the prM-Env vaccine was also able to induce ZIKV-specific nAb and Env-specific CD8 + and CD4 + T lymphocyte responses. Subsequently, a challenge study was performed to determine the protective efficacy of DNA vaccine against a challenge using ZIKV PRVABC59 in addition to Brazil/ZKV2015. Mice that were vaccinated with the complete prM-Env were completely protected following challenge from the two different strains of ZIKV whereas the mice vaccinated with mutated/truncated versions of the DNA vaccine, did not show complete protection against ZIKV-BR. Furthermore, the authors tested the protective efficacy of the DNA vaccine on SJL and C57BL/6 mice which resulted in the same outcome whereby all mice given prM-Env DNA vaccine were fully protected. Subsequently, Larocca and colleagues passively transferred purified IgG sera from BALB/c vaccinated with prM-Env mice to naïve mice. Mice receiving sera Env-specific antibody titers of above log 2.35, were protected against ZIKV-BR challenge. Next, prM-Env vaccinated mice were depleted of CD8 + and CD4 + T lymphocytes prior challenge. Interestingly, depletion of T lymphocytes did not abrogate the protective efficacy of the prM-Env DNA vaccine upon challenge. There was no detectable viremia following challenge with ZIKV-BR however; viral RNA could be detected in the unvaccinated group. Abbink et al., extended this study by testing the protective efficacy of prM-Env DNA vaccine in rhesus macaques [ 46 ]. Monkeys received two doses of the DNA vaccines intramuscularly four weeks apart. Surprisingly, minimal nAb titers were only present on week 4 following primary immunization, however titers increased following the boost. Despite low levels of nAb detected in the animals, the monkeys had a 100% protection following a subcutaneous challenge with 10 3 PFU ZIKV-BR. Muthumani et al. reported on a similar vaccine construct targeting the pre-membrane and envelope proteins (prM-E) of ZIKV [ 47 ]. However, they added an IgE leader sequence to facilitate better expression similar to the concept used by Richner et al. [ 39 ] for their mRNA vaccine. The authors first tested the ability of the DNA vaccine to induce antigen specific T cells following intramuscular vaccination. Splenocytes were harvested and tested for their ability to secrete IFN-γ following ex vivo exposure to various peptide pools. Further evaluation of vaccine immunogenicity found that the vaccination increased the proportion of vaccine-specific T cells which expresses TNF-α and IFN-γ. Subsequently, the authors tested their vaccine efficacy by administering C57BL/6 mice with DNA vaccine intramuscularly in a prime boost regimen. ZIKV-E-specific IgG antibody titers were highest following a boosting, moreover titers were maintained even past day 60 post immunization. Subsequently, the vaccine efficacy was tested on IFNAR-deficient mice. Similarly, sera from vaccinated animals had detectable anti-ZIKV IgG by day 14 and also able to neutralize ZIKV though at relatively lower levels in comparison to C57BL/6 mice. Mice were subsequently challenged with PR209 (Puerto Rico) Zika virus following 2 doses of immunization. By the end of 3 weeks post challenge, all mice that were received the vaccine survived compared to mice control mice which had a 30% survival rate. Furthermore, the vaccinated animals did not show signs of morbidity unlike mice in the control where all animals were presented with either weight loss or struggling with mobility. Subsequently, the authors reduced the dose regimen by subjecting IFNAR-deficient mice to a single immunization prior challenge. Even though mice vaccinated with the DNA vaccine survived, viral RNA could be found in 40% of the mice. Muthumani et al. then tested their vaccine on NHPs. Macaques were immunized using intradermal electroporation with the DNA vaccine twice, two weeks apart. Sera and peripheral blood mononuclear cells (PBMCs) were collected on week 6 following immunization. PBMCs stimulated with prM-E peptide pools showed that ZIKV prM-E immunization was able to induce robust anti-ZIKV T cell responses in vaccinated RM. In addition, all vaccinated animals also had significant increase in nAb titers. The immune sera of macaques were able to prevent ZIKV infection in Vero, neuroblastoma and neural progenitor cells in vitro when tested via IFA. Sera from vaccinated Rhesus macaques were adoptively transferred into IFNAR-deficient mice a day prior to challenge. Mice which received control sera or PBS were all dead by day 8, however 80% of the mice receiving sera from vaccinated macaques survived. The study was further carried by a clinical trial (NCT02809443) [ 48 ], whereby the vaccine was administered intradermally in a prime-boost regimen consisting of 3 doses of vaccine. None of the patients reported any systemic adverse effect following vaccination. At 14 weeks post vaccination, all participants had measurable ZIKV E protein specific antibody titers; however, not all of the vaccinated individuals had nAb. Nonetheless, sera from the participants were passively transferred to IFNAR-deficient mice prior to challenge with a lethal dose of clinical strain ZIKV PR209 (Puerto Rico). Mice that received baseline sera or PBS, died within 9 days of infection. Whereas 92% of the mice receiving sera from vaccinated individuals survived past day 14. Interestingly, mice which survived included those which received sera from individuals with no detectable nAb. This provides further question for the use of nAb, determined in vitro, as a correlate of protection. Another DNA vaccine study conducted by Dowd et al. utilized the prM-E sequences inserted into the cytomegalovirus immediate promoter cloned into plasmid VRC8400 [ 49 ]. However, the difference between the constructs as published by Larocca et al. [ 45 ] and Abbink et al. [ 46 ], Dowd et al. included the whole prM sequence which was based on the French Polynesian isolate H/PF/2013 virus and also replaced the signal sequence of ZIKV prM with analogous region of JEV annotated as VRC5283. Based on VRC5283, Dowd et al. designed a second vaccine construct which replaced the final 98 amino acids of the stem and transmembrane regions of E protein with corresponding JEV sequence (VRC5288). Both vaccines were able to induce of ZIKV-specific nAb production following a single immunization. Subsequently, the authors optimized the dosage and regimen of their vaccine candidates. Macaques were administered either two doses of 1 mg VRC5283 or 4 mg of either VRC5283 or VRC5288. Another group of monkeys received a single 1 mg dose of VRC5288 and control animals were immunized with the empty vector. Animals which were vaccinated with either VRC5283 or VRC5288 had significantly higher nAb titers compared to the control. Moreover, macaques that received two doses of either VRC5283 or VRC5288 had nAb titers which were significantly higher than those that received a single dose. All animals were challenged 8 weeks following the first dose of vaccination with PRVABC59 ZIKV. Only animals which received double doses of vaccine were protected with the exception of one animal which had blips on day 3 and 7. All other animals i.e. those receiving just one vaccination and the control group, showed viremia by day 3. The vaccine constructs VRC5283 and VRC5288 advanced to phase 1 clinical trials (NCT02840487 and NCT02996461) [ 50 ]. None of the healthy volunteers whom were given either a two dose or three dose regimen reported any adverse side effects. Volunteers were either given the vaccine intramuscularly using a needle syringe or a needle free device (Stratis device, Pharmajet). Sera nAb titers could be detected 4 weeks following immunization. However, the only group where all volunteers showed detectable levels of nAb were those given three doses of VRC5283 using the needle-free device. This would have matched the NHP challenge model where the macaques which were administered with VRC5283 showed better protection compared to VRC5288. However, the authors did not utilize the needle free device when administering VRC5288 to any of the patient groups, which makes it difficult to compare the vaccine efficacy of the two different candidates in humans. 6.4. Adenovirus Vector Based Adenovirus vectors have been well studied as a vector for gene and cancer therapy and also vaccines whereby the vector expresses an unknown antigenic protein. Apart from its extensive safety profile, the advantages of utilizing adenovirus vectors are that it is relatively stable, easy to attain high titers and able to infect multiple cell lines which attributes to its potency. Xu et al. manipulated a recombinant chimpanzee adenovirus type 7 (AdC7) by cloning the gene encoding the signal peptide of JEV and full length ZIKV/M/E glycoproteins (AdC7-M/E) [ 51 ]. Vaccine immunogenicity was assessed in two groups of BALB/c mice immunized intramuscularly with either a high or low dose adenovirus particles of recombinant AdC7-M/E. A booster dose was performed 4 weeks following primary immunization. As E-specific antibody titers and nAb titers following primary immunization with high dose of virus particles were considerably high titer. There were no significant difference titers following a booster dose. Subsequently, sera from vaccinated BALB/c mice were tested for its ability to cross neutralize heterologous ZIKV strains, MR766, FSS13025 (Cambodia strain) and SMGC (Asian strain). The authors then tested the efficacy of the vaccine in mice lacking type 1 interferon. IFNAR-deficient mice were given a single dose of AdC7-M/E and antibody titers measured were of similar values as the BALB/c mice. Protection efficacy of the vaccine was evaluated by challenging intraperitoneally with ZIKV-SMGC IFNAR-deficient mice that have been immunized with one dose of vaccines. All sham mice succumbed to infection and died by day 8, however all vaccinated mice (low and high dose) survived with no weight loss. The authors then re-challenged the surviving vaccinated mice with a lethal dose of ZIKV. Astonishingly, there were no viral RNA detected in either the blood or organs (brain, spleen, spinal cord, testes, liver) of immunized mice. In contrast, sham immunized mice showed high levels of viral RNA in the blood and tissues, especially the testes whereby there was fibrosis of the tissues and presence of inflammatory cells. A passive immunization study was conducted to evaluate the protective efficacy of the vaccine. Sera from BALB/c immunized mice were passively transferred intraperitoneally to immunodeficient mice and subsequently challenged later with a lethal dose of ZIKV. All mice receiving sera from vaccinated animals survived, however a slight decrease in weight could be seen between days 5 and 8 post immunization but quickly rebounded and increase steadily after day 8. Sham-immunized mice succumbed to infection and died within 10 days. In an independent study conducted by Guo et al. the authors manipulated a more commonly used adenovirus, human type 5, expressing the pre-membrane and envelope proteins of ZIKV strain MRS_OPY_Martinique_PaRi_2015 [ 52 ]. Three different constructs were designed, with codon optimization to enhance transgene expression; Ad5-Luciferase (control), Ad5-Env and Ad5-Sig-prM-Env. The authors first evaluated the vaccine efficacy by giving mice a single dose of either vaccine or control (Ad5-Luc) intramuscularly. Vaccination with both vaccine constructs were able to induce anti ZIKV-Env-specific antibodies which were significantly higher than Ad5-Luc. Similarly, when the authors phenotyped T cells by flow cytometry and performed ELISPOT assay to evaluate the cellular immune responses elicited by the vaccines, the levels of IFN-γ-, TNF-α-, IL-2-, and CD107a-positive CD8 + T cells and IFN-γ- and TNF-α-positive CD4 + T cells in the vaccinated mice were significantly higher than the control group. Nonetheless, for both immune responses, there were no significant differences in antibody production or cytokine stimulation between the two vaccinated groups. Unlike most studies whereby authors use immunodeficient mice, Guo et al. subjected mice to corticosteroid (dexamethasone) therapy to mimic the scenario of patients with immunosuppression. The two vaccines were able to elicit similarly high levels of anti-ZIKV-Env-specific binding antibodies in sera at 2 weeks post vaccination even under glucocorticoid treatment. Interestingly, mice given Ad5-Sig-prM-Env had significantly higher titers of anti-ZIKV specific nAb which also explain the difference of viral RNA levels between the two vaccinated groups i.e. no detectable viral RNA in sera and tissues of Ad5-Sig-prM-Env vaccinated mice and low levels in the Ad5-Env vaccinated mice following challenge with PRVABC59. A similar result was seen when A129 mice were vaccinated and challenged. Interestingly though, both vaccinated groups including the negative control, showed weight loss, however, only mice vaccinated with Ad5-Sig-prM-Env had no detectable viremia in sera or tissues. In addition, all mice vaccinated with Ad5-Sig-prM-Env survived lethal challenge whereas only 83% of those vaccinated with Ad5-Env survived which could once again be due to higher levels of sera anti-ZIKV specific nAb in Ad5-Sig-prM-Env vaccinated mice. The authors subsequently passively transferred sera of vaccinated mice into naïve A129 mice prior to lethal challenge. Mice receiving sera from Ad5-Sig-prM-Env vaccinated mice had a 100% survival rate with minimal weight loss and zero detectable viral RNA in blood or tissues 6 days post infection. In contrast, mice receiving Ad5-Env only had 66% survival rate, with detectable viral RNA in blood and tissues and weight loss. Nonetheless, levels of detectable viral RNA were still significantly lower than those given sera of sham mice. This study was the only study that tried to mimic vaccination efficacy in immunosuppressed patients which is critical prior to testing the vaccine in phase 2 clinical trials. There were differences in protective efficacy of the two vaccine constructs, adenoviruses expressing prM-EnV and the -Env region, which suggests that the prM region is fairly important which may be resulted from the particulate form antigen (prM-EnV) and the non-particulate antigen (-Env). However, though there were no detectable viremia or viral RNA in the tissues, the mice did suffer weight loss following challenge, which suggests that the animals did succumb to infection but were able to recover. 6.5. Subunit Vaccines Subunit vaccines comprise of a fragment of a pathogen, i.e. protein, or peptides. Unlike live attenuated vaccines, subunit vaccines are generally a safer choice, however they tend to be less immunogenic as well. Hence an adjuvant and/or multiple doses are required. To et al. utilized a recombinant protein derived from the envelope protein of ZIKV (ZIKV E) based on the French-Polynesian strain [ 53 ]. Immunogenicity was tested in various mouse models including Swiss Webster (Swiss), BALB/c and also C57BL/6J mice. The authors initially tested their vaccine candidate, a ZIKV E-based recombinant protein, either alone or formulated with various adjuvants including alum-based adjuvants (Alhydrogel or Imject) or CoVaccine HT in Swiss mice. Mice that were immunized with CoVaccine HT elicited the highest ZIKV E IgG antibody which was significantly higher than those immunized with the ZIKV E together with alum adjuvants. Subsequently, when the authors tested their various vaccine candidates in BALB/c and C57BL/6 mice, both formulations i.e. alum and CoVaccine HT, of ZIKV E protein were able to induce high antibody titers with them showing no significant difference between the two formulations. Neutralizing antibodies were measured using a plaque reduction neutralization (PRNT) test following two doses of immunization of which the authors found comparable titers of nAb in both groups. Subsequently, protection studies were performed using BALB/c mice. Mice were intravenously challenged with PRVABC59 virus, four weeks post immunization. Groups that were given two doses of vaccines were fully protected unlike those given a single dose. Subsequently, authors tested if antisera from immunized mice would confer passive protection against ZIKV challenge. 10 µg of pooled sera, together with the various adjuvants, was administered intraperitoneally into BALB/c mice a day prior to challenge with ZIKV. The authors concluded there was a decrease in ZIKV E specific IgG antibody titers 14 days post infection in the group of mice receiving the antisera. Seroconversion was used as a gauge of virus infection in unprotected mice, therefore the authors looked at antibody production against ZIKV NS1. All animals presented with antibodies against ZIKV NS1 which suggested that mice receiving sera antibodies were unprotected. Another subunit vaccine derived from the ZIKV E protein of strain PRVABC59, was studied by Yang et al. [ 54 ]. Unlike the previous study, Yang et al. utilized a specific portion of the E protein i.e. the Domain III (DIII). ZIKV E DIII was produced in E. coli expression systems and subcutaneously administered into C57BL/6 mice together with adjuvant (alum or TitreMax, a water in oil adjuvant). Mice were given different doses of DIII. There was no significant difference in sera antibody titers between groups of mice given different doses of antigen following the third immunization. The groups of mice immunized with adjuvants were able to induce high titers of both IgG 1 and IgG 2c antibodies with a bias towards IgG 1 suggesting a Th-2 type response. Splenocytes from mice immunized with the vaccine candidates were able to induce IFN-γ, IL-6 and IL-4 production in contrast to PBS immunized mice, when stimulated with either EDIII or non-specifically with Con A. For both vaccine candidates, more than 80% of the virus was neutralized in vitro. Subsequently, the authors tested the possibility of antibody dependent enhancement activity of IgGs by incubating sera of immunized mice with DENV-2 E domain protein. The mixture was subsequently incubated with K562 cells displaying the human FcγR. The authors concluded that the purified IgGs from sera of mice immunized with ZIKV D III vaccine did not display any ADE activity which is comparable to the negative control. However further studies in in vivo models are needed to prove the vaccine is able to confer protection. Tai et al. conducted a long-term vaccination study using various fragments of recombinant ZIKV EDIII (ZikaSPH2015 strain) which were codon optimized as vaccine candidates [ 55 ]. To test the efficacy of their vaccine candidates, BALB/c mice were given four doses of the recombinant protein together with alum or monophosphoryl lipid A (MPLA). Sera from immunized mice were collected and IgG antibody titers were determined. All of the vaccine constructs tested were able to induce nAb following immunization. However, nAb titers following immunization of recombinant protein E298-409 were significantly higher than the other two; E296-406 and E301-404. Subsequently, long term antibody titers were evaluated. Sera obtained at 10 months post immunization showed that all vaccine constructs were able to elicit long term EDIII however, constructs E298-409 and E301-404 were able to significantly induce higher E-specific IgG titers in comparison to E296-406. The authors then tested if sera from immunized mice were able to protect pups against ZIKV infection. Pups born to BALB/c mice 7 months post immunization, were challenged with two human epidemic strains FLR (Colombia) and R103451 (Honduras). All pups which were born to mice immunized with ZIKV EDIII fragments survived except for those which were born to those immunized with E301-404 which only saw 83% survival when challenged with the Honduras strain. In contrast, pups from PBS immunized mice died. Next, the authors conducted an adoptive transfer study using sera from mice immunized with E298-409, which gave the highest titers of nAb. Seven-day old BALB/c pups were given immune sera and challenged with ZIKV FLR and R103451. Unfortunately, only 80% of pups survived in both challenged groups. Consequently, the authors tested the protective efficacy of sera obtained from adult immunized mice transferred to A129 mice. Anti-E298-409 sera equal to 10 5 ZIKV EDIII-specific IgG titer was transferred to A129 mice. A 100% survival was seen when the challenged with both ZIKV strains. However, viral RNA was present in the brain, lung, liver, spleen and kidney 5 days post infection in addition to tissue damage. The overall study identified that E298-409 as a promising vaccine candidate which is able to stimulate high titers of maternal nAb and confer protection to newborns. However, it is not known if maternal Ab are able to confer protection to the fetus if the pregnant female mice were subsequently challenged with ZIKV. This would have been an important study as we know that ZIKV is able to cross the placenta and cause fetal abnormality. 6.6. Combinatorial Vaccines While the development of single ZIKV vaccine is ongoing, a multiple antigenic approach following the measles mumps, rubella and varicella (MMRV) vaccine was being explored by Chattopadhyay and colleagues [ 56 ]. In their combinatorial vaccine against Chikungunya virus (CHIKV) and ZIKV, they utilized recombinant vesicular stomatitis virus (VSV) expressing CHIKV envelope polyprotein and ZIKV E protein. Sera from BALB/C mice immunized with a single dose of 10 7 PFU of recombinant VSV vaccine were able to neutralize 70% of ZIKV (Brazilian strain PE243). Whereas those which were given two doses of the vaccine were able to neutralize 80% of ZIKV. In addition, sera from the single immunization of BALB/C were also able to neutralize 100% of VSVΔG-eGFP/CHIKV pseudotype. In order to determine the protective efficacy of the vaccine, 7 week old A129 mice were immunized with 1 intramuscularly prior to challenge with either MR 766 Zika virus or CHIKV. Interestingly, none of the vaccinated mice showed signs of viremia following infection with either viruses. Although the authors proved that the mice were free from infection, the mice were past 15 weeks old when challenged. In this particular murine model, it is known that ZIKV infection would not have resulted in death of mice, but they would only show transient signs of illness following infection. Nonetheless, the negative control mice given CHIKV succumbed to infection by day 3 whereas all immunized mice survived. Overall, the study managed to prove that the vaccine was able to prevent viremia in immunocompromised mice, though the authors did not mention of any physical signs of infection. However, there was no proof that the vaccine would be useful in inducing production of maternal nAb in pregnant dams which is able to confer protection to newborns. In addition, it would be interesting to also study the possible deviations on the vaccine effects following co-infection with CHIKV and ZIKV. 6.1. Live Attenuated Virus (LAV) Vaccine Live attenuated virus vaccine is arguably one of the favored vaccination strategies due to their previous success with the yellow fever virus vaccine, YF-17D, in the 1930s. A single dose of YF-17D vaccine is able to induce high titers of nAb which confers protection on at least 95% of recipients [ 30 , 31 ]. This strategy has been employed with many other diseases such as polio, measles and mumps [ 18 ]. Moreover, the production of attenuated vaccine is cost effective and fairly simple in comparison to other vaccine strategies. Kwek et al., isolated a small-plaque ZIKV variant by passaging French Polynesian ZIKV isolate in Vero cells and in C6/36 cells prior to infection on BHK-21 cells [ 32 ]. These few passages resulted in a small proportion of genetic variants, of which smaller plaques were selected. The authors picked the smallest variant of ZIKV, encoded DN-2, which had a substitution of alanine to guanine at position 948. This corresponded to a methionine to valine change in the amino acid sequence of the membrane gene. Nonetheless, DN-2 was able to infect monocyte-derived APCs, human umbilical vein endothelial cells, human embryonic stem cell-derived endothelial progenitor cells and HuH-7cells at significantly lower levels in comparison to other ZIKV strains. Subsequently, A129 mice which were inoculated with DN-2 were able to survive a 10 4 PFU challenge without any weight loss or detectable viremia in the peripheral circulation. However, in a subsequent study, 40% of mice had low levels of DN-2 RNA in the testes 15 days post-challenge. The authors also looked at attenuated maternal-fetal transmission in A129 dams following inoculation and challenge. Fetuses of DN-2 infected dams did not show any pathological differences from the uninfected dams however, one of the fetuses did succumb to infection. Another issue concerning the use of LAV in this study is the stability of the virus. The authors noted that DN-2 is stable up to 7 passages, which questions the stability of the LAV that may have resulted in genetic variants in the following passages. Shan et al., manipulated the complementary deoxyribonucleic acid (cDNA) clone of the ZIKV Cambodian strain FSS13025 with various deletions to mutate the local RNA structure of the viral 3'UTR which was consequently transfected into Vero cells [ 33 , 34 , 35 ]. Immunogenicity and efficacy were then tested on AG129 mice where the authors showed that the nAb titers of mice immunized with mutated virus were comparable to those induced by the wildtype. In addition, mice inoculated with the mutated virus showed a delay in the peak viremia which was a 100-fold lower than the wild type (WT) [ 34 ]. Subsequently, the mice were challenged intraperitoneally with the Puerto Rican strain (PRVABC59 ZIKV). Interestingly, the authors found that mice immunized with mutated virus, produced four-fold higher levels of IFN-γ than those immunized with the wild type. Remarkably, none of the A129 mice which received the mutated virus showed any detectable viremia in the periphery circulation or organs following challenge. The results facilitated in narrowing down the groups of mutated viruses for their potential vaccine candidate i.e. virus with 10 nucleotide deletions (del-10). Subsequently, the authors evaluated the neurovirulence of del-10 through intracranial immunization of CD-1 pups with lower virus challenging doses. A dose as low as 10 IFU of WT virus resulted in 25% mortality whereas all the pups given del-10 survived. Finally, the authors exposed Aedes aegypti mosquitoes to artificial blood meals containing either WT virus or del-10. Following 7 days incubation, 56% of the engorged mosquitoes were infected with the WT virus but none with del-10 mutants [ 34 ]. In a follow up study, Shan et al. also tested the efficacy of del-10 to prevent vertical transmission in C57BL/6 female mice, following anti-IFNAR treatment [ 35 ]. At day 28 post immunization, the authors noted that all immunized mice developed high titers of nAb. Subsequently, the mice were mated and monitored for vaginal plugs. Pregnant mice were challenged with mouse adapted ZIKV African strain Dakar 41519 following anti-IFNAR antibody treatment. Mice vaccinated with a single dose of del-10, had a significant reduction in viremia and in viral loads in tissues. As expected, the levels of nAb titers were inversely correlated to the levels of infectious virus. The authors subsequently assessed if del-10 was able to prevent testicular damage in male mice. Three weeks-old A129 male mice were vaccinated with del-10 and subsequently challenged with PRVABC59. No viremia was detected in mice immunized with del-10 unlike the sham vaccinated mice. In addition, there was no reduction in total sperm or motile sperm counts in comparison to healthy uninfected mice. Results with vaccinated male mice were in total contrast with the sham vaccinated mice, which saw a significant reduction in both total and motile sperm counts. In addition, the pathology of the testes also differed whereby mice vaccinated with del-10 did not show any reduction in size or weight unlike the sham treated mice. Interestingly, these results were not reproducible in older male mice as the sham vaccinated mice did not show any weight or size reduction. The others then tested their vaccine candidate on non-human primates. Primarily, the authors evaluated the level of attenuation of their vaccine and found that only one of the four rhesus macaques showed detectable viremia. Subsequently, the authors noted that the levels of nAb titers elicited in macaques vaccinated with del-10, were comparatively lower than those inoculated with the WT ZIKV. However, both groups of macaques did not display any viremia following challenge at day 56. This was in contrast to PBS vaccinated group where high levels of viremia were detected. Neutralizing antibody titers increased between 10–100 folds in mice vaccinated with del-10 however, no significant increase was seen in the group given WT ZIKV. As del-10 was unable to confer sterilizing immunity in macaques, the authors tested another candidate, annotated as del-20 (containing a 20 nucleotide deletion) as a second LAV candidate. The same sets of experiments were performed to test the safety and immunogenicity of the vaccine candidate, del-20. Male A129 mice were immunized with del-20 and by day 6, all mice had exhibited lower virus titers in comparison to WT ZIKV inoculated mice. This was as expected as del-20 was previously shown to be less attenuated than del-10 as it was less sensitive to type I interferon inhibition. Similar to the results of del-10, when young (3-week-old) male A129 mice were challenged with PRVABC59 following immunization, the mice did not show any loss of sperm counts, zero detectable viremia in testis or any changes in testis pathology. Interestingly, when neonate CD-1 mice were subjected to 10 4 FFU of del-20 intracranially, there was a mortality rate of 29% whereas all neonates survived when given 10 3 FFU of del-20. Aedes aegypti mosquitoes were then fed to a blood meal containing WT ZIKV and del-20. 50% of mosquitoes fed with WT were infected with ZIKV whereas none showed any viral RNA when fed with del-20. Finally, the authors proceeded with testing del-20 on macaques. Two of the three macaques that were vaccinated with del-20 subcutaneously exhibited low levels of viremia in various organs. Nonetheless, these NHPs soon developed high nAb titers by day 10 following vaccination. Subsequently, the three macaques were challenged with PRVABC59 on day 56 and displayed no viremia which the authors concluded that del-20 induces sterilizing immunity. In conclusion, a live attenuated vaccine undoubtedly has its advantages of able to elicit strong immune responses with a single dose, however, there are certain drawbacks to using live attenuated vaccines as well. These include its limited use in immunocompromised or pregnant patients due to the risk of adverse effects. Hence there is still a need for other avenues in order to provide complementary options for controlling ZIKV infections. 6.2. Messenger RNA (mRNA) Vaccines The use of mRNA vaccines is a relatively new trend that has gained popularity after it first made headlines in the 1990's. As the minimal genetic construct, mRNA is containing only the elements required for expression of the specific encoded protein region. In addition, mRNA is incapable of interacting with the genome but acts only as a transient carrier of information. Other advantages for its use as vaccine platforms include its safety profile; as it is non-immunogenic, there is no risk of potential infection unlike LAV [ 36 ]. 2017 saw an influx of new mRNA vaccines coming to play in targeting the disease, however, one of the disadvantages of utilizing mRNA in an approach to vaccine design is its rapid degradation by ribonucleases [ 37 ]. This setback however, could be easily tackled with the use of delivery systems as discussed in the examples given below. Early in 2017, Pardi et al. designed a mRNA vaccine against ZIKV utilizing the prM-E glycoproteins of ZIKV H/PF/2013 with a modified nucleoside 1-methypseudourine which abrogate innate sensing and increases mRNA translation in vivo [ 38 ]. mRNA was subsequently packaged into lipid nanoparticles composed of ionizable cationic lipids, phosphatidylcholine, cholesterol and polyethylene glycol. The authors initially tested the efficacy of their vaccine candidate on C57BL/6 mice which were given ZIKV-prM-E mRNA LNP intradermally. By week 8, vaccinated mice were able to elicit high titers of IgG and nAb. Subsequently, similar studies were repeated in BALB/c mice whereby mice showed peak antibody production by week 8. The authors proceeded with a challenge study on the immunized BALB/c mice. Mice were challenged with PRVABC59 on either week 2 or week 20 following immunization. Mice immunized with mRNA LNP did not exhibit any detectable viremia in the blood whereas 8 out of 9 showed presence of viral RNA in the sham group. Consequently, the authors assessed the efficacy of their vaccine candidate on rhesus macaques. Animals were immunized i.d. with varying doses of the vaccine. Interestingly, there were no significant differences in E-protein specific IgG and nAb titers induced by the varying amounts of vaccine given to the various groups of macaques. A challenge with the PRVABC59 virus was conducted 5 weeks following immunization. Negative control animals which did not receive the vaccine were susceptible to infection by day 3 post infection. In contrast, all vaccinated animals, regardless of the amount of vaccine received, were protected against ZIKV infection with one exception. A macaque that received the highest dose of vaccine had low transient viral RNA detected on day 3 however, it resolved by day 5. Unsurprisingly, this macaque exhibited one of the lowest nAb titers amongst the group of vaccinated animals. Further studies need to be conducted with bigger sample size as a single blip in this study, had implications in determining correlates of protection. In 2017, Richner et al., engineered a lipid nanoparticle encapsulating modified mRNA encoding ZIKV prM/M and E genes from the Asian strain together with the signal sequence of human IgE (IgE sig -prM-E LNP) [ 39 ]. The vaccine candidate, IgE sig -prM-E LNP, was initially assessed for its efficacy in AG129 mice. Mice were injected intramuscularly (i.m.) with either a single or double dose of the vaccine. The groups receiving two doses were able to elicit high titers of sera nAb. Subsequently, these mice were challenged with ZIKV P6-740 (Malaysian strain) on day 42. Unsurprisingly, mice which received two doses of the vaccine all survived with the exception of a single animal. Subsequently, the authors tested the vaccine on immunocompetent C57BL/6 mice. Eight week old male mice were given two doses of the vaccine prior to receiving anti-ifnar1 antibody and challenged. All mice survived without detection of viremia post infection. As the fusion loop (FL) in DII of the E protein is immunodominant in humans, there is a possible risk that vaccination with ZIKV full protein could elicit cross-reactive antibodies which may enhance other flavivirus infection through ADE. Hence, the authors generated mutations in the mRNA vaccine (IgE sig -prM-E FL LNP) and a series of Japanese encephalitis virus (JEV) leader sequence replacing the IgE in mRNA LNPs i.e. JEV sig -prM-E-FL LNP, to eliminate the antibody reactivity of FL-specific antibody. BALB/c mice vaccinated with JEV sig -prM-E LNP showed the highest level of protection (one mouse had viral RNA detected in the brain) following challenge with ZIKV Dakar 41,519 after anti-Ifnar1 treatment. Unfortunately, vaccine candidates with FL mutations were unable to confer full protection as viral RNA was detected in some animals albeit at much lower levels in comparison to the placebo LNPs. Similar to other studies, the authors found that levels of nAb were inversely correlated to the levels of ZIKV RNA in the sera and tissues. In order to evaluate whether the mutated vaccines were able to induce cross-reactive antibodies, sera of vaccinated mice were incubated with dengue virus serotype 1 (DENV-1) RVP prior overlaying on Fcγ receptor IIA-expressing K562 cells. As expected, sera from mice immunized with the FL mutation displayed a lower peak sera enhancement titer (PET) i.e., a reduction in cross-reactivity. Subsequently, authors passively transferred sera from vaccinated mice to AG129 mice 1 day prior to challenge with DENV-2. Unsurprisingly, at least 80% of mice that had sera from the FL mutations survived with lower morbidity clinical scores in comparison to those mice given the sera from full IgE sig -prM-E LNP. 6.3. DNA Vaccines DNA vaccines are of the earliest vaccine platforms to be proposed for human clinical trials following the ZIKV outbreak (clinicaltrials.gov). The use of genetically engineered DNA plasmids encoding various antigens to induce both humoral and cellular responses has been explored against various infectious diseases caused by parasites [ 40 , 41 ], bacteria [ 17 , 42 ] and other viruses [ 43 , 44 ]. Here, we explore the recent advances of DNA vaccines against ZIKV infection focusing on those that have reached studies on NHPs and human clinical trials. Larocca et al. designed their DNA vaccine based on the pre-membrane and envelope (prM-Env) of the Brazil strain BeH815744, scrapping the first 93 amino acids of prM [ 45 ]. The authors also tested a few other vaccine candidates, carrying various mutations in the plasmid DNA prM-Env i.e. lacking prM and/or lacking the transmembrane region or the full stem of Env, in parallel with their initial candidate. The immunogenicity of the various DNA plasmids was assessed using BALB/c mice. Mice given prM-Env vaccine intravascularly was able to elicit higher Env-specific IgG titers in comparison to all other groups. In addition, the prM-Env vaccine was also able to induce ZIKV-specific nAb and Env-specific CD8 + and CD4 + T lymphocyte responses. Subsequently, a challenge study was performed to determine the protective efficacy of DNA vaccine against a challenge using ZIKV PRVABC59 in addition to Brazil/ZKV2015. Mice that were vaccinated with the complete prM-Env were completely protected following challenge from the two different strains of ZIKV whereas the mice vaccinated with mutated/truncated versions of the DNA vaccine, did not show complete protection against ZIKV-BR. Furthermore, the authors tested the protective efficacy of the DNA vaccine on SJL and C57BL/6 mice which resulted in the same outcome whereby all mice given prM-Env DNA vaccine were fully protected. Subsequently, Larocca and colleagues passively transferred purified IgG sera from BALB/c vaccinated with prM-Env mice to naïve mice. Mice receiving sera Env-specific antibody titers of above log 2.35, were protected against ZIKV-BR challenge. Next, prM-Env vaccinated mice were depleted of CD8 + and CD4 + T lymphocytes prior challenge. Interestingly, depletion of T lymphocytes did not abrogate the protective efficacy of the prM-Env DNA vaccine upon challenge. There was no detectable viremia following challenge with ZIKV-BR however; viral RNA could be detected in the unvaccinated group. Abbink et al., extended this study by testing the protective efficacy of prM-Env DNA vaccine in rhesus macaques [ 46 ]. Monkeys received two doses of the DNA vaccines intramuscularly four weeks apart. Surprisingly, minimal nAb titers were only present on week 4 following primary immunization, however titers increased following the boost. Despite low levels of nAb detected in the animals, the monkeys had a 100% protection following a subcutaneous challenge with 10 3 PFU ZIKV-BR. Muthumani et al. reported on a similar vaccine construct targeting the pre-membrane and envelope proteins (prM-E) of ZIKV [ 47 ]. However, they added an IgE leader sequence to facilitate better expression similar to the concept used by Richner et al. [ 39 ] for their mRNA vaccine. The authors first tested the ability of the DNA vaccine to induce antigen specific T cells following intramuscular vaccination. Splenocytes were harvested and tested for their ability to secrete IFN-γ following ex vivo exposure to various peptide pools. Further evaluation of vaccine immunogenicity found that the vaccination increased the proportion of vaccine-specific T cells which expresses TNF-α and IFN-γ. Subsequently, the authors tested their vaccine efficacy by administering C57BL/6 mice with DNA vaccine intramuscularly in a prime boost regimen. ZIKV-E-specific IgG antibody titers were highest following a boosting, moreover titers were maintained even past day 60 post immunization. Subsequently, the vaccine efficacy was tested on IFNAR-deficient mice. Similarly, sera from vaccinated animals had detectable anti-ZIKV IgG by day 14 and also able to neutralize ZIKV though at relatively lower levels in comparison to C57BL/6 mice. Mice were subsequently challenged with PR209 (Puerto Rico) Zika virus following 2 doses of immunization. By the end of 3 weeks post challenge, all mice that were received the vaccine survived compared to mice control mice which had a 30% survival rate. Furthermore, the vaccinated animals did not show signs of morbidity unlike mice in the control where all animals were presented with either weight loss or struggling with mobility. Subsequently, the authors reduced the dose regimen by subjecting IFNAR-deficient mice to a single immunization prior challenge. Even though mice vaccinated with the DNA vaccine survived, viral RNA could be found in 40% of the mice. Muthumani et al. then tested their vaccine on NHPs. Macaques were immunized using intradermal electroporation with the DNA vaccine twice, two weeks apart. Sera and peripheral blood mononuclear cells (PBMCs) were collected on week 6 following immunization. PBMCs stimulated with prM-E peptide pools showed that ZIKV prM-E immunization was able to induce robust anti-ZIKV T cell responses in vaccinated RM. In addition, all vaccinated animals also had significant increase in nAb titers. The immune sera of macaques were able to prevent ZIKV infection in Vero, neuroblastoma and neural progenitor cells in vitro when tested via IFA. Sera from vaccinated Rhesus macaques were adoptively transferred into IFNAR-deficient mice a day prior to challenge. Mice which received control sera or PBS were all dead by day 8, however 80% of the mice receiving sera from vaccinated macaques survived. The study was further carried by a clinical trial (NCT02809443) [ 48 ], whereby the vaccine was administered intradermally in a prime-boost regimen consisting of 3 doses of vaccine. None of the patients reported any systemic adverse effect following vaccination. At 14 weeks post vaccination, all participants had measurable ZIKV E protein specific antibody titers; however, not all of the vaccinated individuals had nAb. Nonetheless, sera from the participants were passively transferred to IFNAR-deficient mice prior to challenge with a lethal dose of clinical strain ZIKV PR209 (Puerto Rico). Mice that received baseline sera or PBS, died within 9 days of infection. Whereas 92% of the mice receiving sera from vaccinated individuals survived past day 14. Interestingly, mice which survived included those which received sera from individuals with no detectable nAb. This provides further question for the use of nAb, determined in vitro, as a correlate of protection. Another DNA vaccine study conducted by Dowd et al. utilized the prM-E sequences inserted into the cytomegalovirus immediate promoter cloned into plasmid VRC8400 [ 49 ]. However, the difference between the constructs as published by Larocca et al. [ 45 ] and Abbink et al. [ 46 ], Dowd et al. included the whole prM sequence which was based on the French Polynesian isolate H/PF/2013 virus and also replaced the signal sequence of ZIKV prM with analogous region of JEV annotated as VRC5283. Based on VRC5283, Dowd et al. designed a second vaccine construct which replaced the final 98 amino acids of the stem and transmembrane regions of E protein with corresponding JEV sequence (VRC5288). Both vaccines were able to induce of ZIKV-specific nAb production following a single immunization. Subsequently, the authors optimized the dosage and regimen of their vaccine candidates. Macaques were administered either two doses of 1 mg VRC5283 or 4 mg of either VRC5283 or VRC5288. Another group of monkeys received a single 1 mg dose of VRC5288 and control animals were immunized with the empty vector. Animals which were vaccinated with either VRC5283 or VRC5288 had significantly higher nAb titers compared to the control. Moreover, macaques that received two doses of either VRC5283 or VRC5288 had nAb titers which were significantly higher than those that received a single dose. All animals were challenged 8 weeks following the first dose of vaccination with PRVABC59 ZIKV. Only animals which received double doses of vaccine were protected with the exception of one animal which had blips on day 3 and 7. All other animals i.e. those receiving just one vaccination and the control group, showed viremia by day 3. The vaccine constructs VRC5283 and VRC5288 advanced to phase 1 clinical trials (NCT02840487 and NCT02996461) [ 50 ]. None of the healthy volunteers whom were given either a two dose or three dose regimen reported any adverse side effects. Volunteers were either given the vaccine intramuscularly using a needle syringe or a needle free device (Stratis device, Pharmajet). Sera nAb titers could be detected 4 weeks following immunization. However, the only group where all volunteers showed detectable levels of nAb were those given three doses of VRC5283 using the needle-free device. This would have matched the NHP challenge model where the macaques which were administered with VRC5283 showed better protection compared to VRC5288. However, the authors did not utilize the needle free device when administering VRC5288 to any of the patient groups, which makes it difficult to compare the vaccine efficacy of the two different candidates in humans. 6.4. Adenovirus Vector Based Adenovirus vectors have been well studied as a vector for gene and cancer therapy and also vaccines whereby the vector expresses an unknown antigenic protein. Apart from its extensive safety profile, the advantages of utilizing adenovirus vectors are that it is relatively stable, easy to attain high titers and able to infect multiple cell lines which attributes to its potency. Xu et al. manipulated a recombinant chimpanzee adenovirus type 7 (AdC7) by cloning the gene encoding the signal peptide of JEV and full length ZIKV/M/E glycoproteins (AdC7-M/E) [ 51 ]. Vaccine immunogenicity was assessed in two groups of BALB/c mice immunized intramuscularly with either a high or low dose adenovirus particles of recombinant AdC7-M/E. A booster dose was performed 4 weeks following primary immunization. As E-specific antibody titers and nAb titers following primary immunization with high dose of virus particles were considerably high titer. There were no significant difference titers following a booster dose. Subsequently, sera from vaccinated BALB/c mice were tested for its ability to cross neutralize heterologous ZIKV strains, MR766, FSS13025 (Cambodia strain) and SMGC (Asian strain). The authors then tested the efficacy of the vaccine in mice lacking type 1 interferon. IFNAR-deficient mice were given a single dose of AdC7-M/E and antibody titers measured were of similar values as the BALB/c mice. Protection efficacy of the vaccine was evaluated by challenging intraperitoneally with ZIKV-SMGC IFNAR-deficient mice that have been immunized with one dose of vaccines. All sham mice succumbed to infection and died by day 8, however all vaccinated mice (low and high dose) survived with no weight loss. The authors then re-challenged the surviving vaccinated mice with a lethal dose of ZIKV. Astonishingly, there were no viral RNA detected in either the blood or organs (brain, spleen, spinal cord, testes, liver) of immunized mice. In contrast, sham immunized mice showed high levels of viral RNA in the blood and tissues, especially the testes whereby there was fibrosis of the tissues and presence of inflammatory cells. A passive immunization study was conducted to evaluate the protective efficacy of the vaccine. Sera from BALB/c immunized mice were passively transferred intraperitoneally to immunodeficient mice and subsequently challenged later with a lethal dose of ZIKV. All mice receiving sera from vaccinated animals survived, however a slight decrease in weight could be seen between days 5 and 8 post immunization but quickly rebounded and increase steadily after day 8. Sham-immunized mice succumbed to infection and died within 10 days. In an independent study conducted by Guo et al. the authors manipulated a more commonly used adenovirus, human type 5, expressing the pre-membrane and envelope proteins of ZIKV strain MRS_OPY_Martinique_PaRi_2015 [ 52 ]. Three different constructs were designed, with codon optimization to enhance transgene expression; Ad5-Luciferase (control), Ad5-Env and Ad5-Sig-prM-Env. The authors first evaluated the vaccine efficacy by giving mice a single dose of either vaccine or control (Ad5-Luc) intramuscularly. Vaccination with both vaccine constructs were able to induce anti ZIKV-Env-specific antibodies which were significantly higher than Ad5-Luc. Similarly, when the authors phenotyped T cells by flow cytometry and performed ELISPOT assay to evaluate the cellular immune responses elicited by the vaccines, the levels of IFN-γ-, TNF-α-, IL-2-, and CD107a-positive CD8 + T cells and IFN-γ- and TNF-α-positive CD4 + T cells in the vaccinated mice were significantly higher than the control group. Nonetheless, for both immune responses, there were no significant differences in antibody production or cytokine stimulation between the two vaccinated groups. Unlike most studies whereby authors use immunodeficient mice, Guo et al. subjected mice to corticosteroid (dexamethasone) therapy to mimic the scenario of patients with immunosuppression. The two vaccines were able to elicit similarly high levels of anti-ZIKV-Env-specific binding antibodies in sera at 2 weeks post vaccination even under glucocorticoid treatment. Interestingly, mice given Ad5-Sig-prM-Env had significantly higher titers of anti-ZIKV specific nAb which also explain the difference of viral RNA levels between the two vaccinated groups i.e. no detectable viral RNA in sera and tissues of Ad5-Sig-prM-Env vaccinated mice and low levels in the Ad5-Env vaccinated mice following challenge with PRVABC59. A similar result was seen when A129 mice were vaccinated and challenged. Interestingly though, both vaccinated groups including the negative control, showed weight loss, however, only mice vaccinated with Ad5-Sig-prM-Env had no detectable viremia in sera or tissues. In addition, all mice vaccinated with Ad5-Sig-prM-Env survived lethal challenge whereas only 83% of those vaccinated with Ad5-Env survived which could once again be due to higher levels of sera anti-ZIKV specific nAb in Ad5-Sig-prM-Env vaccinated mice. The authors subsequently passively transferred sera of vaccinated mice into naïve A129 mice prior to lethal challenge. Mice receiving sera from Ad5-Sig-prM-Env vaccinated mice had a 100% survival rate with minimal weight loss and zero detectable viral RNA in blood or tissues 6 days post infection. In contrast, mice receiving Ad5-Env only had 66% survival rate, with detectable viral RNA in blood and tissues and weight loss. Nonetheless, levels of detectable viral RNA were still significantly lower than those given sera of sham mice. This study was the only study that tried to mimic vaccination efficacy in immunosuppressed patients which is critical prior to testing the vaccine in phase 2 clinical trials. There were differences in protective efficacy of the two vaccine constructs, adenoviruses expressing prM-EnV and the -Env region, which suggests that the prM region is fairly important which may be resulted from the particulate form antigen (prM-EnV) and the non-particulate antigen (-Env). However, though there were no detectable viremia or viral RNA in the tissues, the mice did suffer weight loss following challenge, which suggests that the animals did succumb to infection but were able to recover. 6.5. Subunit Vaccines Subunit vaccines comprise of a fragment of a pathogen, i.e. protein, or peptides. Unlike live attenuated vaccines, subunit vaccines are generally a safer choice, however they tend to be less immunogenic as well. Hence an adjuvant and/or multiple doses are required. To et al. utilized a recombinant protein derived from the envelope protein of ZIKV (ZIKV E) based on the French-Polynesian strain [ 53 ]. Immunogenicity was tested in various mouse models including Swiss Webster (Swiss), BALB/c and also C57BL/6J mice. The authors initially tested their vaccine candidate, a ZIKV E-based recombinant protein, either alone or formulated with various adjuvants including alum-based adjuvants (Alhydrogel or Imject) or CoVaccine HT in Swiss mice. Mice that were immunized with CoVaccine HT elicited the highest ZIKV E IgG antibody which was significantly higher than those immunized with the ZIKV E together with alum adjuvants. Subsequently, when the authors tested their various vaccine candidates in BALB/c and C57BL/6 mice, both formulations i.e. alum and CoVaccine HT, of ZIKV E protein were able to induce high antibody titers with them showing no significant difference between the two formulations. Neutralizing antibodies were measured using a plaque reduction neutralization (PRNT) test following two doses of immunization of which the authors found comparable titers of nAb in both groups. Subsequently, protection studies were performed using BALB/c mice. Mice were intravenously challenged with PRVABC59 virus, four weeks post immunization. Groups that were given two doses of vaccines were fully protected unlike those given a single dose. Subsequently, authors tested if antisera from immunized mice would confer passive protection against ZIKV challenge. 10 µg of pooled sera, together with the various adjuvants, was administered intraperitoneally into BALB/c mice a day prior to challenge with ZIKV. The authors concluded there was a decrease in ZIKV E specific IgG antibody titers 14 days post infection in the group of mice receiving the antisera. Seroconversion was used as a gauge of virus infection in unprotected mice, therefore the authors looked at antibody production against ZIKV NS1. All animals presented with antibodies against ZIKV NS1 which suggested that mice receiving sera antibodies were unprotected. Another subunit vaccine derived from the ZIKV E protein of strain PRVABC59, was studied by Yang et al. [ 54 ]. Unlike the previous study, Yang et al. utilized a specific portion of the E protein i.e. the Domain III (DIII). ZIKV E DIII was produced in E. coli expression systems and subcutaneously administered into C57BL/6 mice together with adjuvant (alum or TitreMax, a water in oil adjuvant). Mice were given different doses of DIII. There was no significant difference in sera antibody titers between groups of mice given different doses of antigen following the third immunization. The groups of mice immunized with adjuvants were able to induce high titers of both IgG 1 and IgG 2c antibodies with a bias towards IgG 1 suggesting a Th-2 type response. Splenocytes from mice immunized with the vaccine candidates were able to induce IFN-γ, IL-6 and IL-4 production in contrast to PBS immunized mice, when stimulated with either EDIII or non-specifically with Con A. For both vaccine candidates, more than 80% of the virus was neutralized in vitro. Subsequently, the authors tested the possibility of antibody dependent enhancement activity of IgGs by incubating sera of immunized mice with DENV-2 E domain protein. The mixture was subsequently incubated with K562 cells displaying the human FcγR. The authors concluded that the purified IgGs from sera of mice immunized with ZIKV D III vaccine did not display any ADE activity which is comparable to the negative control. However further studies in in vivo models are needed to prove the vaccine is able to confer protection. Tai et al. conducted a long-term vaccination study using various fragments of recombinant ZIKV EDIII (ZikaSPH2015 strain) which were codon optimized as vaccine candidates [ 55 ]. To test the efficacy of their vaccine candidates, BALB/c mice were given four doses of the recombinant protein together with alum or monophosphoryl lipid A (MPLA). Sera from immunized mice were collected and IgG antibody titers were determined. All of the vaccine constructs tested were able to induce nAb following immunization. However, nAb titers following immunization of recombinant protein E298-409 were significantly higher than the other two; E296-406 and E301-404. Subsequently, long term antibody titers were evaluated. Sera obtained at 10 months post immunization showed that all vaccine constructs were able to elicit long term EDIII however, constructs E298-409 and E301-404 were able to significantly induce higher E-specific IgG titers in comparison to E296-406. The authors then tested if sera from immunized mice were able to protect pups against ZIKV infection. Pups born to BALB/c mice 7 months post immunization, were challenged with two human epidemic strains FLR (Colombia) and R103451 (Honduras). All pups which were born to mice immunized with ZIKV EDIII fragments survived except for those which were born to those immunized with E301-404 which only saw 83% survival when challenged with the Honduras strain. In contrast, pups from PBS immunized mice died. Next, the authors conducted an adoptive transfer study using sera from mice immunized with E298-409, which gave the highest titers of nAb. Seven-day old BALB/c pups were given immune sera and challenged with ZIKV FLR and R103451. Unfortunately, only 80% of pups survived in both challenged groups. Consequently, the authors tested the protective efficacy of sera obtained from adult immunized mice transferred to A129 mice. Anti-E298-409 sera equal to 10 5 ZIKV EDIII-specific IgG titer was transferred to A129 mice. A 100% survival was seen when the challenged with both ZIKV strains. However, viral RNA was present in the brain, lung, liver, spleen and kidney 5 days post infection in addition to tissue damage. The overall study identified that E298-409 as a promising vaccine candidate which is able to stimulate high titers of maternal nAb and confer protection to newborns. However, it is not known if maternal Ab are able to confer protection to the fetus if the pregnant female mice were subsequently challenged with ZIKV. This would have been an important study as we know that ZIKV is able to cross the placenta and cause fetal abnormality. 6.6. Combinatorial Vaccines While the development of single ZIKV vaccine is ongoing, a multiple antigenic approach following the measles mumps, rubella and varicella (MMRV) vaccine was being explored by Chattopadhyay and colleagues [ 56 ]. In their combinatorial vaccine against Chikungunya virus (CHIKV) and ZIKV, they utilized recombinant vesicular stomatitis virus (VSV) expressing CHIKV envelope polyprotein and ZIKV E protein. Sera from BALB/C mice immunized with a single dose of 10 7 PFU of recombinant VSV vaccine were able to neutralize 70% of ZIKV (Brazilian strain PE243). Whereas those which were given two doses of the vaccine were able to neutralize 80% of ZIKV. In addition, sera from the single immunization of BALB/C were also able to neutralize 100% of VSVΔG-eGFP/CHIKV pseudotype. In order to determine the protective efficacy of the vaccine, 7 week old A129 mice were immunized with 1 intramuscularly prior to challenge with either MR 766 Zika virus or CHIKV. Interestingly, none of the vaccinated mice showed signs of viremia following infection with either viruses. Although the authors proved that the mice were free from infection, the mice were past 15 weeks old when challenged. In this particular murine model, it is known that ZIKV infection would not have resulted in death of mice, but they would only show transient signs of illness following infection. Nonetheless, the negative control mice given CHIKV succumbed to infection by day 3 whereas all immunized mice survived. Overall, the study managed to prove that the vaccine was able to prevent viremia in immunocompromised mice, though the authors did not mention of any physical signs of infection. However, there was no proof that the vaccine would be useful in inducing production of maternal nAb in pregnant dams which is able to confer protection to newborns. In addition, it would be interesting to also study the possible deviations on the vaccine effects following co-infection with CHIKV and ZIKV. 7. The Pros and Cons of Various Vaccine Platforms The most common vaccine platform used today is the live attenuated vaccine. This includes the chickenpox, rotavirus and the MMRV vaccines. The advantage of utilizing live attenuated vaccines includes its strong humoral and cellular mediated response triggered. Hence, typically a single dose of vaccine is needed. This proves useful in developing countries where the traveling time and cost of returning to the clinic may be a problem. Nonetheless, the platform is not without its major flaw where the organism may revert and cause more harm or immunosuppression [ 57 ]. This is a crucial point to consider in the context of a ZIKV vaccine as the more susceptible group to vaccinate would be the expectant mothers who already have a weakened immune system. Furthermore, it is known that ZIKV persists for months in the semen and urine [ 58 ], which undoubtedly pose a huge risk of LAV shedding into the semen. Unlike live attenuated vaccines, nucleic acid vaccines are non-immunogenic on their own yet they are also able to induce strong humoral and cellular mediated responses [ 59 ]. The major drawback of ribonucleic acid vaccines is its ability to avoid degradation by circulating ribonucleases. In order to deliver such vaccines, a delivery system is required. One major cause of concern is the possibility of an insertion of foreign genomic sequences into the host genome which may alter the resulting immune responses [ 60 ]. Even so, this is one of the more common vaccine platforms being used in the search for a ZIKV vaccine due to its safety profile. Recombinant subunit vaccines include those manipulating the viral/bacterial vector in order to express the protein(s) of interest. The advantage of utilizing recombinant vaccine lies in its safety profile and low production cost. However, due to its weak immunogenicity, there is a need for these vaccines to be delivered with a potent adjuvant to boost the immune responses triggered by the antigen [ 61 ]. Conversely, with recent advances, we now have live recombinant vectors e.g. adenoviral vectors, which assist in improving the immunogenicity of the vaccine. The recombinant adenoviral vectors are widely used today thanks to its high transduction efficiency and transgene expression. However, as most of the population have been exposed to adenovirus, there is a likelihood for pre-existing immunity against the vector. This has been proven detrimental in a human immunodeficiency virus (HIV-1) phase IIb vaccine trial whereby the vector-based vaccines provided favorable conditions for HIV-1 replication [ 62 ]. The risk of utilizing such recombinant viral vector vaccine may not be of concern for immunocompetent individuals, however, as the vaccine is more important to expectant mothers, the risk may outweigh the benefit. Unfortunately, we are unable to provide a detailed summary of every vaccine study that has been published, however, relevant references are provided in Table 1 . Nonetheless, each individual vaccine platform has its own pros and cons which may not be suitable for the entire population. However, as the essential targets for the vaccine are pregnant mothers, the need to ensure the safety profile is crucial to not harm the mother and fetus in addition to the vaccine being able to induce nAb production and prevent vertical transmission. 8. Conclusions The quest for a vaccine against ZIKV has progressed tremendously since the outbreak of 2014. However, one of the biggest conundrums in vaccine research is finding a vaccine that is applicable to the vast majority. Until today, most of the vaccine study focuses on the protective efficacy without considering the need to vaccinate immunosuppressed patients as well. As Zika research progresses, it is evident that the two different lineages found throughout the world have various infectivity and potency which needs to be considered when performing challenge experiments. The current trends of vaccines against ZIKV mainly concern the pre-membrane, membrane and envelope proteins as the target epitopes for vaccine candidates due to the natural structure of the virus. It is essential to note that most of these studies focus on a single vaccine candidate, with the more comprehensive approach performed by Abbink et al. which compared three different vaccine approaches i.e. purified inactivated vaccine, DNA vaccine and adenovirus vector vaccine [ 44 ] in a single study. Although many dwell on neutralizing antibodies as a predictor for protective efficacy, correlates of protection remain to be defined. As a number of vaccines are already being tested or completed the first phase in clinical trials, it is likely that a vaccine will be available in the coming future.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7065354/

Microecological Koch’s postulates reveal that intestinal microbiota dysbiosis contributes to shrimp white feces syndrome

Background Recently, increasing evidence supports that some complex diseases are not attributed to a given pathogen, but dysbiosis in the host intestinal microbiota (IM). The full intestinal ecosystem alterations, rather than a single pathogen, are associated with white feces syndrome (WFS), a globally severe non-infectious shrimp disease, while no experimental evidence to explore the causality. Herein, we conducted comprehensive metagenomic and metabolomic analysis, and intestinal microbiota transplantation (IMT) to investigate the causal relationship between IM dysbiosis and WFS. Results Compared to the Control shrimp, we found dramatically decreased microbial richness and diversity in WFS shrimp. Ten genera, such as Vibrio , Candidatus Bacilloplasma, Photobacterium , and Aeromonas , were overrepresented in WFS, whereas 11 genera, including Shewanella , Chitinibacter , and Rhodobacter were enriched in control. The divergent changes in these populations might contribute the observation that a decline of pathways conferring lipoic acid metabolism and mineral absorption in WFS. Meanwhile, some sorts of metabolites, especially lipids and organic acids, were found to be related to the IM alteration in WFS. Integrated with multiomics and IMT, we demonstrated that significant alterations in the community composition, functional potentials, and metabolites of IM were closely linked to shrimp WFS. The distinguished metabolites which were attributed to the IM dysbiosis were validated by feed-supplementary challenge. Both homogenous selection and heterogeneous selection process were less pronounced in WFS microbial community assembly. Notably, IMT shrimp from WFS donors eventually developed WFS clinical signs, while the dysbiotic IM can be recharacterized in recipient shrimp. Conclusions Collectively, our findings offer solid evidence of the causality between IM dysbiosis and shrimp WFS, which exemplify the 'microecological Koch's postulates' (an intestinal microbiota dysbiosis, a disease) in disease etiology, and inspire our cogitation on etiology from an ecological perspective. Video abstract Background Recently, increasing evidence supports that some complex diseases are not attributed to a given pathogen, but dysbiosis in the host intestinal microbiota (IM). The full intestinal ecosystem alterations, rather than a single pathogen, are associated with white feces syndrome (WFS), a globally severe non-infectious shrimp disease, while no experimental evidence to explore the causality. Herein, we conducted comprehensive metagenomic and metabolomic analysis, and intestinal microbiota transplantation (IMT) to investigate the causal relationship between IM dysbiosis and WFS. Results Compared to the Control shrimp, we found dramatically decreased microbial richness and diversity in WFS shrimp. Ten genera, such as Vibrio , Candidatus Bacilloplasma, Photobacterium , and Aeromonas , were overrepresented in WFS, whereas 11 genera, including Shewanella , Chitinibacter , and Rhodobacter were enriched in control. The divergent changes in these populations might contribute the observation that a decline of pathways conferring lipoic acid metabolism and mineral absorption in WFS. Meanwhile, some sorts of metabolites, especially lipids and organic acids, were found to be related to the IM alteration in WFS. Integrated with multiomics and IMT, we demonstrated that significant alterations in the community composition, functional potentials, and metabolites of IM were closely linked to shrimp WFS. The distinguished metabolites which were attributed to the IM dysbiosis were validated by feed-supplementary challenge. Both homogenous selection and heterogeneous selection process were less pronounced in WFS microbial community assembly. Notably, IMT shrimp from WFS donors eventually developed WFS clinical signs, while the dysbiotic IM can be recharacterized in recipient shrimp. Conclusions Collectively, our findings offer solid evidence of the causality between IM dysbiosis and shrimp WFS, which exemplify the 'microecological Koch's postulates' (an intestinal microbiota dysbiosis, a disease) in disease etiology, and inspire our cogitation on etiology from an ecological perspective. Video abstract Background The traditional 'Koch's postulates' (a pathogen, a disease) have successfully guided pathologists in identifying the causative agents of diverse infectious diseases [ 1 ]. Recently, increasing human and animal complex diseases, including non-infectious diseases and some diseases with a syndrome, are not fulfill these concepts, thereby prompting researchers to reconsider that the etiology is multifactorial [ 2 ]. Indeed, increasing recognitions on complex diseases (inflammatory bowel disease (IBD), atherosclerotic cardiovascular disease, etc.) illustrate that dysbiosis in intestinal microbiota (IM) contributes to host diseases [ 3 , 4 ]. IM plays fundamental roles in regulating host metabolic homeostasis, physiology, and health [ 3 , 5 , 6 ]. A few animal models have been used to study the microbe-host crosstalk by fecal microbiota transplantation (FMT) with colonization of specific microbial strains [ 7 ]. It is now common knowledge that the intestine is a complex ecosystem with different interacting entities and that infections must be understood in this context rather than isolated as a pathogen and a host. In a very recent review, the ecological Koch's postulates (a gut ecosystem state, a disease) were proposed [ 8 ]. The authors suggest that a whole ecosystem, including host IM, genetic make-up of the host, as well as nutrition and age, etc., forms an entity, ultimately leading to diseases, rather than an isolated microorganism or group of microorganisms [ 8 ]. To some extent, for some complex diseases (e.g., aquatic animal diseases) that lack of obvious evidence to be linked to specific mutations or age, but the microbial alterations and the complexity of surrounding environment, neither the Koch's postulates nor the ecological Koch's postulates are not sufficiently to interpret. We propose yet another interpretation of Koch's postulates, which we have termed 'microecological Koch's postulates' (an intestinal microbiota dysbiosis, a disease). Aquaculture is responsible for the continuing impressive growth in the supply for human consumption, which is the third largest source of animal protein that accounts for 17% protein consumed by the global population [ 9 ]. Pacific white shrimp, Litopenaeus vannamei , represents the largest production in shrimp industry (global production reached 4.1 million tons and valued at over $24 billion) [ 10 ]. However, shrimp production is being threatened by several diseases, such as early mortality syndrome (EMS) [ 11 ], acute hepatopancreatic necrosis disease (AHPND) [ 12 ], hepatopancreas necrosis syndrome (HPNS) [ 13 ], and white feces syndrome (WFS) [ 14 , 15 ], which cumulatively cause a devastating drop (60%) in shrimp production, of which WFS is the most severe and has drawn wide attention. WFS etiology has been widely concerned. The microsporidian was firstly proposed as causality of WFS [ 16 , 17 ], while this hypothesis was not supported by subsequent study [ 18 ]. Another survey on Penaeus monodon demonstrates that several Vibrio species may be the major causative pathogen of WFS, and the occurrence of WFS are related to the total Vibrio count in shrimp intestine [ 19 ]. By comparing the IM between healthy and WFS shrimp, increasing evidences have shown that the IM of WFS shrimp are less diverse and significantly different from those of healthy shrimp, and shrimp WFS is correlated with dysbiosis in IM [19, 20, 21]. Accordingly, WFS is a non-infectious disease that is not attributed by one pathogen, but a bacterial continuum. However, it is still unclear whether IM dysbiosis is a consequence or the causality of WFS. We therefore hypothesized that IM dysbiosis is the causality of WFS. To address the hypothesis above, in the present study, a comprehensive exploration of the IM aberrations in WFS shrimp was achieved by combining compositional, functional, and metabolic data. The causal link of IM dysbiosis and WFS was firstly validated by reciprocal intestinal microbiota transplantation (IMT) between healthy and WFS shrimp, which preferably fit the concept of 'microecological Koch's postulates.' These valuable findings greatly enhanced our understanding on the etiology in host disease from an ecological perspective. Results Clinical signs and histopathology of WFS The comprehensive exploration of the IM aberrations in WFS shrimp was performed following the simple workflow (Additional file 2 : Figure S1). The symptoms of WFS shrimp include weakened activity, no feed intake, and excreting white feces. White floating fecal strings were found at the water surface in ponds with WFS populations. The mid intestine of WFS shrimp was distended and filled with white contents, whereas that of healthy shrimp was brown and filled with feed (Fig. 1 a). Regarding to the histological pathology, the intestine of WFS shrimp contained epithelial cell detachment, reduced or disappeared microvilli, and thinner mid intestine (Fig. 1 b). Fig. 1 Characterization of the clinical signs, histological pathology, and microbial features of WFS. a The diseased shrimp excretes white feces. The intestine of WFS shrimp is distended and filled with white content. b Comparison of histological pathology of shrimp intestine with or without WFS signs. The black arrows point towards pathological features, including dropped epithelial cell, reduced or disappeared microvilli, and thinner mid intestine. c The α -diversity comparison between control group ( n = 75) and WFS group ( n = 84). Shannon index, P = 0.002; Simpson index, P  50% was stratified as WFS shrimp, while  50% was stratified as WFS shrimp, while  1.5, and P value of t test statistics  2.5 and Pi > 0.6; (II) module hubs: nodes with Zi > 2.5 and Pi ≤ 0.6; (III) connectors: nodes with Zi ≤ 2.5 and Pi > 0.6; and (IV) peripheral nodes: nodes with Zi ≤ 2.5 and Pi ≤ 0.665 [ 52 ], in which Zi and Pi respectively indicate how well a node connects to nodes within the same and other modules. The global network properties and individual node's centrality was then displayed by Cytoscape (Version 3.3.0). Estimation of ecological processes The mean nearest taxon distance (MNTD) was calculated to determine which processes govern the assembly of shrimp IM [ 53 ]. The obtained standardized effect size measure (ses.MNTD), which is also known as the negative nearest taxon index (NTI) to determine ecological processes that govern a community in terms of phylogenetic structures [ 54 , 55 ], was calculated with Picante package in R. The mean distance between each taxon and its nearest neighbor (β-MNTD) was computed by random shuffling of OTUs and their abundances across phylogenetic tips, reflecting the dissimilarity between communities [ 53 , 56 ]. Difference between the observed β-MNTD and the mean of the null distribution is referred as β-NTI. The β-NTI in combination with Bray–Curtis-based Raup–Crick (RC Bray ) was further used to quantify the relative contributions of major ecological processes that determine the assembly of IM [ 53 ]. The relative influence of community turnover that was determined by homogeneous and heterogeneous selection was denoted by β-NTI + 2 fractions, respectively [ 57 , 58 ]. The β-NTI and RC Bray were used to estimate the contribution of homogenizing dispersal and dispersal limitation. If |β-NTI| + 0.95 or  1.5, and P value of t test statistics  2.5 and Pi > 0.6; (II) module hubs: nodes with Zi > 2.5 and Pi ≤ 0.6; (III) connectors: nodes with Zi ≤ 2.5 and Pi > 0.6; and (IV) peripheral nodes: nodes with Zi ≤ 2.5 and Pi ≤ 0.665 [ 52 ], in which Zi and Pi respectively indicate how well a node connects to nodes within the same and other modules. The global network properties and individual node's centrality was then displayed by Cytoscape (Version 3.3.0). Estimation of ecological processes The mean nearest taxon distance (MNTD) was calculated to determine which processes govern the assembly of shrimp IM [ 53 ]. The obtained standardized effect size measure (ses.MNTD), which is also known as the negative nearest taxon index (NTI) to determine ecological processes that govern a community in terms of phylogenetic structures [ 54 , 55 ], was calculated with Picante package in R. The mean distance between each taxon and its nearest neighbor (β-MNTD) was computed by random shuffling of OTUs and their abundances across phylogenetic tips, reflecting the dissimilarity between communities [ 53 , 56 ]. Difference between the observed β-MNTD and the mean of the null distribution is referred as β-NTI. The β-NTI in combination with Bray–Curtis-based Raup–Crick (RC Bray ) was further used to quantify the relative contributions of major ecological processes that determine the assembly of IM [ 53 ]. The relative influence of community turnover that was determined by homogeneous and heterogeneous selection was denoted by β-NTI + 2 fractions, respectively [ 57 , 58 ]. The β-NTI and RC Bray were used to estimate the contribution of homogenizing dispersal and dispersal limitation. If |β-NTI| + 0.95 or < − 0.95, community turnover was governed by dispersal limitation or homogenizing dispersal processes [ 53 ]; the fraction of pairwise comparisons with |β-NTI| < 2 and |RC Bray | < 0.95 represented the component of compositional turnover was governed by undominated [ 59 ]. Statistics analysis To compare the difference between groups, statistical analysis was conducted by SPSS (Version 21.0). If data were normally distributed, a two-tailed Student's t test was used to determine significance in experiments with only two groups, and analysis of variance (ANOVA) with Tuckey test was used to determine significant differences between multiple test groups. In cases where data were not normally distributed, a one-way ANOVA with Kruskal-Wallis test was performed. MRPP, ANOSIM, and PerMANOVA were conducted to statistically test whether there is a significant difference between two groups with vegan package in R [ 60 , 61 ]. Permutational analysis of multivariate dispersions (PERMDISP) was for the analysis of multivariate homogeneity of group dispersions by vegan package in R [ 61 ]. Random forests regression was used to regress relative abundances of taxa in the temporal profiles of Control and WFS groups, using the following parameters with randomForest package in R (cv. fold = 10, step = 0.99, replication = 55, ntree = 5000). The POD value was defined as the ratio between the number of decision trees that was voted as 'WFS' and the number of total sampling trees ( n votes / n trees ). Spearman's rank correlation was conducted to measure the correlation of two variables in SPSS. Beta-diversity comparison of metagenomic and metabolome was accomplished by using Procrustes transformations with PCoA based on Bray-Curtis distance with vegan package in R. Supplementary information Additional file 1. Supplementary Tables S1-S10. Additional file 2. Supplementary Figures S1-S14. Additional file 1. Supplementary Tables S1-S10. Additional file 2. Supplementary Figures S1-S14. Supplementary information Supplementary information accompanies this paper at 10.1186/s40168-020-00802-3.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10856655/

Construction of Porphyrin-Based Bimetallic Nanomaterials with Photocatalytic Properties

The efficient synthesis of nanosheets containing two metal ions is currently a formidable challenge. Here, we attempted to dope lanthanide-based bimetals into porphyrin-based metal-organic skeleton materials (MOFs) by microwave-assisted heating. The results of the EDX, ICP, and XPS tests show that we have successfully synthesized porphyrin-based lanthanide bimetallic nanosheets (Tb-Eu-TCPP) using a household microwave oven. In addition, it is tested and experimentally evident that these nanosheets have a thinner thickness, a larger BET surface area, and higher photogenerated carrier separation efficiency than bulk porphyrin-based bimetallic materials, thus exhibiting enhanced photocatalytic activity and n-type semiconductor properties. Furthermore, the prepared Tb-Eu-TCPP nanomaterials are more efficient in generating single-linear state oxygen under visible light irradiation compared to pristine monometallic nanosheets due to the generation of bimetallic nodes. The significant increase in catalytic activity is attributed to the improved separation and transfer efficiency of photogenerated carriers. This study not only deepens our understanding of lanthanide bimetallic nanosheet materials but also introduces an innovative approach to improve the photocatalytic performance of MOFs. 1. Introduction Metal–organic frameworks (MOFs) [ 1 ], also referred to as coordination polymers, are hybrid materials featuring intramolecular pores formed through the self-assembly of organic ligands and metal ions or clusters via coordination bonds [ 2 ]. These MOFs exhibit remarkable potential as multifunctional materials, owing to their high porosity and extensive specific surface area, and have been extensively explored for applications in gas storage and separation [ 3 ], sensing [ 4 ], catalysis [ 5 ], and biomedicine [ 6 ]. Lanthanide metal–organic framework materials are an important branch of MOFs; they are characterized by their sharp emissions, large choke offset, and relatively long luminescence lifetime [ 7 , 8 ]. However, these unique advantages also bring limitations to photocatalysis. When a lanthanide metal–organic skeleton is irradiated by light and enters the excited state, electrons absorb energy and move out of the orbitals, leaving behind holes. Due to a certain coordination pattern between the lanthanide metal and the organic ligand, the photogenerated electrons and holes recombine rapidly, which limits its photocatalytic performance [ 9 , 10 ]. Therefore, the choice of organic ligands is crucial in the synthesis of lanthanide metal–organic framework materials. Porphyrin [ 11 ] is an excellent organic ligand with a unique macrocyclic structure and efficient visible light-trapping ability. In recent years, there has been a strong level of interest in utilizing porphyrin-based MOF photocatalysts for various applications, such as photocatalytic hydrogen evolution [ 12 , 13 ], the photocatalytic reduction of carbon dioxide [ 14 , 15 , 16 ], and photocatalytic organic reactions [ 17 ]. For example, Qi et al. [ 18 ] developed Co-Nd and Ni-Nd PMOF materials for the effective photocatalytic oxidation of benzyl alcohol and benzylamine under moderate air conditions. ARR et al. [ 19 ] reviewed the use of porphyrin-based or porphyrin-containing MOFs and COFs, including nanomaterials as heterogeneous single-linear state oxygen photosensitizers, for antimicrobial applications in recent times. Despite the promise that MOFs hold in terms of structure tuning and composite materials for property enhancement, there remains a need for significant improvements in their performance. On the other hand, since the discovery of materials like graphene [ 20 , 21 , 22 ], it has become evident that 2D materials, characterized by their extensive surface areas and relatively high surface energies, possess unique physicochemical properties unrivaled by bulk crystalline materials. Two-dimensional metal backbone materials [ 23 , 24 ], a recent addition to the realm of 2D materials, have captured significant attention due to their unique attributes. Compared to the typical three-dimensional MOF crystals, 2D MOFs offer wider planar dimensions and ultra-small thicknesses, granting 2D materials an expanded specific surface area and more exposed active sites, particularly beneficial for catalytic and sensing applications [ 25 , 26 , 27 ]. Up until now, 2D MOF-based catalysts have attracted widespread research interest in electrocatalysis, photocatalysis, and thermocatalysis. According to the relevant literature reports, we know that when certain lanthanide ions are doped in the catalysts, they can realize the up-conversion from visible to ultraviolet light of the substances [ 28 ], the extension of the response region [ 29 ], as well as the improvement of the separation efficiency of the photogenerated electron–hole pairs [ 30 , 31 ], which can improve the catalytic activity of the catalysts. Inspired by ultra-thin nanomaterials [ 32 , 33 ] and the MOFs of bimetallic porphyrins [ 34 ], we attempted to incorporate two lanthanide elements into porphyrin nanosheets to investigate their effects on photocatalytic performance. In this study, porphyrin-based nanosheet materials containing two lanthanide metal ions were synthesized using microwave-assisted heating, with acetic acid serving as a moderator. The prepared bimetallic nanosheets exhibit a large BET specific surface area and unique metal sites. Notably, the synthesized lanthanide bimetallic porphyrin nanomaterials demonstrated superior photocatalytic properties compared to both mono-metallic and bulk bimetallic materials. The results indicate that incorporating multi-component metal ions in nanomaterials can effectively enhance the catalytic efficiency of nanomaterial catalysts, offering a promising avenue for improving the photocatalytic performance of nanomaterials. 2. Results and Discussion The crystal structures of the synthesized samples were measured by powder XRD spectroscopy. As shown in Figure 1 a, all the samples showed better crystallinity and similar diffraction peaks in the 2θ (5°–30°) range. Furthermore, comparing them with the XRD of the known crystal Tb-TCPP (CCDC: 1899682), it was found that all the materials have similar diffraction peaks, which suggests that these materials have a similar structure to Tb-TCPP. Since Tb and Eu belong to the same lanthanide metal and have similar chemical properties, they have similar coordination environments during the synthesis of crystals, and during the synthesis process of doping two metals, Tb and Eu may be coordinated at the same position [ 34 , 35 , 36 ]. In addition, the concentrations of Tb and Eu were obtained through ICP testing. The ratio of the two lanthanide elements was calculated to be Tb:Eu = 1.1:1 ( Table S1 ). This result is close to the ratio of the preparation process (Tb:Eu = 1:1). Therefore, when studying the structure, a 1:1 ratio of Tb and Eu was adopted. On this basis, the possible crystal structure formed by Tb-Eu-TCPP is simulated by the known crystal structure [ 37 ], as shown in Figure 1 . The Tb/Eu atom adopts a pseudo-octahedral coordination with the oxygen atoms of four bridging carboxyl groups and two μ 2 -OH groups of four different TCPP linkers to form a 1D chain along the a-axis direction ( Figure 1 b). The porphyrin centers are metal-free, nearly planar, and have no apparent out-of-plane deformation ( Figure 1 c). A one-dimensional channel constructed from each carboxylate in TCPP and two Tb/Eu atoms from the 1D chain is shown in Figure 1 d (the channel is highlighted with a green ball). Subsequently, the morphology and structure of Tb-Eu-TCPP were measured by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The bright spots were clearly presented in a selected area electron diffraction (SAED) pattern, which were attributed to the (020), (021), and (041) planes of the Tb-Eu-TCPP nanosheets, confirming the crystal structure (inset in Figure 2 a). This result is in excellent agreement with the three typical peaks (020), (021), and (041) in the XRD and simulated pattern of Tb-Eu-TCPP and Bulk-TCPP ( Figure 1 a). The contents shown in Figure 2 a,b are the SEM images of Tb-Eu-TCPP and Bulk-TCPP. Unlike the hydrothermal method [ 38 , 39 ], microwave radiation induces localized superheating. The selective adsorption of solvent molecules or moderators on the crystal surface under microwave radiation plays a crucial role in the anisotropic growth of 2D nanosheets, leading to the formation of these distinct 2D nanosheets [ 40 , 41 , 42 ]. Simultaneously, acetic acid was used as a moderating conditioner, which was effective in inhibiting the deprotonation of carboxyl groups within the ligand [ 43 , 44 ]. Therefore, an increase in the amount of acetic acid in the hydrothermal process during the reaction results in a reduced number of nucleation events and increased inhibition of lateral growth in the nanosheets, leading to the thickening of the prepared nanosheets and the formation of a bulkier structure. TEM characterization further revealed the morphology of Tb-Eu-TCPP, with C, O, N, Tb, and Eu elements evenly distributed on the octahedron ( Figure 2 c). In addition, the results of elemental analysis testing further show that the material contains C, H, and N elements ( Table S2 ). According to the test results of EDX ( Figure S3 ) and ICP ( Table S1 ), both Tb and Eu lanthanide metals were present in Tb-Eu-TCPP. The results of TEM and ICP indicate that the bimetallic Tb-Eu-TCPP was successfully synthesized. X-ray photoelectron spectroscopy (XPS) was utilized to explore the chemical state and bonding configuration of H 2 TCPP and Tb-Eu-TCPP. Figure 3 a shows the presence of N, O, and C elements in H 2 TCPP and Tb, Eu, N, O, and C elements in Tb-Eu-TCPP. The binding energy data have been corrected for C-C at 284.8 eV ( Figure 3 d) [ 45 ]. The high-resolution spectrum of Tb 3d was deconvoluted into four peaks ( Figure 3 b). The peaks at 1281.6 eV and 1249.9 eV were satellite signals (abbreviated as "Sat."). The two peaks at 1242.6 eV and 1277.3 eV were attributed to Tb 3d 5/2 and Tb 3d 3/2 , respectively [ 35 ]. The high-resolution spectrum of Eu 3d was deconvoluted into four peaks ( Figure 3 c). The peak at 1143.1 eV was a satellite signal. The peak at 1135.3 eV was attributed to Eu 3+ . Two smaller Eu 3d 5/2 and Eu 3d 3/2 peaks were observed at 1125.2 eV and 1155.8 eV, respectively, attributed to the Eu 2+ oxidation state [ 46 , 47 ]. The corresponding O 1s XPS profiles were deconvoluted into three peaks of 531.6 eV, 532.1 eV, and 533.6 eV, among which the peak at 533.6 eV was associated with C-OH bonds, the peak at 532.6 eV was ascribed to Tb-O bonds, and the peak located at 531.6 eV came from Eu-O bonds [ 48 , 49 ]. The above results all indicated the successful preparation of Tb-Eu-TCPP. Figure 3 f displays the N 1s energy spectrum of the Tb-Eu-TCPP nanosheets alongside the H 2 TCPP ligand. Among the characteristic energy spectrum peaks of the H 2 TCPP ligand, the peaks at 400.0 and 397.8 eV correspond to the C=N-C and C-NH-C groups, respectively [ 50 , 51 ]. Notably, the N 1s energy spectrum of the Tb-Eu-TCPP nanosheets did not exhibit significant changes when compared to the characteristic N 1s peaks of H 2 TCPP. In the characteristic energy spectrum peaks of the H 2 TCPP ligand, the peaks corresponding to the C=N-C and C-NH-C groups are 400.1 and 397.9 eV, respectively. This indicates that the porphyrin ring is not coordinated with Tb and Eu. In addition, the results of XPS demonstrate that during the formation of Tb-Eu-TCPP, Tb/Eu only coordinates with the O in the carboxyl group to form a Tb/Eu-O bond. Meanwhile, Tb/Eu does not bond with the N in the center of the porphyrin ring. Therefore, the results of XPS further demonstrate that in the crystal structure of Tb-Eu-TCPP, Tb/Eu forms a strong coordination with the carboxyl group in the porphyrin, forming a one-dimensional chain structure. Moreover, the Raman spectra of Tb-Eu-TCPP nanosheets and Bulk-TCPP show typical characteristics of H 2 TCPP except for the peaks at 713 cm −1 and 776 cm −1 , which are assigned to Tb-O and Eu-O, respectively ( Figure 4 a) [ 52 , 53 ]. Figure 4 b shows the FTIR spectroscopy results acquired for the aforementioned materials. By analyzing the results obtained from Tb-Eu-TCPP, it can be observed that there are significant peaks at 3321 and 2920 cm −1 in the range of 4000–2000 cm −1 . Among these peaks, the peak at 3321 cm −1 corresponds to the OH-stretching vibration [ 54 , 55 ]. This is consistent with the peak position in the ligand, indicating that there may be less TCPP in the substance in the free state. Notably, the band at 2920 cm −1 was attributed to intergranular water and the intracavity physical absorption of water [ 56 , 57 ]. In the range of 500–2000 cm −1 , the shoulder peaks at 1689 cm −1 and 1604 cm −1 assigned to the asymmetric stretching vibration of C=O in H 2 TCPP were shifted to 1581 cm −1 and 1539 cm −1 , respectively, and significantly decreased in their intensity for the Tb-Eu-TCPP nanosheets. The disappearance of the stretching vibration of C-O at 1261 cm −1 can be also observed in the spectra of Bulk-TCPP and Tb-Eu-TCPP nanosheets [ 7 , 55 , 58 ]. All these results demonstrate that the Tb-Eu-TCPP nanosheets are constructed by the strong coordination of Tb 3+ (Eu 3+ ) with the carboxyl groups in H 2 TCPP. The relative changes in the specific surface area of the Tb-Eu-TCPP nanosheets and Bulk-TCPP were investigated through N 2 adsorption–desorption isotherm experiments. As illustrated in Figure 4 c, the nanosheet material exhibited a significantly higher specific surface area compared to the bulk material. Specifically, the BET specific surface area of the Tb-Eu-TCPP nanosheets was 332.13 m 2 /g, whereas that of Bulk-TCPP was 313.86 m 2 /g. This exceptional specific surface area suggests that these materials have excellent catalytic performance potential [ 59 , 60 , 61 ]. The pore size distribution curves ( Figure S4 ) indicate that the Tb-Eu-TCPP nanosheets have similar micropores as Bulk-TCPP. The thermal stability of the Tb-Eu-TCPP nanosheets and Bulk-TCPP was assessed using thermogravimetric analysis (TGA). As illustrated in Figure 4 d, the TGA curves of the Tb-Eu-TCPP nanosheets and Bulk-TCPP exhibit a high degree of similarity. Within the temperature range of 30–350 °C, the weight loss is approximately 20%. Subsequently, a sharp decline in weight is observed at 500 °C, indicating the onset of nanosheet backbone collapse. These results indicate that both Tb-Eu-TCPP nanosheets and Bulk-TCPP possess high thermal stability, maintaining the integrity of the backbone under nitrogen, and do not undergo thermal decomposition up to 500 °C. Porphyrin-based MOFs are well-established for their applications in singlet oxygen ( 1 O 2 ) generation within the realms of photocatalysis and photodynamic therapy [ 62 , 63 , 64 ]. To determine the production of 1 O 2 species in the photocatalytic process, 9,10-dibenzanthracene (DPA) was used as a probe for 1 O 2 to assess the ability of Tb-Eu-TCPP nanosheets to generate 1 O 2 under visible light irradiation. DPA functions as an 1 O 2 -trapping agent, with the concentration of DPA decreasing as it traps 1 O 2 . The concentration of DPA was monitored by measuring its absorbance at 374 nm using a UV–visible spectrophotometer. Figure 5 illustrates the concentration curves of DPA over time in different solution systems. Subsequently, Figure S5a–d display the absorption spectra of DPA over time for the Tb-TCPP nanosheets, Tb-Eu-TCPP nanosheets, Bulk-TCPP, and the system without added catalyst, respectively. The analysis of the data in Figure 5 leads to the following conclusions: In the presence of the Tb-Eu-TCPP nanosheets, the concentration of DPA consistently decreases over time, indicating the consumption of DPA during the process and providing evidence for the generation of singlet oxygen ( 1 O 2 ). Furthermore, compared to Tb-TCPP, Bulk-TCPP systems, or systems without added catalyst, there is no significant decrease in the concentration of DPA. This result emphasizes the enhanced photocatalytic activity of thinner Tb-Eu-TCPP nanosheets. Additionally, the results demonstrate that bimetallic nanosheets exhibit better catalytic effects compared to monometallic nanosheets. In order to gain a deeper understanding of the mechanism of photocatalytic activity depending on the thickness and metal nodes, photocurrent tests were conducted to evaluate their charge separation efficiency. In Figure 6 a, the plots display electrical impedance data for nanosheets with varying thicknesses. The test results reveal that the radius of the arc in the EIS plots for the Tb-Eu-TCPP nanosheets is smaller than that of Bulk-TCPP. This smaller arc radius suggests that the separation efficiency of photogenerated electron pairs is higher in the Tb-Eu-TCPP nanosheets, indicating an enhanced photocatalytic effect. Figure 6 b presents photocurrent data for nanosheets with different thicknesses. Based on the experimental data, it is evident that the photocurrent density of Tb-Eu-TCPP nanosheets exceeds that of Bulk-TCPP, signifying that thinner nanosheets exhibit a higher separation efficiency of electrons and holes. This phenomenon suggests the potential for increased production of ·O 2 − during the experimental process. Taking into account the previously conducted characterization, it is shown that thinner Tb-Eu-TCPP nanosheets have a greater BET specific surface area, light-trapping capacity, carrier density, and hole separation efficiency, all of which boost photocatalytic activity [ 65 , 66 , 67 ]. Additionally, the UV–Vis diffuse reflectance spectra reveal a similar band gap between the Tb-Eu-TCPP nanosheets and Bulk-TCPP, as depicted in Figure 6 c. This suggests that the electron-leaping ability of both materials is comparable. To determine the conduction band bottom potential (E CB ) of the prepared Tb-Eu-TCPP, Mott–Schottky analysis was employed [ 68 , 69 , 70 ]. The results in Figure 6 a,d reveal that Tb-Eu-TCPP exhibits n-type semiconductor behavior with a flat-band potential E fb = −0.43 V (Ag/AgCl). For n-type semiconductors, E fb is typically 0.1 or 0.2 V more positive than the conduction band bottom potential (E CB ). Consequently, the E CB of Tb-Eu-TCPP is −0.53 V (Ag/AgCl). By applying the conversion equation for the standard hydrogen potential (NHE) and Ag/AgCl potential (E NHE = E Ag/AgCl + 0.197 V), the E NHE of Tb-Eu-TCPP is calculated to be −0.34 V (NHE). From Figure S6 , the band gap (E g ) of the Tb-Eu-TCPP photocatalyst is 2.8 eV with bimetallic 2D properties. The valence band top potential (E VB ) of Tb-Eu-TCPP is calculated to be 2.27 V (NHE) using the equation E VB = E NHE + E g . This establishes a high oxidation–reduction potential for electron/hole pairs, allowing them to engage with dissolved oxygen and generate a substantial quantity of reactive oxygen species [ 71 ], notably 1 O 2 . In the system of bimetallic composites, the combined effect of the bimetals accelerates the charge transfer and inhibits electron and hole complexation, which improves the photocatalytic activity of this catalyst. Meanwhile, in the photocatalytic system of Tb-Eu-TCPP, light energy mainly drives the charge transfer, which leads to the generation of hydrated electrons upon photoexcitation. These hydrated electrons are subsequently trapped by oxygen to produce superoxide anion radicals (·O 2 − ), and ·O 2 − can further react with holes to produce 1 O 2 [ 72 ]. This principle is the same as outlined by Demyanenko et al. [ 73 ], who proposed an alternative mechanism of the photocatalytic generation of singlet oxygen through the photodetachment of an electron from the superoxide radical anion. In other words, once generated through the single-electron reduction of oxygen (Equation (1)), the superoxide radical anion absorbs light, giving rise to singlet oxygen and releasing an electron (Equation (2)), which, in turn, can recombine or reduce another oxygen molecule. The possible reactions are schematically shown in Figure 7 . (1) O 2 + e − → · O 2 − (2) · O 2 − + h v → O 2 1 + e − 3. Materials and Methods 3.1. Materials Tb(NO 3 ) 3 ·6H 2 O, Eu(NO 3 ) 3 ·6H 2 O, meso-Tetra(4-carboxyphenyl) porphyrin (H 2 TCPP), 9,10-diphenylanthracene (DPA), acetonitrile (MeCN), acetic acid (HAc), and N-N dimethylformamide (DMF) were all employed in this study. All the reagents used in the experiments were of commercial grade and did not require additional purification. Unless specified otherwise, ultrapure water was used in the experiments. 3.2. Characterization and Instruments The Midea brand home microwave oven M1-L213B (Midea, Shanghai, China) was used to prepare Tb-Eu-TCPP nanosheet samples. A JSM-7610F (JEOL Ltd., Tokyo, Japan) scanning electron microscope was used to characterize the morphology and size of the prepared samples. A transmission electron microscope was used for the morphological and selected area electron diffraction characterization of the samples with the FEI Tecnai G2 instrument model. An atomic force microscope (Dimension ICON) (Bruker Corporation, NASDAQ, Billerica, MA, United States) was used to characterize the thickness of the samples. A Smartlab 3 X-ray (JEOL Ltd., Tokyo, Japan) diffractometer was used to collect the powder diffraction data of the samples. Nitrogen adsorption–desorption isotherms of the samples were examined using an ASAP 2460 (Micromeritics, Norcross, Norcross, GA, United States) fully automated gas adsorption analyzer. Thermogravimetric analysis of the samples was carried out on an NETZSCH STA 2500 (Netzsch, Selb, Germany) thermogravimetric analyzer. The material was tested for Tb and Eu using an Agilent 7800 ICP-MS tool (Agilent, Santa Clara, CA, United States). The TEM and EDX of the material were tested using Talos F200X G2 (Thermo scientific, Waltham, MA, United States) to characterize the material's morphology and elemental information. The elemental analysis of the materials was conducted using an Elementar Unicube (Elementar, Langenselbold, Germany). A Thermo ESCELAB 250 XI X-ray photoelectron spectrometer (Thermo scientific, Massachusetts, United States) was used to characterize the X-ray photoelectron spectra of the samples. A RENISHAW InVia Raman microscope (Renishaw, London, England) was used to test the Raman spectra of the samples. The Fourier transform infrared spectrometer FTIR Prestige-21 (Shimadzu, Kyoto, Japan) was used to characterize the infrared spectra of the samples. The UV–Vis absorption spectra and solid diffuse reflectance spectra of the samples were tested and recorded using a UV2600 spectrophotometer (Techcomp, Shanghai, China). A 300 W xenon lamp (CEL-HXF300, China Education Au-light Co., Ltd., Beijing, China) was used to provide the light source. All electrochemical measurements were performed on a CHI 760E (CH Instruments, Shanghai, China) electrochemical workstation. 3.3. Synthesis of Tb-Eu-TCPP Nanosheets First, 0.045 mmol Tb(NO 3 ) 3 ·6H 2 O (20.4 mg), 0.045 mmol Eu(NO 3 ) 3 ·6H 2 O (20.1 mg), and 0.03 mmol H 2 TCPP (23.7 mg) were dissolved in 9 mL of DMF. Then, 100 μL of each of the 100 μL of the above-mixed solutions were taken into a 3-mL high-temperature-resistant glass vial. After that, 5 μL of a glacial acetic acid solution was added to it, which was labeled as Tb-Eu-TCPP. The mixed solution was diluted to 1 mL with DMF and then sealed tightly with the cap of the vial and placed in a household microwave oven at medium–low heat (its power was about 231 W) and microwaved for 10 min. After cooling to room temperature, the reaction was washed twice with DMF and anhydrous ethanol by centrifugation at 5000 rpm for 10 min. After washing, they were dried at room temperature. The specific schematic diagram of the synthesis process is shown in Figure S1 . 3.4. Synthesis of Bulk-TCPP Firstly, 0.019 mmol Tb(NO 3 ) 3 ·6H 2 O (8.5 mg) plus 0.013 mmol Eu(NO 3 ) 3 ·6H 2 O (8.6 mg) and 0.03 mmol H 2 TCPP (23.7 mg) were dissolved in 10 mL of DMF; then, 5 mL of the above-mixed was transferred into a PTFE-lined stainless steel autoclave, following which 400 μL of acetic acid was added, and it was placed in a 120 °C blast drying oven for 12 h. After being cooled down to room temperature, the Tb-Eu-TCPP block was washed by centrifugation with DMF and anhydrous ethanol two times. Centrifugation conditions: 5000 rpm for 10 min; after washing, it was dried at room temperature. The specific schematic diagram of the synthesis process is shown in Figure S2 . 3.5. Monitoring of Singlet Oxygen Using DPA as a Probe To validate the generation of singlet oxygen ( 1 O 2 ) under exposure to Tb-Eu-TCPP light, 9,10-diphenylanthracene (DPA) was employed as an 1 O 2 -trapping agent. The experimental procedure was as follows: a solution of DPA (100 μg/mL) was prepared using an oxygen-saturated acetonitrile/water (4:1) solvent mixture. Additionally, a dispersion of Tb-Eu-TCPP nanosheets (10 μg/mL) was prepared. Subsequently, 10 mL of the DPA (100 μg/mL) solution and Tb-Eu-TCPP nanosheet (10 μg/mL) dispersion were mixed in a tube. The mixture was then exposed to a 300 W xenon lamp equipped with a UV cut-off filter, and samples were collected at different exposure times for testing UV–visible absorption spectra. 3.1. Materials Tb(NO 3 ) 3 ·6H 2 O, Eu(NO 3 ) 3 ·6H 2 O, meso-Tetra(4-carboxyphenyl) porphyrin (H 2 TCPP), 9,10-diphenylanthracene (DPA), acetonitrile (MeCN), acetic acid (HAc), and N-N dimethylformamide (DMF) were all employed in this study. All the reagents used in the experiments were of commercial grade and did not require additional purification. Unless specified otherwise, ultrapure water was used in the experiments. 3.2. Characterization and Instruments The Midea brand home microwave oven M1-L213B (Midea, Shanghai, China) was used to prepare Tb-Eu-TCPP nanosheet samples. A JSM-7610F (JEOL Ltd., Tokyo, Japan) scanning electron microscope was used to characterize the morphology and size of the prepared samples. A transmission electron microscope was used for the morphological and selected area electron diffraction characterization of the samples with the FEI Tecnai G2 instrument model. An atomic force microscope (Dimension ICON) (Bruker Corporation, NASDAQ, Billerica, MA, United States) was used to characterize the thickness of the samples. A Smartlab 3 X-ray (JEOL Ltd., Tokyo, Japan) diffractometer was used to collect the powder diffraction data of the samples. Nitrogen adsorption–desorption isotherms of the samples were examined using an ASAP 2460 (Micromeritics, Norcross, Norcross, GA, United States) fully automated gas adsorption analyzer. Thermogravimetric analysis of the samples was carried out on an NETZSCH STA 2500 (Netzsch, Selb, Germany) thermogravimetric analyzer. The material was tested for Tb and Eu using an Agilent 7800 ICP-MS tool (Agilent, Santa Clara, CA, United States). The TEM and EDX of the material were tested using Talos F200X G2 (Thermo scientific, Waltham, MA, United States) to characterize the material's morphology and elemental information. The elemental analysis of the materials was conducted using an Elementar Unicube (Elementar, Langenselbold, Germany). A Thermo ESCELAB 250 XI X-ray photoelectron spectrometer (Thermo scientific, Massachusetts, United States) was used to characterize the X-ray photoelectron spectra of the samples. A RENISHAW InVia Raman microscope (Renishaw, London, England) was used to test the Raman spectra of the samples. The Fourier transform infrared spectrometer FTIR Prestige-21 (Shimadzu, Kyoto, Japan) was used to characterize the infrared spectra of the samples. The UV–Vis absorption spectra and solid diffuse reflectance spectra of the samples were tested and recorded using a UV2600 spectrophotometer (Techcomp, Shanghai, China). A 300 W xenon lamp (CEL-HXF300, China Education Au-light Co., Ltd., Beijing, China) was used to provide the light source. All electrochemical measurements were performed on a CHI 760E (CH Instruments, Shanghai, China) electrochemical workstation. 3.3. Synthesis of Tb-Eu-TCPP Nanosheets First, 0.045 mmol Tb(NO 3 ) 3 ·6H 2 O (20.4 mg), 0.045 mmol Eu(NO 3 ) 3 ·6H 2 O (20.1 mg), and 0.03 mmol H 2 TCPP (23.7 mg) were dissolved in 9 mL of DMF. Then, 100 μL of each of the 100 μL of the above-mixed solutions were taken into a 3-mL high-temperature-resistant glass vial. After that, 5 μL of a glacial acetic acid solution was added to it, which was labeled as Tb-Eu-TCPP. The mixed solution was diluted to 1 mL with DMF and then sealed tightly with the cap of the vial and placed in a household microwave oven at medium–low heat (its power was about 231 W) and microwaved for 10 min. After cooling to room temperature, the reaction was washed twice with DMF and anhydrous ethanol by centrifugation at 5000 rpm for 10 min. After washing, they were dried at room temperature. The specific schematic diagram of the synthesis process is shown in Figure S1 . 3.4. Synthesis of Bulk-TCPP Firstly, 0.019 mmol Tb(NO 3 ) 3 ·6H 2 O (8.5 mg) plus 0.013 mmol Eu(NO 3 ) 3 ·6H 2 O (8.6 mg) and 0.03 mmol H 2 TCPP (23.7 mg) were dissolved in 10 mL of DMF; then, 5 mL of the above-mixed was transferred into a PTFE-lined stainless steel autoclave, following which 400 μL of acetic acid was added, and it was placed in a 120 °C blast drying oven for 12 h. After being cooled down to room temperature, the Tb-Eu-TCPP block was washed by centrifugation with DMF and anhydrous ethanol two times. Centrifugation conditions: 5000 rpm for 10 min; after washing, it was dried at room temperature. The specific schematic diagram of the synthesis process is shown in Figure S2 . 3.5. Monitoring of Singlet Oxygen Using DPA as a Probe To validate the generation of singlet oxygen ( 1 O 2 ) under exposure to Tb-Eu-TCPP light, 9,10-diphenylanthracene (DPA) was employed as an 1 O 2 -trapping agent. The experimental procedure was as follows: a solution of DPA (100 μg/mL) was prepared using an oxygen-saturated acetonitrile/water (4:1) solvent mixture. Additionally, a dispersion of Tb-Eu-TCPP nanosheets (10 μg/mL) was prepared. Subsequently, 10 mL of the DPA (100 μg/mL) solution and Tb-Eu-TCPP nanosheet (10 μg/mL) dispersion were mixed in a tube. The mixture was then exposed to a 300 W xenon lamp equipped with a UV cut-off filter, and samples were collected at different exposure times for testing UV–visible absorption spectra. 4. Conclusions In conclusion, we have successfully synthesized nanosheets of porphyrin-based bimetallic compositions, Tb-Eu-TCPP, using a microwave-assisted method, as well as bulk porphyrin-based bimetallic materials, Bulk-TCPP, using a hydrothermal method. In this study, we succeeded in preparing ultrathin bimetallic nanosheets, and compared with the thicker bimetallic Bulk-TCPP, the prepared Tb-Eu-TCPP nanosheets have a thinner thickness, a larger BET surface area, higher photogenerated carrier separation efficiency, and better photocatalytic activity, thus exhibiting n-type semiconductor properties. This study marks a significant progress in the controlled synthesis of bimetallic-doped nanosheet materials and opens up new possibilities for synthesizing bimetallic nanosheet materials.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4622952/

Increasing the potency of neutralizing single-domain antibodies by functionalization with a CD11b/CD18 binding domain

Recombinant single domain antibodies (nanobodies) constitute an attractive alternative for the production of neutralizing therapeutic agents. Their small size warrants rapid bioavailability and fast penetration to sites of toxin uptake, but also rapid renal clearance, which negatively affects their performance. In this work, we present a new strategy to drastically improve the neutralizing potency of single domain antibodies based on their fusion to a second nanobody specific for the complement receptor CD11b/CD18 (Mac-1). These bispecific antibodies retain a small size (˜30 kDa), but acquire effector functions that promote the elimination of the toxin-immunocomplexes. The principle was demonstrated in a mouse model of lethal toxicity with tetanus toxin. Three anti-tetanus toxin nanobodies were selected and characterized in terms of overlapping epitopes and inhibition of toxin binding to neuron gangliosides. Bispecific constructs of the most promising monodomain antibodies were built using anti Mac-1, CD45 and MHC II nanobodies. When co-administered with the toxin, all bispecific antibodies showed higher toxin-neutralizing capacity than the monomeric ones, but only their fusion to the anti-endocytic receptor Mac-1 nanobody allowed the mice to survive a 10-fold lethal dose. In a model of delayed neutralization of the toxin, the anti- Mac-1 bispecific antibodies outperformed a sheep anti-toxin polyclonal IgG that had shown similar neutralization potency in the co-administration experiments. This strategy should have widespread application in the development of nanobody-based neutralizing therapeutics, which can be produced economically and more safely than conventional antisera.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3298155/

Criteria for Selection of Surrogates Used To Study the Fate and Control of Pathogens in the Environment

This article defines the term surrogate as an organism, particle, or substance used to study the fate of a pathogen in a specific environment. Pathogenic organisms, nonpathogenic organisms, and innocuous particles have been used as surrogates for a variety of purposes, including studies on survival and transport as well as for method development and as "indicators" of certain conditions. This article develops a qualitative surrogate attribute prioritization process and allows investigators to select a surrogate by systematically detailing the experimental process and prioritizing attributes. The results are described through the use of case studies of various laboratories that have used this process. This article also discusses the history of surrogate and microbial indicator use and outlines the method by which surrogates can be used when conducting a quantitative microbial risk assessment. The ultimate goal of selecting a sufficiently representative surrogate is to improve public health through a health-based risk assessment framework. Under- or overestimating the resistance, inactivation, or movement may negatively impact risk assessments that, in turn, will impact health assessments and estimated safety levels. Reducing uncertainty in a risk assessment is one of the objectives of using surrogates and the ultimate motive for any experiment investigating potential exposure of a pathogen.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3625043/

Sorting It Out in Endosomes: An Emerging Concept in Renal Epithelial Cell Transport Regulation

Ion and water transport by the kidney is continually adjusted in response to physiological cues. Selective endocytosis and endosomal trafficking of ion transporters are increasingly appreciated as mechanisms to acutely modulate renal function. Here, we discuss emerging paradigms in this new area of investigation.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5717188/

Stable Chromosomal Expression of Shigella flexneri 2a and 3a O-Antigens in the Live Salmonella Oral Vaccine Vector Ty21a

ABSTRACT We have been exploring the use of the live attenuated Salmonella enterica serovar Typhi Ty21a vaccine strain as a versatile oral vaccine vector for the expression and delivery of multiple foreign antigens, including Shigella O-antigens. In this study, we separately cloned genes necessary for the biosynthesis of the Shigella flexneri serotype 2a and 3a O-antigens, which have been shown to provide broad cross-protection to multiple disease-predominant S. flexneri serotypes. The cloned S. flexneri 2a rfb operon, along with bgt and gtrII , contained on the SfII bacteriophage, was sufficient in Ty21a to express the heterologous S. flexneri 2a O-antigen containing the 3,4 antigenic determinants. Further, this rfb operon, along with gtrA , gtrB , and gtrX contained on the Sfx bacteriophage and oac contained on the Sf6 bacteriophage, was sufficient to express S. flexneri 3a O-antigen containing the 6, 7, and 8 antigenic determinants. Ty21a, with these plasmid-carried or chromosomally inserted genes, demonstrated simultaneous and stable expression of homologous S . Typhi O-antigen plus the heterologous S. flexneri O-antigen. Candidate Ty21a vaccine strains expressing heterologous S. flexneri 2a or 3a lipopolysaccharide (LPS) elicited significant serum antibody responses against both homologous S . Typhi and heterologous Shigella LPS and protected mice against virulent S. flexneri 2a or 3a challenges. These new S. flexneri 2a and 3a O-antigen-expressing Ty21a vaccine strains, together with our previously constructed Ty21a strains expressing Shigella sonnei or Shigella dysenteriae 1 O-antigens, have the potential to be used together for simultaneous protection against the predominant causes of shigellosis worldwide as well as against typhoid fever.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5314968/

Funding and Organization of US Federal Health Security Programs

Recommendations ➢ Establish a dedicated leader at the White House who is responsible for federal health security. Managing US health security programs requires high-level coordination, something that can be done best from the White House and through a dedicated leader focused exclusively on this important mission space. Past administrations have established similar leadership for related issues, such as biodefense, to provide a focal point for coordination of federal programs and policies, as well as a standing base from which to operate when a health security event like Ebola occurs, which is inevitable. The White House, therefore, should establish a dedicated leadership position and office to head federal health security efforts in the executive branch. This measure would greatly improve the Administration's ability to build health security and manage emergencies effectively. ➢ Identify and account for health security programs across government. Creating a strong health security enterprise with programs that improve our national preparedness and response begins with the straightforward, but not simple, task of identifying the many programs and funding streams that contribute to building our defenses. While complete visibility is likely not possible, it is important for the White House to set strategic-level goals and priorities for health security programs, from which policies can be formed and programs informed. This can be done by both the executive agencies and the White House through the budgeting process. Agencies should be asked to identify those programs that contribute to health security and classify them as such in their annual budget-setting process. From that information, the OMB, as an office of the White House, can collect and define the universe of health security programs across agencies. This will enable better accounting and oversight, as well as ease of identifying gaps in our health security enterprise. ➢ Stop reductions in health security funding and invest in systems-building. Support for the health security mission in the executive branch and by Congress through appropriations is critical. These are the programs that will enable recovery if a CBRN attack or pandemic occurs, will prepare us to respond effectively, and will save lives—potentially many thousands or even millions of them—if and when the US experiences such catastrophic health events. Yet, despite their recognized importance and bipartisan support, many health security programs have seen significant funding reductions in the past decade, which have in turn reduced our overall national preparedness to deal with catastrophic health events. In FY2017 alone, funds are estimated to decrease by almost $800 million from the prior year. This is an area in which we cannot afford to be complacent and should not take our past work and success for granted. Health security is a national security imperative and should be treated and funded as such by the new Administration.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1609915/

Human α-defensins neutralize toxins of the mono-ADP-ribosyltransferase family

Various bacterial pathogens secrete toxins, which are not only responsible for fatal pathogenesis of disease, but also facilitate evasion of host defences. One of the best-known bacterial toxin groups is the mono-ADP-ribosyltransferase family. In the present study, we demonstrate that human neutrophil α-defensins are potent inhibitors of the bacterial enzymes, particularly against DT (diphtheria toxin) and ETA ( Pseudomonas exotoxin A). HNP1 (human neutrophil protein 1) inhibited DT- or ETA-mediated ADP-ribosylation of eEF2 (eukaryotic elongation factor 2) and protected HeLa cells against DT- or ETA-induced cell death. Kinetic analysis revealed that inhibition of DT and ETA by HNP1 was competitive with respect to eEF2 and uncompetitive against NAD + substrates. Our results reveal that toxin neutralization represents a novel biological function of HNPs in host defence.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9618875/

Learning from the COVID-19 pandemic for future epidemics and pandemics preparedness and response in Guinea: Findings from a scoping review

The outbreak of the novel coronavirus (SARS-CoV-2) in December 2019 prompted a response from health systems of countries across the globe. The first case of COVID-19 in Guinea was notified on 12 March 2020; however, from January 2020 preparations at policy and implementation preparedness levels had already begun. This study aimed to assess the response triggered in Guinea between 27 th January 2020 and 1 st November 2021 and lessons for future pandemic preparedness and response. We conducted a scoping review using three main data sources: policy documents, research papers and media content. For each of these data sources, a specific search strategy was applied, respectively national websites, PubMed and the Factiva media database. A content analysis was conducted to assess the information found. We found that between January 2020 and November 2021, the response to the COVID-19 pandemic can be divided into five phases: (1) anticipation of the response, (2) a sudden boost of political actions with the implementation of strict restrictive measures, (3) alleviation of restrictive measures, (4) multiple epidemics period and (5) the COVID-19 variants phase, including the strengthening of vaccination activities. This study provides several learning points for countries with similar contexts including: (1) the necessity of setting up, in the pre-epidemic period, an epidemic governance framework that is articulated with the country's health system and epidemiological contexts; (2) the importance of mobilizing, during pre-epidemic period, emergency funds for a rapid health system response whenever epidemics hit; (3) each epidemic is a new experience as previous exposure to similar ones does not necessarily guarantee population and health system resilience; (4) epidemics generate social distress because of the restrictive measures they require for their control, but their excessive securitization is counterproductive. Finally, from a political point of view, decision-making for epidemic control is not always disinterested; it is sometimes rooted in political computations, and health system actors should learn to cope with it while, at the same time, safeguarding trusted and efficient health system responses. We conclude that health system actors anticipated the response to the COVID-19 pandemic and (re-) adapted response strategies as the pandemic evolved in the country. There is a need to rethink epidemics governance and funding mechanisms in Guinea to improve the health system response to epidemics. Introduction An unprecedented outbreak of a novel coronavirus (SARS-CoV-2) that emerged in China in December 2019 and rapidly spread to 114 countries across the globe by 11 March 2020 resulted in the World Health Organization's (WHO's) declaration of a pandemic ( 1 ). In Africa, the spread of SARS-CoV-2 was initially centered in the northern (Algeria, Egypt, Morocco) and southern (South Africa) regions, where significant increases in the number of cases and deaths were reported from March 2020 onwards ( 2 ). In late April 2020, however, the virus had almost spread throughout the continent, and most African countries were experiencing community transmission ( 2 ). Meanwhile, the capacity of the region's health systems to effectively implement pandemic containment measures—travel restrictions, increased testing, contact tracing, isolation of cases and quarantine of contacts—was apparently sparse ( 3 – 5 ). Moreover, the ability of African economic systems to resist and recover from these containment measures represented another tricky situation ( 3 , 4 ). These challenges led to the international prediction of a drastic effect of the pandemic in Africa ( 6 ). In Guinea, the first case of the novel coronavirus disease (COVID-19), imported from Europe, was reported on 12 March 2020 and the first death on 14 April 2020. At the time the pandemic hit Guinea, the country was undergoing a constitutional reform, with a background of socio-political crises. Moreover, no formal national framework for epidemic-preparedness and response existed at the pandemic onset, although the National Agency for Health Security (ANSS)—the former national Ebola response committee—had gained more power and influence in epidemic management in the past years, since its establishment in June 2016 ( 7 ). Therefore, the pandemic declaration resulted in a power struggle among several actors of the health system in Guinea ( 8 , 9 ). These contextual factors certainly contributed to hindering the gains of the post-Ebola health system strengthening reforms for the initial COVID-19 preparedness and response ( 10 , 11 ). From mid-April 2020 onwards, for example, a steep increase in cases was observed in Guinea, compared with Sierra Leone and Liberia, which had also been confronted with the 2014/2016 Ebola epidemic ( 8 , 12 ). Nevertheless, in the same period, the number of reported deaths was much lower in Guinea compared with Sierra Leone, Liberia and many other west African countries ( 12 ). Some authors have documented Guinea's response to the COVID-19 pandemic ( 8 , 9 , 12 – 15 ). However, these studies focused on the country's early response while the pandemic was unfolding. At the time of writing this paper (November 2021), Guinea has experienced three epidemiological waves of the COVID-19 pandemic; the first wave being the longest (from March 2020 to January 2021) and the last the deadliest (43 deaths per month on average compared to 7 and 18 respectively during the first and second wave). It is, therefore, necessary to establish a broader picture of the country's experience during the first and subsequent waves of the pandemic. The national healthcare system is essentially based on the public service with three distinct levels of health services provision: primary, secondary and tertiary. The primary level includes 410 health centers; the secondary level comprises 33 districts hospitals, and 7 regional; and the tertiary level is composed of 3 national hospitals. Households constitute the primary source of health financing through direct payment, with 62% of expenditure, followed by the external funding from multilateral cooperation (27%). The share of the Ministry of Health's budget in the national budget remained below 3% between 2010 and 2014, before rising to 8% in 2016, after the Ebola epidemic. In terms of clinical management of epidemic and pandemic prone diseases, the country has 33 epidemic management centers, distributed among the health districts. At national level, several institutions have been capacitated in the post-Ebola context to improve epidemics preparedness and response in the country. These institutions include the national directorate for epidemiology and disease control (DNELM), the national institute of Public Health (INSP), and the national health security agency (ANSS), with overlapping functions in the epidemics surveillance and control. However, from 2017 onwards, the ANSS has gained more Leadership in epidemics surveillance and control, including sporadic cases of human anthrax, measles, varicella, and yellow fever ( 7 ). The ANSS has deconcentrated structured integrated to regional (regional alert and response team) and district (district team of alert and response) health systems. The ANSS is technically and financially supported by the Ministry of Health and development partners including the world health organization (WHO), the center for disease control and prevention of Atlanta (CDC), and the World Bank ( 7 ). Therefore, this study aimed to describe policy decisions to the preparation and response to the COVID-19 pandemic from January 2020 to November 2021, community responses to these decisions. Such findings would provide lessons for management of future epidemics and pandemics in Guinea, and countries with similar context. Methods Study design This study was conducted using the scoping review guide developed by the Joanna Briggs Institute (JBI) ( 16 ). We opted for this study design because of the scope of our research questions. This scoping review was conducted using the five stages for scoping review outlined by the JBI manual: (1) identification of research question, (2) identification of relevant studies, (3) study selection, (4) data charting, and (5) summarizing and reporting results. Identification of research questions For this review, two main research questions were identified: What policy response was undertaken before, during and after each of the COVID-19 waves registered in Guinea over the period Jan 2020 to Nov 2021? How has the national community reacted to the policy decisions? Search strategy A documentary search plan was developed to integrate five different sources: websites of national institutions, PubMed search engine, Google search engine, Media content search and consultation with key informants for gray literature sources. For each of these sources, a specific search strategy was applied to facilitate the generation of relevant information. The website of the national agency for health security (ANSS) was browsed for the period between January 2020 and November 2021 to identify policy documents (decrees, acts, etc.), and relevant information on the pandemic response (debriefing, meeting notes, etc.). The second search technique consisted of contacting key informants to help get policy documents that were not found on the ANSS website. A Snowball and citation tracking techniques were used for identifying and asking for additional policy documents. Key informants contacted included officials of the ANSS, international migration organization, world health organization, national public health institute, and the scientific committee for COVID-19 response. Three search terms were used for the PubMed search: (1) Coronavirus OR COVID-19 OR Novel Coronavirus; AND (2) preparedness OR response OR governance OR management; AND (3) Guinea. For the Google search, several search strategies, containing multiple combinations of the above search terms, were created. In order to ensure consistency across searches and effective time management, only the first five pages of each search were screened ( 17 ). For the Media content search, the Factiva media database ( https://www.dowjones.com/professional/factiva/ ) was used. This search strategy required defining in advance: (1) search terms, (2) period of the search, (3) targeted media and (4) search Language (French and English). We replicated the search terms used for articles identification on PubMed. This search covered the period between January 2020 and November 2021. Finally, media targeted for our search included Africaguinee.com, Guinée Matin, Guinee7.com, Journal de Conakry.com, Jeune Afrique, Agence de Presse Senegalaise, Agence Ivoirienne de Presse, L'intelligent d'Abidjan, Agence France Presse, BBC, RFI.fr, and FoxNews. Studies selection Our search results were cleaned for duplicates and two members of our research team (DK and FNK) independently screened studies using the eligibility criteria described in Table 1 . The snowball and citation tracking techniques were used to identify additional articles or documents to be added to selected studies. Discrepancies in study selection processes were resolved by a discussion between the two reviewers with the assistance of another team member (AD). Table 1 Eligibility criteria for selection of studies. Inclusion criteria Exclusion criteria Full text written in French or English Unavailable in French or English Most current version of the document Document was a draft or summary version or has been replaced with another document Articles, reports, opinion papers, workshop summaries, briefings, commentaries, blogs, newspaper articles Abstracts Include interventions or strategies for the COVID-19 response in Guinea Include interventions or strategies for the COVID-19 response for other contexts different from Guinea In total, 79 studies were included in this review. Figure 1 shows the search description flow chart of included studies. Table 2 shows the detailed characteristics of studies included and also sources used for their selection. Policy documents represented 46.25% of data sources included in the study followed by media content (38.75%). Figure 1 Study inclusion flow chart. Table 2 Documents characteristics. No . Document Date of publication Search strategies used to identify document Policy documents 1 Technical guidelines for managing the risk of spreading the new coronavirus 27 January 2020 Consultation with policy actors 2 Coronavirus threat management Guidelines 12 February 2020 Consultation with policy actors 3 National COVID-19 Preparedness and Response Plan 19 February 2020 Consultation with policy actors 4 Setting-up of six technical commissions for the response to COVID-19 14 March 2020 Consultation with policy actors 5 Presidential decree on the measures to strengthen Coronavirus management 21 March 2020 Targeted websites, consultation with policy actors 6 Declaration of a state of health emergency 26 March 2020 Targeted websites, consultation with policy actors 7 Act of reinforcement of the prerogatives of the national agency for health security on the management of the COVID-19 27 March 2020 Consultation with policy actors 8 Address to the Nation by the President of the Republic on new measures against COVID-19 in Guinea 30 March 2020 Targeted websites 9 Communiqué on the mobility of people from Conakry to the countryside 31 March 2020 Targeted websites 10 COVID-19 management technical guide April 2020 Consultation with policy actors 11 Plan of economic response to the health crisis of COVID-19 2 April 2020 Google search, consultation with policy actors 12 Creation of the scientific council (task force) of response to COVID-19 10 April 2020 Consultation with policy actors 13 Decree on the reinforcement of the state of emergency measures 13 April 2020 Consultation with policy actors 14 Stop COVID-19 in 60 days Community Response Strategic Plan May 2020 Consultation with policy actors 15 Plan to strengthen the health system and resilience for the continuity of services in the COVID-19 context 19 May 2020 Google search, consultation with policy actors 16 Address to the Nation of the Head of State (Alpha Condé) 15 July 2020 Targeted websites 17 Circular note on the end of home confinement 27 July 2020 Targeted websites, consultation with policy actors 18 Note on the celebration of the Tabaski festival on July 31 2020 23 July 2020 Targeted websites 19 Note on the implementation of COVID-19 tests for air travelers 27 July 2020 Consultation with policy actors 20 Directive on body transfer authorization 19 August 2020 Targeted websites 21 Circular note on the identification and follow-up of COVID-19 contacts 20 August 2020 Targeted websites 22 Guidelines of the reopening of training institutions (schools) October 2020 Consultation with policy actors 23 Guidelines for Reopening Places of Worship October 2020 Consultation with policy actors 24 Active case-finding strategy for COVID-19 coupled with sensitization in Guinea: Stop the COVID-19 "Let's get screened" October 2020 Consultation with policy actors 25 Guideline for the systematic performance of antigenic rapid diagnostic tests in Guinea 8 October 2020 Targeted websites, Consultation with policy actors 26 Guideline for the systematic realization of antigenic rapid diagnostic tests in Guinea September 2020 Consultation with policy actors 27 Government press release on guidelines for the application of restrictive measures 18 September 2020 Targeted websites 28 Alleviation of restrictive measures in the transport, tourism, sports and culture sectors 25 September 2020 Targeted websites 29 National vaccination plan against COVID-19 in Guinea December 2020 Consultation with policy actors 30 Active Case Finding and Outreach Strategy 1 December 2020 Consultation with policy actors 31 Decentralization of COVID-19 response through the strengthening of local coordination and intensification of community contacts follow-up 1 December 2020 Consultation with policy actors 32 Setting-up of the steering committee for the introduction of COVID-19 vaccines 2 December 2020 Consultation with policy actors 33 Temporary authorization for the use of Sputnik V vaccine in Guinea 21 January 2021 Targeted websites 34 Government guidelines on the management of the Ebola virus disease in Guinea 16 February 2021 Targeted websites 35 Decision on the deployment of national and international experts to N'zérékoré for the joint response to Ebola and COVID-19 19 February 2021 Targeted websites 36 Plan to accelerate vaccination against COVID-19 in Guinea, "stop COVID-19, let's vaccinate" September 2021 Consultation with policy actors 37 Introduction of the "Vaccination pass" 1 September 2021 Targeted websites Research articles 38 First-line response to COVID-19: community health centers and doctors' offices in Guinea ( 9 ) April 2020 Google search 39 COVID-19 in Guinea: The first line of health care in South and North get ready for action! ( 15 ) May 2020 Google search 40 Willingness to comply with physical distancing measures against COVID-19 in four African countries ( 13 ) September 2020 PubMed search 41 Tackling the COVID-19 pandemic in West Africa: Have we learned from Ebola in Guinea? ( 8 ) December 2020 PubMed search 42 Evolution of the COVID-19 pandemic over 6 weeks in four French-speaking countries in West Africa ( 12 ) January 2021 PubMed search 43 Ebola and COVID-19 in DR Congo and Guinea ( 18 ) April 2021 PubMed search 44 Ebola Outbreak amid COVID-19 in the Republic of Guinea: Priorities for Achieving Control ( 19 ) April 2021 PubMed search 45 The COVID-19 pandemic in francophone West Africa: from the first cases to responses in seven countries ( 14 ) August 2021 PubMed search 46 Marburg virus amidst COVID-19 pandemic in Guinea: Fighting within the looming cases ( 20 ) September 2021 PubMed search 47 Guinea's response to syndemic hotspots ( 7 ) October 2021 PubMed search 48 Will Guinea's coup interrupt the country's health responses? September 2021 Google search Media content 49 Elections held in Guinea despite COVID-19 19 March 2020 Factiva software search 50 In Guinea, Alpha Conde plays his cards right 21 March 2020 Factiva software search 51 Alpha Conde makes a gesture to the Army on the eve of the elections March 2020 Factiva software search 52 Ali Baba Foundation donate COVID-19 test in Guinea April 2020 Factiva software search 53 Guinea: Deadly protests in several cities against police blockades May 2020 Factiva software search 54 COVID-19: un médecin guéri de la pandémie témoigne 30 May 2020 55 Start of the 2020 university year: our findings in Sonfonia and Mahatma Gandhi of Lambanyi October 2020 Factiva software search 56 The United Arab Emirates donates field hospital to Guinea November 2020 Factiva software search 57 Draft budget 2021: the government announces a budget of more than GNF 27 600 billion October 2020 Factiva software search 58 Guinea: Opposition protesters injured (witnesses, hospital source) October 2020 Factiva software search 59 Back to school: DPE N'Zérékoré gives instructions to school administrators October 2020 Factiva software search 60 National Assembly: deputies adopt the LFR 2020 which amounts to about GNF 27 000 billion November 2020 Factiva software search 61 Jacques Gbonimy, president of the UPG: 'COVID-19 has become a political disease in our country' November 2020 Factiva software search 62 Guinea bans rallies citing virus, opposition cries foul November 2020 Factiva software search 63 Fight against COVID-19: three doctors of the RUSAL company rewarded by the structure Katala 224 November 2020 Factiva software search 64 Council of Ministers: Here are the full minutes November 2020 Factiva software search 65 COVID-19: President Alpha Conde extends the State of Emergency for 30 days November 2020 Factiva software search 66 European Union and Federal Republic of Germany Provide COVID-19 Protection Kits to Guinean Government November 2020 Factiva software search 67 Council of Ministers: here are the full minutes (Press release) November 2020 Factiva software search 68 Fight against COVID-19: Health providers announce a strike from November 13 October 2020 Factiva software search 69 The children must return to school, but the recipes of the past are no longer October 2020 Factiva software search 70 Impact of COVID-19 in Guinea: more than 17 billion in losses suffered by the owners of places of leisure October 2020 Factiva software search 71 Receipt in Conakry of 11 360 doses of Ebola vaccine by February 22, 2021 February 2021 Factiva software search 72 Launch of the vaccination campaign against COVID-19 March 2021 Factiva software search 73 Reception this Saturday, 6 March 2021 at 22:03 of 200 000 doses of Sputnik V vaccine March 2021 Factiva software search 74 Press release on the vaccination against COVID-19 in Conakry March 2021 Factiva software search 75 Guinea's school and the COVID-19 pandemic March 2021 Google search 76 Communiqué from the Ministry of Health on the use of the batch of 69 000 doses of AstraZeneca vaccine received in Guinea on 29 March 2021 March 2021 Factiva software search 77 As part of the COVAX initiative, the Guinean government received its first batch of the AstraZeneca vaccine on Sunday, April 11, 2021. Composed of 194 400 doses April 2021 Factiva software search 78 Receipt of 300 000 doses of #Sinovac vaccine purchased by the Guinean government to increase the country's vaccine capacity and allow many people to be vaccinated April 2021 Factiva software search 79 Communiqué on the administration of the second dose of Sputnik V vaccine April 2021 Factiva software search Data analysis The audio recordings from the national press were completely transcribed. Content analysis of these transcripts, media reports and policy documents were done manually according to a code grid. We compared the data from the different sources (media content, audio records, policy documents) to triangulate the data and thus strengthen the internal validity (credibility) of the findings. We addressed inter-coding bias by having two researchers coding the material and allowing for adjustment upon a team consultation in association with a third researcher. Study design This study was conducted using the scoping review guide developed by the Joanna Briggs Institute (JBI) ( 16 ). We opted for this study design because of the scope of our research questions. This scoping review was conducted using the five stages for scoping review outlined by the JBI manual: (1) identification of research question, (2) identification of relevant studies, (3) study selection, (4) data charting, and (5) summarizing and reporting results. Identification of research questions For this review, two main research questions were identified: What policy response was undertaken before, during and after each of the COVID-19 waves registered in Guinea over the period Jan 2020 to Nov 2021? How has the national community reacted to the policy decisions? Search strategy A documentary search plan was developed to integrate five different sources: websites of national institutions, PubMed search engine, Google search engine, Media content search and consultation with key informants for gray literature sources. For each of these sources, a specific search strategy was applied to facilitate the generation of relevant information. The website of the national agency for health security (ANSS) was browsed for the period between January 2020 and November 2021 to identify policy documents (decrees, acts, etc.), and relevant information on the pandemic response (debriefing, meeting notes, etc.). The second search technique consisted of contacting key informants to help get policy documents that were not found on the ANSS website. A Snowball and citation tracking techniques were used for identifying and asking for additional policy documents. Key informants contacted included officials of the ANSS, international migration organization, world health organization, national public health institute, and the scientific committee for COVID-19 response. Three search terms were used for the PubMed search: (1) Coronavirus OR COVID-19 OR Novel Coronavirus; AND (2) preparedness OR response OR governance OR management; AND (3) Guinea. For the Google search, several search strategies, containing multiple combinations of the above search terms, were created. In order to ensure consistency across searches and effective time management, only the first five pages of each search were screened ( 17 ). For the Media content search, the Factiva media database ( https://www.dowjones.com/professional/factiva/ ) was used. This search strategy required defining in advance: (1) search terms, (2) period of the search, (3) targeted media and (4) search Language (French and English). We replicated the search terms used for articles identification on PubMed. This search covered the period between January 2020 and November 2021. Finally, media targeted for our search included Africaguinee.com, Guinée Matin, Guinee7.com, Journal de Conakry.com, Jeune Afrique, Agence de Presse Senegalaise, Agence Ivoirienne de Presse, L'intelligent d'Abidjan, Agence France Presse, BBC, RFI.fr, and FoxNews. Studies selection Our search results were cleaned for duplicates and two members of our research team (DK and FNK) independently screened studies using the eligibility criteria described in Table 1 . The snowball and citation tracking techniques were used to identify additional articles or documents to be added to selected studies. Discrepancies in study selection processes were resolved by a discussion between the two reviewers with the assistance of another team member (AD). Table 1 Eligibility criteria for selection of studies. Inclusion criteria Exclusion criteria Full text written in French or English Unavailable in French or English Most current version of the document Document was a draft or summary version or has been replaced with another document Articles, reports, opinion papers, workshop summaries, briefings, commentaries, blogs, newspaper articles Abstracts Include interventions or strategies for the COVID-19 response in Guinea Include interventions or strategies for the COVID-19 response for other contexts different from Guinea In total, 79 studies were included in this review. Figure 1 shows the search description flow chart of included studies. Table 2 shows the detailed characteristics of studies included and also sources used for their selection. Policy documents represented 46.25% of data sources included in the study followed by media content (38.75%). Figure 1 Study inclusion flow chart. Table 2 Documents characteristics. No . Document Date of publication Search strategies used to identify document Policy documents 1 Technical guidelines for managing the risk of spreading the new coronavirus 27 January 2020 Consultation with policy actors 2 Coronavirus threat management Guidelines 12 February 2020 Consultation with policy actors 3 National COVID-19 Preparedness and Response Plan 19 February 2020 Consultation with policy actors 4 Setting-up of six technical commissions for the response to COVID-19 14 March 2020 Consultation with policy actors 5 Presidential decree on the measures to strengthen Coronavirus management 21 March 2020 Targeted websites, consultation with policy actors 6 Declaration of a state of health emergency 26 March 2020 Targeted websites, consultation with policy actors 7 Act of reinforcement of the prerogatives of the national agency for health security on the management of the COVID-19 27 March 2020 Consultation with policy actors 8 Address to the Nation by the President of the Republic on new measures against COVID-19 in Guinea 30 March 2020 Targeted websites 9 Communiqué on the mobility of people from Conakry to the countryside 31 March 2020 Targeted websites 10 COVID-19 management technical guide April 2020 Consultation with policy actors 11 Plan of economic response to the health crisis of COVID-19 2 April 2020 Google search, consultation with policy actors 12 Creation of the scientific council (task force) of response to COVID-19 10 April 2020 Consultation with policy actors 13 Decree on the reinforcement of the state of emergency measures 13 April 2020 Consultation with policy actors 14 Stop COVID-19 in 60 days Community Response Strategic Plan May 2020 Consultation with policy actors 15 Plan to strengthen the health system and resilience for the continuity of services in the COVID-19 context 19 May 2020 Google search, consultation with policy actors 16 Address to the Nation of the Head of State (Alpha Condé) 15 July 2020 Targeted websites 17 Circular note on the end of home confinement 27 July 2020 Targeted websites, consultation with policy actors 18 Note on the celebration of the Tabaski festival on July 31 2020 23 July 2020 Targeted websites 19 Note on the implementation of COVID-19 tests for air travelers 27 July 2020 Consultation with policy actors 20 Directive on body transfer authorization 19 August 2020 Targeted websites 21 Circular note on the identification and follow-up of COVID-19 contacts 20 August 2020 Targeted websites 22 Guidelines of the reopening of training institutions (schools) October 2020 Consultation with policy actors 23 Guidelines for Reopening Places of Worship October 2020 Consultation with policy actors 24 Active case-finding strategy for COVID-19 coupled with sensitization in Guinea: Stop the COVID-19 "Let's get screened" October 2020 Consultation with policy actors 25 Guideline for the systematic performance of antigenic rapid diagnostic tests in Guinea 8 October 2020 Targeted websites, Consultation with policy actors 26 Guideline for the systematic realization of antigenic rapid diagnostic tests in Guinea September 2020 Consultation with policy actors 27 Government press release on guidelines for the application of restrictive measures 18 September 2020 Targeted websites 28 Alleviation of restrictive measures in the transport, tourism, sports and culture sectors 25 September 2020 Targeted websites 29 National vaccination plan against COVID-19 in Guinea December 2020 Consultation with policy actors 30 Active Case Finding and Outreach Strategy 1 December 2020 Consultation with policy actors 31 Decentralization of COVID-19 response through the strengthening of local coordination and intensification of community contacts follow-up 1 December 2020 Consultation with policy actors 32 Setting-up of the steering committee for the introduction of COVID-19 vaccines 2 December 2020 Consultation with policy actors 33 Temporary authorization for the use of Sputnik V vaccine in Guinea 21 January 2021 Targeted websites 34 Government guidelines on the management of the Ebola virus disease in Guinea 16 February 2021 Targeted websites 35 Decision on the deployment of national and international experts to N'zérékoré for the joint response to Ebola and COVID-19 19 February 2021 Targeted websites 36 Plan to accelerate vaccination against COVID-19 in Guinea, "stop COVID-19, let's vaccinate" September 2021 Consultation with policy actors 37 Introduction of the "Vaccination pass" 1 September 2021 Targeted websites Research articles 38 First-line response to COVID-19: community health centers and doctors' offices in Guinea ( 9 ) April 2020 Google search 39 COVID-19 in Guinea: The first line of health care in South and North get ready for action! ( 15 ) May 2020 Google search 40 Willingness to comply with physical distancing measures against COVID-19 in four African countries ( 13 ) September 2020 PubMed search 41 Tackling the COVID-19 pandemic in West Africa: Have we learned from Ebola in Guinea? ( 8 ) December 2020 PubMed search 42 Evolution of the COVID-19 pandemic over 6 weeks in four French-speaking countries in West Africa ( 12 ) January 2021 PubMed search 43 Ebola and COVID-19 in DR Congo and Guinea ( 18 ) April 2021 PubMed search 44 Ebola Outbreak amid COVID-19 in the Republic of Guinea: Priorities for Achieving Control ( 19 ) April 2021 PubMed search 45 The COVID-19 pandemic in francophone West Africa: from the first cases to responses in seven countries ( 14 ) August 2021 PubMed search 46 Marburg virus amidst COVID-19 pandemic in Guinea: Fighting within the looming cases ( 20 ) September 2021 PubMed search 47 Guinea's response to syndemic hotspots ( 7 ) October 2021 PubMed search 48 Will Guinea's coup interrupt the country's health responses? September 2021 Google search Media content 49 Elections held in Guinea despite COVID-19 19 March 2020 Factiva software search 50 In Guinea, Alpha Conde plays his cards right 21 March 2020 Factiva software search 51 Alpha Conde makes a gesture to the Army on the eve of the elections March 2020 Factiva software search 52 Ali Baba Foundation donate COVID-19 test in Guinea April 2020 Factiva software search 53 Guinea: Deadly protests in several cities against police blockades May 2020 Factiva software search 54 COVID-19: un médecin guéri de la pandémie témoigne 30 May 2020 55 Start of the 2020 university year: our findings in Sonfonia and Mahatma Gandhi of Lambanyi October 2020 Factiva software search 56 The United Arab Emirates donates field hospital to Guinea November 2020 Factiva software search 57 Draft budget 2021: the government announces a budget of more than GNF 27 600 billion October 2020 Factiva software search 58 Guinea: Opposition protesters injured (witnesses, hospital source) October 2020 Factiva software search 59 Back to school: DPE N'Zérékoré gives instructions to school administrators October 2020 Factiva software search 60 National Assembly: deputies adopt the LFR 2020 which amounts to about GNF 27 000 billion November 2020 Factiva software search 61 Jacques Gbonimy, president of the UPG: 'COVID-19 has become a political disease in our country' November 2020 Factiva software search 62 Guinea bans rallies citing virus, opposition cries foul November 2020 Factiva software search 63 Fight against COVID-19: three doctors of the RUSAL company rewarded by the structure Katala 224 November 2020 Factiva software search 64 Council of Ministers: Here are the full minutes November 2020 Factiva software search 65 COVID-19: President Alpha Conde extends the State of Emergency for 30 days November 2020 Factiva software search 66 European Union and Federal Republic of Germany Provide COVID-19 Protection Kits to Guinean Government November 2020 Factiva software search 67 Council of Ministers: here are the full minutes (Press release) November 2020 Factiva software search 68 Fight against COVID-19: Health providers announce a strike from November 13 October 2020 Factiva software search 69 The children must return to school, but the recipes of the past are no longer October 2020 Factiva software search 70 Impact of COVID-19 in Guinea: more than 17 billion in losses suffered by the owners of places of leisure October 2020 Factiva software search 71 Receipt in Conakry of 11 360 doses of Ebola vaccine by February 22, 2021 February 2021 Factiva software search 72 Launch of the vaccination campaign against COVID-19 March 2021 Factiva software search 73 Reception this Saturday, 6 March 2021 at 22:03 of 200 000 doses of Sputnik V vaccine March 2021 Factiva software search 74 Press release on the vaccination against COVID-19 in Conakry March 2021 Factiva software search 75 Guinea's school and the COVID-19 pandemic March 2021 Google search 76 Communiqué from the Ministry of Health on the use of the batch of 69 000 doses of AstraZeneca vaccine received in Guinea on 29 March 2021 March 2021 Factiva software search 77 As part of the COVAX initiative, the Guinean government received its first batch of the AstraZeneca vaccine on Sunday, April 11, 2021. Composed of 194 400 doses April 2021 Factiva software search 78 Receipt of 300 000 doses of #Sinovac vaccine purchased by the Guinean government to increase the country's vaccine capacity and allow many people to be vaccinated April 2021 Factiva software search 79 Communiqué on the administration of the second dose of Sputnik V vaccine April 2021 Factiva software search Data analysis The audio recordings from the national press were completely transcribed. Content analysis of these transcripts, media reports and policy documents were done manually according to a code grid. We compared the data from the different sources (media content, audio records, policy documents) to triangulate the data and thus strengthen the internal validity (credibility) of the findings. We addressed inter-coding bias by having two researchers coding the material and allowing for adjustment upon a team consultation in association with a third researcher. Results The chronology of events pertaining to the COVID-19 response can be divided into five major phases or periods (described below) according to the national dynamics of the response: policy action or inaction, extensive nature of decisions or not, and the epidemiological context (increased in cases notification or not) ( Figure 2 ). Figure 2 Chronology of key events during the preparation and response to the COVID-19 pandemic in Guinea, January 2020 to November 2021. Phase 1: The anticipation of the pandemic response, 27 January to 20 March 2020 On 9 January 2020, the WHO declared the emergence of a novel coronavirus (SARS-CoV-2) in China. A few days after this event (20 January 2020), the ANSS started communication on the disease. Posts on the ANSS website indicated that this early communication aimed to assure the population that the health authorities were closely following the chronology of events related to this new disease and that, precautions were being taken to anticipate the health system response. On 28 January 2020, upon the demand of the ANSS, several national (Ministries of Health, Livestock, Transport, Environment and Water and Forestry) and international (WHO, The International Organization for Migration (IOM), the Centers for Disease Control and Prevention (CDC), Médecins Sans Frontières (MSF), Expertise France, etc.) actors met to evaluate the risk of the disease spreading in Guinea. During this meeting, an ad hoc committee was set up and entitled, among others, to vulgarize the symptoms and mode of transmission of the disease, and elaborate guidelines for identifying and managing suspected cases. The WHO statement of 30 January 2020 classifying the virus as a threat to human health reinforced the health actors in their approach of anticipating the country's response. Explicably, the WHO declaration triggered the development of the COVID-19 preparedness and response plan (February 2020) [Ministère de la Santé: ( 21 )]. This plan, similar to the Ebola response strategy, was built around seven strategic axes— (1) epidemiological surveillance, (2) laboratory, (3) clinical care, (4) information and communication, (5) logistics, (6) coordination and (7) technical assistance—each of which incorporated two key phases of the pandemic control: alert and response. Measures undertaken in the alert phase, such as strengthening preventive measures at the 44 potential entry points primarily identified (airports, seaports, mining ports, land borders, etc.) intended to avert the introduction of the virus in the country [Ministère de la Santé: ( 21 )]. Meanwhile, the response phase sought to limit the spreading of the virus and included measures such as the reinforcement of diagnostic capacities, training of healthcare providers, dignified and secure burial, and improvements of clinical care conditions through the rehabilitation of epidemic treatment centers ( 21 ). However, some challenges that arose in this period should be noted. First, the cost of implementing this plan, estimated at USD 47 million (Currency of the United States of America), was not secured. At the same moment, the government increased military salaries by 20% (16 March 2020) ( 22 ). This certainly impacted the earlier national response to the pandemic. For instance, a situational analysis undertaken in March 2020 showed that the health system was unprepared to deal with this new crisis. In fact, the analysis showed important insufficiencies in terms of the availability of supplies (masks, gloves, medicines, etc.) and intensive care equipment, the inadequacy of epidemic treatment centers and poor awareness of healthcare workers. Moreover, the COVID-19 threat Management Guidelines, released in February 2020, indicated that in February 2020 itself, more than 15 000 entries were recorded at the national airport. Meanwhile, only three laboratories were available to test for COVID-19; making it challenging to test all suspected cases at arrival or ensure a 14-day follow-up, as planned in the COVID-19 preparedness and response plan of February 2020 ( 23 ). Nevertheless, these national laboratories had the necessary equipment for the diagnosis of pathogens such as SARS-CoV-2 and Ebola ( 24 ). Also, the country had nearly 2,000 health workers trained in community-based epidemiological surveillance ( 21 ). The news of the devastating effects of the disease in Europe and Asia created fear and panic among part of the population, prompting their reactions to call for implementing social restrictive measures and postponing campaigns and elections (scheduled between 15 February and 23 March 2020). A health official interviewed on local media said 'we can't hold elections while there is a pandemic. We have officially recorded two cases; it would be taking too much risk to go to these elections ( 25 ). On 21 March 2020, a presidential decree was broadcasted, banning international travel, mass gathering (more than 20 people) and recommending people freshly coming from Europe and Asia to self-quarantine for 14 days. At the same time, however, legislative and referendum elections were maintained for 22 March 2020 as well as political campaigns ( 26 ). Phase 2: Sudden political action and implementation of strict restrictive measures, 21 March to 14 July 2020 By 31 March 2020, the country accounted for a total of 16 cases and zero death of COVID-19. Thereafter, we observed an explosion of cases with 1,335 new infections and 7 deaths recorded on 30 April 2020. By the end of May 2020, the country reached the first peak of 2355 new infections and 16 deaths. In response to this, the government established three sets of decisions: (1) social restrictive measures, (2) the pandemic's governing framework and (3) initiatives intended to mitigate the pandemic's effects. Social restrictive measures: The one-way response strategy At this stage of the pandemic response, several restrictive measures—similar to most western countries confronted with the acute phase of the disease—were established, giving the impression that the pandemic response strategy is unique regardless of the context. This period was also dominated by the rise of societal responses to the pandemic control decisions. On 30 March 2020, President Alpha Conde, in a solemn speech, called on the Guinean people to take a patriotic surge against the rapid increase in cases and the spread of the epidemic beyond the capital and four rural health districts. Referring to this epidemiological context and the WHO prediction on Africa, he declared a state of emergency and announced a set of social restrictions measures until 15 May 2020: closure of schools, land borders, worship places, bars and other places of recreation; a reduction on the number of people allowed on public transport; and a curfew from 9 p.m. to 5 a.m. The government also declared the compulsory wearing of facemasks from 18 April 2020 onwards, and that offenders would be prevented from movement and would be charged a disobedience fee of USD 3. On 31 March 2020, the Ministry of Health (MoH) announced the isolation of the capital, Conakry, thus banning domestic transport toward the countryside. Nevertheless, some of these decisions generated subsequent tensions between actors, including community resistance; and human and economic consequences. For instance, the limitation of the number of passengers in public transportation led to a 50% increase in fares ( 27 ). Moreover, following the setting up of military checkpoints to ban domestic travel toward the countryside, riots and protests broke out on 12 May 2020 in cities surrounding the capital (Dubreka and Coyah) leading to the death of five people ( 14 , 28 ). Although the contribution of the security department has been essential in the pandemic response— transporting medical supplies, mobilizing military health personnel and facilities for managing COVID-19 cases, and implementing policy decisions—some tensions were observed between the security department and the transporters' union. These tensions were linked to the setup of several "unnecessary" checkpoints and their implication on citizens' movements ( 29 ). Furthermore, several religious leaders and communities reacted to the closure of mosques arguing that this was a violation of their faith ( 9 ). Finally, an estimate of GNF 17 billion (GNF= Guinean currency) in economic losses was revealed by an official of "the Guinean Association of bar, motel, restaurant and nightclub owners" 6 months after the pandemic control measures were adopted in Guinea ( 30 , 31 ). This loss was likely sustained by the prolonged closure of public places ( 31 ). Organization of the health system response This period consisted of establishing governing bodies and standards adapted to the challenges encountered in the response to the first wave. This included establishing the Leadership of the pandemic response and the (re-)organization of care delivery for interrupting the transmission chain. The setting up of a governing framework: The repositioning struggle Centralization of power and decision-making in the response to the first wave was originally considered ( 8 , 9 ). On 27 March 2020, a presidential decree mandated the ANSS as the leading body of the pandemic response. Some national institutions considered it unacceptable to be overhung by the ANSS in their field of expertise ( 9 ). Others, such as teaching hospitals managers, threatened to withdraw from the fight against the pandemic and launched a strike after the ANSS recruitment of 350 medical students and their deployment in the pandemic treatment centers –instead of senior medical doctors and trainees in teaching hospitals ( 32 ). Furthermore, the MoH, the actual affiliated institution of the ANSS, showed its disagreement with this decision, arguing that it was not consulted beforehand. This situation prompted the President to initiate a series of dialogues and consultations with the various protagonists (about 30 people including the Minister of Health and the ANSS General Director) and to make them swear "to serve the Guinean people, and work together to fight the invisible enemy" ( 33 ). In the days that followed, the pandemic response framework was established to involve a large range of national and international actors. This framework also included the scientific committee created by a presidential decree (10 April 2020) and composed of 17 national public health experts and academics with the mandate to monitor, analyse and advise the government on the pandemic response. Care delivery organization for interrupting the transmission chain: Continuous adjustment of policy decisions Strategies for interrupting the chain of community transmission were planned early in the pandemic preparedness and response strategy. However, their implementation experienced a slow pace. As COVID-19 cases increased in April 2020, several measures were taken to readjust the testing, contact tracing and isolation strategies. First, the testing strategy of February 2020 identified three national laboratories in charge of performing RT-PCR antigen tests for COVID-19 suspected cases. Understandably, no rapid diagnostic antigen tests (RDTs) was available at the onset of the pandemic in Guinea, at least until the first donation of 20 000 COVID-19 antigen tests by a humanitarian Foundation on 1 April 2020 ( 34 ). However, as community transmission intensified and as more RDTs became available from mid-April 2020 onwards, several changes in testing strategy were observed including decentralizing sample collection sites in Conakry (the epicenter of the disease with more than 90% of cases in mid-April 2020) and the progressive decentralization and capacity building of other national and regional laboratories (from May 2020 onwards) ( 35 ). Despite this, a slow pace in testing intake was observed, with the delivery of results taking 72 hours or more instead of the 48 hours recommended ( 36 ). As part of the "Stop COVID-19 in 60 Days" initiative (May 2020), mobile sampling units were set up in Conakry, with the involvement of community health workers ( 37 ). This was aimed at alleviating community reluctance and breaking the transmission chain at an accelerated pace ( 37 ). Finally, screening units were set up at bus stations for mandatory testing (free of charge) of all passengers leaving Conakry ( 38 ). A similar scheme was also implemented for foreign travelers in late July 2020, but with a payment of USD 65 per test. Second, the strategy for case isolation originally included two former Ebola treatment centers (one in Conakry and the other in the countryside). These centers were made operational in early February 2020 [Ministère de la Santé: ( 21 )]. However, their combined capacity did not exceed 73 beds. In addition, a video of a press conference published on the ANSS website (March 16, 2020), revealed that the quality of the facilities was considered sub-optimal by the first patients, mostly the Guinean elite. This led to rehabilitating the Donka national hospital premises and the transfer of COVID-19 patients there. Nevertheless, the capacity of the Donka hospital was rapidly exceeded by mid-April 2020 with many patients waiting in the hospital's corridors to receive their first care ( 15 ). Consequently, four other treatment centers were established in Conakry to decentralize COVID-19 care delivery. For instance, in April 2020, following the notification of a wave of contamination in the prisons of Conakry, a treatment center was created at the Alpha Yaya Diallo military camp to primarily receive prisoners of civil rights. Additionally, intensive care services of the Donka and other private hospitals were requested for managing complicated COVID-19 cases. In late April 2020, decentralization of care at the household level was adopted but only officialised in late May 2020 through the "Stop COVID-19 in 60 Days" initiative. This implied that COVID-19 patients would be treated at home, without the need to be admitted to a treatment center. Unfortunately, this decision turned out to be an opportunity for some actors, especially the elite and senior civil servants, to negotiate their home care with private medical staff, or to resist admission to a treatment center—because of their perceived poor quality. Subsequent deaths of high executive officials—including a minister, a chairman of a public institution and several senior public servants—were reported between May and June 2020 and were seemingly related to their late admission to treatment centers ( 15 ). In response to this, an act was introduced by the MoH on 27 July 2020 to prohibit home confinement. This act highlighted that "any person who tests positive would immediately be transferred, either voluntarily or by force, to a treatment center." The re-adaptation of isolation strategies was not limited only to the capital, but also the countryside. For instance, in the countryside, care for COVID-19 patients was initially provided at the regional level, meaning that patients diagnosed in health districts were admitted to regional treatment centers. As the virus spread extensively from mid-April 2020 onwards, epidemic treatment centers at district levels, previously used for routine health care provision, were seized and rehabilitated for COVID-19 patients' care. Third, during the response to this first wave and beyond, the rise of new international players was noted. For instance, the role The Alliance for International Medical Action (ALIMA) was crucial in managing COVID-19 patients. In addition, several initiatives were undertaken by mining companies from China, Russia and the United Arab Emirates (UAE), based in Guinea under bilateral cooperation, during this first wave and beyond. For example, the contribution of the UAE—through Emirates Global Aluminium's Guinea Alumina Corporation—was crucial in building a treatment center in Conakry. It also appeared that three treatment centers located in mining zones in the countryside (Fria and Kindia) were financially and technically supported by Rusal, the Russian aluminum company producer ( 39 ). Another illustration was the involvement of a Cuban medical team, including 21 epidemiologists, in the COVID-19 response upon the government's request ( 40 ). Table 3 provides a list of actors and their responsibilities, as described in the pandemic response framework. Table 3 Health system actors and their roles in the COVID-19 response. Actors Area of responsibilities Roles in the COVID-19 pandemic response Ministry of Health: ANSS, National Directorate for Major Endemics and Disease Control, National Directorate of Community Health and Traditional Medicine Technical and Financial partners: WHO, CDC, IOM Epidemiological surveillance • Define surveillance guidelines and technical documents • Define strategies for identifying and monitoring cases and managing surveillance data • Update the definition of cases according to the evolution of the pandemic • Update the areas of intervention of the prefectures at risk; follow up on the collection of cases and contacts Ministry of Health: ANSS, central pharmacy of Guinea, material and equipment service, National Directorate of Pharmacy and Medicines, health department of Conakry, Expanded Programme on Immunization, National Blood Transfusion Center Ministry for cooperation Ministry of defense: Military Health Service, Customs directorate general Technical and Financial partners: World Food Program, WHO, CRS, Red Cross, UNICEF Logistics • Define the strategy for the supply • Management of supplies • Coordinate the acquisition of supplies as part of the response against COVID-19 • Carry out an inventory of stock requirements • Define the training programme for logistics personnel Ministry of Health : National Directorate of Laboratories, National Institute of Public Health, National Laboratory of Public Health, National Haemorrhagic Fever Laboratory, CERFIG, UAGCP, CREMS, LABOGUI Ministry of Education and Scientific Research : University of Conakry Technical and Financial partners: WHO, Pasteur Institute Laboratory • Determine the nature of the operating protocols to be used • Validate the laboratories to be integrated into the diagnostic system on the basis of conformity examinations • Determine the way in which sampling is organized • Determine the sample distribution circuit and sites • Organize the validity of the proposed tests • Assess the biosafety and biosecurity conditions of the laboratories • Organizing the management of samples • Organizing the training of laboratory staff at national level Ministry of Health: ANSS, national health promotion service, Ministry of Communication: national public and private media Youth Ministry: CENAFOB Ministry of social welfare Prime minister's cabinet: religious affairs secretariat Technical and Financial partners : UNICEF, USAID, UNFPA Communication and social mobilization • Define communication strategy • Design communication plan with a budget • Organize communication and social mobilization activities • Coordinate interventions in the media and communities • Involve leaders and other personalities in favor of the fight • Mobilize and involve community platforms and provide advice to the strategic committee • Collect and manage rumors Ministry of Health: Nationals hospitals of Donka, Ignace Deen and Sino-Guinean friendship, National Directorate of Health Facilities Ministry of Defense: Civil defense Technical and Financial Partners: MSF, ALIMA, JHPIEGO, OMS, France Expertise, Red Cross, International Federation of the Red Cross, IOM Clinical care • Identify and assess management sites according to the evolution of the epidemic • Produce and update management protocols according to the evolution of knowledge about the disease • Assess therapeutic needs • Define the Infection control and prevention protocol in the context of the response to COVID-19 • List the consumable drugs and equipment that can be used in the response to COVID-19 • Establish guidelines for transporting and burying bodies • Produce the training plan for management personnel in the response to COVID-19 Ministry of Health: ANSS, Head of the cabinet, bureau of strategy and development, financial affairs division, human resources department public procurement Ministry of Budget Ministry of Investment Technical and Financial Partners: World Bank, European Union Finances • Carry out the financial assessment of the needs for the implementation of the national response plan • Prepare the files and requests for the financing of the activities; to draw up the financial table of the response • Draw up the manual of procedure for the management of resources according to the conventions and national regulations • Give the status of the mobilization and execution of the national plan budget Inter-departmental committee including the Prime Minister's cabinet Strategic committee: MoH, ANSS, Technical and Financial Partner (WHO, CDC, IOM, UNICEF, etc.) Scientific committee: 17 members including Academics and Researchers Coordination • Provide a framework for consultation, guidance and decision-making for the response to the pandemic • Coordinate activities related to the COVID-19 response Finally, the emergence of several local initiatives was noted during this phase, including local production of hydroalcoholic solutions by several training and research institutions, local manufacture of spraying devices and their installation at the airport and COVID-19 treatment centers, and manufacture of devices for automatic recording of COVID-19 patients' parameters (temperature, blood pressure, heart rate, etc.). Another initiative, not the least, was the self-declaration of elites suffering from COVID-19 on social media, and their admission to public hospitals. This was of particular importance as it occurred in a period when the elites were criticized for not accepting admission to treatment centers. Furthermore, several sewing workshops engaged in producing traditional masks (made from cloth) and sold them at the government-set price (USD 0.50 per unit). Lastly, therapeutic trials based on phytomedicines were undertaken in COVID-19 treatment centers ( 41 ). Mitigation measures overlooking some local realities On 2 April 2020, Guinea adopted an economic contingency plan to mitigate the potential effects of the pandemic. This plan has three main components: (1) the health component, (2) the social component and (3) the economic component (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The health component was aimed at financing the rollout of the pandemic preparedness and response and strengthening the health system capacities (infrastructure, equipment, etc.). The social component included cash transfers to 1.6 million vulnerable people identified nationwide; the full coverage of water, electricity and transport bills for 3 months; and the provision of prevention equipment to almost 200 000 targeted households (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In addition to this, many companies, public departments, social and political movements initiated the distribution of facemasks and handwashing devices to the population as their contribution to the pandemic control. The economic component comprised measures that support the private sector including a three-month postponement of fiscal taxes payment for companies and an exemption from duties and taxes on health equipment dedicated to the COVID-19 response (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). This policy, estimated at USD 400 million, had to be financed through two mechanisms: the national response funds and the contribution of bi- or multilateral partners (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The national funds were sourced from the withholding taxes on fuel prices, savings from the deferral of external public debt servicing and funds reallocation from the MoH and other ministries (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In April 2020, health system actors noticed a sharp decrease in health care utilization between January and March 2020. This was reportedly linked to the stock out of medical supplies, including protective equipment and fear among health personnel and population to provide or seek care ( 42 ). To mitigate this, a "health system resilience strategy for the continuity of care" was developed in May 2020 ( 42 ). This strategy was layered in five specific objectives including ensuring infection prevention and control among healthcare providers and beneficiaries, availability of health services, and effective communication for patients and their caregivers. The cost of rolling out this strategy was estimated at USD 11 million. In the education sector, online courses on the national public media were organized between April and July 2020, for candidates for national exams. This was initiated to mitigate the effects of the closure of schools during the first wave. Finally, in many public services, telework and the lay-off of contractual workers and trainees were undertaken to limit the virus from spreading. However, several questions were raised on the equity of this teaching method. Many media questioned, for instance, how students from poor families managed to follow online courses. Some other media also questioned how the social component of the economic contingency plan was rolled out in the context of political turmoil. Phase 3: Alleviation of restrictive measures: Politics flexibility to social demands, 15 July 2020 to 31 January 2021 On 15 July 2020, President Alpha Conde, in an address to the nation, announced the alleviation of certain restrictions: the gradual reopening of air and land borders in accordance with the reciprocity measures between countries; the curfew was also pushed back from midnight to 4 a.m. (as against 9 p.m. to 5 a.m. initially). On 17 July 2020, a protocol for the reopening of air borders was adopted by the Ministries of Transport and Health. This protocol describes the arrangements in place for health control at air entry points (e.g., certificate of RT-PCR test negativity valid for 5 days) and community-based follow-up of travelers' contacts (e.g., self-quarantine for 14 days) in Guinea. This alleviation of restrictive measures came on the eve of a series of statements by some religious leaders, asking the authorities to facilitate the celebration of "The Tabaski" scheduled for 31 July 2020. The decree of isolating the capital was also repealed to facilitate the movement of people to the countryside. During this period, the country was still confronted with the first wave of the pandemic with increases in confirmed and hospital fatalities cases, at least until September 2020 (June 2020: 1645 cases, 8 deaths; July 2020: 1895 cases, 15 deaths; August 2020: 2009 cases, 13 deaths; September 2020: 1297 cases, 7 deaths). Despite this, the government pursued the adoption of pandemic alleviation measures including gradual schools and public places reopening (October 2020), the launch of the presidential election campaigns (September to October 2020) and the holding of presidential elections (18 October 2020). The ANSS, in response to the intensification of mass movement in the country, adopted "active case finding" as a strategy for interrupting the COVID-19 transmission chain (October 2020). A key element of this strategy consisted of a campaign for systematic screening of 80% of high-risk groups including patients with influenza-like syndromes, febrile patients with a negative RDT for malaria, health workers, security forces and people over 60 years old. In September 2020, a bi-annual immunization plan was developed and adopted in December 2020, with the objective of vaccinating 95% of targeted people (18 years and older) before January 2022. In December 2020, using 60 doses of Sputnik V, Guinea was among the countries that initiated vaccination against COVID-19 early ( 43 ). Despite these, the delivery of additional COVID-19 vaccines was not effective until the country had experienced a new wave of the COVID-19 pandemic in February 2021. Phase 4: The multiple epidemics period, 1 February to 30 June 2021 By February 2021, the country experienced a new wave of the pandemic; 1,006 confirmed cases and 5 hospital deaths in February 2021, 4,198 cases and 30 deaths in March 2021, and 799 cases and 14 deaths in June 2021. This period coincided with the emergence of Ebola (on 12 February 2021), Lassa Fever (on 18 May 2021), and many other recurrent infectious diseases including measles, yellow fever, poliomyelitis and meningitis ( 7 ). The concomitant emergence of these epidemics overstretched the ANSS, the principal national public health institute, and had an impact on COVID-19 surveillance and response ( 18 , 44 ). It was noted also that the surveillance and response to recurrent epidemics (measles, meningitis, poliomyelitis, etc.) were neglected in favor of Ebola ( 7 ). Although the national vaccination plan has been in place since December 2020, it was not until 3 March 2021 that the country received the first bulk quantity of vaccines (~152 000 doses of Sinopharm). Consequently, the mass vaccination campaign against COVID-19 was only launched on 5 March 2021. Thereafter, 20 000 doses of Sputnik (8 March 2021), 69 000 doses of AstraZeneca (29 March 2021), 23 000 doses of Sinovac (19 April 2021), 194 000 doses of AstraZeneca (11 April 2021) and 300 000 doses of Sinovac (18 April 2021) were delivered to ANSS. These vaccines were sourced through three main channels: (1) bilateral cooperation—Sinopharm and Sinovac vaccines were purchased by the government from China while Sputnik vaccines from Russia; (2) regional cooperation-−69 000 doses of the AstraZeneca vaccine were offered by the African Union; and (3) the Covax initiative provided 194 000 doses of the AstraZeneca vaccine ( 7 ). Despite efforts deployed, the vaccination coverage remained low as of 31 May 2021. Indeed, by this date, only 4% of the target population had received their first dose of the vaccine. Additionally, reinforcing the response to Ebola and Lassa fever that involved graduates of the CDC-led Field Epidemiology Training Programme ( 45 , 46 ). Furthermore, some experiences of the COVID-19 response were applied for Ebola control. For example, all national and international workforce deployed in N'zérékoré for the Ebola response were administered an Ebola vaccine. Phase 5: The COVID-19 variants period, 1 July to 31 October 2021 The news of the end of Ebola (19 June 2021) and the second wave of COVID-19 (29 June 2021) lasted only a few weeks since in July 2021 a new wave started. This wave—the deadliest in the country since the emergence of COVID-19—coincided with the notification of Delta, Alpha, Beta, Gamma and Eta COVID-19 variants in the country ( 47 ). This wave bore, respectively, 25 and 59% of COVID-confirmed cases and deaths cases registered in the country from March 2020 to November 2021. The emergence of SARS-CoV-2 variants coincided with low vaccination rate of the population. For example, as of August 2021, only 41% of the 2 900 000 doses of vaccines supplied to Guinea had been used, and full immunization coverage of the population was only 3.56%. This full immunization coverage was five times lower than the 20% projection of the national COVID-19 immunization plan. In addition, strong disparities were noted between the country's health regions in terms of complete vaccination coverage against COVID-19. The complete vaccination coverage of the Conakry health region was 15.29% compared to 1.23% for the Labé, Kindia and Boké regions, and 1% for the Mamou, Kankan, Faranah and N'zérékoré regions ( Figure 3 ). Figure 3 Full vaccination coverage against COVID-19 per administrative regions in the country, August 2021, Guinea. In response to this, in September 2021, the Guinean government adopted the "accelerated COVID-19 vaccination plan" with the aim of "improving equity of access to vaccines and contributing to the reduction of mortality and morbidity related to COVID-19, through the complete vaccination of 20% of the population before the end of December 2021" ( 48 ). This plan was coupled with the adoption of the "vaccination pass" which made movements between cities and access to public services conditional to the presentation of a vaccination card against COVID-19 ( 47 ). This plan was based on 11 acceleration measures, including (1) the use of mobile and semi-mobile strategies (identification of neighborhoods by communes/health districts for sequential immunization), (2) the involvement of certain public and private health centers in immunization and (3) the involvement of civil society associations and corporations in immunization ( 48 ). In addition, through this strategy, the supply of COVID-19 vaccines increased from 2.9 to 5.7 million doses between August and November 2021. Also, during the same period, full immunization coverage of the general population increased from 3.56 to 6.95%. However, the objectives of complete vaccination of 20% of the general population by the end of December were far from being achieved. Similarly, disparities persisted in terms of access to COVID-19 vaccines—on 22 November 2021, full vaccination coverage in Conakry was 25.34%, compared with 3.47% for the Labé, Kindia and Boké regions, and 2.64% for the Mamou, Kankan, Faranah and N'zérékoré regions. COVID-19-related activities, including vaccination, experienced a slower pace after the military coup of September 2021 as a result of, among others: the freezing of ANSS bank accounts; the reshuffling of management positions at the ANSS and the MoH; the abolition of the vaccine pass and the lifting of the curfew, and the state of health emergency ( 49 ). All these raised questioned on how health system actors will reorganize to pursue the COVID-19 control in Guinea. Phase 1: The anticipation of the pandemic response, 27 January to 20 March 2020 On 9 January 2020, the WHO declared the emergence of a novel coronavirus (SARS-CoV-2) in China. A few days after this event (20 January 2020), the ANSS started communication on the disease. Posts on the ANSS website indicated that this early communication aimed to assure the population that the health authorities were closely following the chronology of events related to this new disease and that, precautions were being taken to anticipate the health system response. On 28 January 2020, upon the demand of the ANSS, several national (Ministries of Health, Livestock, Transport, Environment and Water and Forestry) and international (WHO, The International Organization for Migration (IOM), the Centers for Disease Control and Prevention (CDC), Médecins Sans Frontières (MSF), Expertise France, etc.) actors met to evaluate the risk of the disease spreading in Guinea. During this meeting, an ad hoc committee was set up and entitled, among others, to vulgarize the symptoms and mode of transmission of the disease, and elaborate guidelines for identifying and managing suspected cases. The WHO statement of 30 January 2020 classifying the virus as a threat to human health reinforced the health actors in their approach of anticipating the country's response. Explicably, the WHO declaration triggered the development of the COVID-19 preparedness and response plan (February 2020) [Ministère de la Santé: ( 21 )]. This plan, similar to the Ebola response strategy, was built around seven strategic axes— (1) epidemiological surveillance, (2) laboratory, (3) clinical care, (4) information and communication, (5) logistics, (6) coordination and (7) technical assistance—each of which incorporated two key phases of the pandemic control: alert and response. Measures undertaken in the alert phase, such as strengthening preventive measures at the 44 potential entry points primarily identified (airports, seaports, mining ports, land borders, etc.) intended to avert the introduction of the virus in the country [Ministère de la Santé: ( 21 )]. Meanwhile, the response phase sought to limit the spreading of the virus and included measures such as the reinforcement of diagnostic capacities, training of healthcare providers, dignified and secure burial, and improvements of clinical care conditions through the rehabilitation of epidemic treatment centers ( 21 ). However, some challenges that arose in this period should be noted. First, the cost of implementing this plan, estimated at USD 47 million (Currency of the United States of America), was not secured. At the same moment, the government increased military salaries by 20% (16 March 2020) ( 22 ). This certainly impacted the earlier national response to the pandemic. For instance, a situational analysis undertaken in March 2020 showed that the health system was unprepared to deal with this new crisis. In fact, the analysis showed important insufficiencies in terms of the availability of supplies (masks, gloves, medicines, etc.) and intensive care equipment, the inadequacy of epidemic treatment centers and poor awareness of healthcare workers. Moreover, the COVID-19 threat Management Guidelines, released in February 2020, indicated that in February 2020 itself, more than 15 000 entries were recorded at the national airport. Meanwhile, only three laboratories were available to test for COVID-19; making it challenging to test all suspected cases at arrival or ensure a 14-day follow-up, as planned in the COVID-19 preparedness and response plan of February 2020 ( 23 ). Nevertheless, these national laboratories had the necessary equipment for the diagnosis of pathogens such as SARS-CoV-2 and Ebola ( 24 ). Also, the country had nearly 2,000 health workers trained in community-based epidemiological surveillance ( 21 ). The news of the devastating effects of the disease in Europe and Asia created fear and panic among part of the population, prompting their reactions to call for implementing social restrictive measures and postponing campaigns and elections (scheduled between 15 February and 23 March 2020). A health official interviewed on local media said 'we can't hold elections while there is a pandemic. We have officially recorded two cases; it would be taking too much risk to go to these elections ( 25 ). On 21 March 2020, a presidential decree was broadcasted, banning international travel, mass gathering (more than 20 people) and recommending people freshly coming from Europe and Asia to self-quarantine for 14 days. At the same time, however, legislative and referendum elections were maintained for 22 March 2020 as well as political campaigns ( 26 ). Phase 2: Sudden political action and implementation of strict restrictive measures, 21 March to 14 July 2020 By 31 March 2020, the country accounted for a total of 16 cases and zero death of COVID-19. Thereafter, we observed an explosion of cases with 1,335 new infections and 7 deaths recorded on 30 April 2020. By the end of May 2020, the country reached the first peak of 2355 new infections and 16 deaths. In response to this, the government established three sets of decisions: (1) social restrictive measures, (2) the pandemic's governing framework and (3) initiatives intended to mitigate the pandemic's effects. Social restrictive measures: The one-way response strategy At this stage of the pandemic response, several restrictive measures—similar to most western countries confronted with the acute phase of the disease—were established, giving the impression that the pandemic response strategy is unique regardless of the context. This period was also dominated by the rise of societal responses to the pandemic control decisions. On 30 March 2020, President Alpha Conde, in a solemn speech, called on the Guinean people to take a patriotic surge against the rapid increase in cases and the spread of the epidemic beyond the capital and four rural health districts. Referring to this epidemiological context and the WHO prediction on Africa, he declared a state of emergency and announced a set of social restrictions measures until 15 May 2020: closure of schools, land borders, worship places, bars and other places of recreation; a reduction on the number of people allowed on public transport; and a curfew from 9 p.m. to 5 a.m. The government also declared the compulsory wearing of facemasks from 18 April 2020 onwards, and that offenders would be prevented from movement and would be charged a disobedience fee of USD 3. On 31 March 2020, the Ministry of Health (MoH) announced the isolation of the capital, Conakry, thus banning domestic transport toward the countryside. Nevertheless, some of these decisions generated subsequent tensions between actors, including community resistance; and human and economic consequences. For instance, the limitation of the number of passengers in public transportation led to a 50% increase in fares ( 27 ). Moreover, following the setting up of military checkpoints to ban domestic travel toward the countryside, riots and protests broke out on 12 May 2020 in cities surrounding the capital (Dubreka and Coyah) leading to the death of five people ( 14 , 28 ). Although the contribution of the security department has been essential in the pandemic response— transporting medical supplies, mobilizing military health personnel and facilities for managing COVID-19 cases, and implementing policy decisions—some tensions were observed between the security department and the transporters' union. These tensions were linked to the setup of several "unnecessary" checkpoints and their implication on citizens' movements ( 29 ). Furthermore, several religious leaders and communities reacted to the closure of mosques arguing that this was a violation of their faith ( 9 ). Finally, an estimate of GNF 17 billion (GNF= Guinean currency) in economic losses was revealed by an official of "the Guinean Association of bar, motel, restaurant and nightclub owners" 6 months after the pandemic control measures were adopted in Guinea ( 30 , 31 ). This loss was likely sustained by the prolonged closure of public places ( 31 ). Organization of the health system response This period consisted of establishing governing bodies and standards adapted to the challenges encountered in the response to the first wave. This included establishing the Leadership of the pandemic response and the (re-)organization of care delivery for interrupting the transmission chain. The setting up of a governing framework: The repositioning struggle Centralization of power and decision-making in the response to the first wave was originally considered ( 8 , 9 ). On 27 March 2020, a presidential decree mandated the ANSS as the leading body of the pandemic response. Some national institutions considered it unacceptable to be overhung by the ANSS in their field of expertise ( 9 ). Others, such as teaching hospitals managers, threatened to withdraw from the fight against the pandemic and launched a strike after the ANSS recruitment of 350 medical students and their deployment in the pandemic treatment centers –instead of senior medical doctors and trainees in teaching hospitals ( 32 ). Furthermore, the MoH, the actual affiliated institution of the ANSS, showed its disagreement with this decision, arguing that it was not consulted beforehand. This situation prompted the President to initiate a series of dialogues and consultations with the various protagonists (about 30 people including the Minister of Health and the ANSS General Director) and to make them swear "to serve the Guinean people, and work together to fight the invisible enemy" ( 33 ). In the days that followed, the pandemic response framework was established to involve a large range of national and international actors. This framework also included the scientific committee created by a presidential decree (10 April 2020) and composed of 17 national public health experts and academics with the mandate to monitor, analyse and advise the government on the pandemic response. Care delivery organization for interrupting the transmission chain: Continuous adjustment of policy decisions Strategies for interrupting the chain of community transmission were planned early in the pandemic preparedness and response strategy. However, their implementation experienced a slow pace. As COVID-19 cases increased in April 2020, several measures were taken to readjust the testing, contact tracing and isolation strategies. First, the testing strategy of February 2020 identified three national laboratories in charge of performing RT-PCR antigen tests for COVID-19 suspected cases. Understandably, no rapid diagnostic antigen tests (RDTs) was available at the onset of the pandemic in Guinea, at least until the first donation of 20 000 COVID-19 antigen tests by a humanitarian Foundation on 1 April 2020 ( 34 ). However, as community transmission intensified and as more RDTs became available from mid-April 2020 onwards, several changes in testing strategy were observed including decentralizing sample collection sites in Conakry (the epicenter of the disease with more than 90% of cases in mid-April 2020) and the progressive decentralization and capacity building of other national and regional laboratories (from May 2020 onwards) ( 35 ). Despite this, a slow pace in testing intake was observed, with the delivery of results taking 72 hours or more instead of the 48 hours recommended ( 36 ). As part of the "Stop COVID-19 in 60 Days" initiative (May 2020), mobile sampling units were set up in Conakry, with the involvement of community health workers ( 37 ). This was aimed at alleviating community reluctance and breaking the transmission chain at an accelerated pace ( 37 ). Finally, screening units were set up at bus stations for mandatory testing (free of charge) of all passengers leaving Conakry ( 38 ). A similar scheme was also implemented for foreign travelers in late July 2020, but with a payment of USD 65 per test. Second, the strategy for case isolation originally included two former Ebola treatment centers (one in Conakry and the other in the countryside). These centers were made operational in early February 2020 [Ministère de la Santé: ( 21 )]. However, their combined capacity did not exceed 73 beds. In addition, a video of a press conference published on the ANSS website (March 16, 2020), revealed that the quality of the facilities was considered sub-optimal by the first patients, mostly the Guinean elite. This led to rehabilitating the Donka national hospital premises and the transfer of COVID-19 patients there. Nevertheless, the capacity of the Donka hospital was rapidly exceeded by mid-April 2020 with many patients waiting in the hospital's corridors to receive their first care ( 15 ). Consequently, four other treatment centers were established in Conakry to decentralize COVID-19 care delivery. For instance, in April 2020, following the notification of a wave of contamination in the prisons of Conakry, a treatment center was created at the Alpha Yaya Diallo military camp to primarily receive prisoners of civil rights. Additionally, intensive care services of the Donka and other private hospitals were requested for managing complicated COVID-19 cases. In late April 2020, decentralization of care at the household level was adopted but only officialised in late May 2020 through the "Stop COVID-19 in 60 Days" initiative. This implied that COVID-19 patients would be treated at home, without the need to be admitted to a treatment center. Unfortunately, this decision turned out to be an opportunity for some actors, especially the elite and senior civil servants, to negotiate their home care with private medical staff, or to resist admission to a treatment center—because of their perceived poor quality. Subsequent deaths of high executive officials—including a minister, a chairman of a public institution and several senior public servants—were reported between May and June 2020 and were seemingly related to their late admission to treatment centers ( 15 ). In response to this, an act was introduced by the MoH on 27 July 2020 to prohibit home confinement. This act highlighted that "any person who tests positive would immediately be transferred, either voluntarily or by force, to a treatment center." The re-adaptation of isolation strategies was not limited only to the capital, but also the countryside. For instance, in the countryside, care for COVID-19 patients was initially provided at the regional level, meaning that patients diagnosed in health districts were admitted to regional treatment centers. As the virus spread extensively from mid-April 2020 onwards, epidemic treatment centers at district levels, previously used for routine health care provision, were seized and rehabilitated for COVID-19 patients' care. Third, during the response to this first wave and beyond, the rise of new international players was noted. For instance, the role The Alliance for International Medical Action (ALIMA) was crucial in managing COVID-19 patients. In addition, several initiatives were undertaken by mining companies from China, Russia and the United Arab Emirates (UAE), based in Guinea under bilateral cooperation, during this first wave and beyond. For example, the contribution of the UAE—through Emirates Global Aluminium's Guinea Alumina Corporation—was crucial in building a treatment center in Conakry. It also appeared that three treatment centers located in mining zones in the countryside (Fria and Kindia) were financially and technically supported by Rusal, the Russian aluminum company producer ( 39 ). Another illustration was the involvement of a Cuban medical team, including 21 epidemiologists, in the COVID-19 response upon the government's request ( 40 ). Table 3 provides a list of actors and their responsibilities, as described in the pandemic response framework. Table 3 Health system actors and their roles in the COVID-19 response. Actors Area of responsibilities Roles in the COVID-19 pandemic response Ministry of Health: ANSS, National Directorate for Major Endemics and Disease Control, National Directorate of Community Health and Traditional Medicine Technical and Financial partners: WHO, CDC, IOM Epidemiological surveillance • Define surveillance guidelines and technical documents • Define strategies for identifying and monitoring cases and managing surveillance data • Update the definition of cases according to the evolution of the pandemic • Update the areas of intervention of the prefectures at risk; follow up on the collection of cases and contacts Ministry of Health: ANSS, central pharmacy of Guinea, material and equipment service, National Directorate of Pharmacy and Medicines, health department of Conakry, Expanded Programme on Immunization, National Blood Transfusion Center Ministry for cooperation Ministry of defense: Military Health Service, Customs directorate general Technical and Financial partners: World Food Program, WHO, CRS, Red Cross, UNICEF Logistics • Define the strategy for the supply • Management of supplies • Coordinate the acquisition of supplies as part of the response against COVID-19 • Carry out an inventory of stock requirements • Define the training programme for logistics personnel Ministry of Health : National Directorate of Laboratories, National Institute of Public Health, National Laboratory of Public Health, National Haemorrhagic Fever Laboratory, CERFIG, UAGCP, CREMS, LABOGUI Ministry of Education and Scientific Research : University of Conakry Technical and Financial partners: WHO, Pasteur Institute Laboratory • Determine the nature of the operating protocols to be used • Validate the laboratories to be integrated into the diagnostic system on the basis of conformity examinations • Determine the way in which sampling is organized • Determine the sample distribution circuit and sites • Organize the validity of the proposed tests • Assess the biosafety and biosecurity conditions of the laboratories • Organizing the management of samples • Organizing the training of laboratory staff at national level Ministry of Health: ANSS, national health promotion service, Ministry of Communication: national public and private media Youth Ministry: CENAFOB Ministry of social welfare Prime minister's cabinet: religious affairs secretariat Technical and Financial partners : UNICEF, USAID, UNFPA Communication and social mobilization • Define communication strategy • Design communication plan with a budget • Organize communication and social mobilization activities • Coordinate interventions in the media and communities • Involve leaders and other personalities in favor of the fight • Mobilize and involve community platforms and provide advice to the strategic committee • Collect and manage rumors Ministry of Health: Nationals hospitals of Donka, Ignace Deen and Sino-Guinean friendship, National Directorate of Health Facilities Ministry of Defense: Civil defense Technical and Financial Partners: MSF, ALIMA, JHPIEGO, OMS, France Expertise, Red Cross, International Federation of the Red Cross, IOM Clinical care • Identify and assess management sites according to the evolution of the epidemic • Produce and update management protocols according to the evolution of knowledge about the disease • Assess therapeutic needs • Define the Infection control and prevention protocol in the context of the response to COVID-19 • List the consumable drugs and equipment that can be used in the response to COVID-19 • Establish guidelines for transporting and burying bodies • Produce the training plan for management personnel in the response to COVID-19 Ministry of Health: ANSS, Head of the cabinet, bureau of strategy and development, financial affairs division, human resources department public procurement Ministry of Budget Ministry of Investment Technical and Financial Partners: World Bank, European Union Finances • Carry out the financial assessment of the needs for the implementation of the national response plan • Prepare the files and requests for the financing of the activities; to draw up the financial table of the response • Draw up the manual of procedure for the management of resources according to the conventions and national regulations • Give the status of the mobilization and execution of the national plan budget Inter-departmental committee including the Prime Minister's cabinet Strategic committee: MoH, ANSS, Technical and Financial Partner (WHO, CDC, IOM, UNICEF, etc.) Scientific committee: 17 members including Academics and Researchers Coordination • Provide a framework for consultation, guidance and decision-making for the response to the pandemic • Coordinate activities related to the COVID-19 response Finally, the emergence of several local initiatives was noted during this phase, including local production of hydroalcoholic solutions by several training and research institutions, local manufacture of spraying devices and their installation at the airport and COVID-19 treatment centers, and manufacture of devices for automatic recording of COVID-19 patients' parameters (temperature, blood pressure, heart rate, etc.). Another initiative, not the least, was the self-declaration of elites suffering from COVID-19 on social media, and their admission to public hospitals. This was of particular importance as it occurred in a period when the elites were criticized for not accepting admission to treatment centers. Furthermore, several sewing workshops engaged in producing traditional masks (made from cloth) and sold them at the government-set price (USD 0.50 per unit). Lastly, therapeutic trials based on phytomedicines were undertaken in COVID-19 treatment centers ( 41 ). Mitigation measures overlooking some local realities On 2 April 2020, Guinea adopted an economic contingency plan to mitigate the potential effects of the pandemic. This plan has three main components: (1) the health component, (2) the social component and (3) the economic component (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The health component was aimed at financing the rollout of the pandemic preparedness and response and strengthening the health system capacities (infrastructure, equipment, etc.). The social component included cash transfers to 1.6 million vulnerable people identified nationwide; the full coverage of water, electricity and transport bills for 3 months; and the provision of prevention equipment to almost 200 000 targeted households (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In addition to this, many companies, public departments, social and political movements initiated the distribution of facemasks and handwashing devices to the population as their contribution to the pandemic control. The economic component comprised measures that support the private sector including a three-month postponement of fiscal taxes payment for companies and an exemption from duties and taxes on health equipment dedicated to the COVID-19 response (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). This policy, estimated at USD 400 million, had to be financed through two mechanisms: the national response funds and the contribution of bi- or multilateral partners (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The national funds were sourced from the withholding taxes on fuel prices, savings from the deferral of external public debt servicing and funds reallocation from the MoH and other ministries (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In April 2020, health system actors noticed a sharp decrease in health care utilization between January and March 2020. This was reportedly linked to the stock out of medical supplies, including protective equipment and fear among health personnel and population to provide or seek care ( 42 ). To mitigate this, a "health system resilience strategy for the continuity of care" was developed in May 2020 ( 42 ). This strategy was layered in five specific objectives including ensuring infection prevention and control among healthcare providers and beneficiaries, availability of health services, and effective communication for patients and their caregivers. The cost of rolling out this strategy was estimated at USD 11 million. In the education sector, online courses on the national public media were organized between April and July 2020, for candidates for national exams. This was initiated to mitigate the effects of the closure of schools during the first wave. Finally, in many public services, telework and the lay-off of contractual workers and trainees were undertaken to limit the virus from spreading. However, several questions were raised on the equity of this teaching method. Many media questioned, for instance, how students from poor families managed to follow online courses. Some other media also questioned how the social component of the economic contingency plan was rolled out in the context of political turmoil. Social restrictive measures: The one-way response strategy At this stage of the pandemic response, several restrictive measures—similar to most western countries confronted with the acute phase of the disease—were established, giving the impression that the pandemic response strategy is unique regardless of the context. This period was also dominated by the rise of societal responses to the pandemic control decisions. On 30 March 2020, President Alpha Conde, in a solemn speech, called on the Guinean people to take a patriotic surge against the rapid increase in cases and the spread of the epidemic beyond the capital and four rural health districts. Referring to this epidemiological context and the WHO prediction on Africa, he declared a state of emergency and announced a set of social restrictions measures until 15 May 2020: closure of schools, land borders, worship places, bars and other places of recreation; a reduction on the number of people allowed on public transport; and a curfew from 9 p.m. to 5 a.m. The government also declared the compulsory wearing of facemasks from 18 April 2020 onwards, and that offenders would be prevented from movement and would be charged a disobedience fee of USD 3. On 31 March 2020, the Ministry of Health (MoH) announced the isolation of the capital, Conakry, thus banning domestic transport toward the countryside. Nevertheless, some of these decisions generated subsequent tensions between actors, including community resistance; and human and economic consequences. For instance, the limitation of the number of passengers in public transportation led to a 50% increase in fares ( 27 ). Moreover, following the setting up of military checkpoints to ban domestic travel toward the countryside, riots and protests broke out on 12 May 2020 in cities surrounding the capital (Dubreka and Coyah) leading to the death of five people ( 14 , 28 ). Although the contribution of the security department has been essential in the pandemic response— transporting medical supplies, mobilizing military health personnel and facilities for managing COVID-19 cases, and implementing policy decisions—some tensions were observed between the security department and the transporters' union. These tensions were linked to the setup of several "unnecessary" checkpoints and their implication on citizens' movements ( 29 ). Furthermore, several religious leaders and communities reacted to the closure of mosques arguing that this was a violation of their faith ( 9 ). Finally, an estimate of GNF 17 billion (GNF= Guinean currency) in economic losses was revealed by an official of "the Guinean Association of bar, motel, restaurant and nightclub owners" 6 months after the pandemic control measures were adopted in Guinea ( 30 , 31 ). This loss was likely sustained by the prolonged closure of public places ( 31 ). Organization of the health system response This period consisted of establishing governing bodies and standards adapted to the challenges encountered in the response to the first wave. This included establishing the Leadership of the pandemic response and the (re-)organization of care delivery for interrupting the transmission chain. The setting up of a governing framework: The repositioning struggle Centralization of power and decision-making in the response to the first wave was originally considered ( 8 , 9 ). On 27 March 2020, a presidential decree mandated the ANSS as the leading body of the pandemic response. Some national institutions considered it unacceptable to be overhung by the ANSS in their field of expertise ( 9 ). Others, such as teaching hospitals managers, threatened to withdraw from the fight against the pandemic and launched a strike after the ANSS recruitment of 350 medical students and their deployment in the pandemic treatment centers –instead of senior medical doctors and trainees in teaching hospitals ( 32 ). Furthermore, the MoH, the actual affiliated institution of the ANSS, showed its disagreement with this decision, arguing that it was not consulted beforehand. This situation prompted the President to initiate a series of dialogues and consultations with the various protagonists (about 30 people including the Minister of Health and the ANSS General Director) and to make them swear "to serve the Guinean people, and work together to fight the invisible enemy" ( 33 ). In the days that followed, the pandemic response framework was established to involve a large range of national and international actors. This framework also included the scientific committee created by a presidential decree (10 April 2020) and composed of 17 national public health experts and academics with the mandate to monitor, analyse and advise the government on the pandemic response. Care delivery organization for interrupting the transmission chain: Continuous adjustment of policy decisions Strategies for interrupting the chain of community transmission were planned early in the pandemic preparedness and response strategy. However, their implementation experienced a slow pace. As COVID-19 cases increased in April 2020, several measures were taken to readjust the testing, contact tracing and isolation strategies. First, the testing strategy of February 2020 identified three national laboratories in charge of performing RT-PCR antigen tests for COVID-19 suspected cases. Understandably, no rapid diagnostic antigen tests (RDTs) was available at the onset of the pandemic in Guinea, at least until the first donation of 20 000 COVID-19 antigen tests by a humanitarian Foundation on 1 April 2020 ( 34 ). However, as community transmission intensified and as more RDTs became available from mid-April 2020 onwards, several changes in testing strategy were observed including decentralizing sample collection sites in Conakry (the epicenter of the disease with more than 90% of cases in mid-April 2020) and the progressive decentralization and capacity building of other national and regional laboratories (from May 2020 onwards) ( 35 ). Despite this, a slow pace in testing intake was observed, with the delivery of results taking 72 hours or more instead of the 48 hours recommended ( 36 ). As part of the "Stop COVID-19 in 60 Days" initiative (May 2020), mobile sampling units were set up in Conakry, with the involvement of community health workers ( 37 ). This was aimed at alleviating community reluctance and breaking the transmission chain at an accelerated pace ( 37 ). Finally, screening units were set up at bus stations for mandatory testing (free of charge) of all passengers leaving Conakry ( 38 ). A similar scheme was also implemented for foreign travelers in late July 2020, but with a payment of USD 65 per test. Second, the strategy for case isolation originally included two former Ebola treatment centers (one in Conakry and the other in the countryside). These centers were made operational in early February 2020 [Ministère de la Santé: ( 21 )]. However, their combined capacity did not exceed 73 beds. In addition, a video of a press conference published on the ANSS website (March 16, 2020), revealed that the quality of the facilities was considered sub-optimal by the first patients, mostly the Guinean elite. This led to rehabilitating the Donka national hospital premises and the transfer of COVID-19 patients there. Nevertheless, the capacity of the Donka hospital was rapidly exceeded by mid-April 2020 with many patients waiting in the hospital's corridors to receive their first care ( 15 ). Consequently, four other treatment centers were established in Conakry to decentralize COVID-19 care delivery. For instance, in April 2020, following the notification of a wave of contamination in the prisons of Conakry, a treatment center was created at the Alpha Yaya Diallo military camp to primarily receive prisoners of civil rights. Additionally, intensive care services of the Donka and other private hospitals were requested for managing complicated COVID-19 cases. In late April 2020, decentralization of care at the household level was adopted but only officialised in late May 2020 through the "Stop COVID-19 in 60 Days" initiative. This implied that COVID-19 patients would be treated at home, without the need to be admitted to a treatment center. Unfortunately, this decision turned out to be an opportunity for some actors, especially the elite and senior civil servants, to negotiate their home care with private medical staff, or to resist admission to a treatment center—because of their perceived poor quality. Subsequent deaths of high executive officials—including a minister, a chairman of a public institution and several senior public servants—were reported between May and June 2020 and were seemingly related to their late admission to treatment centers ( 15 ). In response to this, an act was introduced by the MoH on 27 July 2020 to prohibit home confinement. This act highlighted that "any person who tests positive would immediately be transferred, either voluntarily or by force, to a treatment center." The re-adaptation of isolation strategies was not limited only to the capital, but also the countryside. For instance, in the countryside, care for COVID-19 patients was initially provided at the regional level, meaning that patients diagnosed in health districts were admitted to regional treatment centers. As the virus spread extensively from mid-April 2020 onwards, epidemic treatment centers at district levels, previously used for routine health care provision, were seized and rehabilitated for COVID-19 patients' care. Third, during the response to this first wave and beyond, the rise of new international players was noted. For instance, the role The Alliance for International Medical Action (ALIMA) was crucial in managing COVID-19 patients. In addition, several initiatives were undertaken by mining companies from China, Russia and the United Arab Emirates (UAE), based in Guinea under bilateral cooperation, during this first wave and beyond. For example, the contribution of the UAE—through Emirates Global Aluminium's Guinea Alumina Corporation—was crucial in building a treatment center in Conakry. It also appeared that three treatment centers located in mining zones in the countryside (Fria and Kindia) were financially and technically supported by Rusal, the Russian aluminum company producer ( 39 ). Another illustration was the involvement of a Cuban medical team, including 21 epidemiologists, in the COVID-19 response upon the government's request ( 40 ). Table 3 provides a list of actors and their responsibilities, as described in the pandemic response framework. Table 3 Health system actors and their roles in the COVID-19 response. Actors Area of responsibilities Roles in the COVID-19 pandemic response Ministry of Health: ANSS, National Directorate for Major Endemics and Disease Control, National Directorate of Community Health and Traditional Medicine Technical and Financial partners: WHO, CDC, IOM Epidemiological surveillance • Define surveillance guidelines and technical documents • Define strategies for identifying and monitoring cases and managing surveillance data • Update the definition of cases according to the evolution of the pandemic • Update the areas of intervention of the prefectures at risk; follow up on the collection of cases and contacts Ministry of Health: ANSS, central pharmacy of Guinea, material and equipment service, National Directorate of Pharmacy and Medicines, health department of Conakry, Expanded Programme on Immunization, National Blood Transfusion Center Ministry for cooperation Ministry of defense: Military Health Service, Customs directorate general Technical and Financial partners: World Food Program, WHO, CRS, Red Cross, UNICEF Logistics • Define the strategy for the supply • Management of supplies • Coordinate the acquisition of supplies as part of the response against COVID-19 • Carry out an inventory of stock requirements • Define the training programme for logistics personnel Ministry of Health : National Directorate of Laboratories, National Institute of Public Health, National Laboratory of Public Health, National Haemorrhagic Fever Laboratory, CERFIG, UAGCP, CREMS, LABOGUI Ministry of Education and Scientific Research : University of Conakry Technical and Financial partners: WHO, Pasteur Institute Laboratory • Determine the nature of the operating protocols to be used • Validate the laboratories to be integrated into the diagnostic system on the basis of conformity examinations • Determine the way in which sampling is organized • Determine the sample distribution circuit and sites • Organize the validity of the proposed tests • Assess the biosafety and biosecurity conditions of the laboratories • Organizing the management of samples • Organizing the training of laboratory staff at national level Ministry of Health: ANSS, national health promotion service, Ministry of Communication: national public and private media Youth Ministry: CENAFOB Ministry of social welfare Prime minister's cabinet: religious affairs secretariat Technical and Financial partners : UNICEF, USAID, UNFPA Communication and social mobilization • Define communication strategy • Design communication plan with a budget • Organize communication and social mobilization activities • Coordinate interventions in the media and communities • Involve leaders and other personalities in favor of the fight • Mobilize and involve community platforms and provide advice to the strategic committee • Collect and manage rumors Ministry of Health: Nationals hospitals of Donka, Ignace Deen and Sino-Guinean friendship, National Directorate of Health Facilities Ministry of Defense: Civil defense Technical and Financial Partners: MSF, ALIMA, JHPIEGO, OMS, France Expertise, Red Cross, International Federation of the Red Cross, IOM Clinical care • Identify and assess management sites according to the evolution of the epidemic • Produce and update management protocols according to the evolution of knowledge about the disease • Assess therapeutic needs • Define the Infection control and prevention protocol in the context of the response to COVID-19 • List the consumable drugs and equipment that can be used in the response to COVID-19 • Establish guidelines for transporting and burying bodies • Produce the training plan for management personnel in the response to COVID-19 Ministry of Health: ANSS, Head of the cabinet, bureau of strategy and development, financial affairs division, human resources department public procurement Ministry of Budget Ministry of Investment Technical and Financial Partners: World Bank, European Union Finances • Carry out the financial assessment of the needs for the implementation of the national response plan • Prepare the files and requests for the financing of the activities; to draw up the financial table of the response • Draw up the manual of procedure for the management of resources according to the conventions and national regulations • Give the status of the mobilization and execution of the national plan budget Inter-departmental committee including the Prime Minister's cabinet Strategic committee: MoH, ANSS, Technical and Financial Partner (WHO, CDC, IOM, UNICEF, etc.) Scientific committee: 17 members including Academics and Researchers Coordination • Provide a framework for consultation, guidance and decision-making for the response to the pandemic • Coordinate activities related to the COVID-19 response Finally, the emergence of several local initiatives was noted during this phase, including local production of hydroalcoholic solutions by several training and research institutions, local manufacture of spraying devices and their installation at the airport and COVID-19 treatment centers, and manufacture of devices for automatic recording of COVID-19 patients' parameters (temperature, blood pressure, heart rate, etc.). Another initiative, not the least, was the self-declaration of elites suffering from COVID-19 on social media, and their admission to public hospitals. This was of particular importance as it occurred in a period when the elites were criticized for not accepting admission to treatment centers. Furthermore, several sewing workshops engaged in producing traditional masks (made from cloth) and sold them at the government-set price (USD 0.50 per unit). Lastly, therapeutic trials based on phytomedicines were undertaken in COVID-19 treatment centers ( 41 ). Mitigation measures overlooking some local realities On 2 April 2020, Guinea adopted an economic contingency plan to mitigate the potential effects of the pandemic. This plan has three main components: (1) the health component, (2) the social component and (3) the economic component (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The health component was aimed at financing the rollout of the pandemic preparedness and response and strengthening the health system capacities (infrastructure, equipment, etc.). The social component included cash transfers to 1.6 million vulnerable people identified nationwide; the full coverage of water, electricity and transport bills for 3 months; and the provision of prevention equipment to almost 200 000 targeted households (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In addition to this, many companies, public departments, social and political movements initiated the distribution of facemasks and handwashing devices to the population as their contribution to the pandemic control. The economic component comprised measures that support the private sector including a three-month postponement of fiscal taxes payment for companies and an exemption from duties and taxes on health equipment dedicated to the COVID-19 response (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). This policy, estimated at USD 400 million, had to be financed through two mechanisms: the national response funds and the contribution of bi- or multilateral partners (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The national funds were sourced from the withholding taxes on fuel prices, savings from the deferral of external public debt servicing and funds reallocation from the MoH and other ministries (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In April 2020, health system actors noticed a sharp decrease in health care utilization between January and March 2020. This was reportedly linked to the stock out of medical supplies, including protective equipment and fear among health personnel and population to provide or seek care ( 42 ). To mitigate this, a "health system resilience strategy for the continuity of care" was developed in May 2020 ( 42 ). This strategy was layered in five specific objectives including ensuring infection prevention and control among healthcare providers and beneficiaries, availability of health services, and effective communication for patients and their caregivers. The cost of rolling out this strategy was estimated at USD 11 million. In the education sector, online courses on the national public media were organized between April and July 2020, for candidates for national exams. This was initiated to mitigate the effects of the closure of schools during the first wave. Finally, in many public services, telework and the lay-off of contractual workers and trainees were undertaken to limit the virus from spreading. However, several questions were raised on the equity of this teaching method. Many media questioned, for instance, how students from poor families managed to follow online courses. Some other media also questioned how the social component of the economic contingency plan was rolled out in the context of political turmoil. The setting up of a governing framework: The repositioning struggle Centralization of power and decision-making in the response to the first wave was originally considered ( 8 , 9 ). On 27 March 2020, a presidential decree mandated the ANSS as the leading body of the pandemic response. Some national institutions considered it unacceptable to be overhung by the ANSS in their field of expertise ( 9 ). Others, such as teaching hospitals managers, threatened to withdraw from the fight against the pandemic and launched a strike after the ANSS recruitment of 350 medical students and their deployment in the pandemic treatment centers –instead of senior medical doctors and trainees in teaching hospitals ( 32 ). Furthermore, the MoH, the actual affiliated institution of the ANSS, showed its disagreement with this decision, arguing that it was not consulted beforehand. This situation prompted the President to initiate a series of dialogues and consultations with the various protagonists (about 30 people including the Minister of Health and the ANSS General Director) and to make them swear "to serve the Guinean people, and work together to fight the invisible enemy" ( 33 ). In the days that followed, the pandemic response framework was established to involve a large range of national and international actors. This framework also included the scientific committee created by a presidential decree (10 April 2020) and composed of 17 national public health experts and academics with the mandate to monitor, analyse and advise the government on the pandemic response. Care delivery organization for interrupting the transmission chain: Continuous adjustment of policy decisions Strategies for interrupting the chain of community transmission were planned early in the pandemic preparedness and response strategy. However, their implementation experienced a slow pace. As COVID-19 cases increased in April 2020, several measures were taken to readjust the testing, contact tracing and isolation strategies. First, the testing strategy of February 2020 identified three national laboratories in charge of performing RT-PCR antigen tests for COVID-19 suspected cases. Understandably, no rapid diagnostic antigen tests (RDTs) was available at the onset of the pandemic in Guinea, at least until the first donation of 20 000 COVID-19 antigen tests by a humanitarian Foundation on 1 April 2020 ( 34 ). However, as community transmission intensified and as more RDTs became available from mid-April 2020 onwards, several changes in testing strategy were observed including decentralizing sample collection sites in Conakry (the epicenter of the disease with more than 90% of cases in mid-April 2020) and the progressive decentralization and capacity building of other national and regional laboratories (from May 2020 onwards) ( 35 ). Despite this, a slow pace in testing intake was observed, with the delivery of results taking 72 hours or more instead of the 48 hours recommended ( 36 ). As part of the "Stop COVID-19 in 60 Days" initiative (May 2020), mobile sampling units were set up in Conakry, with the involvement of community health workers ( 37 ). This was aimed at alleviating community reluctance and breaking the transmission chain at an accelerated pace ( 37 ). Finally, screening units were set up at bus stations for mandatory testing (free of charge) of all passengers leaving Conakry ( 38 ). A similar scheme was also implemented for foreign travelers in late July 2020, but with a payment of USD 65 per test. Second, the strategy for case isolation originally included two former Ebola treatment centers (one in Conakry and the other in the countryside). These centers were made operational in early February 2020 [Ministère de la Santé: ( 21 )]. However, their combined capacity did not exceed 73 beds. In addition, a video of a press conference published on the ANSS website (March 16, 2020), revealed that the quality of the facilities was considered sub-optimal by the first patients, mostly the Guinean elite. This led to rehabilitating the Donka national hospital premises and the transfer of COVID-19 patients there. Nevertheless, the capacity of the Donka hospital was rapidly exceeded by mid-April 2020 with many patients waiting in the hospital's corridors to receive their first care ( 15 ). Consequently, four other treatment centers were established in Conakry to decentralize COVID-19 care delivery. For instance, in April 2020, following the notification of a wave of contamination in the prisons of Conakry, a treatment center was created at the Alpha Yaya Diallo military camp to primarily receive prisoners of civil rights. Additionally, intensive care services of the Donka and other private hospitals were requested for managing complicated COVID-19 cases. In late April 2020, decentralization of care at the household level was adopted but only officialised in late May 2020 through the "Stop COVID-19 in 60 Days" initiative. This implied that COVID-19 patients would be treated at home, without the need to be admitted to a treatment center. Unfortunately, this decision turned out to be an opportunity for some actors, especially the elite and senior civil servants, to negotiate their home care with private medical staff, or to resist admission to a treatment center—because of their perceived poor quality. Subsequent deaths of high executive officials—including a minister, a chairman of a public institution and several senior public servants—were reported between May and June 2020 and were seemingly related to their late admission to treatment centers ( 15 ). In response to this, an act was introduced by the MoH on 27 July 2020 to prohibit home confinement. This act highlighted that "any person who tests positive would immediately be transferred, either voluntarily or by force, to a treatment center." The re-adaptation of isolation strategies was not limited only to the capital, but also the countryside. For instance, in the countryside, care for COVID-19 patients was initially provided at the regional level, meaning that patients diagnosed in health districts were admitted to regional treatment centers. As the virus spread extensively from mid-April 2020 onwards, epidemic treatment centers at district levels, previously used for routine health care provision, were seized and rehabilitated for COVID-19 patients' care. Third, during the response to this first wave and beyond, the rise of new international players was noted. For instance, the role The Alliance for International Medical Action (ALIMA) was crucial in managing COVID-19 patients. In addition, several initiatives were undertaken by mining companies from China, Russia and the United Arab Emirates (UAE), based in Guinea under bilateral cooperation, during this first wave and beyond. For example, the contribution of the UAE—through Emirates Global Aluminium's Guinea Alumina Corporation—was crucial in building a treatment center in Conakry. It also appeared that three treatment centers located in mining zones in the countryside (Fria and Kindia) were financially and technically supported by Rusal, the Russian aluminum company producer ( 39 ). Another illustration was the involvement of a Cuban medical team, including 21 epidemiologists, in the COVID-19 response upon the government's request ( 40 ). Table 3 provides a list of actors and their responsibilities, as described in the pandemic response framework. Table 3 Health system actors and their roles in the COVID-19 response. Actors Area of responsibilities Roles in the COVID-19 pandemic response Ministry of Health: ANSS, National Directorate for Major Endemics and Disease Control, National Directorate of Community Health and Traditional Medicine Technical and Financial partners: WHO, CDC, IOM Epidemiological surveillance • Define surveillance guidelines and technical documents • Define strategies for identifying and monitoring cases and managing surveillance data • Update the definition of cases according to the evolution of the pandemic • Update the areas of intervention of the prefectures at risk; follow up on the collection of cases and contacts Ministry of Health: ANSS, central pharmacy of Guinea, material and equipment service, National Directorate of Pharmacy and Medicines, health department of Conakry, Expanded Programme on Immunization, National Blood Transfusion Center Ministry for cooperation Ministry of defense: Military Health Service, Customs directorate general Technical and Financial partners: World Food Program, WHO, CRS, Red Cross, UNICEF Logistics • Define the strategy for the supply • Management of supplies • Coordinate the acquisition of supplies as part of the response against COVID-19 • Carry out an inventory of stock requirements • Define the training programme for logistics personnel Ministry of Health : National Directorate of Laboratories, National Institute of Public Health, National Laboratory of Public Health, National Haemorrhagic Fever Laboratory, CERFIG, UAGCP, CREMS, LABOGUI Ministry of Education and Scientific Research : University of Conakry Technical and Financial partners: WHO, Pasteur Institute Laboratory • Determine the nature of the operating protocols to be used • Validate the laboratories to be integrated into the diagnostic system on the basis of conformity examinations • Determine the way in which sampling is organized • Determine the sample distribution circuit and sites • Organize the validity of the proposed tests • Assess the biosafety and biosecurity conditions of the laboratories • Organizing the management of samples • Organizing the training of laboratory staff at national level Ministry of Health: ANSS, national health promotion service, Ministry of Communication: national public and private media Youth Ministry: CENAFOB Ministry of social welfare Prime minister's cabinet: religious affairs secretariat Technical and Financial partners : UNICEF, USAID, UNFPA Communication and social mobilization • Define communication strategy • Design communication plan with a budget • Organize communication and social mobilization activities • Coordinate interventions in the media and communities • Involve leaders and other personalities in favor of the fight • Mobilize and involve community platforms and provide advice to the strategic committee • Collect and manage rumors Ministry of Health: Nationals hospitals of Donka, Ignace Deen and Sino-Guinean friendship, National Directorate of Health Facilities Ministry of Defense: Civil defense Technical and Financial Partners: MSF, ALIMA, JHPIEGO, OMS, France Expertise, Red Cross, International Federation of the Red Cross, IOM Clinical care • Identify and assess management sites according to the evolution of the epidemic • Produce and update management protocols according to the evolution of knowledge about the disease • Assess therapeutic needs • Define the Infection control and prevention protocol in the context of the response to COVID-19 • List the consumable drugs and equipment that can be used in the response to COVID-19 • Establish guidelines for transporting and burying bodies • Produce the training plan for management personnel in the response to COVID-19 Ministry of Health: ANSS, Head of the cabinet, bureau of strategy and development, financial affairs division, human resources department public procurement Ministry of Budget Ministry of Investment Technical and Financial Partners: World Bank, European Union Finances • Carry out the financial assessment of the needs for the implementation of the national response plan • Prepare the files and requests for the financing of the activities; to draw up the financial table of the response • Draw up the manual of procedure for the management of resources according to the conventions and national regulations • Give the status of the mobilization and execution of the national plan budget Inter-departmental committee including the Prime Minister's cabinet Strategic committee: MoH, ANSS, Technical and Financial Partner (WHO, CDC, IOM, UNICEF, etc.) Scientific committee: 17 members including Academics and Researchers Coordination • Provide a framework for consultation, guidance and decision-making for the response to the pandemic • Coordinate activities related to the COVID-19 response Finally, the emergence of several local initiatives was noted during this phase, including local production of hydroalcoholic solutions by several training and research institutions, local manufacture of spraying devices and their installation at the airport and COVID-19 treatment centers, and manufacture of devices for automatic recording of COVID-19 patients' parameters (temperature, blood pressure, heart rate, etc.). Another initiative, not the least, was the self-declaration of elites suffering from COVID-19 on social media, and their admission to public hospitals. This was of particular importance as it occurred in a period when the elites were criticized for not accepting admission to treatment centers. Furthermore, several sewing workshops engaged in producing traditional masks (made from cloth) and sold them at the government-set price (USD 0.50 per unit). Lastly, therapeutic trials based on phytomedicines were undertaken in COVID-19 treatment centers ( 41 ). Mitigation measures overlooking some local realities On 2 April 2020, Guinea adopted an economic contingency plan to mitigate the potential effects of the pandemic. This plan has three main components: (1) the health component, (2) the social component and (3) the economic component (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The health component was aimed at financing the rollout of the pandemic preparedness and response and strengthening the health system capacities (infrastructure, equipment, etc.). The social component included cash transfers to 1.6 million vulnerable people identified nationwide; the full coverage of water, electricity and transport bills for 3 months; and the provision of prevention equipment to almost 200 000 targeted households (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In addition to this, many companies, public departments, social and political movements initiated the distribution of facemasks and handwashing devices to the population as their contribution to the pandemic control. The economic component comprised measures that support the private sector including a three-month postponement of fiscal taxes payment for companies and an exemption from duties and taxes on health equipment dedicated to the COVID-19 response (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). This policy, estimated at USD 400 million, had to be financed through two mechanisms: the national response funds and the contribution of bi- or multilateral partners (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). The national funds were sourced from the withholding taxes on fuel prices, savings from the deferral of external public debt servicing and funds reallocation from the MoH and other ministries (Office of the Prime Minister: Plan of economic response to the health crisis of COVID-19, April 2020). In April 2020, health system actors noticed a sharp decrease in health care utilization between January and March 2020. This was reportedly linked to the stock out of medical supplies, including protective equipment and fear among health personnel and population to provide or seek care ( 42 ). To mitigate this, a "health system resilience strategy for the continuity of care" was developed in May 2020 ( 42 ). This strategy was layered in five specific objectives including ensuring infection prevention and control among healthcare providers and beneficiaries, availability of health services, and effective communication for patients and their caregivers. The cost of rolling out this strategy was estimated at USD 11 million. In the education sector, online courses on the national public media were organized between April and July 2020, for candidates for national exams. This was initiated to mitigate the effects of the closure of schools during the first wave. Finally, in many public services, telework and the lay-off of contractual workers and trainees were undertaken to limit the virus from spreading. However, several questions were raised on the equity of this teaching method. Many media questioned, for instance, how students from poor families managed to follow online courses. Some other media also questioned how the social component of the economic contingency plan was rolled out in the context of political turmoil. Phase 3: Alleviation of restrictive measures: Politics flexibility to social demands, 15 July 2020 to 31 January 2021 On 15 July 2020, President Alpha Conde, in an address to the nation, announced the alleviation of certain restrictions: the gradual reopening of air and land borders in accordance with the reciprocity measures between countries; the curfew was also pushed back from midnight to 4 a.m. (as against 9 p.m. to 5 a.m. initially). On 17 July 2020, a protocol for the reopening of air borders was adopted by the Ministries of Transport and Health. This protocol describes the arrangements in place for health control at air entry points (e.g., certificate of RT-PCR test negativity valid for 5 days) and community-based follow-up of travelers' contacts (e.g., self-quarantine for 14 days) in Guinea. This alleviation of restrictive measures came on the eve of a series of statements by some religious leaders, asking the authorities to facilitate the celebration of "The Tabaski" scheduled for 31 July 2020. The decree of isolating the capital was also repealed to facilitate the movement of people to the countryside. During this period, the country was still confronted with the first wave of the pandemic with increases in confirmed and hospital fatalities cases, at least until September 2020 (June 2020: 1645 cases, 8 deaths; July 2020: 1895 cases, 15 deaths; August 2020: 2009 cases, 13 deaths; September 2020: 1297 cases, 7 deaths). Despite this, the government pursued the adoption of pandemic alleviation measures including gradual schools and public places reopening (October 2020), the launch of the presidential election campaigns (September to October 2020) and the holding of presidential elections (18 October 2020). The ANSS, in response to the intensification of mass movement in the country, adopted "active case finding" as a strategy for interrupting the COVID-19 transmission chain (October 2020). A key element of this strategy consisted of a campaign for systematic screening of 80% of high-risk groups including patients with influenza-like syndromes, febrile patients with a negative RDT for malaria, health workers, security forces and people over 60 years old. In September 2020, a bi-annual immunization plan was developed and adopted in December 2020, with the objective of vaccinating 95% of targeted people (18 years and older) before January 2022. In December 2020, using 60 doses of Sputnik V, Guinea was among the countries that initiated vaccination against COVID-19 early ( 43 ). Despite these, the delivery of additional COVID-19 vaccines was not effective until the country had experienced a new wave of the COVID-19 pandemic in February 2021. Phase 4: The multiple epidemics period, 1 February to 30 June 2021 By February 2021, the country experienced a new wave of the pandemic; 1,006 confirmed cases and 5 hospital deaths in February 2021, 4,198 cases and 30 deaths in March 2021, and 799 cases and 14 deaths in June 2021. This period coincided with the emergence of Ebola (on 12 February 2021), Lassa Fever (on 18 May 2021), and many other recurrent infectious diseases including measles, yellow fever, poliomyelitis and meningitis ( 7 ). The concomitant emergence of these epidemics overstretched the ANSS, the principal national public health institute, and had an impact on COVID-19 surveillance and response ( 18 , 44 ). It was noted also that the surveillance and response to recurrent epidemics (measles, meningitis, poliomyelitis, etc.) were neglected in favor of Ebola ( 7 ). Although the national vaccination plan has been in place since December 2020, it was not until 3 March 2021 that the country received the first bulk quantity of vaccines (~152 000 doses of Sinopharm). Consequently, the mass vaccination campaign against COVID-19 was only launched on 5 March 2021. Thereafter, 20 000 doses of Sputnik (8 March 2021), 69 000 doses of AstraZeneca (29 March 2021), 23 000 doses of Sinovac (19 April 2021), 194 000 doses of AstraZeneca (11 April 2021) and 300 000 doses of Sinovac (18 April 2021) were delivered to ANSS. These vaccines were sourced through three main channels: (1) bilateral cooperation—Sinopharm and Sinovac vaccines were purchased by the government from China while Sputnik vaccines from Russia; (2) regional cooperation-−69 000 doses of the AstraZeneca vaccine were offered by the African Union; and (3) the Covax initiative provided 194 000 doses of the AstraZeneca vaccine ( 7 ). Despite efforts deployed, the vaccination coverage remained low as of 31 May 2021. Indeed, by this date, only 4% of the target population had received their first dose of the vaccine. Additionally, reinforcing the response to Ebola and Lassa fever that involved graduates of the CDC-led Field Epidemiology Training Programme ( 45 , 46 ). Furthermore, some experiences of the COVID-19 response were applied for Ebola control. For example, all national and international workforce deployed in N'zérékoré for the Ebola response were administered an Ebola vaccine. Phase 5: The COVID-19 variants period, 1 July to 31 October 2021 The news of the end of Ebola (19 June 2021) and the second wave of COVID-19 (29 June 2021) lasted only a few weeks since in July 2021 a new wave started. This wave—the deadliest in the country since the emergence of COVID-19—coincided with the notification of Delta, Alpha, Beta, Gamma and Eta COVID-19 variants in the country ( 47 ). This wave bore, respectively, 25 and 59% of COVID-confirmed cases and deaths cases registered in the country from March 2020 to November 2021. The emergence of SARS-CoV-2 variants coincided with low vaccination rate of the population. For example, as of August 2021, only 41% of the 2 900 000 doses of vaccines supplied to Guinea had been used, and full immunization coverage of the population was only 3.56%. This full immunization coverage was five times lower than the 20% projection of the national COVID-19 immunization plan. In addition, strong disparities were noted between the country's health regions in terms of complete vaccination coverage against COVID-19. The complete vaccination coverage of the Conakry health region was 15.29% compared to 1.23% for the Labé, Kindia and Boké regions, and 1% for the Mamou, Kankan, Faranah and N'zérékoré regions ( Figure 3 ). Figure 3 Full vaccination coverage against COVID-19 per administrative regions in the country, August 2021, Guinea. In response to this, in September 2021, the Guinean government adopted the "accelerated COVID-19 vaccination plan" with the aim of "improving equity of access to vaccines and contributing to the reduction of mortality and morbidity related to COVID-19, through the complete vaccination of 20% of the population before the end of December 2021" ( 48 ). This plan was coupled with the adoption of the "vaccination pass" which made movements between cities and access to public services conditional to the presentation of a vaccination card against COVID-19 ( 47 ). This plan was based on 11 acceleration measures, including (1) the use of mobile and semi-mobile strategies (identification of neighborhoods by communes/health districts for sequential immunization), (2) the involvement of certain public and private health centers in immunization and (3) the involvement of civil society associations and corporations in immunization ( 48 ). In addition, through this strategy, the supply of COVID-19 vaccines increased from 2.9 to 5.7 million doses between August and November 2021. Also, during the same period, full immunization coverage of the general population increased from 3.56 to 6.95%. However, the objectives of complete vaccination of 20% of the general population by the end of December were far from being achieved. Similarly, disparities persisted in terms of access to COVID-19 vaccines—on 22 November 2021, full vaccination coverage in Conakry was 25.34%, compared with 3.47% for the Labé, Kindia and Boké regions, and 2.64% for the Mamou, Kankan, Faranah and N'zérékoré regions. COVID-19-related activities, including vaccination, experienced a slower pace after the military coup of September 2021 as a result of, among others: the freezing of ANSS bank accounts; the reshuffling of management positions at the ANSS and the MoH; the abolition of the vaccine pass and the lifting of the curfew, and the state of health emergency ( 49 ). All these raised questioned on how health system actors will reorganize to pursue the COVID-19 control in Guinea. Discussion This study analyses Guinea's preparedness and response to the COVID-19 pandemic during the first 22 months. This study provides five key learning points for future epidemics and pandemic preparedness and response in Guinea, and beyond. These include, in the pre-epidemic period, the necessity of setting up an epidemic governance framework that is articulated with the country's health system and epidemiological contexts; and the importance of mobilizing emergency funds to support a rapid health system response whenever epidemics or pandemics hit. This study also teaches that each epidemic is unique, and previous exposure to similar ones does not necessarily guarantee population and health system resilience. Moreover, the study analyses that epidemics generate social distress because of the restrictive measures they require for their control, but the excessive securitization of epidemics response is counterproductive. Finally, from a political point of view, decision-making for epidemic control is not always disinterested; it is sometimes rooted in political computations, and health system actors should learn to cope with it while at the same time, safeguard trusted and efficient health system responses. First, the analysis of pandemic preparedness reveals an early response, almost 47 days before the virus was introduced into Guinea. This early response was coordinated by the ANSS, the former Ebola coordination body, which has strengthened its leadership in the response to recurrent epidemics including anthrax, measles and yellow fever from 2017 onwards. This recent role of ANSS in the response to epidemics in Guinea had implicitly played its value, particularly in obtaining support from several actors, including decision-makers and development partners, for the coordination of preparedness and response to COVID-19. Despite this, however, we noted tensions between several health system actors when it came to formally assigning, through a presidential decree, the coordination of the response to the ANSS. These tensions were likely entertained by the absence, before COVID-19, of a formal framework for epidemic governance, with a clear distinction between the roles and responsibilities of the actors. The difficulties in coordinating the COVID-19 response mechanisms at the beginning of the pandemic might explain the increase in incidence cases of disease during this period in Guinea, compared with west African countries like Liberia, Sierra Leone, Mali, and Burkina Faso ( 8 , 12 ). This is why we argue that national stakeholders should formalize epidemic governance framework, in pre-epidemic period, in order to ensure effective coordination of epidemic or pandemic response. Such a governance framework should take into account local specificities. In Guinea, for example, two options can be envisaged with their pros and cons. A first option would be to establish the ANSS as the leading organization for the coordination and implementation of epidemics and pandemics preparedness and response interventions. Such a measure, because of its centralized nature, would facilitate decision making and operationalisation of health interventions in the face of epidemics or pandemics. However, as noted during this study, the concurrent emergence of epidemics would substantially reduce the response capacity of such an arrangement ( 7 , 44 ). Also, such a vertical response mechanism would raise questions about the sustainable strengthening of the health system, including routine immunization services and interventions since the latter has not yet been integrated with COVID-19 immunization in Guinea. Also, in a context of political instability—with its implication for change in leadership positions in national institutions—the risk that institutional memory for epidemics preparedness and response is altered could be high. Specially, because with the recurrence of epidemics and pandemics in Africa, the mobility of such experienced health workers in managing epidemics could also be high. A second option in Guinea would be to decentralize epidemics and pandemics response functions. For example, in the pre-epidemic period, the DNELM could be responsible for routine surveillance of diseases with epidemic potential. In the event of an epidemic or pandemic, this surveillance role could be devolved to the ANSS while the INSP would be responsible for their diagnosis. The EPI/DNELM would be responsible for the vaccination pillar in the response to epidemics. Such a mechanism therefore implies setting up the ANSS as an "agence de veille sanitaire" in the pre-epidemic period. However, such a scheme would give the advantage to several national institutions to test their capacities during epidemics and draw lessons for their continuous strengthening. Second, this study highlights the unpredictability of epidemics including their occurrence in a difficult socio-political and economical context. In Guinea, the pandemic occurred in Guinea in a period when donor countries were confronted with an acute phase of the disease. Importantly, the COVID-19 pandemic also occurred in the middle of an electoral process for the renewal of legislative and presidential institutions. In this context, despite the anticipation of the response by health system actors, the implementation of the pandemic preparedness and response interventions was difficult, due to among others the low mobilization of financial resources to support the COVID-19 response strategies. This finding raises the question on the need of putting in place sustainable funding mechanisms for improving national health system readiness toward epidemic-prone diseases. One way of doing this may be creating, in pre-epidemic situations, an emergency fund that can be used by health system actors as soon as an epidemic-prone disease occurs on the national territory. Such funding mechanisms could avoid unnecessary delays in responding to emergencies. Third, this study also analyses that each epidemic is unique, and previous exposure to similar ones does not necessarily guarantee population and health system resilience. In 2014, Guinea was the epicenter of the largest and deadliest Ebola outbreak in human history. This outbreak triggered several health system reforms to improve the national health system resilience. These efforts yielded positive results in the rapid response to the recurrent epidemics in Guinea, including the 2016 Ebola outbreak ( 7 ). However, our findings showed that the country was not well-prepared to deal with the COVID-19 pandemic. First, Ebola treatment centers were proven inappropriate to manage COVID-19 cases due to lack of adequate equipment—for example, lack of oxygen tanks—for COVID-19 treatment. Second, the testing capacity of national reference laboratories were very limited, given the high infectivity of COVID-19 compared to Ebola. Similarly, the bed capacities of Ebola treatment centers were very limited during the early phases the COVID-19 pandemic. Third, the health services utilization was compromised during the first wave of the pandemic due to a lack of personal protective equipment for health workers, and people's fear of visiting health facilities. All of these point to the fact that the Ebola experience did not guarantee effective preparedness and response to the COVID-19 pandemic. Delamou et al. in their rapid assessment of health system preparedness and response to the COVID-19 pandemic undertaken in 13 health districts in Guinea, reported that most of healthcare facilities visited did not have intensive care units (95%), oxygenators (98%), and respirators (94.4%) ( 50 ). Fourth, the finding shows that epidemics generate social distress because of the restrictive measures they require for their control, but their excessive securitization is counterproductive. The securization of the COVID-19 pandemic response has been documented to negatively impact populations health with implication on injuries and deaths as results of policy interventions, but also an increase in their resistance to government response measures ( 7 ). Fifth, the analysis of the COVID-19 pandemic response process in Guinea reveals that decision-making was sometimes underpinned by political considerations. At times, COVID-19 was used by politicians as a springboard to push through decisions that were favorable to them. For example, bans on mass gatherings or movements on the eve and aftermath of elections were surely taking as a dissuading measure for political protests. Such situations have also been reported in other countries such as France ( 51 ). However, health system actors should learn to cope with it while at the same time, safeguard trusted and efficient health system responses. In Guinea, health system actors adapted to this situation by emphasizing the importance of facemask wearing and compliance with barrier measures during political campaigns, appealing to political leaders for their responsibility to enforce health guidelines related to COVID-19, but also by undertaking active case finding and systemic testing among high-risk groups following periods of intensified mass movements. Furthermore, referring to Kruk definition of resilient health system, this study analyzed that the Guinean health system resilience capacities have been improved ( 52 ). First, the Guinean health system actors adopted an earlier response, about 47 days, in face of the SARS-CoV-2 introduction on the national territory, notably through the planning of epidemiological surveillance, and case management. This result could be explained by the experience heath system actors have acquired in managing the 2014–16 Ebola epidemic, the gains of the post-Ebola health system reforms that led to the establishment of a central public health emergency operation department networked with 38 health districts and the training of biosecurity workforce ( 45 ). Authors have reported that exposure to major epidemics in the past is associated with earlier responses, especially in implementing epidemiological surveillance strategies ( 53 ). In Liberia and Sierra Leone, the two other West-African countries severely affected by Ebola in 2014–16, rapid readiness of health systems have also been reported ( 54 , 55 ). Second, the declaration of the COVID-19 pandemic triggered the undertaking and implementation of several innovative approaches in Guinea. These initiatives include a phytomedicine therapeutic trial; manufacturing of hydroalcoholic solutions and traditional facemasks; and COVID-19-patient monitoring and spraying machines. These initiatives have certainly helped the country to overcome the global challenges of health resources (e.g., masks, hydroalcoholic solutions), especially observed in developing countries, for the control of the acute phase of the pandemic ( 56 , 57 ). Some authors define attitudes that help societies to resist shocks as an indication of improved collective resilience ( 58 , 59 ). Third, the COVID-19 pandemic response in Guinea acknowledge the rise of new players in the health system, compared with Ebola in 2014/2016. Despite the positives discussed, some challenges have been noted in the COVID-19 pandemic response processes in Guinea. This include the indirect effects of the COVID-19 pandemic control measures on populations socio-economic and psychological wellbeing. For instance, increases in transportation fare and loss of income for many private institutions in Guinea such as restaurants, bars and hotels were noted as results of, unnecessarily prolonged restrictive measures. Authors in South Africa have reported that the prolonged effects of restrictive measures such as lockdowns have resulted in acute panic, depression, anxiety and social unrest ( 60 ). The present review has some limitations. First, the methodological design of this study—scoping review—did not allow any quality (or risk of bias) assessment of included studies. This could limit the internal validity of the findings reported in the present study. Second, this review covers the period between January 2020 and November 2021. Given the highly dynamic and changing prospects of the COVID-19 pandemic, and the high political instability in Guinea, conclusion of the study may not be applicable to next phases of COVID-19 response. However, this study utilizes diverse data sources to allow a better understanding of the responses from the government and other civil actors in the country, during the 22 first months of the pandemic. It therefore provides insights on challenges health system actors have been confronted with, and proposed solutions for an improved preparedness and response strategies of future epidemics and pandemics in Guinea. Conclusion This scoping review aimed to assess the response triggered in Guinea between January 2020 and November 2021. It provides several learning points for future epidemics and pandemics preparedness and response in Guinea, and beyond. These learning points include the necessity of setting up, in the pre-epidemic period, an epidemic governance framework that is articulated with the country's health system and epidemiological contexts; the importance of mobilizing, during pre-epidemic period, emergency funds for a rapid health system response whenever epidemics hit; each epidemic is a new experience as previous exposure to similar ones does not necessarily guarantee population and health system resilience; epidemics generate social distress because of the restrictive measures they require for their control, but their excessive securitization is counterproductive; and that decision-making for epidemic control is not always disinterested, from a political perspective, and that health system actors should learn to cope with it while safeguarding trusted and efficient health system responses. Health system actors anticipated the response to the COVID-19 pandemic and (re-) adapted response strategies as the pandemic evolved. There is a need to work toward rethinking the epidemics governance framework and funding mechanisms in Guinea for improved health system efficacy toward their management. Author contributions The study protocol was developed by DK, FK, and AD and reviewed by WV and IA. DK and FK did the data analysis, performed the literature search, screening, and selection of articles. The first draft of the manuscript was written by DK. The manuscript was critically reviewed by WV, J-PD, and IA. All authors were involved with interpretation, read, and agreed to the final version of this manuscript. Funding This study was funded by the International Development Research Centre (IDRC), Canada through the Grant number (CATALYSE PROJECT 109479-001). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8336673/

Bacteria-Inspired Nanomedicine

The natural world has provided a host of materials and inspiration for the field of nanomedicine. By taking design cues from naturally occurring systems, the nanoengineering of advanced biomimetic platforms has significantly accelerated over the past decade. In particular, the biomimicry of bacteria, with their motility, taxis, immunomodulation, and overall dynamic host interactions, has elicited substantial interest and opened up exciting avenues of research. More recently, advancements in genetic engineering have given way to more complex and elegant systems with tunable control characteristics. Furthermore, bacterial derivatives such as membrane ghosts, extracellular vesicles, spores, and toxins have proven advantageous for use in nanotherapeutic applications, as they preserve many of the features from the original bacteria while also offering distinct advantages. Overall, bacteria-inspired nanomedicines can be employed in a range of therapeutic settings, from payload delivery to immunotherapy, and have proven successful in combatting both cancer and infectious disease.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6290263/

Analysis of Whole-Genome Sequences for the Prediction of Penicillin Resistance and β-Lactamase Activity in Bacillus anthracis

Determination of antimicrobial susceptibility of B. anthracis is essential for the appropriate distribution of antimicrobial agents for postexposure prophylaxis (PEP) and treatment of anthrax. Analysis of WGS data allows for the rapid detection of mutations in antimicrobial resistance (AMR) genes in an isolate, but the presence of a mutation in an AMR gene does not always accurately predict resistance. As mutations in the anti-sigma factor RsiP have been previously associated with high-level penicillin resistance in a limited number of strains, we investigated WGS assemblies from 374 strains to determine the frequency of mutations and performed functional antimicrobial susceptibility testing. Of the five strains that contained mutations in rsiP , only four were PEN-R by functional antimicrobial susceptibility testing. We conclude that while sequence analysis of this region is useful for AMR prediction in B. anthracis , genetic analysis should not be used exclusively and phenotypic susceptibility testing remains essential. INTRODUCTION Ciprofloxacin, doxycycline, and β-lactam antibiotics (including ampicillin, penicillin G, and penicillin VK) are recommended by the U.S. Centers for Disease Control and Prevention for postexposure prophylaxis (PEP) of inhalation anthrax in adults following exposure to Bacillus anthracis ( 1 , 2 ). In developing countries where anthrax is endemic, penicillin is considered a drug of choice for treatment because it is effective, widely available, and low in cost ( 3 ). Penicillin susceptibility is a B. anthracis characteristic that is commonly used to differentiate it from Bacillus cereus and Bacillus thuringiensis , which display inducible β-lactam resistance ( 4 ). While most B. anthracis strains are susceptible to penicillin, surveys of clinical and environmental isolates indicate that penicillin resistance occurs in 2% to 16% of strains ( 5 ). Penicillin treatment failures have been reported for anthrax ( 6 , 7 ), and use of this antibiotic for PEP in experimental animals had variable results ( 8 , 9 ). Therefore, antimicrobial susceptibility testing (AST) is recommended prior to treatment with penicillin ( 1 ). All B. anthracis strains analyzed to date have two chromosomal β-lactamase genes: bla1 (which encodes a penicillinase) and bla2 (which encodes a cephalosporinase) ( 5 ). When transferred to other organisms, such as Escherichia coli , both genes are complete and functional, but in B. anthracis , bla1 and bla2 are poorly transcribed and gene expression is not sufficient to confer resistance to β-lactam antibiotics ( 5 , 10 ). Furthermore, induction of β-lactamase activity or penicillin resistance was not observed following growth in sublethal levels of β-lactam antibiotics ( 4 , 10 ). β-Lactamase expression and penicillin resistance in B. anthracis were characterized in studies of penicillin-resistant (PEN-R) strain 32 ( 4 , 5 , 10 , 11 ), which was originally isolated in 1974 from a fatal anthrax case in Northampton, England ( 12 , 13 ). While the β-lactamase genes of a typical penicillin-susceptible (PEN-S) B. anthracis strain are transcriptionally silent, bla1 and bla2 are expressed constitutively in strain 32, and bla1 was identified as the major contributor to PEN resistance ( 10 ). Another naturally occurring, PEN-R B. anthracis isolate, strain SK57, has been described previously ( 14 ) and was isolated in England in November 1975; however, few details about the strain's source are available. DNA sequence analysis and details of the β-lactamase expression of SK57 have also not been published previously. Whole-genome sequencing (WGS) data for strains 32 and SK57 are publicly available under GenBank accession numbers QPKO00000000 and QPKQ00000000 , respectively ( 15 ). In B. anthracis and other B. cereus group species, an extracytoplasmic function (ECF) sigma factor, SigP, and its cognate anti-sigma factor, RsiP, regulate bla1 and bla2 transcription ( 4 ). ECF sigma factors represent a diverse subfamily of alternative sigma factors that typically activate gene expression in response to extracellular signals, including agents that threaten cell envelope integrity ( 4 , 16 , 17 ). The mechanism of signal perception and the basis for anti-sigma factor inactivation are not well understood for the majority of ECF sigma factor/anti-sigma factor pairs ( 16 ), including SigP and RsiP. Although associated with inducible β-lactam resistance in B. cereus and B. thuringiensis , the presence of SigP and RsiP is not sufficient for β-lactamase gene expression in B. anthracis ( 4 ). In PEN-R strain 32, a nucleotide deletion that results in a frameshift mutation and an amino-terminally truncated RsiP was described as the basis of high-level PEN resistance ( 4 ). B. anthracis strain 32 also contains a single nucleotide polymorphism (SNP) in sigP (an A-to-G transition at position 183) that results in a single amino acid difference (an aspartic acid in the PEN-S reference strain and a glycine in strain 32 at position 24). This mutation occurs within a conserved sigma factor domain that is important for interactions with both the RNA polymerase and the −10 promoter element and is predicted to affect protein activity ( 4 ). Mutations in rsiP and sigP in several PEN-R strains isolated from cattle following a 2011 anthrax outbreak were also previously described ( 18 ). Transcriptome analysis of one of the PEN-R isolates from that study revealed that the frameshift mutation in rsiP led to upregulation of five genes, rsiP , sigP , bla1 , bla2 , and a predicted penicillin-binding protein (PBP) transpeptidase gene that is located immediately upstream of bla1 ( 18 ). Detection of phenotypic penicillin resistance by AST is essential for the distribution of appropriate antimicrobial agents for PEP and treatment during a public health emergency involving anthrax. Conventional broth microdilution (BMD) is considered the gold standard laboratory method for AST and requires a 16- to 20-h incubation period for B. anthracis based on Clinical and Laboratory Standards Institute (CLSI) guidelines ( 19 ). Several functional phenotypic methods have been developed to reduce the time required for antimicrobial susceptibility profiling of B. anthracis , including real-time PCR to detect growth in the presence of antimicrobial agents ( 20 ), bioluminescent reporter phage analysis ( 21 ), laser light scattering technology ( 22 ), and optical screening ( 23 ). The analysis of WGS data from a suspect isolate can complement these phenotypic AST methods, as they can detect the introduction of mutations, genes, and/or plasmids associated with antimicrobial resistance and provide details on the mechanisms of resistance that are critical for an effective public health response. For penicillin resistance in B. anthracis , the presence of β-lactamase genes cannot predict whether an isolate would be penicillin resistant; however, the analysis of genes that regulate β-lactamase expression may serve as a more accurate predictor. Here, we analyze WGS data for genetic markers that predict penicillin resistance in B. anthracis . The chromosomal regions containing sigP , rsiP , and bla1 ( sigP - bla1 region) were compared in a collection of B. anthracis strains at CDC in order to (i) determine the frequency of mutations (SNPs, insertions, or deletions) in this region; (ii) identify mutations associated with penicillin resistance; and (iii) evaluate the usefulness of WGS for predicting penicillin resistance. Here, we show that the coding regions of sigP , rsiP , and bla1 have low sequence variability among B. anthracis strains, with only 3.5% (13/374) of strains containing mutations compared to the Ames Ancestor reference genome. When a mutation was identified in the anti-sigma factor rsiP , penicillin resistance was detected in only four of the five strains. Therefore, analysis of the sigP-bla1 region of B. anthracis has shown that it is a useful locus to analyze for prediction of penicillin resistance in B. anthracis . However, to accurately assess β-lactam resistance in B. anthracis , a conventional method such as BMD remains essential. RESULTS BLAST analysis of the sigP-bla1 region in B. anthracis strains. The Ames Ancestor reference sequence contains five predicted open reading frames (ORF) in the sigP-bla1 region ( Fig. 1 ). Together with sigP (ORF 2502), rsiP (ORF 2503), and bla1 (ORF 2507), there are two additional ORFs (2504 and 2506) predicted to encode PBPs within the approximately 5-kb region between rsiP and bla1. A local BLAST search was performed by querying the sigP-bla1 region, 6,892 bp ( Fig. 1 ), from the PEN-S Ames Ancestor reference against genomes from 374 B. anthracis strains in the CDC WGS database. Forty-two additional B. anthracis whole-genome sequences, for which there are no BMD data available, were in GenBank at the time of analysis and included in this screen (total, 416 strains). Over half of the strains, 235/416 (56%), were identical to the Ames Ancestor reference strain, and 185/416 (44%) of the strains had at least one mutation in the sigP-bla1 region. These mutations were distributed among 42 positions. The majority of mutations were synonymous substitutions and/or shared across multiple strains (see Fig. S1 in the supplemental material). bla2 is located ∼900 kb away from bla1 on the chromosome and was not included in the screen. The sequence of bla2 and its promoter region were analyzed in all 13 strains from the CDC collection (see below) that contained a sigP , rsiP , or bla1 mutation and were identical to those of the Ames Ancestor reference. FIG 1 Diagram of the B. anthracis sigP-bla1 region. The numbers refer to ORFs in the B. anthracis Ames Ancestor reference sequence. The asterisk (*) indicates that pbp2 contains a frameshift that results in two predicted ORFs; the first contains 124 amino acids of the predicted PBP2, and the second contains the remaining 586 amino acids of the predicted PBP2. (Adapted from reference 4 with permission). 10.1128/mSystems.00154-18.1 FIG S1 B. anthracis strain BLAST Analysis Heat map. A total of 185/416 (44%) strains had at least one mutation. There were 42 mutation positions across the 6,892-bp region, and the majority of mutations were synonymous substitutions and/or were common across multiple strains; e.g., the C-to-T transition at position 2241 (C > T, 2241) and the C-to-A transversion at position 2727 (C > A, 2727) are synonymous substitutions, and Ins 1568 and Ins 3972 are insertions. Download FIG S1, TIF file, 2.7 MB . This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. Sequence analysis and AST of strains with sigP , rsiP , and bla1 coding region mutations. No mutations were identified in the promoter or coding regions of sigP , rsiP , or bla1 in the 42 B. anthracis genome sequences in GenBank. Analysis of the 374 sequenced B. anthracis strains from the CDC collection identified 13 (3.5%) strains with mutations in the sigP , rsiP , or bla1 coding region ( Table 1 ). As previously described by Ross et al. in 2009 (4), the promoter regions are highly conserved. These regions were also analyzed for the 13 strains. TABLE 1 B. anthracis strains identified from the WGS screen as containing mutations within sigP , rsiP , and bla1 a Strain Gene nt mutation position(s) Confirmed by Sanger sequencing Mutation type Predicted effect PEN MIC (µg/ml) Mutation previously described (reference[s]) 2002013094 sigP 119 Yes SNP; transition (G) ▶ (D) ≤0.015 No 2000031021 sigP 119 Yes SNP; transition (G) ▶ (D) 0.3 No 2000031052 sigP 119, 395 Yes SNP; transitions (G) ▶ (D), (L) ▶ (S) 0.3 No 2002734089 bla1 93 Yes SNP; transition Silent mutation (G) ▶ (G) 0.06 No 2000031048 bla1 −64†Yes SNP; transversion in −35 binding element 5′-ATGGAA C AAA-3′ ▶ 5′-ATGGAA A AAA-3′ 0.3 No 2002013017 rsiP 60 No SNP; transversion (H) ▶ (Q) ≤0.015 No 2002013007 rsiP 827 No Insertion of AAAAAG Deletion of (K) + (R) 0.03 No 2002013011 rsiP 505 Yes SNP; transition (Y) ▶ (H) 0.03 No 2000031038 rsiP 39 Yes SNP; deletion FS; truncates RsiP to 30 aa >512 No 2000032823 rsiP 471 Yes SNP; insertion FS; truncates RsiP to 163 aa 256 Yes ( 18 ) 2002734065 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 ) 2002734039 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 ) 2002013027 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa 0.06 Yes ( 4 ) Pen-R strains not included in screen* 32 sigP 183 Yes SNP; transition (D) ▶ (G) >512 Yes ( 4 , 15 ) rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes UT308 sigP 183 Yes SNP; transition (D) ▶ (G) >512 Yes ( 4 , 15 ) rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 , 15 ) SK57 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes a Nucleotide (nt) mutation positions representing the position within the corresponding gene in the Ames Ancestor reference sequence (NC_007530.2). The last column indicates whether the corresponding mutation has been previously described in the literature. Amino acid abbreviations are as follows: (G), glycine; (D), aspartic acid; (L), leucine; (S), serine, (H), histidine; (Q), glutamine; (Y), tyrosine. †, nt position upstream of the ATG start codon for bla1 . *, penicillin-resistant strains not included in this screen but included in this table for reference purposes. FS, frameshift. sigP . Three of the 13 strains, 2002013094 ( B. anthracis 3094 [Ba3094]), 2000031021 (Ba1021), and 2000031052 (Ba1052), contained mutations in sigP , and all mutations were confirmed by Sanger sequencing. Ba3094, Ba1021, and Ba1052 were shown to belong to clade C by multilocus variable-number tandem-repeat 8 (MLVA-8) genotyping ( 24 ) and to share the same transition mutation at nucleotide position 119. Strain Ba1052 contained an additional transition mutation in sigP at position 395. None of the strains contained the sigP SNP previously described in PEN-R strain 32 that resulted in a single amino acid difference predicted to decrease SigP activity ( 4 ). All study strains with sigP mutations were PEN-S by conventional BMD AST ( Table 1 ). bla1 . Two of 13 strains contained mutations in the bla1 coding region, or in the bla1 promoter region, and all of the mutations were confirmed by Sanger sequencing. Strain 2002734089 (Ba4089) contained a transition mutation and belonged to clade A by MLVA-8 genotyping, and strain 2000031048 (Ba1048) contained a transversion in the −35 promoter element and belonged to clade B. All strains with mutations in the bla1 coding region or promoter region were PEN-S by conventional BMD AST. rsiP . Eight of the 13 strains contained mutations in the anti-sigma factor gene, rsiP ( Table 1 ). Mutations were identified in each WGS assembly for strains 2002013017 (Ba3017) and 2002013007 (Ba3007), but Sanger sequencing did not confirm these changes. Analysis of the WGS data revealed errors in the WGS assemblies, and, as a result, these strains were removed from further analysis. The remaining six strains with Sanger-confirmed rsiP mutations belonged to clade A (MLVA-8). Strain 2002013011 (Ba3011) contained a novel SNP that is predicted to lead to an amino acid substitution in RsiP at position 170 of 275 and was PEN-S by AST. Strain 2000031038 (Ba1038) contained a novel frameshift mutation predicted to result in an amino-terminally truncated RsiP protein (30 amino acids [aa]). Strain Ba1038 was isolated from an environmental surface sample in 1957; however, the source of the sample is unknown. Strain 2000032823 (Ba2823) contained a frameshift mutation predicted to result in an amino-terminally truncated RsiP protein (163 aa). This mutation was also previously described in PEN-R strains isolated in 2011 from cattle ( 18 ). Strain Ba2823 was isolated prior to 2011, but additional details about the source of the strain are unknown ( 25 ). Both Ba1038 and Ba2823 were PEN-R by conventional BMD AST. Three strains, 2002734065 (Ba4065 or SK57A), 2002734039 (Ba4039 or SK57C), and 2002013027 (Ba3027), contained the same frameshift predicted to result in a 12-amino-acid, amino-terminally truncated RsiP previously described in PEN-R strain 32 (4) and in PEN-R strains isolated from cattle ( 18 ). Strains Ba4065 (SK57A) and Ba4039 (SK57C) were collected in England in November 1975. Both are likely related to previously described PEN-R strain SK57 ( 14 ). Despite possessing the same rsiP mutation as PEN-R strains SK57A, SK57C, and strain 32, conventional AST confirmed that Ba3027 is PEN-S (MIC = 0.06 µg/ml). ORF 2506 and ORF 2504. In addition to bla1 and bla2 , SigP-induced transcription of ORF 2506 in a PEN-R B. anthracis strain was described previously ( 18 ). BLASTX analysis revealed that ORF 2506 shared similarity with B. cereus group PBP2 (99% coverage, 100% identity; accession number WP_000662966.1) and contained the conserved protein domain family FtsI, which is associated with cell cycle control, cell division, chromosome partitioning, and cell wall/membrane/envelope biogenesis. The majority (∼93.75%) of strains in this study, including all PEN-R strains, contained a frameshift in a homopolymer region of this gene that is also found in the Ames Ancestor reference strain (NCBI accession number AE017334 ) and that resulted in 2 predicted ORFs. The first ORF consisted of the N-terminal 124 amino acids of the predicted PBP2 and contained a predicted PBP dimerization domain. The second ORF contained the last 586 amino acids of the predicted PBP2, which includes the transpeptidase domain. Only 6.25% (26/416) of the strains did not contain this frameshift and had a single ORF predicted to represent the full-length PBP2 (710 aa). BLASTX analysis revealed that ORF 2504 also shared similarity with a B. cereus group PBP2 (99% coverage, 100% identity; accession number WP_000903320.1). However, ORF 2504 is translated in the opposite direction from the other ORFs in the sigP-bla1 region and is not regulated by SigP ( 18 ). Sequencing of PEN-R strains and phylogeny based on whole-genome SNP calling. The de novo assemblies for all study strains were each ≥99.8% identical to the Ames Ancestor reference genome (data not shown). A phylogeny based on whole-genome SNP calling was created for all strains with mutations in the coding regions of sigP , rsiP , and/or bla1 ( Fig. 2 ). Strains 32, UT308, and SK57 ( 15 ) contained mutations in sigP and/or rsiP ( Table 1 ) and were included in the phylogenetic analysis as reference strains. All strains with a mutation in rsiP belonged to clade A. Two of the strains identified in this screen with rsiP mutations (Ba4065 and Ba4039) fell into the same clade as the three PEN-R reference strains. In comparison to SK57, Ba4065 (SK57A), Ba4039 (SK57C), and strain 32 and its derivative UT308 have an additional mutation in sigP that is predicted to affect protein activity ( 4 ). PEN-S Ba3027 harbors the same rsiP mutation as strains 32, SK57, Ba4065 (SK57A), and Ba4039 (SK57C) but is not in the same branch as the PEN-R strains. Two other PEN-R isolates with rsiP mutations, Ba2823 and Ba1038, were in different branches. The two strains with mutations in bla1 were members of clade A, and the three strains containing sigP mutations fell within clade C. FIG 2 Neighbor-joining tree of B. anthracis strains containing mutations within sigP , rsiP , and bla1 identified from the WGS screen. Strains from the three major B. anthracis clades (A, B, and C) were identified in the WGS screen (color coded) ( 24 ). PEN-R strains are colored in red. Asterisks (*) indicate strains containing an rsiP mutation. Control strains SK57, UT308, 2000031103 (strain 32), and Sterne, as well as the Ames Ancestor reference strain, were included for comparison. Characterization of Ba3027: Growth, colony morphology, β-lactamase activity, and bla gene expression. Strain Ba3027 exhibited a slower growth rate in broth culture than strain SK57, which contains the same rsiP mutation, and 2007740878 (Ba0878), a prototypical PEN-S strain without the rsiP mutation ( Fig. 3A ). After 24 h of culture on agar, single isolated colonies of strain Ba0878 were nearly half the size of strain SK57 colonies, with average colony diameters of 2.82 mm and 5.40 mm, respectively ( Fig. 3B ). Strain Ba3027 formed smaller colonies (average colony diameter, 0.85 mm) than Ba0878 and SK57. The β-lactamase activity of culture supernatants was measured in a quantitative nitrocefin assay to evaluate extracellular β-lactamase production in Ba3027. Similarly to PEN-S strains Ba0878 and Sterne, β-lactamase activity was not detected in the culture supernatant from Ba3027 ( Fig. 4A ). A statistically significant difference in activity levels ( P = 0.001 to 0.01) was observed for these three strains compared to β-lactamase-producing strain UT308. SK57 had significantly higher β-lactamase activity (223 ± 4.35 mU/ml) than UT308 (61 ± 3.15 mU/ml) ( P  T, 2241) and the C-to-A transversion at position 2727 (C > A, 2727) are synonymous substitutions, and Ins 1568 and Ins 3972 are insertions. Download FIG S1, TIF file, 2.7 MB . This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. Sequence analysis and AST of strains with sigP , rsiP , and bla1 coding region mutations. No mutations were identified in the promoter or coding regions of sigP , rsiP , or bla1 in the 42 B. anthracis genome sequences in GenBank. Analysis of the 374 sequenced B. anthracis strains from the CDC collection identified 13 (3.5%) strains with mutations in the sigP , rsiP , or bla1 coding region ( Table 1 ). As previously described by Ross et al. in 2009 (4), the promoter regions are highly conserved. These regions were also analyzed for the 13 strains. TABLE 1 B. anthracis strains identified from the WGS screen as containing mutations within sigP , rsiP , and bla1 a Strain Gene nt mutation position(s) Confirmed by Sanger sequencing Mutation type Predicted effect PEN MIC (µg/ml) Mutation previously described (reference[s]) 2002013094 sigP 119 Yes SNP; transition (G) ▶ (D) ≤0.015 No 2000031021 sigP 119 Yes SNP; transition (G) ▶ (D) 0.3 No 2000031052 sigP 119, 395 Yes SNP; transitions (G) ▶ (D), (L) ▶ (S) 0.3 No 2002734089 bla1 93 Yes SNP; transition Silent mutation (G) ▶ (G) 0.06 No 2000031048 bla1 −64†Yes SNP; transversion in −35 binding element 5′-ATGGAA C AAA-3′ ▶ 5′-ATGGAA A AAA-3′ 0.3 No 2002013017 rsiP 60 No SNP; transversion (H) ▶ (Q) ≤0.015 No 2002013007 rsiP 827 No Insertion of AAAAAG Deletion of (K) + (R) 0.03 No 2002013011 rsiP 505 Yes SNP; transition (Y) ▶ (H) 0.03 No 2000031038 rsiP 39 Yes SNP; deletion FS; truncates RsiP to 30 aa >512 No 2000032823 rsiP 471 Yes SNP; insertion FS; truncates RsiP to 163 aa 256 Yes ( 18 ) 2002734065 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 ) 2002734039 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 ) 2002013027 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa 0.06 Yes ( 4 ) Pen-R strains not included in screen* 32 sigP 183 Yes SNP; transition (D) ▶ (G) >512 Yes ( 4 , 15 ) rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes UT308 sigP 183 Yes SNP; transition (D) ▶ (G) >512 Yes ( 4 , 15 ) rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes ( 4 , 15 ) SK57 rsiP 10 Yes SNP; deletion FS; truncates RsiP to 12 aa >512 Yes a Nucleotide (nt) mutation positions representing the position within the corresponding gene in the Ames Ancestor reference sequence (NC_007530.2). The last column indicates whether the corresponding mutation has been previously described in the literature. Amino acid abbreviations are as follows: (G), glycine; (D), aspartic acid; (L), leucine; (S), serine, (H), histidine; (Q), glutamine; (Y), tyrosine. †, nt position upstream of the ATG start codon for bla1 . *, penicillin-resistant strains not included in this screen but included in this table for reference purposes. FS, frameshift. sigP . Three of the 13 strains, 2002013094 ( B. anthracis 3094 [Ba3094]), 2000031021 (Ba1021), and 2000031052 (Ba1052), contained mutations in sigP , and all mutations were confirmed by Sanger sequencing. Ba3094, Ba1021, and Ba1052 were shown to belong to clade C by multilocus variable-number tandem-repeat 8 (MLVA-8) genotyping ( 24 ) and to share the same transition mutation at nucleotide position 119. Strain Ba1052 contained an additional transition mutation in sigP at position 395. None of the strains contained the sigP SNP previously described in PEN-R strain 32 that resulted in a single amino acid difference predicted to decrease SigP activity ( 4 ). All study strains with sigP mutations were PEN-S by conventional BMD AST ( Table 1 ). bla1 . Two of 13 strains contained mutations in the bla1 coding region, or in the bla1 promoter region, and all of the mutations were confirmed by Sanger sequencing. Strain 2002734089 (Ba4089) contained a transition mutation and belonged to clade A by MLVA-8 genotyping, and strain 2000031048 (Ba1048) contained a transversion in the −35 promoter element and belonged to clade B. All strains with mutations in the bla1 coding region or promoter region were PEN-S by conventional BMD AST. rsiP . Eight of the 13 strains contained mutations in the anti-sigma factor gene, rsiP ( Table 1 ). Mutations were identified in each WGS assembly for strains 2002013017 (Ba3017) and 2002013007 (Ba3007), but Sanger sequencing did not confirm these changes. Analysis of the WGS data revealed errors in the WGS assemblies, and, as a result, these strains were removed from further analysis. The remaining six strains with Sanger-confirmed rsiP mutations belonged to clade A (MLVA-8). Strain 2002013011 (Ba3011) contained a novel SNP that is predicted to lead to an amino acid substitution in RsiP at position 170 of 275 and was PEN-S by AST. Strain 2000031038 (Ba1038) contained a novel frameshift mutation predicted to result in an amino-terminally truncated RsiP protein (30 amino acids [aa]). Strain Ba1038 was isolated from an environmental surface sample in 1957; however, the source of the sample is unknown. Strain 2000032823 (Ba2823) contained a frameshift mutation predicted to result in an amino-terminally truncated RsiP protein (163 aa). This mutation was also previously described in PEN-R strains isolated in 2011 from cattle ( 18 ). Strain Ba2823 was isolated prior to 2011, but additional details about the source of the strain are unknown ( 25 ). Both Ba1038 and Ba2823 were PEN-R by conventional BMD AST. Three strains, 2002734065 (Ba4065 or SK57A), 2002734039 (Ba4039 or SK57C), and 2002013027 (Ba3027), contained the same frameshift predicted to result in a 12-amino-acid, amino-terminally truncated RsiP previously described in PEN-R strain 32 (4) and in PEN-R strains isolated from cattle ( 18 ). Strains Ba4065 (SK57A) and Ba4039 (SK57C) were collected in England in November 1975. Both are likely related to previously described PEN-R strain SK57 ( 14 ). Despite possessing the same rsiP mutation as PEN-R strains SK57A, SK57C, and strain 32, conventional AST confirmed that Ba3027 is PEN-S (MIC = 0.06 µg/ml). ORF 2506 and ORF 2504. In addition to bla1 and bla2 , SigP-induced transcription of ORF 2506 in a PEN-R B. anthracis strain was described previously ( 18 ). BLASTX analysis revealed that ORF 2506 shared similarity with B. cereus group PBP2 (99% coverage, 100% identity; accession number WP_000662966.1) and contained the conserved protein domain family FtsI, which is associated with cell cycle control, cell division, chromosome partitioning, and cell wall/membrane/envelope biogenesis. The majority (∼93.75%) of strains in this study, including all PEN-R strains, contained a frameshift in a homopolymer region of this gene that is also found in the Ames Ancestor reference strain (NCBI accession number AE017334 ) and that resulted in 2 predicted ORFs. The first ORF consisted of the N-terminal 124 amino acids of the predicted PBP2 and contained a predicted PBP dimerization domain. The second ORF contained the last 586 amino acids of the predicted PBP2, which includes the transpeptidase domain. Only 6.25% (26/416) of the strains did not contain this frameshift and had a single ORF predicted to represent the full-length PBP2 (710 aa). BLASTX analysis revealed that ORF 2504 also shared similarity with a B. cereus group PBP2 (99% coverage, 100% identity; accession number WP_000903320.1). However, ORF 2504 is translated in the opposite direction from the other ORFs in the sigP-bla1 region and is not regulated by SigP ( 18 ). Sequencing of PEN-R strains and phylogeny based on whole-genome SNP calling. The de novo assemblies for all study strains were each ≥99.8% identical to the Ames Ancestor reference genome (data not shown). A phylogeny based on whole-genome SNP calling was created for all strains with mutations in the coding regions of sigP , rsiP , and/or bla1 ( Fig. 2 ). Strains 32, UT308, and SK57 ( 15 ) contained mutations in sigP and/or rsiP ( Table 1 ) and were included in the phylogenetic analysis as reference strains. All strains with a mutation in rsiP belonged to clade A. Two of the strains identified in this screen with rsiP mutations (Ba4065 and Ba4039) fell into the same clade as the three PEN-R reference strains. In comparison to SK57, Ba4065 (SK57A), Ba4039 (SK57C), and strain 32 and its derivative UT308 have an additional mutation in sigP that is predicted to affect protein activity ( 4 ). PEN-S Ba3027 harbors the same rsiP mutation as strains 32, SK57, Ba4065 (SK57A), and Ba4039 (SK57C) but is not in the same branch as the PEN-R strains. Two other PEN-R isolates with rsiP mutations, Ba2823 and Ba1038, were in different branches. The two strains with mutations in bla1 were members of clade A, and the three strains containing sigP mutations fell within clade C. FIG 2 Neighbor-joining tree of B. anthracis strains containing mutations within sigP , rsiP , and bla1 identified from the WGS screen. Strains from the three major B. anthracis clades (A, B, and C) were identified in the WGS screen (color coded) ( 24 ). PEN-R strains are colored in red. Asterisks (*) indicate strains containing an rsiP mutation. Control strains SK57, UT308, 2000031103 (strain 32), and Sterne, as well as the Ames Ancestor reference strain, were included for comparison. Characterization of Ba3027: Growth, colony morphology, β-lactamase activity, and bla gene expression. Strain Ba3027 exhibited a slower growth rate in broth culture than strain SK57, which contains the same rsiP mutation, and 2007740878 (Ba0878), a prototypical PEN-S strain without the rsiP mutation ( Fig. 3A ). After 24 h of culture on agar, single isolated colonies of strain Ba0878 were nearly half the size of strain SK57 colonies, with average colony diameters of 2.82 mm and 5.40 mm, respectively ( Fig. 3B ). Strain Ba3027 formed smaller colonies (average colony diameter, 0.85 mm) than Ba0878 and SK57. The β-lactamase activity of culture supernatants was measured in a quantitative nitrocefin assay to evaluate extracellular β-lactamase production in Ba3027. Similarly to PEN-S strains Ba0878 and Sterne, β-lactamase activity was not detected in the culture supernatant from Ba3027 ( Fig. 4A ). A statistically significant difference in activity levels ( P = 0.001 to 0.01) was observed for these three strains compared to β-lactamase-producing strain UT308. SK57 had significantly higher β-lactamase activity (223 ± 4.35 mU/ml) than UT308 (61 ± 3.15 mU/ml) ( P < 0.001). Whole-cell lysates of SK57, Ba0878, and Ba3027 were prepared and tested using the quantitative nitrocefin assay to determine if β-lactamase is produced in Ba3027 but is not exported to the culture supernatant or associated with the cell wall. β-Lactamase activity was detected in the PEN-R SK57 lysates but not in the PEN-S Ba3027 or Ba0878 lysates (data not shown). FIG 3 Growth characteristics of Ba3027. (A) Growth kinetics of strains SK57 (PEN-R, rsiP 10 mutation), Ba3027 (PEN-S, rsiP 10 mutation), and Ba0878 (PEN-S, wild-type strain) were evaluated over a 12-h incubation at 35°C in broth. Growth was measured by the Segmentation and Extraction of Surface Area (SESA) algorithm. Graphs represent the average growth value ± standard deviations from three replicate wells. (B) Microscope images (×8) of single colonies were taken following an 18-h incubation at 35°C in ambient air on SBA (top); optical screen images represent bacterial growth in a 100-µl cell suspension after 7 h (bottom). FIG 4 β-Lactamase production and semiquantitative RT-PCR analysis of bla1 , bla2 , and 16S transcripts in B. anthracis strains. (A) β-Lactamase activity of culture supernatants from strains SK57 (PEN-R, rsiP 10 mutation), UT308 (PEN-R, sigP 183 mutation, rsiP 10 mutation), Ba3027 (PEN-S, rsiP 10 mutation), Sterne (PEN-S), and Ba0878 (PEN-S, wild-type strain) was measured using nitrocefin. Error bars represent averages ± standard deviations. **, P = 0.001 to 0.01 (statistical significance compared to β-lactamase-producing strain UT308); ***, P < 0.001 (statistical significance of UT308 compared to SK57). (B) Expression of SK57, UT308, Ba3027, Sterne, and Ba0878 bla1 , bla2 , and 16S genes was analyzed by semiquantitative RT-PCR after 20 cycles. The molecular marker was run in lane M. Semiquantitative reverse transcriptase PCR (RT-PCR) analysis of bla1 and bla2 was performed to measure bla1 and bla2 transcription in Ba3027 at the exponential phase of growth in Luria-Bertani (LB) broth. Transcripts of bla1 and bla2 were detected in Ba3027 but at lower levels than were seen with PEN-R strains SK57 and UT308. No bla1 or bla2 transcript was detected in PEN-S control strains Sterne and Ba0878 ( Fig. 4B ). Controls for these assays confirmed that no amplification product was detectable in the no-transcriptase or no-template reactions for each reaction set (data not shown). Genomic mutations unique to Ba3027. Despite harboring the same RsiP truncation mutation as four other PEN-R B. anthracis strains, PEN-S strain Ba3027 clustered with one other PEN-S strain in the WGS SNP-based tree ( Fig. 2 ). To identify chromosomal mutations unique to Ba3027, the coding region mutations in Ba3027 were compared to coding region mutations found in the B. anthracis isolates used to generate the phylogenetic tree. Forty-six frameshift and missense mutations were unique to Ba3027. These nucleotide differences were found in a variety of genes, including those predicted to be involved in cell growth and cell division, peptidoglycan biosynthesis processes, regulation of transcription, and DNA-directed RNA polymerase activity (see Table S1 in the supplemental material). 10.1128/mSystems.00154-18.2 TABLE S1 Coding region mutations unique to Ba3027, including gene ontology (GO, Gene Ontology Consortium) identifications (IDs) and descriptions. Download Table S1, PDF file, 0.1 MB . This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply. DISCUSSION WGS can be used to detect the mutations and genes most frequently associated with drug resistance for bacterial pathogens such as Mycobacterium tuberculosis and methicillin-resistant Staphylococcus aureus ( 26 – 29 ). For M. tuberculosis , the WGS data are valuable because it offers a more rapidly determined AST profile for first-line and second-line anti-TB drugs than the conventional BMD method ( 28 , 30 ). In the event of a deliberate release of B. anthracis , antimicrobial susceptibility results would be essential for the distribution of appropriate antimicrobial agents for PEP and treatment. WGS of the implicated strain(s) could reveal known, novel, and/or engineered genetic modifications, including indels and SNPs related to antibiotic resistance. For example, SNPs located in the quinolone resistance-determining region of B. anthracis can be analyzed to predict functional ciprofloxacin resistance ( 31 , 32 ) and can serve as useful targets during genomic analysis. The presence of β-lactamase genes is considered predictive of penicillin resistance for many bacterial species ( 5 ). However, B. anthracis is characteristically susceptible to penicillin despite containing two chromosomal β-lactamase genes. A B. anthracis anti-sigma factor ( rsiP ) mutation can lead to β-lactamase gene expression and penicillin resistance ( 4 ). Mutations in sigP and rsiP have been reported only in strain 32 ( 4 ) and in PEN-R strains isolated from cattle following a 2011 anthrax outbreak ( 18 ). To more accurately predict resistance phenotypes from bacterial sequence data, it is critical to identify mutations that exist among different strains and to determine the impact that these variations have on the observed PEN-R phenotype ( 33 ). In this work, we analyzed the sigP-bla1 region of 374 strains of B. anthracis from a strain collection at the CDC to determine the frequency of mutations across the ∼6.9-kb region. This collection includes B. anthracis strains isolated from human, animal, and environmental sources worldwide from the 1950s to 2013 ( 34 ). Functional antimicrobial susceptibility testing was then performed to determine the accuracy of penicillin resistance prediction. A limitation of this study is that BMD AST was not performed on every study strain, and we therefore do not know if there are any strains in the CDC collection that are PEN-R and lack a mutation in the sigP-bla1 region. While it was not feasible to perform AST for all 374 strains for which WGS data are available, we screened the WGS data to identify strains containing mutations in the sigP-bla1 region with the goal of identifying strains similar to strain 32. Few strains (3.5% [13/374]) contained a mutation(s) (SNPs, insertions, or deletions) in the sigP , rsiP , or bla1 coding region. None of the strains with mutations in sigP or bla1 were resistant to penicillin, indicating that the sigma factor and anti-sigma factor supression systems are likely fully functional in these strains. Only 5 strains (1.3% [5/374]) contained a mutation in the rsiP coding region that was predicted to result in a truncation of the anti-sigma factor. Four of those five strains that contained an rsiP mutation were resistant to penicillin, indicating that the anti-sigma factor is likely not functional. A single PEN-S strain with a rsiP mutation, Ba3027, displayed a smaller colony size and a slower growth rate than the wild-type PEN-R and PEN-S strains. While β-lactamase activity was not detected by quantitative nitrocefin assays, transcripts of both bla1 and bla2 were detected in Ba3027. This indicates that bla1 and bla2 are expressed in Ba3027 to some extent but suggests that the production level is not sufficient to detect β-lactamase activity using the nitrocefin assay or resistance by BMD AST. The PEN-S phenotype and lack of β-lactamase activity in Ba3027 were unexpected. To assess whether this phenotype/genotype discrepancy could be explained using WGS data, we compared sequences of all strains included in the phylogenetic tree to the Ames Ancestor reference sequence. All coding region mutations unique to Ba3027 are listed in Table S1 in the supplemental material. A total of 46 frameshift and missense mutations were found, any of which might contribute to the unusual growth characteristics and/or PEN-S phenotype of Ba3027. For example, one mutation was found in ftsA , a gene related to the cell cycle. The loss of this gene has been reported to result in impaired cell division and sporulation in B. subtilis ( 35 ). While the specific genetic basis of this unexpected PEN-S phenotype of Ba3027 was not immediately identified, these unique mutations represent potential candidates for future investigation. Mutations identified at three positions that led to predicted truncated RsiP proteins of 12, 30, and 163 amino acids were all located in homopolymer regions of the B. anthracis chromosome, suggesting that these regions are hot spots for insertion or deletion events ( 18 ). We identified two strains with deletions in these homopolymer regions, but Sanger sequencing data did not confirm the presence of these mutations. Subsequent AST testing revealed that both strains were PEN-S. Some next-generation sequencing (NGS) technologies have difficulty resolving homopolymer regions of DNA sequence ( 36 ). This emphasizes the importance of sequence accuracy, especially in homopolymer regions, for high-confidence detection of mutations. Depending on the NGS technology used, confirmatory sequencing of these repetitive regions by the Sanger method may be necessary. While both SK57 and UT308 are PEN-R by AST, the β-lactamase activity in the culture supernatant of SK57 (223 mU/ml) was higher than in that of UT308 (61 mU/ml). β-Lactamases in Gram-positive bacteria are predominantly located extracellularly, but β-lactamases can also adhere to the peptidoglycan layer (or capsule) or diffuse away ( 37 ). Whole-cell lysates of Ba3027 were tested using the quantitative nitrocefin assay to assess whether β-lactamase is produced in Ba3027 cells but not exported to the culture supernatant and is instead associated with the cell wall. However, no β-lactamase activity was detected (data not shown). This indicates that the level of production of Bla1 and Bla2 by Ba3027 was too low for detection of cell wall-associated or extracellular β-lactamase activity using the quantitative nitrocefin assay. The sigP sequence of SK57 is identical to that of the Ames Ancestor reference, but UT308 contains the sigP mutation that is predicted to be associated with decreased SigP activity ( 4 ). The elevated β-lactamase activity in SK57 compared to UT308 is likely due to this predicted decreased activity of SigP in UT308. Ross et al. ( 4 ) suggested that constitutive expression of wild-type SigP is detrimental to B. anthracis growth and that the strain 32 SigP mutation is acting to reduce the activity of the sigma factor and to alleviate this toxicity. Wild-type SigP activity was noted to be tolerated in the PEN-R strains isolated from the anthrax outbreak in cattle ( 18 ). Our findings indicate that SigP activity was also tolerated in the 4 PEN-R strains that contained wild-type SigP in this study. Agren et al. ( 18 ) also described a subpopulation that was isolated from the frozen stock of a PEN-R strain that not only contained the rsiP mutation but also contained a sigP mutation that abolished SigP activity and resulted in a penicillin-susceptible phenotype. This was described as a counteracting mutation. Both Ross et al. ( 4 ) and Agren et al. ( 18 ) speculated that strains expressing wild-type SigP with rsiP mutations acquired detrimental sigP mutations since SigP expression results in growth defects. Downregulation of β-lactamase genes could offer these strains a fitness advantage ( 4 , 18 ). None of the PEN-R strains isolated in this study contained both a sigP mutation and a rsiP mutation; therefore, all of the PEN-R strains are presumed to express the wild-type sigP gene. Not only was a smaller colony size observed for Ba3027 ( Fig. 3 ), but PEN-R strains Ba1038, Ba2823, and Ba4065 (SK57A) also had smaller colony diameters than PEN-R SK57 and Ba4039 (SK57C) as well as other PEN-S wild-type strains (data not shown). However, it is not uncommon to see variability in growth characteristics and colony morphologies in wild-type B. anthracis strains ( 23 ). Because not all PEN-R strains in this study exhibited a reduced growth rate and smaller colony size, additional work is needed to establish whether expression of wild-type SigP is directly associated with growth defects in B. anthracis . Work by Ross et al. ( 4 ) showed that transforming a B. anthracis sigP-rsiP null mutant with sigP and rsiP genes from PEN-R B. cereus or B. thuringiensis strains resulted in β-lactamase activity. This suggests that B. anthracis contains the genes required for sensing β-lactam antibiotics but that the presence of wild-type (prototypical) SigP and RsiP is not sufficient for bla induction. Ross et al. propose that the B. cereus and B. thuringiensis RsiP proteins can respond to the signal when a β-lactam antibiotic is present and that those species are in turn characteristically PEN-R. The defective anti-sigma factor could explain why RsiP in prototypical PEN-S B. anthracis strains does not respond to this signal ( 4 ). While the B. cereus and B. thuringiensis rsiP genes are 91% to 99% identical to the corresponding B. anthracis gene on the nucleotide level, future work will determine whether any sequence differences are involved in the inability of B. anthracis RsiP to respond to the β-lactam antibiotic signal. Molecular phylogenies and comparative genome sequencing have revealed that B. cereus species group bacteria are closely related and could be classified as a single species ( 38 , 39 ). The de novo assemblies for every strain included in this study were ≥99.8% identical to those in the Ames Ancestor reference genome. Analysis of the whole-genome SNP phylogeny revealed that every strain with a mutation in rsiP belonged to clade A; however, the limited number of clade B and clade C strains included in this study precluded a definitive association of rsiP mutations or PEN-R genotypes with clade A strains. Two of the strains with rsiP mutations (SK57A and SK57C) were located in the same group as the three PEN-R reference strains (strain 32, SK57, and UT308). The results revealing this monophyletic group, along with the common frameshift mutation in rsiP , strongly suggest a common origin for all of these isolates. However, three isolates with rsiP mutations (Ba3011, Ba3027, and Ba1038) grouped differently, indicating that these isolates evolved independently of the monophyletic group. Strains SK57A and SK57C were isolated from an archival collection of CDC strains that were originally stored on agar slants overlaid with mineral oil at room temperature ( 40 ). Both strains were isolated in November 1975, but the CDC records associated with these strains contain few details about the source and there is no clear association with strain 32 ( 12 ). Despite the underlying genomic similarity, B. anthracis isolates are phenotypically diverse because of altered gene expression rather than gene content ( 4 , 38 ). The control of β-lactamase expression by SigP and RsiP is an example of how trans -acting factors differentially affect transcription of genes in the B. cereus group species. Here, we show that analysis of the sigP-bla1 region in B. anthracis was useful in predicting penicillin resistance in the majority of B. anthracis strains that contained a mutation in rsiP . This locus should be included in analyses of WGS data to predict antimicrobial resistance of B. anthracis . However, it cannot be used exclusively, as only four of the five strains with rsiP mutations were PEN-R by conventional BMD testing. There are a limited number of PEN-R B. anthracis strains available for analysis, and this makes it difficult to accurately assess concordance between phenotypic susceptibility and resistance genotypes. To accurately assess clinically relevant β-lactamase production in B. anthracis , use of a conventional phenotypic method, like BMD AST, remains essential. MATERIALS AND METHODS Biosafety procedures. B. anthracis is subject to select agent regulations (42 CFR part 73). All procedures using the attenuated, select agent-excluded strains, B. anthracis UT308 and Sterne, were performed in a biosafety level 2 (BSL2) laboratory by trained personnel wearing appropriate personal protective equipment (PPE). All procedures involving wild-type B. anthracis strains were performed by trained personnel wearing PPE (including a powered air-purifying respirator [PAPR] and protective laboratory clothing) in a class II type A2 biological safety cabinet located in a BSL3 laboratory registered with the U.S. Federal Select Agent Program. Additional information regarding the facility and equipment and procedural guidelines for BSL2 and BSL3 laboratories can be found in the CDC/NIH publication "Biosafety in Microbiological and Biomedical Laboratories," 5th edition ( 41 ). Bacterial strains. All B. anthracis strains identified in the sigP-bla1 screen and control strains used in this study are listed in Table 2 . PEN-R strains 32, UT308, and SK57 and the PEN-S Sterne and Ba0878 strains were used as control strains in this study. TABLE 2 B. anthracis strains used in this study a Strain Alternative ID Clade/MLVA-8 genotype Plasmid content Origin Source NCBI accession no. Reference Sterne* 34F2 A/— pX01 + , pX02 − South Africa Animal CP009541.1 51 UT308* NA A/— pX01 + , pX02 − Derived from strain 32 Human QPKP00000000 4 2007740878* Ba0878, BA0018 A/7 pX01 + , pX02 + Canada Unknown NA 52 2000031656 Ba1656, Ames A/62 pX01 + , pX02 + USA (Texas) Animal NC_007530.2 53 2000031103 Ba1103, ASC 32, strain 32 A/53 pX01 + , pX02 + England Human QPKO00000000 12 2007740863 Ba0863, SK57 A/48 pX01 + , pX02 + England Unknown QPKQ00000000 14 2002734065 Ba4065, SK57A A/— pX01 − , pX02 − England Unknown SRR5811123 This study 2002734039 Ba4039, SK57C A/116 pX01 + , pX02 + England Unknown SRR2340304 25 2002013027 Ba3027, AO427 A/4 pX01 + , pX02 + Unknown Unknown SRR2339620 This study 2000031038 Ba1038, 300, Dole 111 A/71 pX01 + , pX02 + Unknown Environmental SRR5811163 54 2002013017 Ba3017, AO412 A/4 pX01 + , pX02 + Unknown Unknown SRR2339614 This study 2000032823 Ba2823, A0048d A/— pX01 − , pX02 + Unknown Unknown SRR5811071 25 2002013011 Ba3011, SPU A0423 A/118 pX01 + , pX02 + Unknown Unknown SRR2340484 25 2002013007 Ba3007, AO461 A/4 pX01 + , pX02 + Unknown Unknown SRR2340462 This study 2000031048 Ba1048, 305, tannery 42 B/107 pX01 + , pX02 + Unknown Environmental SRR5811217 25 2002734089 Ba4089, SK83, C2291 A/71 pX01 + , pX02 + USA (New Jersey) Environmental SRR2340461 This study 2002013094 Ba3094, 240 C/133 pX01 + , pX02 + Unknown Environmental SRR5947106 25 2000031021 Ba1021, 239, LA164B C/— pX01 − , pX02 + USA (Louisiana) Environmental SRR5947105 This study 2000031052 Ba1052, 278, #25600 C/— pX01 − , pX02 + USA (Wyoming) Animal SRR5811214 This study a Asterisks (*) indicate reference strains not included in the sigP-bla1 screen. ID, identifier(s); NA, not available; —, MLVA-8 genotype determination not performed due to lack of a plasmid. Growth conditions. B. anthracis strains were cultured on BD BBL Trypticase soy agar II with 5% sheep blood (SBA) (Thermo Fisher Scientific, Waltham, MA, USA) at 35°C in ambient air from glycerol stocks stored at −70°C. For microscopy and sizing of single colony isolates, each strain was subcultured at 35°C in ambient air for 18 h on SBA. For AST, strains were grown overnight (16 to 24 h) at 35°C on SBA in ambient air, following CLSI guidelines. To assess β-lactamase activity ( 4 ) and for RNA isolation ( 5 , 10 ), each strain was cultured at 37°C in ambient air on SBA overnight to replicate the culturing conditions described previously. From an overnight growth culture, a cell suspension equivalent to a 0.5 McFarland density standard was prepared in fresh LB broth (BD Difco Miller Luria-Bertani; Thermo Fisher Scientific, Waltham, MA, USA) for RNA isolations or in LB broth containing 0.5% glycerol for β-lactamase activity assays. Sequencing and analysis. Genomic DNA was sequenced from 374 B. anthracis strains from the CDC collection. DNA from isolates was extracted using one of two technologies, a QIAamp DNA blood minikit (Qiagen, Valencia, CA) or a Maxwell 16 instrument (Promega, Madison, WI). For the QIAamp extraction, one 10-µl loopful of cells from overnight growth on SBA was inoculated in heart infusion broth (Remel, Lenexa, KS) and incubated for 18 to 24 h. Cells were harvested by centrifugation for 10 min at 5,000 × g . After removal of broth, DNA was extracted from remaining cells using a Qiagen QIAamp DNA blood minikit following the manufacturer's protocol for isolating Gram-positive bacteria. For DNA extractions performed on the Maxwell instrument, four to five colonies of overnight growth from SBA were mechanically disrupted by vortex mixing for 2 min in a suspension of silica beads and Tris-EDTA (TE) buffer. The suspension was centrifuged for 30 s at 10,000 × g . A 300-µl volume of the resulting supernatant was used for DNA extraction following the manufacturer's protocol for blood and cells. Sequencing was performed on an Illumina GAIIx system using TruSeq chemistry (Illumina, San Diego, CA). Paired-end reads were trimmed, adaptor sequences were removed, and the reads were subjected to quality checking using SolexaQA++ ( 42 ) with a Phred quality score threshold of 20 and minimum length of 50 bp, Scythe with default parameters, and FastQC, respectively. Paired-end reads for which both reads passed the quality control were then assembled using IDBA-UD ( 43 ) with precorrection and default parameters, and the resulting scaffolds that were shorter than 500 bp were discarded. A local BLAST ( 44 ) search was performed by querying the sigP-bla1 region (6,892 bp) from a PEN-S reference strain, the Ames Ancestor (accession number AE017334 ), against the 374 B. anthracis strains in the CDC WGS database to identify mutations. Forty-two B. anthracis strains in GenBank for which WGS data were available at the time of analysis were also included in the BLAST screen (total, 416 strains). The sigP-rsiP , bla1 , and bla2 promoter regions were also analyzed for the 13 strains that contained a sigP , rsiP , or bla1 mutation. Sanger sequencing of the sigP-bla1 region was performed using an ABI 3130xl or 3500xl Genetic Analyzer (Applied Biosystems, by Thermo Fisher Scientific, Waltham, MA, USA) as described by Hakovirta et al. ( 45 ) to confirm mutations. The primers used for PCR amplification and sequencing were as follows: sigP-rsiP FwdPCR (5′-GGAGAACTCGAACTAAATGG-3′), sigP-rsiP RevPCR (5′-GCTGCTCTCGTTACATCA-3′), sigP-rsiP IntFwd1 (5′-TGATAAACAAACTCTGTCGG-3′), sigP-rsiP IntFwd2 (5′-CCTAAAAAGCACCGTGA-3′), sigP-rsiP IntFwd3 (5′-CTGCTCAAGATCCAACAT-3′), sigP-rsiP IntFwd4 (5′-CTGAACCAAAGCGAGAAT-3′), sigP-rsiP IntRev1 (5′-GCCGACAGAGTTTGTTTA-3′), sigP-rsiP IntRev2 (5′-AATGGTCTTGTATGTTCCC-3′), sigP-rsiP IntRev3 (5′-CTTTTGATTCTCGCTTTGGT-3′), bla1 FwdPCR (5′-AATAAGAGATAGCAGCGG-3′), bla1 RevPCR (5′-GGTTTTTCACGTATCTGG-3′), and bla1 IntRev1 (5′-ACACCTAATCGAGCATCA-3′). BLAST and Sanger sequence data were analyzed using Geneious R8 software, version 8.1.4, and CLC Genomics Workbench software, version 7.5.1. Mutations in wild-type strains were identified by comparing the genome assemblies of the wild-type strains to that of the Ames Ancestor reference sequence using the dnadiff utility from the MUMmer package ( 46 ). The SnpEff utility ( 47 ) was used to predict the effects of mutations identified as unique to Ba3027. The Harvest suite of tools ( 48 ) was used to determine the numbers of SNPs in comparisons between all wild-type genomes. These differences were represented as a distance matrix and used to create a neighbor-joining tree using MEGA 6 ( 49 ). MLVA-8 subtyping. MLVA-8 genotyping was performed as described by Keim et al. ( 24 ) and Sue et al. ( 25 ). Briefly, six chromosomal loci ( vrrA , vrrB1 , vrrB2 , vrrC1 , vrrC2 , and CG3) and two plasmid loci (pXO1- aat and pXO2- at ) were amplified by PCR and the resulting DNA fragments were separated on an ABI 3130xl instrument or an ABI 3500xl instrument (Applied Biosystems, by Thermo Fisher Scientific, Waltham, MA, USA). Antimicrobial susceptibility testing (AST). Broth microdilution (BMD) was performed to determine antimicrobial susceptibility for penicillin following the CLSI guidelines ( 19 ). Cells from four to six isolated colonies of an overnight culture were suspended in saline solution (Beckman Coulter, Brea, CA) and mixed using a vortex mixer to a turbidity equivalent to a 0.5 McFarland standard as measured with a MicroScan turbidity meter (Siemens, Munich, Germany). Each suspension was then diluted 1:20. BMD AST panels prepared in-house with cation-adjusted Mueller-Hinton broth were inoculated and incubated at 35°C in ambient air for 16 to 20 h. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212 were used as control strains. The MIC of penicillin for each B. anthracis strain was recorded as the concentration at the first well where there was no visible growth. Microscopy and sizing of single colony isolates. Images of individual colonies on SBA were acquired with a Leica EZ4 HD digital stereo microscope (Leica Microsystems, Wetzlar, Germany). Colony diameters were measured using a Digimatic Solar Caliper (Mitutoyo America, IL, USA). Imaging and analysis of bacterial growth in broth culture by optical screening. B. anthracis cell suspensions equivalent to a 0.5 McFarland density standard were prepared as described previously ( 23 ) from colonies grown overnight on SBA. The cell suspension was diluted 1:100 in cation-adjusted Mueller-Hinton broth with TES [ N -tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid] (CAMHBT; Remel Inc., Lenexa, KS). A 100-µl aliquot of each diluted cell suspension was transferred into each well of a 96-well plate. Optical screening images were generated from scans through a fluid sample using digital time-lapse microscopy with an oCelloScope instrument (Biosense Solutions ApS, Farum, Denmark) as described previously ( 23 , 50 ). The instrument-derived growth values were obtained using the Segmentation and Extraction Surface Area (SESA) normalized algorithm. Growth kinetic data in Fig. 3 represent the means of triplicate values ± standard deviations. Nitrocefin β-lactamase assays. Broth cultures in the late-exponential-growth phase (10 5 CFU/ml) of B. anthracis strains SK57, Ba3027, Ba0878, Sterne (PEN-S, select agent-excluded strain), and UT308 (PEN-R, select agent-excluded derivative of strain 32) were tested for β-lactamase activity using a nitrocefin-based quantitative β-lactamase activity assay (β-lactamase activity assay kit, MAK221; Sigma-Aldrich, St. Louis, MO) according to the manufacturer's instructions. Supernatants from 500 µl of culture from each strain were collected by centrifugation at 8,000 × g for 2 min through a 0.1-μm-pore-size polyvinylidene difluoride (PVDF) Ultrafree-MC spin-filter column (Millipore, Billerica, MA, USA). The absorbance ( A 490 ) was measured in a 96-well microplate reader (SpectraMax i3; Molecular Devices, Sunnyvale, CA) in kinetic mode for 60 min at room temperature. Two biological replicates were performed for each sample, and three technical replicates were analyzed for each biological replicate. The statistical significance of differences in levels of β-lactamase production between strains was calculated with a two-tailed t test ( n = 3). RNA isolation. Total RNA was isolated as previously described ( 5 , 10 ). Briefly, total RNA was purified from exponential-phase cultures (10 5 CFU/ml) grown in LB broth at 37°C in ambient air, without shaking. Cells were collected on 0.22-µM-pore-size PVDF Ultrafree-MC spin-filter columns (Millipore, Billerica, MA, USA) by centrifugation at 8,000 × g for 2 min, resuspended in RNAprotect bacterial reagent (Qiagen, Hilden, Germany), and stored at −70°C until the time of extraction. RNA was isolated and purified using a RiboPure RNA purification kit for bacteria, including the DNase I treatment to remove traces of contaminating DNA, according to the instructions of the manufacturer (Thermo Fisher Scientific, USA). RNA quantity was assessed using a Qubit RNA HS assay kit and a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). The RNA quality was assessed using an Agilent 2100 Bioanalyzer with an RNA 6000 Nano kit (Agilent, Santa Clara, CA, USA). Semiquantitative RT-PCR. Purified RNA (1.0 µg) was subjected to reverse transcription with random decamers using a RETROscript kit (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions. The resulting reverse transcriptase (RT) mixtures were used as the template for PCRs with internal primers BLA1-5 and BLA1-6 for bla1 , BLA2-5 and BLA2-6 for bla2 , and 16S-1 and 16S-2 for the 16S rRNA housekeeping gene, which were previously described by Chen et al. ( 10 ). Semiquantitative PCRs were carried out with a RETROscript Kit as described by the manufacturer with 2 U of SuperTaq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) in 50-μl reaction mixtures containing 100 ng of RT mixture template. PCR amplification conditions were as follows: 1 cycle of denaturation at 94°C for 4 min, followed by 20 cycles of denaturation at 94°C for 30 s, primer annealing at 55°C for 30 s and template extension at 72°C for 1 min, and 1 cycle of final extension at 72°C for 5 min. Genomic DNA extracted from B. anthracis Sterne was used as a positive control. A no-RT sample and a no-template sample were included as negative controls for each reaction set. Amplification products were detected by gel electrophoresis on a 1% agarose gel, and size was assessed with a DNA ladder (Invitrogen low-DNA mass ladder; Thermo Fisher Scientific, Waltham, MA, USA). Accession number(s). All reads were submitted to the NCBI Sequence Read Archive, and the accession numbers are listed in Table 2 . Biosafety procedures. B. anthracis is subject to select agent regulations (42 CFR part 73). All procedures using the attenuated, select agent-excluded strains, B. anthracis UT308 and Sterne, were performed in a biosafety level 2 (BSL2) laboratory by trained personnel wearing appropriate personal protective equipment (PPE). All procedures involving wild-type B. anthracis strains were performed by trained personnel wearing PPE (including a powered air-purifying respirator [PAPR] and protective laboratory clothing) in a class II type A2 biological safety cabinet located in a BSL3 laboratory registered with the U.S. Federal Select Agent Program. Additional information regarding the facility and equipment and procedural guidelines for BSL2 and BSL3 laboratories can be found in the CDC/NIH publication "Biosafety in Microbiological and Biomedical Laboratories," 5th edition ( 41 ). Bacterial strains. All B. anthracis strains identified in the sigP-bla1 screen and control strains used in this study are listed in Table 2 . PEN-R strains 32, UT308, and SK57 and the PEN-S Sterne and Ba0878 strains were used as control strains in this study. TABLE 2 B. anthracis strains used in this study a Strain Alternative ID Clade/MLVA-8 genotype Plasmid content Origin Source NCBI accession no. Reference Sterne* 34F2 A/— pX01 + , pX02 − South Africa Animal CP009541.1 51 UT308* NA A/— pX01 + , pX02 − Derived from strain 32 Human QPKP00000000 4 2007740878* Ba0878, BA0018 A/7 pX01 + , pX02 + Canada Unknown NA 52 2000031656 Ba1656, Ames A/62 pX01 + , pX02 + USA (Texas) Animal NC_007530.2 53 2000031103 Ba1103, ASC 32, strain 32 A/53 pX01 + , pX02 + England Human QPKO00000000 12 2007740863 Ba0863, SK57 A/48 pX01 + , pX02 + England Unknown QPKQ00000000 14 2002734065 Ba4065, SK57A A/— pX01 − , pX02 − England Unknown SRR5811123 This study 2002734039 Ba4039, SK57C A/116 pX01 + , pX02 + England Unknown SRR2340304 25 2002013027 Ba3027, AO427 A/4 pX01 + , pX02 + Unknown Unknown SRR2339620 This study 2000031038 Ba1038, 300, Dole 111 A/71 pX01 + , pX02 + Unknown Environmental SRR5811163 54 2002013017 Ba3017, AO412 A/4 pX01 + , pX02 + Unknown Unknown SRR2339614 This study 2000032823 Ba2823, A0048d A/— pX01 − , pX02 + Unknown Unknown SRR5811071 25 2002013011 Ba3011, SPU A0423 A/118 pX01 + , pX02 + Unknown Unknown SRR2340484 25 2002013007 Ba3007, AO461 A/4 pX01 + , pX02 + Unknown Unknown SRR2340462 This study 2000031048 Ba1048, 305, tannery 42 B/107 pX01 + , pX02 + Unknown Environmental SRR5811217 25 2002734089 Ba4089, SK83, C2291 A/71 pX01 + , pX02 + USA (New Jersey) Environmental SRR2340461 This study 2002013094 Ba3094, 240 C/133 pX01 + , pX02 + Unknown Environmental SRR5947106 25 2000031021 Ba1021, 239, LA164B C/— pX01 − , pX02 + USA (Louisiana) Environmental SRR5947105 This study 2000031052 Ba1052, 278, #25600 C/— pX01 − , pX02 + USA (Wyoming) Animal SRR5811214 This study a Asterisks (*) indicate reference strains not included in the sigP-bla1 screen. ID, identifier(s); NA, not available; —, MLVA-8 genotype determination not performed due to lack of a plasmid. Growth conditions. B. anthracis strains were cultured on BD BBL Trypticase soy agar II with 5% sheep blood (SBA) (Thermo Fisher Scientific, Waltham, MA, USA) at 35°C in ambient air from glycerol stocks stored at −70°C. For microscopy and sizing of single colony isolates, each strain was subcultured at 35°C in ambient air for 18 h on SBA. For AST, strains were grown overnight (16 to 24 h) at 35°C on SBA in ambient air, following CLSI guidelines. To assess β-lactamase activity ( 4 ) and for RNA isolation ( 5 , 10 ), each strain was cultured at 37°C in ambient air on SBA overnight to replicate the culturing conditions described previously. From an overnight growth culture, a cell suspension equivalent to a 0.5 McFarland density standard was prepared in fresh LB broth (BD Difco Miller Luria-Bertani; Thermo Fisher Scientific, Waltham, MA, USA) for RNA isolations or in LB broth containing 0.5% glycerol for β-lactamase activity assays. Sequencing and analysis. Genomic DNA was sequenced from 374 B. anthracis strains from the CDC collection. DNA from isolates was extracted using one of two technologies, a QIAamp DNA blood minikit (Qiagen, Valencia, CA) or a Maxwell 16 instrument (Promega, Madison, WI). For the QIAamp extraction, one 10-µl loopful of cells from overnight growth on SBA was inoculated in heart infusion broth (Remel, Lenexa, KS) and incubated for 18 to 24 h. Cells were harvested by centrifugation for 10 min at 5,000 × g . After removal of broth, DNA was extracted from remaining cells using a Qiagen QIAamp DNA blood minikit following the manufacturer's protocol for isolating Gram-positive bacteria. For DNA extractions performed on the Maxwell instrument, four to five colonies of overnight growth from SBA were mechanically disrupted by vortex mixing for 2 min in a suspension of silica beads and Tris-EDTA (TE) buffer. The suspension was centrifuged for 30 s at 10,000 × g . A 300-µl volume of the resulting supernatant was used for DNA extraction following the manufacturer's protocol for blood and cells. Sequencing was performed on an Illumina GAIIx system using TruSeq chemistry (Illumina, San Diego, CA). Paired-end reads were trimmed, adaptor sequences were removed, and the reads were subjected to quality checking using SolexaQA++ ( 42 ) with a Phred quality score threshold of 20 and minimum length of 50 bp, Scythe with default parameters, and FastQC, respectively. Paired-end reads for which both reads passed the quality control were then assembled using IDBA-UD ( 43 ) with precorrection and default parameters, and the resulting scaffolds that were shorter than 500 bp were discarded. A local BLAST ( 44 ) search was performed by querying the sigP-bla1 region (6,892 bp) from a PEN-S reference strain, the Ames Ancestor (accession number AE017334 ), against the 374 B. anthracis strains in the CDC WGS database to identify mutations. Forty-two B. anthracis strains in GenBank for which WGS data were available at the time of analysis were also included in the BLAST screen (total, 416 strains). The sigP-rsiP , bla1 , and bla2 promoter regions were also analyzed for the 13 strains that contained a sigP , rsiP , or bla1 mutation. Sanger sequencing of the sigP-bla1 region was performed using an ABI 3130xl or 3500xl Genetic Analyzer (Applied Biosystems, by Thermo Fisher Scientific, Waltham, MA, USA) as described by Hakovirta et al. ( 45 ) to confirm mutations. The primers used for PCR amplification and sequencing were as follows: sigP-rsiP FwdPCR (5′-GGAGAACTCGAACTAAATGG-3′), sigP-rsiP RevPCR (5′-GCTGCTCTCGTTACATCA-3′), sigP-rsiP IntFwd1 (5′-TGATAAACAAACTCTGTCGG-3′), sigP-rsiP IntFwd2 (5′-CCTAAAAAGCACCGTGA-3′), sigP-rsiP IntFwd3 (5′-CTGCTCAAGATCCAACAT-3′), sigP-rsiP IntFwd4 (5′-CTGAACCAAAGCGAGAAT-3′), sigP-rsiP IntRev1 (5′-GCCGACAGAGTTTGTTTA-3′), sigP-rsiP IntRev2 (5′-AATGGTCTTGTATGTTCCC-3′), sigP-rsiP IntRev3 (5′-CTTTTGATTCTCGCTTTGGT-3′), bla1 FwdPCR (5′-AATAAGAGATAGCAGCGG-3′), bla1 RevPCR (5′-GGTTTTTCACGTATCTGG-3′), and bla1 IntRev1 (5′-ACACCTAATCGAGCATCA-3′). BLAST and Sanger sequence data were analyzed using Geneious R8 software, version 8.1.4, and CLC Genomics Workbench software, version 7.5.1. Mutations in wild-type strains were identified by comparing the genome assemblies of the wild-type strains to that of the Ames Ancestor reference sequence using the dnadiff utility from the MUMmer package ( 46 ). The SnpEff utility ( 47 ) was used to predict the effects of mutations identified as unique to Ba3027. The Harvest suite of tools ( 48 ) was used to determine the numbers of SNPs in comparisons between all wild-type genomes. These differences were represented as a distance matrix and used to create a neighbor-joining tree using MEGA 6 ( 49 ). MLVA-8 subtyping. MLVA-8 genotyping was performed as described by Keim et al. ( 24 ) and Sue et al. ( 25 ). Briefly, six chromosomal loci ( vrrA , vrrB1 , vrrB2 , vrrC1 , vrrC2 , and CG3) and two plasmid loci (pXO1- aat and pXO2- at ) were amplified by PCR and the resulting DNA fragments were separated on an ABI 3130xl instrument or an ABI 3500xl instrument (Applied Biosystems, by Thermo Fisher Scientific, Waltham, MA, USA). Antimicrobial susceptibility testing (AST). Broth microdilution (BMD) was performed to determine antimicrobial susceptibility for penicillin following the CLSI guidelines ( 19 ). Cells from four to six isolated colonies of an overnight culture were suspended in saline solution (Beckman Coulter, Brea, CA) and mixed using a vortex mixer to a turbidity equivalent to a 0.5 McFarland standard as measured with a MicroScan turbidity meter (Siemens, Munich, Germany). Each suspension was then diluted 1:20. BMD AST panels prepared in-house with cation-adjusted Mueller-Hinton broth were inoculated and incubated at 35°C in ambient air for 16 to 20 h. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212 were used as control strains. The MIC of penicillin for each B. anthracis strain was recorded as the concentration at the first well where there was no visible growth. Microscopy and sizing of single colony isolates. Images of individual colonies on SBA were acquired with a Leica EZ4 HD digital stereo microscope (Leica Microsystems, Wetzlar, Germany). Colony diameters were measured using a Digimatic Solar Caliper (Mitutoyo America, IL, USA). Imaging and analysis of bacterial growth in broth culture by optical screening. B. anthracis cell suspensions equivalent to a 0.5 McFarland density standard were prepared as described previously ( 23 ) from colonies grown overnight on SBA. The cell suspension was diluted 1:100 in cation-adjusted Mueller-Hinton broth with TES [ N -tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid] (CAMHBT; Remel Inc., Lenexa, KS). A 100-µl aliquot of each diluted cell suspension was transferred into each well of a 96-well plate. Optical screening images were generated from scans through a fluid sample using digital time-lapse microscopy with an oCelloScope instrument (Biosense Solutions ApS, Farum, Denmark) as described previously ( 23 , 50 ). The instrument-derived growth values were obtained using the Segmentation and Extraction Surface Area (SESA) normalized algorithm. Growth kinetic data in Fig. 3 represent the means of triplicate values ± standard deviations. Nitrocefin β-lactamase assays. Broth cultures in the late-exponential-growth phase (10 5 CFU/ml) of B. anthracis strains SK57, Ba3027, Ba0878, Sterne (PEN-S, select agent-excluded strain), and UT308 (PEN-R, select agent-excluded derivative of strain 32) were tested for β-lactamase activity using a nitrocefin-based quantitative β-lactamase activity assay (β-lactamase activity assay kit, MAK221; Sigma-Aldrich, St. Louis, MO) according to the manufacturer's instructions. Supernatants from 500 µl of culture from each strain were collected by centrifugation at 8,000 × g for 2 min through a 0.1-μm-pore-size polyvinylidene difluoride (PVDF) Ultrafree-MC spin-filter column (Millipore, Billerica, MA, USA). The absorbance ( A 490 ) was measured in a 96-well microplate reader (SpectraMax i3; Molecular Devices, Sunnyvale, CA) in kinetic mode for 60 min at room temperature. Two biological replicates were performed for each sample, and three technical replicates were analyzed for each biological replicate. The statistical significance of differences in levels of β-lactamase production between strains was calculated with a two-tailed t test ( n = 3). RNA isolation. Total RNA was isolated as previously described ( 5 , 10 ). Briefly, total RNA was purified from exponential-phase cultures (10 5 CFU/ml) grown in LB broth at 37°C in ambient air, without shaking. Cells were collected on 0.22-µM-pore-size PVDF Ultrafree-MC spin-filter columns (Millipore, Billerica, MA, USA) by centrifugation at 8,000 × g for 2 min, resuspended in RNAprotect bacterial reagent (Qiagen, Hilden, Germany), and stored at −70°C until the time of extraction. RNA was isolated and purified using a RiboPure RNA purification kit for bacteria, including the DNase I treatment to remove traces of contaminating DNA, according to the instructions of the manufacturer (Thermo Fisher Scientific, USA). RNA quantity was assessed using a Qubit RNA HS assay kit and a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). The RNA quality was assessed using an Agilent 2100 Bioanalyzer with an RNA 6000 Nano kit (Agilent, Santa Clara, CA, USA). Semiquantitative RT-PCR. Purified RNA (1.0 µg) was subjected to reverse transcription with random decamers using a RETROscript kit (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions. The resulting reverse transcriptase (RT) mixtures were used as the template for PCRs with internal primers BLA1-5 and BLA1-6 for bla1 , BLA2-5 and BLA2-6 for bla2 , and 16S-1 and 16S-2 for the 16S rRNA housekeeping gene, which were previously described by Chen et al. ( 10 ). Semiquantitative PCRs were carried out with a RETROscript Kit as described by the manufacturer with 2 U of SuperTaq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) in 50-μl reaction mixtures containing 100 ng of RT mixture template. PCR amplification conditions were as follows: 1 cycle of denaturation at 94°C for 4 min, followed by 20 cycles of denaturation at 94°C for 30 s, primer annealing at 55°C for 30 s and template extension at 72°C for 1 min, and 1 cycle of final extension at 72°C for 5 min. Genomic DNA extracted from B. anthracis Sterne was used as a positive control. A no-RT sample and a no-template sample were included as negative controls for each reaction set. Amplification products were detected by gel electrophoresis on a 1% agarose gel, and size was assessed with a DNA ladder (Invitrogen low-DNA mass ladder; Thermo Fisher Scientific, Waltham, MA, USA). Accession number(s). All reads were submitted to the NCBI Sequence Read Archive, and the accession numbers are listed in Table 2 .

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4915566/

Effect of Vaccine Administration Modality on Immunogenicity and Efficacy

Summary The many factors impacting the efficacy of a vaccine can be broadly divided into three categories: (1) features of the vaccine itself, including immunogen design, vaccine type, formulation, adjuvant, and dosing; (2) individual variations among vaccine recipients; and (3) vaccine administration-related parameters. While much literature exists related to vaccines, and recently systems biology has started to dissect the impact of individual subject variation on vaccine efficacy, few studies have focused on the role of vaccine administration-related parameters on vaccine efficacy. Parenteral and mucosal vaccinations are traditional approaches for licensed vaccines; novel vaccine delivery approaches, including needless injection and adjuvant formulations, are being developed to further improve vaccine safety and efficacy. This review provides a brief summary of vaccine administration-related factors, including vaccination approach, delivery route, and method of administration, to gain a better understanding of their potential impact on the safety and immunogenicity of candidate vaccines.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8242524/

How are healthcare provider systems preparing for health emergency situations?

Abstract Natural disasters, disease outbreaks, famine, and human conflict have strained communities everywhere over the course of human existence. However, modern changes in climate, human mobility, and other factors have increased the global community's vulnerability to widespread emergencies. We are in the midst of a disruptive health event, with the COVID‐19 pandemic testing our health provider systems globally. This study presents a qualitative analysis of published literature, obtained systematically, to examine approaches health providers are taking to prepare for and respond to mass casualty incidents around the globe. The research reveals emerging trends in the weaknesses of systems' disaster responses while highlighting proposed solutions, so that others may better prepare for future disasters. Additionally, the research examines gaps in the literature, to foster more targeted and actionable contributions to the literature. INTRODUCTION Natural disasters, disease outbreaks, famine, and human conflict have strained communities everywhere over the course of human existence. However, modern changes in climate, human mobility, and other factors have increased the global community's vulnerability to widespread emergencies. We are currently in the midst of such a disruptive health event. The COVID‐19 pandemic is currently testing health provider systems globally, impacting all sectors and corners of the globe. Disasters of various types present unique challenges to providers who are on the front lines. Providers, equipped only with their training and the limited, unverified information available to them, have to meet the needs of the surge of patients that present to their facilities. At this point, the literature is saturated with needs assessments, which suggests a general consensus: many medical providers feel unprepared to respond to emergency situations. Though emergencies by their very nature are unpredictable, we can still influence the preparation and resources of healthcare providers. This paper qualitatively assesses systematically obtained primary reports and meta‐analyses published in English from 1998 to 2019, asking "How are Healthcare Provider Systems Preparing for Health Emergency Situations?" The objectives are to learn from our collective mistakes, gain from the literature's proposed solutions, and illuminate areas for improvement or further research. METHODS The authors conducted a qualitative assessment of articles systematically obtained from PubMed, Google Scholar, and JSTOR using the search terms in Table 1 in the spring of 2019. In the search, "and" was used between the columns and "or" was used in the columns. Table 1 Review Search Terms Keyword 1 Terms Keyword 2 Terms Keyword 3 Terms Keyword 4 Terms Keyword 5 Terms Keyword 6 Terms Emergencies Planning Techniques Health Personnel Preceptorship Health Facilities Mortality Disease Outbreaks Strategic planning Nurses Education, Medical Bed Occupancy Days of Lost work Epidemics Civil Defense Physicians Education, Medical, Continuing Hospital Bed Capacity Days of lost services Pandemics Relief Work Emergency Medical Technicians Teaching Rounds Health Facility Size Mass Casualty Incidents Quarantine Health Facility Administrator Health Education Hospitals Disasters Disaster Planning Medical Staff Disaster Medicine Community Health Centers Military Personnel John Wiley & Sons, Ltd. To be incorporated into the analysis, the publication must meet the following criteria: (1) Date of publication after 1998, (2) discuss emergency preparation/response, (3) involve healthcare providers, (4) incorporate clinical management and patient‐centered outcomes, (5) have an English version available, and (6) be a primary report/study or meta‐analysis. The authors conducted a search of the PubMed (MEDLINE), Google Scholar, and Journal Storage (JSTOR) digital libraries using the search terms in Table 1 in the spring of 2019. All PubMed results were screened for inclusion due to its relative density of articles that met inclusion criteria, as well as the first 100 results from Google Scholar and JSTOR. We utilized the Rayyan website and mobile app through the Qatar Computing Research Institute of Hamad Bin Khalifa University for the inclusion/exclusion process, screening conducted based on abstracts and keywords (Ouzzani et al., 2016 ). Once the initial screening process was complete, the authors read the included articles in entirety for analysis. A systematic review was initially designed to emulate the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) methodology. A mixed‐methods analytic approach was anticipated, beginning with qualitative analysis to be followed by quantitative assessment and potential meta‐analysis. However, once the lack of quantitative patient‐centered outcome data in the literature was apparent, direct extraction of data and quantitative assessment to satisfy the PRISMA checklist was not feasible. Thus, the methodology shifted to a predominantly qualitative analysis of the literature, only quantifying proportions of major trends in the literature. RESULTS The authors screened 509 citations from PubMed, and 100 from each Google Scholar and JSTOR, for a total of 709 abstractions and analysis. Only 2% of Google Scholar and 6% of JSTOR results met the inclusion criteria, compared to 31% of those identified via PubMed, resulting in 141 articles included for full analysis. The flowchart of the article collection process is seen in Figure 1 . The geographic distribution of the included articles is presented in Figure 2 , and the distribution of articles by disaster type is presented in Table 2 . Figure 2 depicts a large geographic disparity in the articles, with the bulk of articles arising from North America and Asia. Figure 1 Systematic review flowchart Figure 2 Geographic distribution of articles Table 2 Types of disasters described Type of disaster % of articles Natural disasters 56 Terror/Intentional events 17 Accidents 13 Infectious disease outbreaks 10 War 4 John Wiley & Sons, Ltd. Qualitative assessment of the included articles revealed distinct themes in which each study could be categorized. Half of the articles retrospectively analyzed the shortcomings of a true emergency, categorized as "Primary Report/Study" (74 articles). The other half of the included articles focused on disaster preparation. Their varied approaches can be subcategorized as "Simulation and training" (23), "Formal Protocols" (19), "Supply, Staffing and Capacity" (14), and "Post‐Acute Response" (11). The second half of the articles, which focused on preparation, had an even larger geographic disparity of articles than the total articles that Figure 2 depicts, with 76% of articles originating in North America and 15% from Asia. The distribution of the articles by year is presented in Figure 3 . Over the past two decades, the number of articles peaked in the mid 2000s following the SARS epidemic, Indian Ocean tsunami, and Hurricane Katrina. Our subsequent analysis of the articles that meet inclusion criteria are by no means comprehensive of available work regarding preparation, response, and recovery to disasters. Rather, the themes reflect the trends in the included articles. Figure 3 Distribution of articles by year Disaster response: Primary report/study Approximately half of the included articles (74 articles, 52%) were primary reports or studies that hospitals published following a disaster, assessing their training and response. The authors of these articles had endured various true emergency situations and retrospectively analyzed their experiences in either narrative (58 articles, 78%) or study (16 articles, 22%) format. The majority (63 articles, 85%) of the articles described emergency operations in existing medical infrastructure. The remainder of the articles (11 articles, 15%) described mobile medical operations that were set up in response to a disaster, the majority (8 articles, 71%) of which were led by the United States. The vast majority of the primary reports explicitly mentioned the existence of an emergency action plan that was activated during an emergency (61 articles, 83%) and presented lessons learned after the activation of said plans (58 articles, 78%). These common shortcomings, listed by percent of included articles, are presented in Figure 4 . Figure 4 Common issues in disaster response For these common problems, the papers collectively offered solutions that they planned to implement within their respective systems in preparation for another emergency. For the issue of telecommunication, the authors repeatedly described the massive number of phone calls to the hospital/emergency department, increased radio "traffic," lack of cellphone signal close to the epicenter of emergencies or in health facilities, as well as the inability to contact other leaders within the healthcare facility. Solutions offered were to equip emergency departments and ambulances with satellite radio transmission, to re‐issue pagers or radios for Hospital Incident Command System leadership in the event of an emergency, and to have a second landline number for community partners (G. G. Lavery & Horan, 2005 ). Additionally, many articles described the issue of crowd control in emergency departments, as many types of health emergencies lead to a surge of patients. This surge of patients, volunteers, and family members is exacerbated by the fact that a majority of people self‐present to the emergency department (ED) rather than by ambulance, which leads to a maldistribution of patients across a community's hospitals (Zoraster et al., 2007 ). As a solution, the hospitals sealed ED entrances, triaged outside, limited ED presence to essential staff, and kept additional/volunteer staff out of the ED but close in case they were needed (Lee et al., 2016 ). One article advocated for hospital incident command system leaders to wear vests to clearly denote their role, and even labeling providers with "Airway doctor," "Trauma nurse," "Trauma surgeon," and so on, to avoid confusion in a crowded space (Lee et al., 2016 ). Multiple articles encouraged establishing family assistance centers for next of kin looking for victims, which can be expected on a magnitude of three to five people per patient (Lynn et al., 2006 ). More ED space can be made available by discharging stable patients home earlier than normal from the wards. Some articles described disrupted road infrastructure, during which health staff coordinated with other first responders, such as fire departments or military to create alternative means to transport patients to and from hospitals (Chen et al., 2010 ; Jenkins et al., 2010). In this way, patients admitted from the ED could rapidly be moved up and ED boarding and crowding minimized. With regard to insufficient stockpiling before a foreseeable infrastructure‐disrupting event like a hurricane, hospitals discussed stockpiling water to last 5 days and minimizing the number of people in the hospital beforehand (Bovender & Carey, 2006 ). Additionally, they described how stockpiles of medications in areas surrounding an emergency can be exhausted as people evacuate and seek medications for chronic diseases in the EDs and pharmacies of surrounding regions (Hogue et al., 2009 ). The majority of the articles (61 articles, 83%) mentioned that their facility had an emergency plan that was activated during the emergency. Nonetheless, there was a recurring theme of staff being unaware of such plans or protocols as well as generalized confusion when such plans were activated. Many articles called for more disaster drills with clear direction. One article studied its disaster drills, showing they had no significant impact on regular patient care (Timm & Kennebeck, 2008 ). In drills, it is important to create a clear identification scheme for both hospital incident command system leaders (vests) or patients (paper record/handwritten wrist band with triage status) in the event of the electronic medical record (EMR) systems becoming unavailable (Lee et al., 2016 ). Additionally, clear protocols for staff attendance/leave in the event of a disaster should be established to mitigate staff elopement and subsequent overstressing of the remaining staff (Kodama et al., 2014 ; Laditka et al., 2009 ; Santibañez et al., 2016 ). Multiple articles expressed the need for increased collaboration with other disaster responding organizations, such as local public health agencies, police, EMT/paramedics, fire brigades, and militaries. Some formed a regional emergency response coalition, meeting regularly and forming a regional central command structure in the event of an emergency (Cyganik, 2003 ). A consideration for action would be to create a memorandum of understanding outlining joint planning; all‐level workforce training; and the sharing of supplies, manpower, and data such as patient census in the event of a disaster (Werner et al., 2005 ). Disaster preparation: Simulation and training A contingent of articles (23 articles, 16%) addressed disaster simulation and training in their health systems. The largest proportion of articles that addressed simulation and training assessed the competency of existing training measures in place (eight articles, 33%). Unfortunately, the majority of the assessments were based on qualitative data, with the only quantitative data coming from pre‐ and post‐training assessments that are vulnerable to response shift bias. Another 33% of the articles (eight articles) described actual simulation exercises, with half of them in the clinical setting and half in a tabletop format. The remainder of the articles discussed needs assessments for training (three articles, 13%), systematic reviews of training (two articles, 9%), and descriptions of training programs (two articles, 9%). Many primary reports called for more disaster drilling to increase staff knowledge of disaster protocols, refine emergent decision‐making at various levels in emergency health systems, and foster oral and written communication between such levels. The drills described by the articles in this section—from in‐hospital drills with staff to tabletop exercises with clinical leaders and community partners—were universally seen as helpful from a qualitative standpoint. Analysis of pre‐ and post‐tests universally supported that assessment. However, two systematic reviews attempted to provide more robust evidence for or against the utility of such exercises. They agreed that there is limited quantitative evidence proving the effectiveness of such drilling and that those studies have "significant limitations in design and evaluation methods" (Hsu et al., 2004 ). For this reason, the reviews both concluded that there is insufficient evidence to draw a valid conclusion (Hsu et al., 2004 ; Williams et al., 2008 ). Nonetheless, one noted that the drills are useful to "improve familiarity with disaster procedures, identify problems in different components of response (e.g., incident command, communication, triage, patient flow, materials and resources, and security) and provide the opportunity to apply the lessons learned to disaster response" (Hsu et al., 2004 ). Disaster preparation: Proposed plans Nineteen articles (13%) focused on delineating plans for future disasters. Some advocated for identifying multiple hospitals in a region to be involved in planning (Lynn et al., 2006 ; Potter et al., 2005 ). They emphasized that when planning, it is important to estimate the expected wounded in various types of mass casualty incidents (Lynn et al., 2006 ). One can subsequently use those figures to calculate the maximum number of patients able to be absorbed per hospital, as well as the necessary supplies and staff. Staff needs should be overestimated by 30%, taking into consideration that 30% of called‐in personnel will not be available. Alternative spaces for triage, treatment, and stretcher routes should be considered (Lynn et al., 2006 ). Hospital and regional command centers, clear leadership roles and duties, and postactivation checklist should be outlined. Lastly, staff call‐in/report protocols should be published. Past infectious disease outbreaks such as SARS have shown high rates of nosocomial infection, with 21% of SARS patients being healthcare workers. To address this known issue, an article suggested the following hierarchy: (1) Engineering controls, followed by (2) administrative and work practice controls, and supplemented by (3) personal protective equipment (PPE) (Thorne et al., 2004 ). Engineering controls include temporary structures, outside facilities for screening, and negative pressure rooms. Administrative controls include restricting patient contact, staff/patient movement, control measures for high‐risk procedures, and monitoring staff adherence to isolation procedures. PPE should be seen as a supplement to a more important means of transmission prevention. A portion of the articles (four articles, 21%) discussed individualized departmental task forces that have been formed to take on emergency situations. Such task forces complement the overarching hospital incident command system and lead to individualized response models, increased staff morale, and staff "buy‐in" about the importance of disaster preparedness within individual departments (Zavotsky et al., 2004 ). The task forces corresponded and trained with outside first response organizations, and some even planned to assign staff members to leave the hospital to join such organizations in the field in the event of a mass casualty situation (R. F. Lavery et al., 2000 ). The remaining articles focused on best practices in the care of a specific patient group (three articles, 16%). For example, obese, elderly, and perinatal women have increased health needs and limited mobility. Thus, all should be considered special‐needs and at‐risk patient populations, and all patient plans should consider them (Geiling, 2010 ; Orlando et al., 2010 ). Disaster preparation: Post‐acute response (recovery) The smallest proportion of articles focused on post‐acute phase disaster response (11 articles, 8%), taking varying approaches. The majority of the articles (seven articles, 64%) focused on the issues providers face in the post‐acute/recovery period, when regular health infrastructure remains disrupted. Articles called for a more coordinated international response to disasters, ensuring the qualification of those deployed and prioritizing the autonomy and needs of the host nation's ministry of health (Carballo et al., 2005 ). Other articles examined the issues that patients face in the event of prolonged infrastructure failure, both acute and chronic. For example, when infrastructure fails, patients cannot access their chronic medications or healthy food and present with related complications of their chronic illnesses (Hogue et al., 2009 ). One article found increased rates of dental caries in the months following earthquakes in two countries, Japan and Haiti (Hosokawa et al., 2012 ). Interestingly, rates of caries were particularly high in areas with international responders, which the authors attributed to international teams distributing sugary food and candy, which have a long shelf life, to survivors. Oral hygiene can be a nidus of infection, particularly if elderly patients aspirate, and thus the article suggested distributing toothbrushes/toothpaste, educating patients, involving dentists, and distributing healthy food. The remainder of the articles (four articles, 36%) took aim at the impact the mass casualty incident has on healthcare workers over time. Some discussed medical morbidity, but all discussed psychosocial distress from social distancing‐associated social isolation, increased work, stigmatization of healthcare workers during infectious disease outbreaks, and the fear of contracting the disease and/or spreading it to family. For example, a study during the SARS epidemic found that two‐thirds of the hospital staff reported SARS‐related concerns for their own or their family's health (Nickell et al., 2004 ). More alarmingly, 29% scored above the threshold point on the GHQ‐12, indicating probable emotional distress, social dysfunction, anxiety, or loss of confidence. The rate among nurses was even higher at 45%. Such obstacles threaten staff absenteeism and dysfunction, thus challenging the sustainability of post‐acute disaster response. Therefore, early and aggressive campaigns to boost morale, treat depressive symptoms, and foster social connections are important. A second article suggested educating staff that they are at risk for symptoms of depression/anxiety, normalizing seeking psychotherapy/pharmaceutical treatment, proactively creating phone/video chat networks, and providing financial support for those in quarantine (Johal, 2009 ). Disaster preparation: Supply, staffing, and capacity Fourteen articles (10%) addressed supply, staffing, and capacity concerns in disaster planning. The largest portion of this group (five articles, 36%) discussed surge capacity—the ability of a medical facility to care for the "surge" of new patients in the event of a mass casualty incident. To minimize the surge, papers described methods to facilitate interhospital communication of key data—supplies, manpower, and the number of open hospital beds—in the event of a disaster to distribute supplies and staff to meet patient needs (Tadmor et al., 2006 ). Similarly, health provider systems can work with the media to supply the public with instructive information to deter panic and control patient flow; mitigating unequal burden between hospitals (Tadmor et al., 2006 ). In response to the surge, the articles outlined various means to redistributing patient burden, such as discharging stable patients home early or to subacute/short‐term rehabilitation facilities and using typically nonclinical space for patient care. Some took it a step further, discussing more proactive measures, such as planning with the region's outpatient community health centers, closed hospitals, or retirement homes to take advantage of their space for reserve capabilities (Koh et al., 2006 ; Phillips, 2006 ). Another portion of the articles (four articles, 29%) discussed means of incidence control. For infectious disease or chemical outbreaks, articles advocated for mass vaccination/prophylaxis and PPE for medical providers. They highlighted the lack of and need for improved PPE for first responders such as EMT/paramedics in the event of a chemical/nuclear/biological disaster (Migl & Powell, 2010 ; Phelps, 2007 ). Some of the articles (three articles, 21%) discussed means to secure additional medical supplies, discussing strategic stockpiling and formulas to calculate supply needs to better inform suppliers in the early response. For example, one article formulated the blood needs from the number of casualties in a mass casualty incident (Beekley et al., 2009 ). The remainder of the articles (two articles, 14%) discussed staffing, highlighting systems to credential out of hospital clinician volunteers rapidly and an app that calls/verifies reinforcement without sacrificing critical manpower. For example, one article created an app that circumvents a phone tree and provides the institution with real‐time provider responses and estimated times of arrival (K. Tanaka et al., 2017 ). The other article supported advanced provider credentialing in a region so that hospitals can accept volunteers as needed to assist in post‐disaster care (Schultz & Stratton, 2007 ). Disaster response: Primary report/study Approximately half of the included articles (74 articles, 52%) were primary reports or studies that hospitals published following a disaster, assessing their training and response. The authors of these articles had endured various true emergency situations and retrospectively analyzed their experiences in either narrative (58 articles, 78%) or study (16 articles, 22%) format. The majority (63 articles, 85%) of the articles described emergency operations in existing medical infrastructure. The remainder of the articles (11 articles, 15%) described mobile medical operations that were set up in response to a disaster, the majority (8 articles, 71%) of which were led by the United States. The vast majority of the primary reports explicitly mentioned the existence of an emergency action plan that was activated during an emergency (61 articles, 83%) and presented lessons learned after the activation of said plans (58 articles, 78%). These common shortcomings, listed by percent of included articles, are presented in Figure 4 . Figure 4 Common issues in disaster response For these common problems, the papers collectively offered solutions that they planned to implement within their respective systems in preparation for another emergency. For the issue of telecommunication, the authors repeatedly described the massive number of phone calls to the hospital/emergency department, increased radio "traffic," lack of cellphone signal close to the epicenter of emergencies or in health facilities, as well as the inability to contact other leaders within the healthcare facility. Solutions offered were to equip emergency departments and ambulances with satellite radio transmission, to re‐issue pagers or radios for Hospital Incident Command System leadership in the event of an emergency, and to have a second landline number for community partners (G. G. Lavery & Horan, 2005 ). Additionally, many articles described the issue of crowd control in emergency departments, as many types of health emergencies lead to a surge of patients. This surge of patients, volunteers, and family members is exacerbated by the fact that a majority of people self‐present to the emergency department (ED) rather than by ambulance, which leads to a maldistribution of patients across a community's hospitals (Zoraster et al., 2007 ). As a solution, the hospitals sealed ED entrances, triaged outside, limited ED presence to essential staff, and kept additional/volunteer staff out of the ED but close in case they were needed (Lee et al., 2016 ). One article advocated for hospital incident command system leaders to wear vests to clearly denote their role, and even labeling providers with "Airway doctor," "Trauma nurse," "Trauma surgeon," and so on, to avoid confusion in a crowded space (Lee et al., 2016 ). Multiple articles encouraged establishing family assistance centers for next of kin looking for victims, which can be expected on a magnitude of three to five people per patient (Lynn et al., 2006 ). More ED space can be made available by discharging stable patients home earlier than normal from the wards. Some articles described disrupted road infrastructure, during which health staff coordinated with other first responders, such as fire departments or military to create alternative means to transport patients to and from hospitals (Chen et al., 2010 ; Jenkins et al., 2010). In this way, patients admitted from the ED could rapidly be moved up and ED boarding and crowding minimized. With regard to insufficient stockpiling before a foreseeable infrastructure‐disrupting event like a hurricane, hospitals discussed stockpiling water to last 5 days and minimizing the number of people in the hospital beforehand (Bovender & Carey, 2006 ). Additionally, they described how stockpiles of medications in areas surrounding an emergency can be exhausted as people evacuate and seek medications for chronic diseases in the EDs and pharmacies of surrounding regions (Hogue et al., 2009 ). The majority of the articles (61 articles, 83%) mentioned that their facility had an emergency plan that was activated during the emergency. Nonetheless, there was a recurring theme of staff being unaware of such plans or protocols as well as generalized confusion when such plans were activated. Many articles called for more disaster drills with clear direction. One article studied its disaster drills, showing they had no significant impact on regular patient care (Timm & Kennebeck, 2008 ). In drills, it is important to create a clear identification scheme for both hospital incident command system leaders (vests) or patients (paper record/handwritten wrist band with triage status) in the event of the electronic medical record (EMR) systems becoming unavailable (Lee et al., 2016 ). Additionally, clear protocols for staff attendance/leave in the event of a disaster should be established to mitigate staff elopement and subsequent overstressing of the remaining staff (Kodama et al., 2014 ; Laditka et al., 2009 ; Santibañez et al., 2016 ). Multiple articles expressed the need for increased collaboration with other disaster responding organizations, such as local public health agencies, police, EMT/paramedics, fire brigades, and militaries. Some formed a regional emergency response coalition, meeting regularly and forming a regional central command structure in the event of an emergency (Cyganik, 2003 ). A consideration for action would be to create a memorandum of understanding outlining joint planning; all‐level workforce training; and the sharing of supplies, manpower, and data such as patient census in the event of a disaster (Werner et al., 2005 ). Disaster preparation: Simulation and training A contingent of articles (23 articles, 16%) addressed disaster simulation and training in their health systems. The largest proportion of articles that addressed simulation and training assessed the competency of existing training measures in place (eight articles, 33%). Unfortunately, the majority of the assessments were based on qualitative data, with the only quantitative data coming from pre‐ and post‐training assessments that are vulnerable to response shift bias. Another 33% of the articles (eight articles) described actual simulation exercises, with half of them in the clinical setting and half in a tabletop format. The remainder of the articles discussed needs assessments for training (three articles, 13%), systematic reviews of training (two articles, 9%), and descriptions of training programs (two articles, 9%). Many primary reports called for more disaster drilling to increase staff knowledge of disaster protocols, refine emergent decision‐making at various levels in emergency health systems, and foster oral and written communication between such levels. The drills described by the articles in this section—from in‐hospital drills with staff to tabletop exercises with clinical leaders and community partners—were universally seen as helpful from a qualitative standpoint. Analysis of pre‐ and post‐tests universally supported that assessment. However, two systematic reviews attempted to provide more robust evidence for or against the utility of such exercises. They agreed that there is limited quantitative evidence proving the effectiveness of such drilling and that those studies have "significant limitations in design and evaluation methods" (Hsu et al., 2004 ). For this reason, the reviews both concluded that there is insufficient evidence to draw a valid conclusion (Hsu et al., 2004 ; Williams et al., 2008 ). Nonetheless, one noted that the drills are useful to "improve familiarity with disaster procedures, identify problems in different components of response (e.g., incident command, communication, triage, patient flow, materials and resources, and security) and provide the opportunity to apply the lessons learned to disaster response" (Hsu et al., 2004 ). Disaster preparation: Proposed plans Nineteen articles (13%) focused on delineating plans for future disasters. Some advocated for identifying multiple hospitals in a region to be involved in planning (Lynn et al., 2006 ; Potter et al., 2005 ). They emphasized that when planning, it is important to estimate the expected wounded in various types of mass casualty incidents (Lynn et al., 2006 ). One can subsequently use those figures to calculate the maximum number of patients able to be absorbed per hospital, as well as the necessary supplies and staff. Staff needs should be overestimated by 30%, taking into consideration that 30% of called‐in personnel will not be available. Alternative spaces for triage, treatment, and stretcher routes should be considered (Lynn et al., 2006 ). Hospital and regional command centers, clear leadership roles and duties, and postactivation checklist should be outlined. Lastly, staff call‐in/report protocols should be published. Past infectious disease outbreaks such as SARS have shown high rates of nosocomial infection, with 21% of SARS patients being healthcare workers. To address this known issue, an article suggested the following hierarchy: (1) Engineering controls, followed by (2) administrative and work practice controls, and supplemented by (3) personal protective equipment (PPE) (Thorne et al., 2004 ). Engineering controls include temporary structures, outside facilities for screening, and negative pressure rooms. Administrative controls include restricting patient contact, staff/patient movement, control measures for high‐risk procedures, and monitoring staff adherence to isolation procedures. PPE should be seen as a supplement to a more important means of transmission prevention. A portion of the articles (four articles, 21%) discussed individualized departmental task forces that have been formed to take on emergency situations. Such task forces complement the overarching hospital incident command system and lead to individualized response models, increased staff morale, and staff "buy‐in" about the importance of disaster preparedness within individual departments (Zavotsky et al., 2004 ). The task forces corresponded and trained with outside first response organizations, and some even planned to assign staff members to leave the hospital to join such organizations in the field in the event of a mass casualty situation (R. F. Lavery et al., 2000 ). The remaining articles focused on best practices in the care of a specific patient group (three articles, 16%). For example, obese, elderly, and perinatal women have increased health needs and limited mobility. Thus, all should be considered special‐needs and at‐risk patient populations, and all patient plans should consider them (Geiling, 2010 ; Orlando et al., 2010 ). Disaster preparation: Post‐acute response (recovery) The smallest proportion of articles focused on post‐acute phase disaster response (11 articles, 8%), taking varying approaches. The majority of the articles (seven articles, 64%) focused on the issues providers face in the post‐acute/recovery period, when regular health infrastructure remains disrupted. Articles called for a more coordinated international response to disasters, ensuring the qualification of those deployed and prioritizing the autonomy and needs of the host nation's ministry of health (Carballo et al., 2005 ). Other articles examined the issues that patients face in the event of prolonged infrastructure failure, both acute and chronic. For example, when infrastructure fails, patients cannot access their chronic medications or healthy food and present with related complications of their chronic illnesses (Hogue et al., 2009 ). One article found increased rates of dental caries in the months following earthquakes in two countries, Japan and Haiti (Hosokawa et al., 2012 ). Interestingly, rates of caries were particularly high in areas with international responders, which the authors attributed to international teams distributing sugary food and candy, which have a long shelf life, to survivors. Oral hygiene can be a nidus of infection, particularly if elderly patients aspirate, and thus the article suggested distributing toothbrushes/toothpaste, educating patients, involving dentists, and distributing healthy food. The remainder of the articles (four articles, 36%) took aim at the impact the mass casualty incident has on healthcare workers over time. Some discussed medical morbidity, but all discussed psychosocial distress from social distancing‐associated social isolation, increased work, stigmatization of healthcare workers during infectious disease outbreaks, and the fear of contracting the disease and/or spreading it to family. For example, a study during the SARS epidemic found that two‐thirds of the hospital staff reported SARS‐related concerns for their own or their family's health (Nickell et al., 2004 ). More alarmingly, 29% scored above the threshold point on the GHQ‐12, indicating probable emotional distress, social dysfunction, anxiety, or loss of confidence. The rate among nurses was even higher at 45%. Such obstacles threaten staff absenteeism and dysfunction, thus challenging the sustainability of post‐acute disaster response. Therefore, early and aggressive campaigns to boost morale, treat depressive symptoms, and foster social connections are important. A second article suggested educating staff that they are at risk for symptoms of depression/anxiety, normalizing seeking psychotherapy/pharmaceutical treatment, proactively creating phone/video chat networks, and providing financial support for those in quarantine (Johal, 2009 ). Disaster preparation: Supply, staffing, and capacity Fourteen articles (10%) addressed supply, staffing, and capacity concerns in disaster planning. The largest portion of this group (five articles, 36%) discussed surge capacity—the ability of a medical facility to care for the "surge" of new patients in the event of a mass casualty incident. To minimize the surge, papers described methods to facilitate interhospital communication of key data—supplies, manpower, and the number of open hospital beds—in the event of a disaster to distribute supplies and staff to meet patient needs (Tadmor et al., 2006 ). Similarly, health provider systems can work with the media to supply the public with instructive information to deter panic and control patient flow; mitigating unequal burden between hospitals (Tadmor et al., 2006 ). In response to the surge, the articles outlined various means to redistributing patient burden, such as discharging stable patients home early or to subacute/short‐term rehabilitation facilities and using typically nonclinical space for patient care. Some took it a step further, discussing more proactive measures, such as planning with the region's outpatient community health centers, closed hospitals, or retirement homes to take advantage of their space for reserve capabilities (Koh et al., 2006 ; Phillips, 2006 ). Another portion of the articles (four articles, 29%) discussed means of incidence control. For infectious disease or chemical outbreaks, articles advocated for mass vaccination/prophylaxis and PPE for medical providers. They highlighted the lack of and need for improved PPE for first responders such as EMT/paramedics in the event of a chemical/nuclear/biological disaster (Migl & Powell, 2010 ; Phelps, 2007 ). Some of the articles (three articles, 21%) discussed means to secure additional medical supplies, discussing strategic stockpiling and formulas to calculate supply needs to better inform suppliers in the early response. For example, one article formulated the blood needs from the number of casualties in a mass casualty incident (Beekley et al., 2009 ). The remainder of the articles (two articles, 14%) discussed staffing, highlighting systems to credential out of hospital clinician volunteers rapidly and an app that calls/verifies reinforcement without sacrificing critical manpower. For example, one article created an app that circumvents a phone tree and provides the institution with real‐time provider responses and estimated times of arrival (K. Tanaka et al., 2017 ). The other article supported advanced provider credentialing in a region so that hospitals can accept volunteers as needed to assist in post‐disaster care (Schultz & Stratton, 2007 ). DISCUSSION Disaster preparation It is clear that healthcare provider systems around the world differ in their approach to disaster preparation, targeting an array of simulation and training, formal protocols, supplies, staffing, and capacity. Hospital incident command system leaders can consider implementing a hierarchy of engineering controls, administrative/work practice controls, and PPE with the needs of all staff and at‐risk patients kept in mind. Additionally, hospital system leadership can consider creating stockpiles that are intended to last at least five days, establishing memorandums of understanding between hospitals/departments and local public health agencies, writing evacuation protocols, defining staff attendance protocols during disasters, training of staff outside of the Emergency Department (Accident and Emergency) to assist during disasters, and including all levels of providers in the protocol‐making process. In terms of the literature, which is saturated with needs assessments, further research is needed on disaster preparation. A specific contribution to the literature would be to examine the effectiveness of preparation activities such as disaster drills, computerized scenarios, simulation, and tabletop exercises. Hospitals/departments could more rigorously evaluate their drill efforts, monitoring their training against the following outcomes measures: Time to first provider contact, time to laboratory/radiological study, time to diagnosis, time to treatment, length of stay in department/hospital, rates of over‐triage and under‐triage, length of time until backup staff called/arrived, appropriateness of decisions made by hospital incident command, and so forth. One systematic review by Williams et al. ( 2008 ) called for studies with better scientific rigor and objective measures, ideally randomized controlled trials involving control groups of untrained individuals. Though it seems counterintuitive to abstain from training a portion of the workforce, they argue, "control groups could receive training once the studied intervention is shown to be effective." Through studies of greater rigor, preparation work can be more targeted, evidence‐based, and efficient use of resources. Disaster response As other hospital systems, departments, and providers plan their disaster response disaster, they can learn from their colleagues' lessons learned in Figure 4 to prevent similar mistakes in the future. Important considerations for future disaster response would be for providers to incorporate satellite/radio communication when telecommunications are disrupted, initiate crowd control and triage outside of the ED, institute clear demarcation of hospital incident command system leadership, and organize volunteer manpower from outside departments/institutions. Additionally, more quantitative data is needed in the literature. This study was initially designed with the intention of extracting quantitative outcome data from articles to present a robust argument of which disaster preparation interventions work, but rarely do articles publish quantitative data on disaster response interventions. Even fewer examine patient‐centered outcomes, such as mortality data (12 articles, 9%). This lack of outcome data is a major limitation of the individual articles examined, as well as this review article, which was based on qualitative trends. Thus, we encourage the collection and presentation of quantitative measurement in clinical disaster response. Aside from mortality and morbidity, it would be important to include the following measures: Number of patients over time (surge), number of patients by chief complaint, average length of stay, hospital capacity, bed occupancy over time, days of lost work, days of lost services. Systems could also use published Utstein‐style templates for publishing uniform data following acute response (Debacker et al., 2012 ). Through comparison of measurable outcomes, future recommendations can be evidence‐based. Disaster sustainability and recovery In terms of disaster sustainability and recovery, hospital systems should plan for the repercussions of infrastructure failure following the acute disaster, such as chronic disease exacerbation and new acute infection. Hospital system leadership can consider incorporating technology, creating regional provider partnerships, and utilizing the media to equilibrate the supply and staffing strain in a region during a disaster. Additionally, as medical staff are vulnerable to psychological strain, particularly during infectious disease outbreaks, their psychological needs should be addressed early and aggressively. As relatively few articles in the literature focused on disaster recovery, future contributions about this phase are warranted. Limitations Quantitative assessment and meta‐analysis using PRISMA guidelines was unable to be pursued as anticipated due to the subjective nature of current literature. The qualitative analysis that ensued is limited by its dependence upon investigator designation of identified themes. The authors attempted to mitigate potential resulting bias by quantifying proportions of major trends in the literature—first the proportions of articles responding to disasters and those planning for them and subsequently quantifying both common lessons learned and different areas of disaster planning focus. The available data was contextual to the events with which they were associated, and conclusions not prospectively tested. Consequently, the developments and lessons learned that were common in the articles are not necessarily generalizable. For example, the feasibility of stockpiling is limited by storage capacity, the expense of storage, obsolescence, support of local or national governments, and so forth. Disaster preparation It is clear that healthcare provider systems around the world differ in their approach to disaster preparation, targeting an array of simulation and training, formal protocols, supplies, staffing, and capacity. Hospital incident command system leaders can consider implementing a hierarchy of engineering controls, administrative/work practice controls, and PPE with the needs of all staff and at‐risk patients kept in mind. Additionally, hospital system leadership can consider creating stockpiles that are intended to last at least five days, establishing memorandums of understanding between hospitals/departments and local public health agencies, writing evacuation protocols, defining staff attendance protocols during disasters, training of staff outside of the Emergency Department (Accident and Emergency) to assist during disasters, and including all levels of providers in the protocol‐making process. In terms of the literature, which is saturated with needs assessments, further research is needed on disaster preparation. A specific contribution to the literature would be to examine the effectiveness of preparation activities such as disaster drills, computerized scenarios, simulation, and tabletop exercises. Hospitals/departments could more rigorously evaluate their drill efforts, monitoring their training against the following outcomes measures: Time to first provider contact, time to laboratory/radiological study, time to diagnosis, time to treatment, length of stay in department/hospital, rates of over‐triage and under‐triage, length of time until backup staff called/arrived, appropriateness of decisions made by hospital incident command, and so forth. One systematic review by Williams et al. ( 2008 ) called for studies with better scientific rigor and objective measures, ideally randomized controlled trials involving control groups of untrained individuals. Though it seems counterintuitive to abstain from training a portion of the workforce, they argue, "control groups could receive training once the studied intervention is shown to be effective." Through studies of greater rigor, preparation work can be more targeted, evidence‐based, and efficient use of resources. Disaster response As other hospital systems, departments, and providers plan their disaster response disaster, they can learn from their colleagues' lessons learned in Figure 4 to prevent similar mistakes in the future. Important considerations for future disaster response would be for providers to incorporate satellite/radio communication when telecommunications are disrupted, initiate crowd control and triage outside of the ED, institute clear demarcation of hospital incident command system leadership, and organize volunteer manpower from outside departments/institutions. Additionally, more quantitative data is needed in the literature. This study was initially designed with the intention of extracting quantitative outcome data from articles to present a robust argument of which disaster preparation interventions work, but rarely do articles publish quantitative data on disaster response interventions. Even fewer examine patient‐centered outcomes, such as mortality data (12 articles, 9%). This lack of outcome data is a major limitation of the individual articles examined, as well as this review article, which was based on qualitative trends. Thus, we encourage the collection and presentation of quantitative measurement in clinical disaster response. Aside from mortality and morbidity, it would be important to include the following measures: Number of patients over time (surge), number of patients by chief complaint, average length of stay, hospital capacity, bed occupancy over time, days of lost work, days of lost services. Systems could also use published Utstein‐style templates for publishing uniform data following acute response (Debacker et al., 2012 ). Through comparison of measurable outcomes, future recommendations can be evidence‐based. Disaster sustainability and recovery In terms of disaster sustainability and recovery, hospital systems should plan for the repercussions of infrastructure failure following the acute disaster, such as chronic disease exacerbation and new acute infection. Hospital system leadership can consider incorporating technology, creating regional provider partnerships, and utilizing the media to equilibrate the supply and staffing strain in a region during a disaster. Additionally, as medical staff are vulnerable to psychological strain, particularly during infectious disease outbreaks, their psychological needs should be addressed early and aggressively. As relatively few articles in the literature focused on disaster recovery, future contributions about this phase are warranted. Limitations Quantitative assessment and meta‐analysis using PRISMA guidelines was unable to be pursued as anticipated due to the subjective nature of current literature. The qualitative analysis that ensued is limited by its dependence upon investigator designation of identified themes. The authors attempted to mitigate potential resulting bias by quantifying proportions of major trends in the literature—first the proportions of articles responding to disasters and those planning for them and subsequently quantifying both common lessons learned and different areas of disaster planning focus. The available data was contextual to the events with which they were associated, and conclusions not prospectively tested. Consequently, the developments and lessons learned that were common in the articles are not necessarily generalizable. For example, the feasibility of stockpiling is limited by storage capacity, the expense of storage, obsolescence, support of local or national governments, and so forth. CONCLUSIONS In summary, our review found the following: The literature is saturated with needs assessments for disaster training and primary reports following disasters. More research addressing innovation in hospital disaster simulation, protocols, post‐acute response and supply, and staffing is warranted. More literature is needed documenting the successes and failures of medical provider disaster response in Europe, S. America, and Africa. The majority of articles were reactive, rather than proactive. Expansion of articles delineating efficacy‐tested disaster preparation work is warranted for future evidence‐based practice. Relatively few articles have been written on infectious disease outbreaks in the context of provider and hospital system emergency response. More quantitative research is needed, particularly showing mortality and morbidity outcomes from disaster interventions at the provider and hospital system level. Common shortcomings in hospital disaster plans are telecommunication failure, plans for surge capacity, insufficient stockpiles, coordination with other hospitals/disaster responders, and staff knowledge of disaster protocols. Hospital Incident Command Systems can consider the hierarchy of controls including engineering controls and administrative/work practices when designing disaster protocols controls. Infection prevention and control needs, and those controls that apply to all staff and particularly at‐risk patients need special attention. Hospital system leadership should plan for the repercussions of disruptions in usual care during the disaster, such as chronic disease exacerbation and new acute infections during recovery. Disaster plans can incorporate technology, regional provider partnerships, and the media to equilibrate the supply and staffing strain in a region during a disaster. Healthcare providers and staff are vulnerable to psychological strain, particularly during infectious disease outbreaks, and this should be addressed early and aggressively. As hospitals learn from the COVID‐19 experience as well as other future disasters, they will hopefully start filling in these blanks with more targeted and actionable contributions to the literature. CONFLICT OF INTERESTS The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. ETHICS STATEMENT The authors declare that human ethics approval was not needed for this study.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4482429/

A Computational, Tissue-Realistic Model of Pressure Ulcer Formation in Individuals with Spinal Cord Injury

People with spinal cord injury (SCI) are predisposed to pressure ulcers (PU). PU remain a significant burden in cost of care and quality of life despite improved mechanistic understanding and advanced interventions. An agent-based model (ABM) of ischemia/reperfusion-induced inflammation and PU (the PUABM) was created, calibrated to serial images of post-SCI PU, and used to investigate potential treatments in silico . Tissue-level features of the PUABM recapitulated visual patterns of ulcer formation in individuals with SCI. These morphological features, along with simulated cell counts and mediator concentrations, suggested that the influence of inflammatory dynamics caused simulations to be committed to "better" vs. "worse" outcomes by 4 days of simulated time and prior to ulcer formation. Sensitivity analysis of model parameters suggested that increasing oxygen availability would reduce PU incidence. Using the PUABM, in silico trials of anti-inflammatory treatments such as corticosteroids and a neutralizing antibody targeted at Damage-Associated Molecular Pattern molecules (DAMPs) suggested that, at best, early application at a sufficiently high dose could attenuate local inflammation and reduce pressure-associated tissue damage, but could not reduce PU incidence. The PUABM thus shows promise as an adjunct for mechanistic understanding, diagnosis, and design of therapies in the setting of PU. Introduction Pressure ulcers (PU) affect 2.5 million US acute care patients and cost up to $1 billion per year [ 1 ]. They are a significant source of morbidity in both hospitalized patients and community-dwelling individuals with impaired mobility. PUs are especially common in individuals with spinal cord injury (SCI), occurring in up to 80% of this population at some point during their lifetime [ 2 ]. Spinal cord injury is a condition associated with decreased functional mobility, acutely increased oxidative activity in leukocytes, and chronic elevation of systemic inflammatory markers [ 3 – 5 ]. Pressure ulcers are thought to arise from pressure-induced ischemia, reperfusion injury, and/or deformation-induced cellular damage [ 6 ]. The pathogenesis of PU involves activation of the acute inflammatory response [ 7 , 8 ], a highly conserved cascade of events mediated by a set of specialized cells (e.g. platelets, mast cells, macrophages, and neutrophils) and molecules (inflammatory cytokines, free radicals, and Damage-Associated Molecular Pattern molecules [DAMPs]) that demarcate stressed or damaged tissue, and alert and recruit other cells and molecules. The inflammatory response can either restore the tissue to equilibrium (healing) and resolve, or become self-maintaining inflammation that causes and is caused by ancillary tissue damage. This excessive inflammation prevents the body from initiating wound healing and can lead to PU incidence [ 9 ]. The dynamic interplay of the inflammatory and healing cascades determines the ultimate success or failure of the healing process. These intracellular signaling networks and their products, including diffusible molecular mediators, are possible targets for diagnosis or therapeutic intervention. However, the complexity of the process as a whole, and the dependence of any given pathway or mediator on timing and context, complicates such translational approaches. The "translational dilemma" centers on the inability of traditional, reductionist approaches to yield better diagnostics and novel drug targets for complex diseases [ 10 – 12 ]. Despite increased understanding of the underlying mechanisms and improved clinical vigilance, PUs remain a prevalent problem in hospitalized patients and people with chronic conditions such as diabetes and SCI [ 13 – 15 ]. While wound healing is well studied in animal systems [ 16 – 18 ], these animal models generally do not recapitulate the complex etiology of impaired wound healing such as that which occurs in PU. Only recently have experimental methodologies emerged that may allow for the study of the time courses of wound healing in humans [ 17 ], but these approaches are limited in that time courses of primary samples from humans with chronic wounds are difficult to collect without disturbing the very process being measured. As an alternative diagnostic approach, digital photographs of developing ulcers are, in theory, both plentiful and non-invasive—unlike wound biopsies. Given the dire need for new therapeutic avenues for complex diseases, a platform combining in silico approaches with easily- and inexpensively-obtained clinical samples (such as photographic images), may yield novel diagnostic and therapeutic modalities. We hypothesize that wound images could be used to calibrate mechanistic simulations of inflammation and healing that could then be interrogated to predict how and when a small irritation might resolve or progress to become a chronic ulcer. Agent-Based Models (ABMs) allow for the investigation of both space- and time-dependent dynamics of complex systems via mechanistic simulations. The modeler provides behavioral rules that allow the model to proceed stepwise through discrete space and time. Unlike differential equation models, ABM simulations are stochastic and thereby able to replicate the randomness of biological processes. Furthermore, ABMs produce visual outputs whose morphological features evolve throughout a simulation and provide a rich set of spatio-temporal data that can be leveraged to probe underlying dynamics [ 19 ]. We report herein on the creation of an ABM of post-SCI PU formation (the Pressure Ulcer Agent Based Model, or PUABM) that incorporates key inflammation mechanisms. Explicitly included is the forward feedback loop of inflammation to damage to inflammation that has served as the core motif of our prior simulations of inflammation in both systemic and local contexts [ 9 , 20 , 21 ]. The PUABM replicates visual morphology associated with the development and resolution of post-SCI PU by simulating vascularized soft tissue overlaying a bony prominence (the clinically recognized "pressure points" at which PU typically develop) and the effects of repeated ischemia/reperfusion (representing the turning of a person with SCI in bed) on such an area of tissue. Also recapitulating clinical outcomes, the model reaches two distinct endpoints when simulated from the same initial parameters. We leverage thousands of model simulations to explore the root of this phenomenon and predict at what time an ulcer's fate is determined. We also demonstrate the utility of the PUABM as a platform for in silico clinical trials of strategies for prevention and therapies for treatment of PU post-SCI. This model suggests that the reason treatments thus far have been ineffective is that they have been applied too late. Materials and Methods Ethics statement This study was approved by the University of Pittsburgh Institutional Review Board, approval number PRO08010011. Written informed consent was obtained from subjects participating in the study. If the subject was unable to provide written consent due to medical condition, proxy consent was obtained from the subject's authorized representative, and the subject provided written consent for continued research participation as soon as they were able to do so. Clinical data collection Photographic data were obtained from pressure ulcers sustained by patients enrolled in the Rehabilitation Engineering Research Center on Spinal Cord Injury at the University of Pittsburgh. Subjects were recruited from a single tertiary care center if they were 18 years of age or older and had sustained a traumatic spinal cord injury. Subjects were excluded if they had a history of pre-existing diseases that affect the inflammatory response to SCI (e.g. autoimmune disease, demyelinating diseases) or a history of previous SCI or other neurological diseases affecting motor and sensory function. Pressure ulcers were initially photographed when identified, and serial photographs were taken three times per week while patients remained in the acute care and/or weekly while in the inpatient rehabilitation hospital to monitor progression and/or healing, and during their outpatient visits or 6 month intervals after discharge. Photos were obtained using a Canon Power Shot SD 750 camera with 3X optical zoom lens with flash enabled, at a distance of 12 inches from the ulcer, resolution of 2048 x 1536 pixels. Pressure ulcer site, severity (stage 1–4, unstageable, or deep tissue injury), general appearance, size and shape were also recorded. Work flow The PUABM was built using an iterative approach ( Fig 1A ). Hereafter, when we use words or phrases that usually specify biological reality, such as "pressure," "neutrophil," "TNF-α," "inflammatory mediators," "cell," "oxygen," "ischemic," etc. we are referring to PUABM components or generated phenomena unless clearly stated otherwise. A simpler model of pressure ulcer formation [ 22 ] was altered to increase mechanistic detail and create clinically relevant model output. Rational improvements were based on domain knowledge and data from the literature. First, the area of tissue simulated in the field was extended. Whereas in the previous version of the model pressure was applied evenly across the entire field, in this version, it was applied maximally (value determined by the parameter, pressure-intensity ) to a circular area in the center of the field and decreasing radially outward. This allowed representation of pressure over a bony prominence and the surrounding tissue, which also experiences pressure, but to a lessening degree [ 23 , 24 ]. Next, the PUABM code was altered so that neutrophils and macrophages enter the tissue in a resting state and can be activated by one of two circulating mediators (TNF-α or TGF-β1). In the previous version of the model, all entering leukocytes were in the active state. After this modification, the activation state of neutrophils and monocytes/ macrophages was determined by the local concentrations of inflammatory mediators. These thresholds are explained further in the section below entitled, "Rules: Tissue Damage." 10.1371/journal.pcbi.1004309.g001 Fig 1 Work flow and schematic of the PUABM. ( A ) Work Flow utilized in constructing the PUABM. The ABM was built iteratively, incorporating domain knowledge and data from the literature, then validated with clinical data. Initially, pressure directly injured tissue, thereby inciting inflammation. Ischemia/reperfusion injury was next added as a cause of injury, and the complexity of the inflammatory response was increased. Clinical data were next used to calibrate parameters in the model, and the model was subjected to sensitivity analysis and in silico trials. ( B ) A schematic illustrates how model components interact to simulate two mechanisms of injury: ischemia/reperfusion injury and damage due to inflammation. A tissue cell (muscle, fat, or skin) is situated over a bony prominence. When pressure is applied, oxygen supply is reduced and the cell becomes ischemic, leading to tissue damage. Simultaneously, the enzyme xanthine dehydrogenase is converted into xanthine oxidase. Thus, reactive oxygen species (ROS) are produced when pressure is released and oxygen flows back to the cell, causing further damage. Tissue damage causes the cell to release DAMPs, which, along with local concentrations of pro-inflammatory cytokines activate both neutrophils (N) and macrophages (M1, M2) when these mediators are present above a given threshold parameter. For example, a sufficiently high local concentration of DAMPs activates neutrophils to secrete TNF-α, which can activate macrophages to a pro- inflammatory M1 phenotype. Tissue damage is ameliorated by anti-inflammatory mediators [TGF-β1]. In the previous version of the ABM, all cells and mediators appeared in a single visualization window. Epithelial cells ranged in color from green to red, depending on their level of health. Other cells were rendered in layers over the grid of epithelial cells[ 22 , 25 ]. In the PUABM presented herein, circulating inflammatory cells and mediators were each removed to a separate viewing window, allowing for comparison of spatio-temporal pattern of each individual component with any other. In the main window, stationary epithelial cells remained, but the colors indicating tissue health were altered to increase the realism: healthy tissue was now rendered as peach, mimicking the most common skin tones in the clinical cohort. Unhealthy tissue continued to be rendered as red, and cells that died disappeared from the grid, leaving behind a white empty space. Thus, each component of the model was represented according to the rules in the ABM, but by separating the viewing windows, we were able to view and compare spatial details of individual components (damage, mediators, etc.). Furthermore, we were able to compare epithelial damage patterns between simulations and clinical images, as described in later sections. Mechanistic details in the PUABM were augmented by incorporating damage to tissue via ischemic injury and its counterpart, reperfusion injury. Damage directly resulting from pressure (as previously simulated [ 22 , 25 ]) was removed and tissue health was penalized for tissue cells receiving inadequate oxygen levels. (In the previous version, oxygen had positive effects on tissue health, but lack of oxygen was not damaging [ 22 ]). A mechanism based on the conversion of xanthine dehydrogenase to xanthine oxidase during ischemia [ 26 , 27 ] was also implemented. The accumulation of xanthine oxidase in ischemic cells represents the potential of a cell to experience reperfusion injury, due to formation of ROS when oxygen reperfuses [ 26 ]. The complexity of the inflammatory response was increased by creating two subpopulations of activated macrophages, one with pro-inflammatory (M1 macrophages) and another with anti-inflammatory (M2 macrophages) phenotype [ 28 , 29 ]. A fourth mediator was also added. Representing a canonical later-acting pro-inflammatory mediator, it is released by pro-inflammatory macrophages and labeled IL-1β. After the mechanisms comprising the PUABM were set, model behaviors were explored over wide ranges of parameter values and default values were tweaked into ranges producing behavior that qualitatively matched the clinical data. Finally, we further investigated dynamics encoded in the model by simulating hypothetical and existing therapies. Corticosteroids were incorporated as a data layer that could be introduced to the tissue via blood vessels (as in an intravenous injection), while antibodies to DAMPs were simulated as a topical cream applied to the entire field. The dose and timing of administrations of each were varied to assess the viability of these treatment options. Model implementation The PUABM was built using SPARK, a platform designed for agent-based modeling of biological systems, freely available for download at: www.pitt.edu/~cirm/spark [ 25 , 30 , 31 ]. SPARK models are written in a logo-like language called SPARK-PL and run on a java platform [ 32 ]. SPARK models contain Agents—autonomous entities that interact with each other and the environment, called Space—and Data Layers, corresponding to individual species in the environment that can diffuse, etc. The behavior of the model is determined by rules that govern how and when agents interact and react. These rules are generally written to be interpreted by one agent at a time, and therefore are necessarily restricted in scope (both time and space). A rule specifies how much an agent should move, produce, change an internal variable, etc. when it encounters a certain amount or type of data layer or agent in its immediate neighborhood. Model rules can be probabilistic in nature, allowing the model to evolve in a stochastic manner. Therefore, the behaviors and patterns produced by simulating several ticks (time steps) of the model in succession arise as emergent phenomena resulting from the accumulated actions of a population of agents over time. SPARK has several built-in standard methods that allow for convenient coding of common biological processes. For example, diffusion is encoded by the function diffuse , which implements a simple discrete approximation in which each data layer cell shares a given percentage of its value with its eight neighbors. Examples of other methods that were used in this model include wiggle and jump to approximate undirected random movement and sniff , for chemotaxis. See pseudocode ( S1 Text ) for all methods used. The components of the PUABM are described in Table 1 , and the operant biological mechanisms and tissue structures are shown graphically in Fig 1B . In the PUABM, simulation time is linked to actual time using the lifespans of cellular agents. A simulated macrophage lifespan ranges from 100–150 model time steps (ticks), corresponding to 100–150 hours (4–6 days) of real time. Neutrophil lifespans range from 10–20 ticks (hours) in the tissue, but increase when neutrophils become activated. Each simulation takes roughly five minutes to compute on a supercomputing node containing 32 process cores, each running at 4.7 GHz, making it possible to complete approximately 1000 simulations per day. The model contains 68 numerical parameters that are set by the modeler and 11 random variables (whose values are drawn from a corresponding uniform distribution when necessary during the course of a simulation). At any given tick, the maximum number of agents computed in the model is on the order of 10 5 . 10.1371/journal.pcbi.1004309.t001 Table 1 Summary of agents, data layers, and their interactions encoded in the ABM. Substances in ABM Cell Sources in ABM Biological Functions in ABM DAMPs secreted by damaged Tissue cells Chemo-attract and Activate Neutrophils TNF-α secreted by Activated Neutrophils Chemo-attract resting and M1 Macrophages, activate Macrophages to M1 state IL-1β secreted by M1 Macrophages Chemo-attract resting and M1 Macrophages, activate Macrophages to M1 state, at high levels activate Macrophages to M2 state TGF-β1 secreted by M2 Macrophages Chemo-attract resting and M2 Macrophages, activate Macrophages to M2 state ROS secreted by Tissue Cells on Reperfusion, by Neutrophils on death Injure Tissue Cells Oxidase produced inside Tissue Cells during Ischemia React with Oxygen to produce ROS Oxygen released by Blood Vessels Necessary for Tissue Cell health, react with Oxidase to produce ROS Substances are all encoded as Data Layers; Cells are all encoded as Agents, each with its own type. See Fig 1B for schematic. Rules: Homeostasis/baseline architecture The pseudocode in S1 Text details all model rules. In brief, the architecture of the PUABM is abstracted from healthy tissue. The PUABM consists of a layer of tissue cells fed by blood vessels throughout the tissue that carry oxygen and inflammatory cells. Instead of representing red blood cells explicitly, their presence is implicit since oxygen flows unaccompanied through blood vessels. Inflammatory cells arrive to the tissue in their native resting state and are not activated unless danger signals are released from damaged tissue. Resting inflammatory cells move randomly, but can be chemo-attracted to local mediators via the SPARK function, sniff . The agents in the PUABM represent either cell types or cellular structures: neutrophils, macrophages, tissue cells, and blood vessels. Data layers are employed to represent mediators in a computationally efficient manner (e.g., diffusible cytokines, free radicals, oxygen, xanthine oxidase, and exogenously-administered drugs). As mentioned above, each tick (time step) of the model represents one hour of real time. On each tick, the agents behave according to rules specific to their type: they produce and consume oxygen and mediators, are chemo-attracted to and activated (undergo a state change) by mediators; and they die, according to their lifespans. Mediators, represented in data layers, can be consumed or produced by individual agents or in reactions with other mediators. They undergo diffusion and degradation at each tick. To mimic natural patient-to-patient and cell-to-cell variability, some pseudo-random variables were built into the model. These values were drawn as necessary as simulations progressed, using the Mersenne twister (as implemented in the Colt library, http://acs.lbl.gov/ACSSoftware/colt/readme.html ). Some examples of random variables are size of each blood vessel agent (the blood vessels are all sized within a range but not exactly the same), and probabilities governing state changes of macrophages and neutrophils (a threshold must be met for the agent to be eligible for state change, but the random variable decides whether the transition actually happens). The function of the thresholds was to partition the interval (0,1). Accordingly, a uniform distribution was the basis of pseudo-random number generation. Similarly, blood vessels were assumed to be sized uniformly across a narrow range: the largest was less than twice the size of the smallest. Rules: Pressure on/off The turning of people with SCI while they are lying in bed was simulated as alternating cycles of pressure (on/off). In the PUABM, pressure is simulated by constriction of blood vessels, decreasing the amount of material that can flow through them. Without oxygen, tissue cell health in the PUABM begins to decline. Ischemic tissue cells increase xanthine oxidase enzyme activity, which represent the capacity of the cell to produce damaging free radicals upon reintroduction of oxygen (pressure turned off). During a simulation, when pressure is released (turned off), blood vessel sizes increase, and oxygen again perfuses these cells, reactive oxygen species (ROS) are formed proportionally to the concentration of xanthine oxidase present in that cell. Free radicals cause damage to the immediate cell and those they encounter via diffusion, but they do so in a stepwise manner: tissue cells show no sign of damage from radicals until they have accumulated a certain threshold of insults. At that time, their health is reduced drastically. Rules: Tissue damage Each tissue cell in a simulation has an intrinsic variable called "life," which ranges from 0 (dead) to 100 (perfect health). Tissue damage for a single cell is the opposite of life (100-life) and the measure for the whole field, total tissue damage, is simply the sum of each cell's damage score. This number gives a general sense of how much injury has accrued throughout the course of a simulation. T i s s u e D a m a g e ( t o t a l ) = ∑ A l l T i s s u e C e l l s ( i ) 100 − l i f e ( i ) A cell's life score is determined by local concentrations of oxygen, TNF-α, and TGF-β1. These represent the contributions of ischemia and inflammation to epithelial cell health. To mimic ischemic injury and recovery upon reperfusion, oxygen can be either beneficial or detrimental. We choose to model ischemia by comparing local concentrations of oxygen to a threshold, above which tissue health improved and below which ischemic damage was incurred. To mimic both damaging and pro-healing effects of inflammation, in the PUABM TNF-α decreases cell health while TGF-β1 supports healing [ 29 ]. Tissue cells that are stressed (have decreasing life scores) release diffusible danger signals (Damage-Associated Molecular Patterns [DAMPs]) that stimulate the inflammatory response by triggering the secretion of cytokines by inflammatory cells [ 33 ] Three diffusible mediators represent the canonical early pro-inflammatory response, the canonical slower pro-inflammatory response, and the canonical anti-inflammatory response, each of which is secreted by activated neutrophils or M1 or M2 macrophages. Neutrophils and macrophages are initially in a resting state, and are activated by local concentrations of mediators in a threshold-dependent manner. A local concentration of DAMPs above a certain value will activate nearby neutrophils to produce early pro-inflammatory mediators (abstracted as the pro-inflammatory cytokine TNF-α). At a certain threshold of local TNF-α concentration, resting macrophages will be activated to a M1 phenotype and begin secreting longer-acting pro-inflammatory mediators[ 34 ] (abstracted as IL-1β). TNF-α also causes damage to nearby tissue cells, thus re-stimulating the pro-inflammatory response. Local concentration of IL-1β above a threshold will activate macrophages to a M1 (pro-inflammatory) phenotype, and above a higher threshold, IL-1β will induce macrophages to a M2 (anti-inflammatory/reparative) phenotype[ 35 ]. These phenotype changes are reversible, meaning a M2 macrophage could be induced to switch to a M1 phenotype, depending on the local concentrations of mediators as outlined above (see Table 1 and pseudocode in S1 Text for rules governing macrophage state changes). Active M2 macrophages produce anti-inflammatory mediators (abstracted as TGF-β1), which above a threshold will cause further activation of M2 macrophages ( Fig 1B ). TGF-β1 also increases health of nearby tissue cells. Model calibration A default set of parameters was based on a model time step of 1 hour, which informed the lifespans of Neutrophils and Macrophages [ 28 , 29 , 33 , 34 ](see S1 Table for parameter values). Due to the lack of appropriate data, many parameter values were selected to achieve appropriate qualitative behavior at the cellular or molecular scales [ 36 ]. Accordingly, parameter values were modified one at a time until the model behavior aligned qualitatively with the rules to which they applied and baseline expected behaviors. The first test was to ensure that the "healthy" tissue was stable, i.e. the tissue was able to remain healthy when unperturbed. When damage appeared in simulations of undisturbed tissue after several thousand ticks (simulating nearly 7 months of real time), it was noted that oxygen levels at the edges of the grid were below the threshold at which tissue health began to decline, leading to tissue damage. Since unperturbed tissue should have plentiful oxygen and would not spontaneously degrade, diffusion rates were increased to ensure a more constant oxygen level across the grid. Similarly, the parameters controlling steroid inhibition of neutrophils and macrophages were tuned so that different doses of steroids produced a range of neutrophil or macrophage inhibition. Other parameters were adjusted so that simulated ulcers evolved comparably to those in clinical images. Some properties that were compared included size of ulcers over time relative to first clinical observation or moment of ulceration in the model and extent of tissue damage in simulations or discoloration in clinical subjects beyond the outer edge of ulceration. For uniformity of appearance, only sacral ulcers within the clinical cohort were considered. As clinical images were used to define the baseline conditions of the PUABM, consequently this model is calibrated to the unique properties of SCI tissue (as we have suggested in a prior modeling study [ 25 ]). By altering a subset of the default parameters in the PUABM, the model could be calibrated to match time courses from a different cohort of patients or type of ulcer. Several parameters represented lumped processes, and therefore it was necessary to make reasonable assumptions as to their values, accounting for relationships to other parameters. For example, TNF-α in the PUABM represents the collective actions of several pro-inflammatory species. Therefore, it would be inappropriate to base physical parameters (e.g. diffusion constants) on those known for TNF-α protein. Other parameters were tuned such that key behaviors of the PUABM were upheld. For example, in order to maintain reversibility of macrophage phenotypes, mediator thresholds of M1 and M2 activation were tuned so that neither population could too quickly or easily dominate without a significant and sustained source of the appropriate mediator. Certainly, many other default parameterizations could also have met these criteria, but they likely exist within ranges close to the values chosen in our parameterization. As explored later, model sensitivity analysis revealed several nearby regions of parameter space that yielded irrelevant or non-biological outcomes. In examining and plotting the inflammatory dynamics resulting from the PUABM, some adjustments were necessary in order to bring the cell populations all into comparable ranges. Because the agents in the model are represented as unit-less numbers rather than specific numbers of cells, [ 36 ] absolute population sizes from the PUABM are meaningless. For a given cell type at time t, its population count was normalized by dividing by the Euclidean norm of the vector containing counts of that cell type for each tick throughout that simulation. This procedure allowed us to compare the relative timing of the peaks of each population. Presumably, it is not only the number of cells present that determines the level of the response, but also how sensitive those cells might be to their environment, and vice versa . Therefore, instead of calibrating for total number of cells, we focused on achieving reasonable qualitative behaviors in the tissue as a whole. Simulations of acute inflammation incited by initial injuries of varying degree Pressure was a key factor in simulations of the PUABM. In simulations of the acute inflammatory response in the absence of pressure, as in Fig 2C and 2D , the differences in outcomes were most dramatic, as they either did or did not result in an ulcer. In contrast, when pressure was added, the differences in predicted outcome were more subtle: an ulcer always formed, but it was associated with varying degrees of overall damage. To better define the mechanisms influencing this bifurcated outcome, we initially carried out simulations of acute inflammation in the absence of pressure. 10.1371/journal.pcbi.1004309.g002 Fig 2 Model Verification: Mechanisms Lead to Expected Behaviors in Baseline Conditions. ( A ) Negative controls verify that the model behaves as expected in various situations. Green: undisturbed, tissue health is stable for >5000 hours (data past 1000 h not shown). Yellow: an initial period of 12 h of ischemia causes damage to the tissue, but after release no further damage was incurred. Red: a characteristic damage curve for an ulcer caused by acute inflammation after 40% initial injury (similar to 2C). Blue: a characteristic damage curve for a pressure ulcer resulting from the default parameters for the model. ( B ) The ischemia/reperfusion injury mechanism was validated by varying the period of pressure (on/off) cycles. Increasing the length of a pressure cycle allowed us to decrease the number of reperfusion events over the same length of ischemia. Pressure switched from on to off (or vice versa) once every 2, 6, or 12 h. y-axis is total damage, x-axis is time (h). Pressure cycle length did not seem to affect the total amount of damage until just after t = 400 h. At that time point, the simulations with the shortest cycles (black) show a sudden increase in damage, which visually corresponds to the formation of an ulcer (see inset). Eventually, ischemic injury in simulations with the longest cycle length (green) causes these simulations to incur more overall damage. ( C ) A 35% initial injury without pressure is sufficient to induce self-perpetuating, damaging inflammation, leading to an ulcer. The relative dynamics of the response are as expected from the literature: an initial influx and activation of neutrophils, followed by M1 macrophages, and then followed by M2 macrophages. ( D ) In 10–20% of simulations with the same parameters and starting conditions, though the inflammatory response was incited, it did not become self-sustaining and consequently no ulcer formed. The trajectories of inflammatory cells are characteristic of each of these outcomes. Data were normalized per cell type, and so quantities are not relative. Acute inflammatory dynamics were incited by an initial injury to the center of the tissue. We first varied the intensity of the initial injury in order to determine over what range of injury both outcomes persisted. Because we found that the frequency of ulcerative inflammation was correlated to the intensity of initial injury, we focused on simulations with a 30% initial injury—the level at which 50% of simulations resolved and 50% formed an ulcer (see S2 Fig ). Approximating damage distribution in simulations We approximated the distributions of total tissue damage in PUABM simulations using a Gaussian Mixture Model fit using the Expectation-Maximization algorithm, as implemented in the Matlab function, gmmfit . We repeated this process varying the number of Gaussians in the model from 1 to 4, and then compared the goodness-of fit for each one, using Aikikae and Bayesian Information Criteria. This step allowed us to determine quantitatively the number of underlying Gaussian models that give rise to the pattern we see, an important step when the total damage from each outcome does not vary significantly. We selected the model that was a mixture of two independent Gaussian distributions because this model yielded the lowest scores for both Aikikae Information Criteria (AIC) and Bayesian Information Criteria (BIC), shown in S2 Table . 1-Nearest Neighbor analysis of simulation results Following the work of Xing et al. [ 37 ], we employed a 1-nearest neighbor (1NN) approach to automatically segregate simulations ending in ulcers from those that displayed resolving inflammation. The training set consisted of data from 100 simulations, labeled according to which endpoint was reached (resolved or ulcerated). For each simulation in the test set, a pairwise distance was computed between itself and every simulation in the training set. The unlabeled test sequence was assigned the same class label as the training sequence that was closest in distance to it, its "nearest neighbor." This method relies heavily on the choice of distance metric, in this case the Euclidean distance between sequences. To take advantage of the time dependence of the features, we considered each simulation to be a sequence, wherein every entry in the sequence was a time point from a simulation. Each of those entries consisted of either a single value (e.g. total oxygen at tick t) or a 13-dimensional vector of model component values. For univariate time series, we calculated the Euclidean distance between each pair of vectors consisting of measurements of a single feature through time. For multivariate time series, we calculated the Euclidean distance between two feature vectors at each time step and took the Euclidean norm of those distances to be the distance between the two simulations. To equalize the contributions of all features, their values were normalized to fall into similar ranges before calculating distance. Sensitivity analysis of model parameters We next sought to define the model parameters that most affected simulation outcomes in the PUABM. Because there are more than 50 free parameters in the model, it was impractical to examine the sensitivity of key model outputs to all of parameter space at once. Instead, modules of rules in the model that impacted tissue health were identified and used to create groups of parameters. Parameters within these groups were then prioritized according to the mathematical degree of their effect on the system. For example, a parameter that sets the value of an exponent produces a more dramatic effect than one that sets a scalar multiple. Therefore, the first parameters varied were those controlling threshold values. Simulations initially varied parameters over coarse-grained and then finer-grained threshold value ranges (100 simulations per parameter value). Total tissue damage at time t served as a quantitative output measure. From this analysis, a sensitivity index could be calculated for each parameter, taking the ratio of change in damage to change in threshold value. A second level of sensitivity analysis was designed to examine the interplay between two potentially related parameter values. This analysis allowed for a direct comparison of the sensitivities of two parameters, and also revealed any secondary effects that occurred when the two parameters changed in a combinatorial way. Parameters were chosen from the same "damage module" in order to get a sense of the relationships between "sub-processes" in each module. Snapshots of the tissue layer in the model at a fixed time point served as the output in order to assess overall damage, presence and size of ulcer, and other qualitative features. In silico clinical trials Treatments with corticosteroids or neutralizing anti-DAMP antibodies were simulated as in silico clinical trials. These trials consisted of sets of model simulations in which parameters controlling drug dose and timing and tissue response were varied, each independently. Corticosteroid administration was simulated as an intravenous injection. Therefore, steroid molecules (implemented as a data layer) were introduced to the tissue via blood vessels (and were restricted when pressure was applied). The mechanism of steroid action was to kill inflammatory cells, regardless of their state (active/ resting), as illustrated in S6 Fig . When neutrophils were killed, additional ROS was released by the dying cell (see S1 Text ). Anti-DAMPs were simulated as a topical cream administration. A uniform layer of this molecule was introduced onto the field as a data layer at the tick specified by the parameter designating time of onset. The method of action of the antibodies was controlled by a quenching reaction, wherein local concentrations of DAMPs were reduced by an amount proportional to the smaller concentration of the two molecules present: antibody or DAMPs (see S1 Text for pseudocode and also S7 Fig ). Ethics statement This study was approved by the University of Pittsburgh Institutional Review Board, approval number PRO08010011. Written informed consent was obtained from subjects participating in the study. If the subject was unable to provide written consent due to medical condition, proxy consent was obtained from the subject's authorized representative, and the subject provided written consent for continued research participation as soon as they were able to do so. Clinical data collection Photographic data were obtained from pressure ulcers sustained by patients enrolled in the Rehabilitation Engineering Research Center on Spinal Cord Injury at the University of Pittsburgh. Subjects were recruited from a single tertiary care center if they were 18 years of age or older and had sustained a traumatic spinal cord injury. Subjects were excluded if they had a history of pre-existing diseases that affect the inflammatory response to SCI (e.g. autoimmune disease, demyelinating diseases) or a history of previous SCI or other neurological diseases affecting motor and sensory function. Pressure ulcers were initially photographed when identified, and serial photographs were taken three times per week while patients remained in the acute care and/or weekly while in the inpatient rehabilitation hospital to monitor progression and/or healing, and during their outpatient visits or 6 month intervals after discharge. Photos were obtained using a Canon Power Shot SD 750 camera with 3X optical zoom lens with flash enabled, at a distance of 12 inches from the ulcer, resolution of 2048 x 1536 pixels. Pressure ulcer site, severity (stage 1–4, unstageable, or deep tissue injury), general appearance, size and shape were also recorded. Work flow The PUABM was built using an iterative approach ( Fig 1A ). Hereafter, when we use words or phrases that usually specify biological reality, such as "pressure," "neutrophil," "TNF-α," "inflammatory mediators," "cell," "oxygen," "ischemic," etc. we are referring to PUABM components or generated phenomena unless clearly stated otherwise. A simpler model of pressure ulcer formation [ 22 ] was altered to increase mechanistic detail and create clinically relevant model output. Rational improvements were based on domain knowledge and data from the literature. First, the area of tissue simulated in the field was extended. Whereas in the previous version of the model pressure was applied evenly across the entire field, in this version, it was applied maximally (value determined by the parameter, pressure-intensity ) to a circular area in the center of the field and decreasing radially outward. This allowed representation of pressure over a bony prominence and the surrounding tissue, which also experiences pressure, but to a lessening degree [ 23 , 24 ]. Next, the PUABM code was altered so that neutrophils and macrophages enter the tissue in a resting state and can be activated by one of two circulating mediators (TNF-α or TGF-β1). In the previous version of the model, all entering leukocytes were in the active state. After this modification, the activation state of neutrophils and monocytes/ macrophages was determined by the local concentrations of inflammatory mediators. These thresholds are explained further in the section below entitled, "Rules: Tissue Damage." 10.1371/journal.pcbi.1004309.g001 Fig 1 Work flow and schematic of the PUABM. ( A ) Work Flow utilized in constructing the PUABM. The ABM was built iteratively, incorporating domain knowledge and data from the literature, then validated with clinical data. Initially, pressure directly injured tissue, thereby inciting inflammation. Ischemia/reperfusion injury was next added as a cause of injury, and the complexity of the inflammatory response was increased. Clinical data were next used to calibrate parameters in the model, and the model was subjected to sensitivity analysis and in silico trials. ( B ) A schematic illustrates how model components interact to simulate two mechanisms of injury: ischemia/reperfusion injury and damage due to inflammation. A tissue cell (muscle, fat, or skin) is situated over a bony prominence. When pressure is applied, oxygen supply is reduced and the cell becomes ischemic, leading to tissue damage. Simultaneously, the enzyme xanthine dehydrogenase is converted into xanthine oxidase. Thus, reactive oxygen species (ROS) are produced when pressure is released and oxygen flows back to the cell, causing further damage. Tissue damage causes the cell to release DAMPs, which, along with local concentrations of pro-inflammatory cytokines activate both neutrophils (N) and macrophages (M1, M2) when these mediators are present above a given threshold parameter. For example, a sufficiently high local concentration of DAMPs activates neutrophils to secrete TNF-α, which can activate macrophages to a pro- inflammatory M1 phenotype. Tissue damage is ameliorated by anti-inflammatory mediators [TGF-β1]. In the previous version of the ABM, all cells and mediators appeared in a single visualization window. Epithelial cells ranged in color from green to red, depending on their level of health. Other cells were rendered in layers over the grid of epithelial cells[ 22 , 25 ]. In the PUABM presented herein, circulating inflammatory cells and mediators were each removed to a separate viewing window, allowing for comparison of spatio-temporal pattern of each individual component with any other. In the main window, stationary epithelial cells remained, but the colors indicating tissue health were altered to increase the realism: healthy tissue was now rendered as peach, mimicking the most common skin tones in the clinical cohort. Unhealthy tissue continued to be rendered as red, and cells that died disappeared from the grid, leaving behind a white empty space. Thus, each component of the model was represented according to the rules in the ABM, but by separating the viewing windows, we were able to view and compare spatial details of individual components (damage, mediators, etc.). Furthermore, we were able to compare epithelial damage patterns between simulations and clinical images, as described in later sections. Mechanistic details in the PUABM were augmented by incorporating damage to tissue via ischemic injury and its counterpart, reperfusion injury. Damage directly resulting from pressure (as previously simulated [ 22 , 25 ]) was removed and tissue health was penalized for tissue cells receiving inadequate oxygen levels. (In the previous version, oxygen had positive effects on tissue health, but lack of oxygen was not damaging [ 22 ]). A mechanism based on the conversion of xanthine dehydrogenase to xanthine oxidase during ischemia [ 26 , 27 ] was also implemented. The accumulation of xanthine oxidase in ischemic cells represents the potential of a cell to experience reperfusion injury, due to formation of ROS when oxygen reperfuses [ 26 ]. The complexity of the inflammatory response was increased by creating two subpopulations of activated macrophages, one with pro-inflammatory (M1 macrophages) and another with anti-inflammatory (M2 macrophages) phenotype [ 28 , 29 ]. A fourth mediator was also added. Representing a canonical later-acting pro-inflammatory mediator, it is released by pro-inflammatory macrophages and labeled IL-1β. After the mechanisms comprising the PUABM were set, model behaviors were explored over wide ranges of parameter values and default values were tweaked into ranges producing behavior that qualitatively matched the clinical data. Finally, we further investigated dynamics encoded in the model by simulating hypothetical and existing therapies. Corticosteroids were incorporated as a data layer that could be introduced to the tissue via blood vessels (as in an intravenous injection), while antibodies to DAMPs were simulated as a topical cream applied to the entire field. The dose and timing of administrations of each were varied to assess the viability of these treatment options. Model implementation The PUABM was built using SPARK, a platform designed for agent-based modeling of biological systems, freely available for download at: www.pitt.edu/~cirm/spark [ 25 , 30 , 31 ]. SPARK models are written in a logo-like language called SPARK-PL and run on a java platform [ 32 ]. SPARK models contain Agents—autonomous entities that interact with each other and the environment, called Space—and Data Layers, corresponding to individual species in the environment that can diffuse, etc. The behavior of the model is determined by rules that govern how and when agents interact and react. These rules are generally written to be interpreted by one agent at a time, and therefore are necessarily restricted in scope (both time and space). A rule specifies how much an agent should move, produce, change an internal variable, etc. when it encounters a certain amount or type of data layer or agent in its immediate neighborhood. Model rules can be probabilistic in nature, allowing the model to evolve in a stochastic manner. Therefore, the behaviors and patterns produced by simulating several ticks (time steps) of the model in succession arise as emergent phenomena resulting from the accumulated actions of a population of agents over time. SPARK has several built-in standard methods that allow for convenient coding of common biological processes. For example, diffusion is encoded by the function diffuse , which implements a simple discrete approximation in which each data layer cell shares a given percentage of its value with its eight neighbors. Examples of other methods that were used in this model include wiggle and jump to approximate undirected random movement and sniff , for chemotaxis. See pseudocode ( S1 Text ) for all methods used. The components of the PUABM are described in Table 1 , and the operant biological mechanisms and tissue structures are shown graphically in Fig 1B . In the PUABM, simulation time is linked to actual time using the lifespans of cellular agents. A simulated macrophage lifespan ranges from 100–150 model time steps (ticks), corresponding to 100–150 hours (4–6 days) of real time. Neutrophil lifespans range from 10–20 ticks (hours) in the tissue, but increase when neutrophils become activated. Each simulation takes roughly five minutes to compute on a supercomputing node containing 32 process cores, each running at 4.7 GHz, making it possible to complete approximately 1000 simulations per day. The model contains 68 numerical parameters that are set by the modeler and 11 random variables (whose values are drawn from a corresponding uniform distribution when necessary during the course of a simulation). At any given tick, the maximum number of agents computed in the model is on the order of 10 5 . 10.1371/journal.pcbi.1004309.t001 Table 1 Summary of agents, data layers, and their interactions encoded in the ABM. Substances in ABM Cell Sources in ABM Biological Functions in ABM DAMPs secreted by damaged Tissue cells Chemo-attract and Activate Neutrophils TNF-α secreted by Activated Neutrophils Chemo-attract resting and M1 Macrophages, activate Macrophages to M1 state IL-1β secreted by M1 Macrophages Chemo-attract resting and M1 Macrophages, activate Macrophages to M1 state, at high levels activate Macrophages to M2 state TGF-β1 secreted by M2 Macrophages Chemo-attract resting and M2 Macrophages, activate Macrophages to M2 state ROS secreted by Tissue Cells on Reperfusion, by Neutrophils on death Injure Tissue Cells Oxidase produced inside Tissue Cells during Ischemia React with Oxygen to produce ROS Oxygen released by Blood Vessels Necessary for Tissue Cell health, react with Oxidase to produce ROS Substances are all encoded as Data Layers; Cells are all encoded as Agents, each with its own type. See Fig 1B for schematic. Rules: Homeostasis/baseline architecture The pseudocode in S1 Text details all model rules. In brief, the architecture of the PUABM is abstracted from healthy tissue. The PUABM consists of a layer of tissue cells fed by blood vessels throughout the tissue that carry oxygen and inflammatory cells. Instead of representing red blood cells explicitly, their presence is implicit since oxygen flows unaccompanied through blood vessels. Inflammatory cells arrive to the tissue in their native resting state and are not activated unless danger signals are released from damaged tissue. Resting inflammatory cells move randomly, but can be chemo-attracted to local mediators via the SPARK function, sniff . The agents in the PUABM represent either cell types or cellular structures: neutrophils, macrophages, tissue cells, and blood vessels. Data layers are employed to represent mediators in a computationally efficient manner (e.g., diffusible cytokines, free radicals, oxygen, xanthine oxidase, and exogenously-administered drugs). As mentioned above, each tick (time step) of the model represents one hour of real time. On each tick, the agents behave according to rules specific to their type: they produce and consume oxygen and mediators, are chemo-attracted to and activated (undergo a state change) by mediators; and they die, according to their lifespans. Mediators, represented in data layers, can be consumed or produced by individual agents or in reactions with other mediators. They undergo diffusion and degradation at each tick. To mimic natural patient-to-patient and cell-to-cell variability, some pseudo-random variables were built into the model. These values were drawn as necessary as simulations progressed, using the Mersenne twister (as implemented in the Colt library, http://acs.lbl.gov/ACSSoftware/colt/readme.html ). Some examples of random variables are size of each blood vessel agent (the blood vessels are all sized within a range but not exactly the same), and probabilities governing state changes of macrophages and neutrophils (a threshold must be met for the agent to be eligible for state change, but the random variable decides whether the transition actually happens). The function of the thresholds was to partition the interval (0,1). Accordingly, a uniform distribution was the basis of pseudo-random number generation. Similarly, blood vessels were assumed to be sized uniformly across a narrow range: the largest was less than twice the size of the smallest. Rules: Pressure on/off The turning of people with SCI while they are lying in bed was simulated as alternating cycles of pressure (on/off). In the PUABM, pressure is simulated by constriction of blood vessels, decreasing the amount of material that can flow through them. Without oxygen, tissue cell health in the PUABM begins to decline. Ischemic tissue cells increase xanthine oxidase enzyme activity, which represent the capacity of the cell to produce damaging free radicals upon reintroduction of oxygen (pressure turned off). During a simulation, when pressure is released (turned off), blood vessel sizes increase, and oxygen again perfuses these cells, reactive oxygen species (ROS) are formed proportionally to the concentration of xanthine oxidase present in that cell. Free radicals cause damage to the immediate cell and those they encounter via diffusion, but they do so in a stepwise manner: tissue cells show no sign of damage from radicals until they have accumulated a certain threshold of insults. At that time, their health is reduced drastically. Rules: Tissue damage Each tissue cell in a simulation has an intrinsic variable called "life," which ranges from 0 (dead) to 100 (perfect health). Tissue damage for a single cell is the opposite of life (100-life) and the measure for the whole field, total tissue damage, is simply the sum of each cell's damage score. This number gives a general sense of how much injury has accrued throughout the course of a simulation. T i s s u e D a m a g e ( t o t a l ) = ∑ A l l T i s s u e C e l l s ( i ) 100 − l i f e ( i ) A cell's life score is determined by local concentrations of oxygen, TNF-α, and TGF-β1. These represent the contributions of ischemia and inflammation to epithelial cell health. To mimic ischemic injury and recovery upon reperfusion, oxygen can be either beneficial or detrimental. We choose to model ischemia by comparing local concentrations of oxygen to a threshold, above which tissue health improved and below which ischemic damage was incurred. To mimic both damaging and pro-healing effects of inflammation, in the PUABM TNF-α decreases cell health while TGF-β1 supports healing [ 29 ]. Tissue cells that are stressed (have decreasing life scores) release diffusible danger signals (Damage-Associated Molecular Patterns [DAMPs]) that stimulate the inflammatory response by triggering the secretion of cytokines by inflammatory cells [ 33 ] Three diffusible mediators represent the canonical early pro-inflammatory response, the canonical slower pro-inflammatory response, and the canonical anti-inflammatory response, each of which is secreted by activated neutrophils or M1 or M2 macrophages. Neutrophils and macrophages are initially in a resting state, and are activated by local concentrations of mediators in a threshold-dependent manner. A local concentration of DAMPs above a certain value will activate nearby neutrophils to produce early pro-inflammatory mediators (abstracted as the pro-inflammatory cytokine TNF-α). At a certain threshold of local TNF-α concentration, resting macrophages will be activated to a M1 phenotype and begin secreting longer-acting pro-inflammatory mediators[ 34 ] (abstracted as IL-1β). TNF-α also causes damage to nearby tissue cells, thus re-stimulating the pro-inflammatory response. Local concentration of IL-1β above a threshold will activate macrophages to a M1 (pro-inflammatory) phenotype, and above a higher threshold, IL-1β will induce macrophages to a M2 (anti-inflammatory/reparative) phenotype[ 35 ]. These phenotype changes are reversible, meaning a M2 macrophage could be induced to switch to a M1 phenotype, depending on the local concentrations of mediators as outlined above (see Table 1 and pseudocode in S1 Text for rules governing macrophage state changes). Active M2 macrophages produce anti-inflammatory mediators (abstracted as TGF-β1), which above a threshold will cause further activation of M2 macrophages ( Fig 1B ). TGF-β1 also increases health of nearby tissue cells. Model calibration A default set of parameters was based on a model time step of 1 hour, which informed the lifespans of Neutrophils and Macrophages [ 28 , 29 , 33 , 34 ](see S1 Table for parameter values). Due to the lack of appropriate data, many parameter values were selected to achieve appropriate qualitative behavior at the cellular or molecular scales [ 36 ]. Accordingly, parameter values were modified one at a time until the model behavior aligned qualitatively with the rules to which they applied and baseline expected behaviors. The first test was to ensure that the "healthy" tissue was stable, i.e. the tissue was able to remain healthy when unperturbed. When damage appeared in simulations of undisturbed tissue after several thousand ticks (simulating nearly 7 months of real time), it was noted that oxygen levels at the edges of the grid were below the threshold at which tissue health began to decline, leading to tissue damage. Since unperturbed tissue should have plentiful oxygen and would not spontaneously degrade, diffusion rates were increased to ensure a more constant oxygen level across the grid. Similarly, the parameters controlling steroid inhibition of neutrophils and macrophages were tuned so that different doses of steroids produced a range of neutrophil or macrophage inhibition. Other parameters were adjusted so that simulated ulcers evolved comparably to those in clinical images. Some properties that were compared included size of ulcers over time relative to first clinical observation or moment of ulceration in the model and extent of tissue damage in simulations or discoloration in clinical subjects beyond the outer edge of ulceration. For uniformity of appearance, only sacral ulcers within the clinical cohort were considered. As clinical images were used to define the baseline conditions of the PUABM, consequently this model is calibrated to the unique properties of SCI tissue (as we have suggested in a prior modeling study [ 25 ]). By altering a subset of the default parameters in the PUABM, the model could be calibrated to match time courses from a different cohort of patients or type of ulcer. Several parameters represented lumped processes, and therefore it was necessary to make reasonable assumptions as to their values, accounting for relationships to other parameters. For example, TNF-α in the PUABM represents the collective actions of several pro-inflammatory species. Therefore, it would be inappropriate to base physical parameters (e.g. diffusion constants) on those known for TNF-α protein. Other parameters were tuned such that key behaviors of the PUABM were upheld. For example, in order to maintain reversibility of macrophage phenotypes, mediator thresholds of M1 and M2 activation were tuned so that neither population could too quickly or easily dominate without a significant and sustained source of the appropriate mediator. Certainly, many other default parameterizations could also have met these criteria, but they likely exist within ranges close to the values chosen in our parameterization. As explored later, model sensitivity analysis revealed several nearby regions of parameter space that yielded irrelevant or non-biological outcomes. In examining and plotting the inflammatory dynamics resulting from the PUABM, some adjustments were necessary in order to bring the cell populations all into comparable ranges. Because the agents in the model are represented as unit-less numbers rather than specific numbers of cells, [ 36 ] absolute population sizes from the PUABM are meaningless. For a given cell type at time t, its population count was normalized by dividing by the Euclidean norm of the vector containing counts of that cell type for each tick throughout that simulation. This procedure allowed us to compare the relative timing of the peaks of each population. Presumably, it is not only the number of cells present that determines the level of the response, but also how sensitive those cells might be to their environment, and vice versa . Therefore, instead of calibrating for total number of cells, we focused on achieving reasonable qualitative behaviors in the tissue as a whole. Rules: Homeostasis/baseline architecture The pseudocode in S1 Text details all model rules. In brief, the architecture of the PUABM is abstracted from healthy tissue. The PUABM consists of a layer of tissue cells fed by blood vessels throughout the tissue that carry oxygen and inflammatory cells. Instead of representing red blood cells explicitly, their presence is implicit since oxygen flows unaccompanied through blood vessels. Inflammatory cells arrive to the tissue in their native resting state and are not activated unless danger signals are released from damaged tissue. Resting inflammatory cells move randomly, but can be chemo-attracted to local mediators via the SPARK function, sniff . The agents in the PUABM represent either cell types or cellular structures: neutrophils, macrophages, tissue cells, and blood vessels. Data layers are employed to represent mediators in a computationally efficient manner (e.g., diffusible cytokines, free radicals, oxygen, xanthine oxidase, and exogenously-administered drugs). As mentioned above, each tick (time step) of the model represents one hour of real time. On each tick, the agents behave according to rules specific to their type: they produce and consume oxygen and mediators, are chemo-attracted to and activated (undergo a state change) by mediators; and they die, according to their lifespans. Mediators, represented in data layers, can be consumed or produced by individual agents or in reactions with other mediators. They undergo diffusion and degradation at each tick. To mimic natural patient-to-patient and cell-to-cell variability, some pseudo-random variables were built into the model. These values were drawn as necessary as simulations progressed, using the Mersenne twister (as implemented in the Colt library, http://acs.lbl.gov/ACSSoftware/colt/readme.html ). Some examples of random variables are size of each blood vessel agent (the blood vessels are all sized within a range but not exactly the same), and probabilities governing state changes of macrophages and neutrophils (a threshold must be met for the agent to be eligible for state change, but the random variable decides whether the transition actually happens). The function of the thresholds was to partition the interval (0,1). Accordingly, a uniform distribution was the basis of pseudo-random number generation. Similarly, blood vessels were assumed to be sized uniformly across a narrow range: the largest was less than twice the size of the smallest. Rules: Pressure on/off The turning of people with SCI while they are lying in bed was simulated as alternating cycles of pressure (on/off). In the PUABM, pressure is simulated by constriction of blood vessels, decreasing the amount of material that can flow through them. Without oxygen, tissue cell health in the PUABM begins to decline. Ischemic tissue cells increase xanthine oxidase enzyme activity, which represent the capacity of the cell to produce damaging free radicals upon reintroduction of oxygen (pressure turned off). During a simulation, when pressure is released (turned off), blood vessel sizes increase, and oxygen again perfuses these cells, reactive oxygen species (ROS) are formed proportionally to the concentration of xanthine oxidase present in that cell. Free radicals cause damage to the immediate cell and those they encounter via diffusion, but they do so in a stepwise manner: tissue cells show no sign of damage from radicals until they have accumulated a certain threshold of insults. At that time, their health is reduced drastically. Rules: Tissue damage Each tissue cell in a simulation has an intrinsic variable called "life," which ranges from 0 (dead) to 100 (perfect health). Tissue damage for a single cell is the opposite of life (100-life) and the measure for the whole field, total tissue damage, is simply the sum of each cell's damage score. This number gives a general sense of how much injury has accrued throughout the course of a simulation. T i s s u e D a m a g e ( t o t a l ) = ∑ A l l T i s s u e C e l l s ( i ) 100 − l i f e ( i ) A cell's life score is determined by local concentrations of oxygen, TNF-α, and TGF-β1. These represent the contributions of ischemia and inflammation to epithelial cell health. To mimic ischemic injury and recovery upon reperfusion, oxygen can be either beneficial or detrimental. We choose to model ischemia by comparing local concentrations of oxygen to a threshold, above which tissue health improved and below which ischemic damage was incurred. To mimic both damaging and pro-healing effects of inflammation, in the PUABM TNF-α decreases cell health while TGF-β1 supports healing [ 29 ]. Tissue cells that are stressed (have decreasing life scores) release diffusible danger signals (Damage-Associated Molecular Patterns [DAMPs]) that stimulate the inflammatory response by triggering the secretion of cytokines by inflammatory cells [ 33 ] Three diffusible mediators represent the canonical early pro-inflammatory response, the canonical slower pro-inflammatory response, and the canonical anti-inflammatory response, each of which is secreted by activated neutrophils or M1 or M2 macrophages. Neutrophils and macrophages are initially in a resting state, and are activated by local concentrations of mediators in a threshold-dependent manner. A local concentration of DAMPs above a certain value will activate nearby neutrophils to produce early pro-inflammatory mediators (abstracted as the pro-inflammatory cytokine TNF-α). At a certain threshold of local TNF-α concentration, resting macrophages will be activated to a M1 phenotype and begin secreting longer-acting pro-inflammatory mediators[ 34 ] (abstracted as IL-1β). TNF-α also causes damage to nearby tissue cells, thus re-stimulating the pro-inflammatory response. Local concentration of IL-1β above a threshold will activate macrophages to a M1 (pro-inflammatory) phenotype, and above a higher threshold, IL-1β will induce macrophages to a M2 (anti-inflammatory/reparative) phenotype[ 35 ]. These phenotype changes are reversible, meaning a M2 macrophage could be induced to switch to a M1 phenotype, depending on the local concentrations of mediators as outlined above (see Table 1 and pseudocode in S1 Text for rules governing macrophage state changes). Active M2 macrophages produce anti-inflammatory mediators (abstracted as TGF-β1), which above a threshold will cause further activation of M2 macrophages ( Fig 1B ). TGF-β1 also increases health of nearby tissue cells. Model calibration A default set of parameters was based on a model time step of 1 hour, which informed the lifespans of Neutrophils and Macrophages [ 28 , 29 , 33 , 34 ](see S1 Table for parameter values). Due to the lack of appropriate data, many parameter values were selected to achieve appropriate qualitative behavior at the cellular or molecular scales [ 36 ]. Accordingly, parameter values were modified one at a time until the model behavior aligned qualitatively with the rules to which they applied and baseline expected behaviors. The first test was to ensure that the "healthy" tissue was stable, i.e. the tissue was able to remain healthy when unperturbed. When damage appeared in simulations of undisturbed tissue after several thousand ticks (simulating nearly 7 months of real time), it was noted that oxygen levels at the edges of the grid were below the threshold at which tissue health began to decline, leading to tissue damage. Since unperturbed tissue should have plentiful oxygen and would not spontaneously degrade, diffusion rates were increased to ensure a more constant oxygen level across the grid. Similarly, the parameters controlling steroid inhibition of neutrophils and macrophages were tuned so that different doses of steroids produced a range of neutrophil or macrophage inhibition. Other parameters were adjusted so that simulated ulcers evolved comparably to those in clinical images. Some properties that were compared included size of ulcers over time relative to first clinical observation or moment of ulceration in the model and extent of tissue damage in simulations or discoloration in clinical subjects beyond the outer edge of ulceration. For uniformity of appearance, only sacral ulcers within the clinical cohort were considered. As clinical images were used to define the baseline conditions of the PUABM, consequently this model is calibrated to the unique properties of SCI tissue (as we have suggested in a prior modeling study [ 25 ]). By altering a subset of the default parameters in the PUABM, the model could be calibrated to match time courses from a different cohort of patients or type of ulcer. Several parameters represented lumped processes, and therefore it was necessary to make reasonable assumptions as to their values, accounting for relationships to other parameters. For example, TNF-α in the PUABM represents the collective actions of several pro-inflammatory species. Therefore, it would be inappropriate to base physical parameters (e.g. diffusion constants) on those known for TNF-α protein. Other parameters were tuned such that key behaviors of the PUABM were upheld. For example, in order to maintain reversibility of macrophage phenotypes, mediator thresholds of M1 and M2 activation were tuned so that neither population could too quickly or easily dominate without a significant and sustained source of the appropriate mediator. Certainly, many other default parameterizations could also have met these criteria, but they likely exist within ranges close to the values chosen in our parameterization. As explored later, model sensitivity analysis revealed several nearby regions of parameter space that yielded irrelevant or non-biological outcomes. In examining and plotting the inflammatory dynamics resulting from the PUABM, some adjustments were necessary in order to bring the cell populations all into comparable ranges. Because the agents in the model are represented as unit-less numbers rather than specific numbers of cells, [ 36 ] absolute population sizes from the PUABM are meaningless. For a given cell type at time t, its population count was normalized by dividing by the Euclidean norm of the vector containing counts of that cell type for each tick throughout that simulation. This procedure allowed us to compare the relative timing of the peaks of each population. Presumably, it is not only the number of cells present that determines the level of the response, but also how sensitive those cells might be to their environment, and vice versa . Therefore, instead of calibrating for total number of cells, we focused on achieving reasonable qualitative behaviors in the tissue as a whole. Simulations of acute inflammation incited by initial injuries of varying degree Pressure was a key factor in simulations of the PUABM. In simulations of the acute inflammatory response in the absence of pressure, as in Fig 2C and 2D , the differences in outcomes were most dramatic, as they either did or did not result in an ulcer. In contrast, when pressure was added, the differences in predicted outcome were more subtle: an ulcer always formed, but it was associated with varying degrees of overall damage. To better define the mechanisms influencing this bifurcated outcome, we initially carried out simulations of acute inflammation in the absence of pressure. 10.1371/journal.pcbi.1004309.g002 Fig 2 Model Verification: Mechanisms Lead to Expected Behaviors in Baseline Conditions. ( A ) Negative controls verify that the model behaves as expected in various situations. Green: undisturbed, tissue health is stable for >5000 hours (data past 1000 h not shown). Yellow: an initial period of 12 h of ischemia causes damage to the tissue, but after release no further damage was incurred. Red: a characteristic damage curve for an ulcer caused by acute inflammation after 40% initial injury (similar to 2C). Blue: a characteristic damage curve for a pressure ulcer resulting from the default parameters for the model. ( B ) The ischemia/reperfusion injury mechanism was validated by varying the period of pressure (on/off) cycles. Increasing the length of a pressure cycle allowed us to decrease the number of reperfusion events over the same length of ischemia. Pressure switched from on to off (or vice versa) once every 2, 6, or 12 h. y-axis is total damage, x-axis is time (h). Pressure cycle length did not seem to affect the total amount of damage until just after t = 400 h. At that time point, the simulations with the shortest cycles (black) show a sudden increase in damage, which visually corresponds to the formation of an ulcer (see inset). Eventually, ischemic injury in simulations with the longest cycle length (green) causes these simulations to incur more overall damage. ( C ) A 35% initial injury without pressure is sufficient to induce self-perpetuating, damaging inflammation, leading to an ulcer. The relative dynamics of the response are as expected from the literature: an initial influx and activation of neutrophils, followed by M1 macrophages, and then followed by M2 macrophages. ( D ) In 10–20% of simulations with the same parameters and starting conditions, though the inflammatory response was incited, it did not become self-sustaining and consequently no ulcer formed. The trajectories of inflammatory cells are characteristic of each of these outcomes. Data were normalized per cell type, and so quantities are not relative. Acute inflammatory dynamics were incited by an initial injury to the center of the tissue. We first varied the intensity of the initial injury in order to determine over what range of injury both outcomes persisted. Because we found that the frequency of ulcerative inflammation was correlated to the intensity of initial injury, we focused on simulations with a 30% initial injury—the level at which 50% of simulations resolved and 50% formed an ulcer (see S2 Fig ). Approximating damage distribution in simulations We approximated the distributions of total tissue damage in PUABM simulations using a Gaussian Mixture Model fit using the Expectation-Maximization algorithm, as implemented in the Matlab function, gmmfit . We repeated this process varying the number of Gaussians in the model from 1 to 4, and then compared the goodness-of fit for each one, using Aikikae and Bayesian Information Criteria. This step allowed us to determine quantitatively the number of underlying Gaussian models that give rise to the pattern we see, an important step when the total damage from each outcome does not vary significantly. We selected the model that was a mixture of two independent Gaussian distributions because this model yielded the lowest scores for both Aikikae Information Criteria (AIC) and Bayesian Information Criteria (BIC), shown in S2 Table . 1-Nearest Neighbor analysis of simulation results Following the work of Xing et al. [ 37 ], we employed a 1-nearest neighbor (1NN) approach to automatically segregate simulations ending in ulcers from those that displayed resolving inflammation. The training set consisted of data from 100 simulations, labeled according to which endpoint was reached (resolved or ulcerated). For each simulation in the test set, a pairwise distance was computed between itself and every simulation in the training set. The unlabeled test sequence was assigned the same class label as the training sequence that was closest in distance to it, its "nearest neighbor." This method relies heavily on the choice of distance metric, in this case the Euclidean distance between sequences. To take advantage of the time dependence of the features, we considered each simulation to be a sequence, wherein every entry in the sequence was a time point from a simulation. Each of those entries consisted of either a single value (e.g. total oxygen at tick t) or a 13-dimensional vector of model component values. For univariate time series, we calculated the Euclidean distance between each pair of vectors consisting of measurements of a single feature through time. For multivariate time series, we calculated the Euclidean distance between two feature vectors at each time step and took the Euclidean norm of those distances to be the distance between the two simulations. To equalize the contributions of all features, their values were normalized to fall into similar ranges before calculating distance. Sensitivity analysis of model parameters We next sought to define the model parameters that most affected simulation outcomes in the PUABM. Because there are more than 50 free parameters in the model, it was impractical to examine the sensitivity of key model outputs to all of parameter space at once. Instead, modules of rules in the model that impacted tissue health were identified and used to create groups of parameters. Parameters within these groups were then prioritized according to the mathematical degree of their effect on the system. For example, a parameter that sets the value of an exponent produces a more dramatic effect than one that sets a scalar multiple. Therefore, the first parameters varied were those controlling threshold values. Simulations initially varied parameters over coarse-grained and then finer-grained threshold value ranges (100 simulations per parameter value). Total tissue damage at time t served as a quantitative output measure. From this analysis, a sensitivity index could be calculated for each parameter, taking the ratio of change in damage to change in threshold value. A second level of sensitivity analysis was designed to examine the interplay between two potentially related parameter values. This analysis allowed for a direct comparison of the sensitivities of two parameters, and also revealed any secondary effects that occurred when the two parameters changed in a combinatorial way. Parameters were chosen from the same "damage module" in order to get a sense of the relationships between "sub-processes" in each module. Snapshots of the tissue layer in the model at a fixed time point served as the output in order to assess overall damage, presence and size of ulcer, and other qualitative features. In silico clinical trials Treatments with corticosteroids or neutralizing anti-DAMP antibodies were simulated as in silico clinical trials. These trials consisted of sets of model simulations in which parameters controlling drug dose and timing and tissue response were varied, each independently. Corticosteroid administration was simulated as an intravenous injection. Therefore, steroid molecules (implemented as a data layer) were introduced to the tissue via blood vessels (and were restricted when pressure was applied). The mechanism of steroid action was to kill inflammatory cells, regardless of their state (active/ resting), as illustrated in S6 Fig . When neutrophils were killed, additional ROS was released by the dying cell (see S1 Text ). Anti-DAMPs were simulated as a topical cream administration. A uniform layer of this molecule was introduced onto the field as a data layer at the tick specified by the parameter designating time of onset. The method of action of the antibodies was controlled by a quenching reaction, wherein local concentrations of DAMPs were reduced by an amount proportional to the smaller concentration of the two molecules present: antibody or DAMPs (see S1 Text for pseudocode and also S7 Fig ). Results Model description and work flow The PUABM represents vascularized soft tissue (skin, fat, muscle, or a composite thereof) overlaying a bony prominence. The central hypothesis of this model is that when this tissue is damaged (which can happen in a variety of ways), further complications can evolve, including inciting further tissue damage. See the Materials and Methods for a more detailed description of model rules. Pseudocode is provided in the online supplement (S1). Epithelial cell health in the PUABM depends on the continuous flow of oxygen to the field of simulated tissue. Pressure—such as that created by shifting a person's weight—is simulated as compressing the tissue and vasculature over a bony prominence, indirectly causing damage to the tissue by restricting the flow of oxygen to the tissue and secondarily by formation of oxygen species (ROS) upon reperfusion (when pressure is relieved). Injured tissue incites activation of inflammatory cells, which bring mediators to the field, some of which injure the tissue further. As simulated cycles of pressure on/off are repeated, these mechanisms are sufficient to generate pressure ulcers in silico . The molecular and cellular components of the model are enumerated in Table 1 . Model verification with literature-derived data Simulations of negative controls After tuning model parameters so that we were studying behaviors in realistic physiological ranges (see Materials and Methods ), there were a few specific mechanisms whose behavior we sought to verify. It was important to verify that the damage accumulated throughout a simulation is in fact due to the mechanisms of injury encoded in the model, and not to other spurious mechanisms. Unperturbed, the tissue remained "healthy" for thousands of simulated hours ( Fig 2A "No Pressure" data shown up to 1000 h; this behavior remained unchanged up to 5000 h). Some damage was observed when tissue was subjected to 12 h of simulated ischemia ( Fig 2A "12 h constant pressure"), but no additional injury to the tissue was observed after release of pressure. By comparison, starting with a 35% injury over a patch of tissue at time 0 h ( Fig 2A "35% Initial Injury") was sufficient to incite the inflammatory response, ultimately leading to PU formation. However, this injury was not as damaging as repeated cycles of ischemia and reperfusion carried out throughout the complete time course ( Fig 2A "Default Parameters"). Simulations of ischemia and reperfusion injury If tissue damage in the model is caused both by ischemia and by reperfusion, then, all else being equal, a period of ischemia would be expected to cause less tissue damage than the same period of ischemia followed by reperfusion. This ischemia-reperfusion (I/R) damage hypothesis was verified by varying the length of pressure cycles (turning rate) in PUABM simulations while measuring total tissue damage ( Fig 2B ). As expected, both ischemia and reperfusion caused tissue damage. Holding the amount of ischemic time constant while increasing the number of reperfusion events led to earlier ulceration and a more dramatic increase in overall tissue damage upon ulceration. With lower turning frequencies (and thus longer ischemic periods), the amount of damage increased steadily until ulceration became an inevitable outcome, whereas with shorter ischemic periods followed by short recovery, the tissue was able to maintain more health until ROS levels crossed the injury threshold, leading to an ulcer (see S1 Fig ). In both cases, the level of damage incurred long term reached similar levels ( Fig 2B , inset). These in silico experiments were concordant with in vivo studies in which rats were subjected to varying amounts of ischemia followed by reperfusion injury [ 38 ]. As in our simulations, rats with more reperfusion events incurred more damage overall. Simulations of acute inflammation yield two predominant outcomes The rules governing inflammatory mechanisms in the PUABM are based on dynamics of acute inflammation. These dynamics were initially tested in a simulated acute wound, without repeated pressure cycles (Fig 2C and 2D ). As each mediator in the simulation represents an amalgam of several mediators, the dynamics were validated at the cellular level. In a successful implementation of these mechanisms, we hypothesized that tracking the activation of neutrophils and macrophages would reveal cellular dynamics similar to those found in settings of acute inflammation. Interestingly, two characteristic patterns of inflammatory cell dynamics emerged spontaneously from the same set of default parameter values. We explore the potential causes of these two patterns below, but we initially sought to verify that both patterns agree with inflammatory dynamics reported in the literature. As observed in situations of self-sustaining inflammation [ 39 ], all cell counts in the first pattern reached a maximum and were still increasing at the end of the simulation ( Fig 2C ). Conversely, as one would expect to see in a resolving inflammation scenario [ 40 ], every cell type except M2 macrophages peaked well before the end of the simulations, displaying the second characteristic pattern ( Fig 2D ). In such simulations, neutrophil counts dropped to zero relatively early in the time course and did not rise again. Tissue cell counts from simulations corresponding to each of the characteristic patterns suggested that every simulation that had the first pattern of inflammatory dynamics resulted in an ulcer, while all simulations corresponding to the second pattern did not. This was also verified by examining visual outputs from simulations with both types of inflammatory dynamics ( Fig 3 ). 10.1371/journal.pcbi.1004309.g003 Fig 3 Simulations of the PUABM match key features of clinical images. ( A ) Simulations achieved visual appearances with characteristics similar to each stage of PU development. The first row of clinical images come from the National Pressure Ulcer Advisory Panel (images from npuap.org, used with permission) and are of different subjects. Images in the second row are from people with SCI enrolled in a prospective study of PU at each stage. Irregular shapes and increasing nearby damage are observed in both sets of clinical data. ( B ) Numbers indicate days post-injury. Simulated ulcers evolve with visual characteristics that match PU progression observed in people with SCI. Two simulation time courses are matched against one patient from our study. We match key features: irregular shapes, nearby satellite ulcers (open arrows), jagged edges (solid arrows), and decreasing tissue health across the field. Model validation with clinical data In simulations of the PUABM, tissue health begins to decline within the first few hours of pressure cycling, first in the area where pressure intensity is greatest, directly over a simulated bony prominence. This is manifested by a change in color from peach to red (similar to the erythematous appearance of inflamed tissue; notice the area of redness developing over time in the simulations [ Fig 3B , top 2 rows)]. While all of the patients in our cohort who had detectable ulcers at this early stage had light skin tones (and therefore simulations did as well), model parameters could be adjusted to match darker skin tones. The simulated tissue is able to recover some health in the very first rounds of pressure, during periods of reperfusion, but after a certain point, the PUABM tissue field remains red (damaged) despite reflow of oxygen and leukocytes to the region. This is comparable to a region of un-blanchable skin erythema that is the diagnostic criteria for a Stage I PU [ 14 ]. Tissue health declines further as the simulation progresses, increasing the intensity and radius of redness. Eventually, tissue cells begin to die and exit the simulation, leaving behind a white patch to indicate lack of cellular activity at that position (see 4 th panel [day 25] of the top 2 rows of Fig 3B ). The first instance of tissue cell death was considered to be the opening of a PU. From this time onward in the simulation, the PU was observed to grow outward while cells near the edge continued to decline in health. Simulations of the PUABM were initially compared to reference clinical images of PU severity (see Fig 3A ). The National Pressure Ulcer Advisory Panel (NPUAP) has issued guidelines classifying ulcer severity by depth [ 41 ], i.e. layer of tissue affected: skin, fat, muscle, or bone. The prevailing notion is that some ulcers begin as deep tissue injury before opening to the epidermal surface [ 42 ]. While the PUABM is two-dimensional, simulations were nonetheless able to capture the appearance of a variety of PU of various degrees of severity. As shown in Fig 3A , simulations achieved appearances similar to all four stages of ulcer progression. Specifically, the pattern of damaged tissue surrounding the ulcers was similar in simulations and clinical images. In both simulations and clinical ulcers, tissue immediately surrounding the ulcer showed the most damage, while tissue farther away from the pressure point generally appeared healthier (though not without some smaller areas of damage). The simulations also recapitulated evolution of PU development observed in a prospective patient cohort of 49 patients (see Materials and Methods ). The average time to ulceration post-injury was 20 ± 12 days among the RERC subjects. With default parameters in the PUABM, the time to ulceration was 405 ± 6 hours (17 ± 1 days); there was no statistically significant difference between the simulation predictions and the actual RERC ulceration times (p = 0.606 by Mann-Whitney comparison on ranks). We take this agreement to be one measure of validation of the mechanisms encoded in the PUABM and their associated parameter values. The timing of ulceration in the PUABM was not encoded explicitly; in fact, ulceration itself was not encoded in the model. Rather, the progression to ulcer occurred as a result of the accumulation of many smaller tissue-damaging events in a larger context (field of tissue). The parameter values that allowed this progression to agree with clinical data do not govern ulceration itself, but only control these smaller events (e.g. macrophage lifespan, cytokine secretion rates, etc.). When taken together across the whole simulation field, we conclude that PUABM-generated patterns are comparable to images of tissue in human patients. The reference clinical images are static snapshots of a dynamic process; moreover, the images are from different individuals. Accordingly, individual simulations of the PUABM were compared to the dynamics of PU formation in a single individual. Note that clinical days are approximate, and that we sought to determine if the PUABM would result in qualitative concordance with clinical images. Therefore, we allowed for alignment of simulation images to clinical snapshots within a several-day window, especially early in the process of ulcer formation (when it is quite difficult for caregivers to define nascent ulcers). Fig 3B shows snapshots from a single simulation in which PU reached visual appearances with striking similarities to time courses of individual patients. Key characteristics of the PU arise in both simulations and patients. Though the simulated bony prominence is circular while in reality the sacral pressure area is approximately a triangle, both simulated and actual patients develop ulcers with irregular shapes. This is especially noticeable from day 33–34 onward in Fig 3B , for both simulations and clinical images. Furthermore, in both simulations and clinical subjects, once an ulcer formed, a second nearby ulcer was more likely to develop. Secondary ulcers are marked with open arrows; they are apparent in RERC subject showing a stage IV ulcer and in the last simulation panel of Fig 3A , though the ulcer also appears in the third simulation panel in a less severe state. In our simulations, decreased oxygen and increased tissue damage surrounding the primary ulcer contribute to weakened surrounding tissue that is more vulnerable to ulceration than tissue that is farther away. The model also recapitulated irregularly shaped ulcer boundaries: compare the last panel of Fig 3A and both rows of Fig 3B to images from stage IV NPUAP and the clinical time course in Fig 3B . Jagged edges were also noted in both simulations and clinical images, which are indicated with solid arrows and appear in the rightmost panels of Fig 3A and across time courses in Fig 3B . Nearest neighbor analysis suggests that pro-inflammatory mediators drive tissue damage in the PUABM In order to gain insights that may at some point be of diagnostic utility, we sought to understand the range of behaviors that the PUABM can achieve, as well as the conditions that accompany these behaviors. To inspire new treatment strategies, we investigated the likely mechanism by which those outcomes are reached. While the structure of a given computational model can, of course, dictate its ultimate behaviors, much of the behavior of a model can be determined by the values of its parameters [ 43 – 49 ]. Accordingly, model parameters (including initial conditions, rates, thresholds, etc.) were varied alone and in combination with each other in order to gain an understanding of how variations in parameters relate to variations in output (time-evolving health of tissue). Tissue health was quantified by assessing overall damage as indicated by the life score attributed to each tissue cell, as detailed in the Materials and Methods . A second metric for tissue health was presence or absence and size of an ulcer, which was marked by the death of the first tissue cell. Taken together, total tissue damage and presence/ size of an ulcer allowed us to compare one simulation to another and qualify whether the outcome of one was better or worse than the other. Bimodal behavior in simulations with pressure cycles We observed that simulations could reach two distinct endpoints, even when initiated with the exact same parameter values. This could be a significant feature of the model, as a similar dichotomy is also observed in the clinic: two patients with otherwise similar health may have drastically different responses to the same apparent injury and similar subsequent treatment [ 50 ]. An understanding of how the model could arrive stochastically at two outcomes might lead to information that could help a clinician predict which patients might be headed for which outcomes. The dichotomous model behavior was most prominent in simulations of the acute inflammatory response in the absence of pressure, as in Fig 2C and 2D (see also S4 Fig ), but was also observed when pressure cycles were applied ( Fig 4A ). With pressure applied, and other parameters set to default ranges, all simulations led to an ulcer; however, the amount of damage to the tissue surrounding the ulcer differed from simulation to simulation. Reducing this damage would be a desirable clinical outcome, because such damage reduction may prevent the worsening of that ulcer and/or lower the risk of developing secondary ulcers in nearby tissue. Histograms of total damage at t = 1000 from 500 simulations illustrate the complexity of this problem ( Fig 4B , grey bars). Using the Expectation-Maximum algorithm, we found that the data were best approximated by a mixture of two Gaussian models, which suggested that there were in fact two levels of damage among the simulation outputs. In Fig 4B , the peaks estimated by the mixture of two Gaussians model are plotted below the tissue damage histogram for visual comparison. We notice substantial overlap between the two outcomes, and the mixing proportions of these two distributions estimated by the GMM fit are unequal. Using this parameterization of the model, the lower damage outcomes occur in only approximately 10% of the simulations. As with bimodal inflammatory responses to an acute initial injury, this mixing proportion could be tuned by using a different parameterization. However, it is difficult to calibrate this number to clinical outcomes because ulcers that resolve before they open at the surface often go unnoticed [ 51 , 52 ]. 10.1371/journal.pcbi.1004309.g004 Fig 4 Bimodal outcomes are determined by 125 ticks (5 simulated days) prior to the appearance of post-SCI ulcers. Panel ( A ) shows two simulations that were initiated with the same set of parameter values but evolved different levels of damage over time. In panel ( B ) these levels are characterized by a bimodal distribution of tissue damage at tick t = 1000 (day 42). The upper panel is a histogram of these values and the lower panel shows the bimodal distribution estimated from 1000 simulations with the same parameters and initial conditions. ( C ) Some individual features display very different trajectories between resolved and ulcerated simulation outcomes. Means and standard deviations of 1000 simulations are plotted. Upper 3 panels are the features plotted over the first 200 ticks (8.3 days); lower panels show the full time course of 1000 ticks. ( D ) 1-Nearest Neighbor performance versus length of sequences in training set was determined for various numbers of included features. ( E ) Percent of simulations resulting in ulceration after restarting from saved states at indicated restart tick. The saved states were derived from simulations that originally resulted in either ulceration or resolving inflammation. Pro-inflammatory mediators produced during simulations correlate with bimodal tissue damage Once we verified that the bimodal behavior was quantifiable, we aimed to define the variables in the PUABM most predictive of the amount of damage. Accordingly, we examined both visual and statistical patterns in model components (e.g. each cell type and mediator). Using counts of each component as a time-dependent readout, we determined which of these features was most predictive of a particular level of overall damage. At each tick, recordings were made of total amounts of tissue damage, each mediator, oxygen, and counts of each type of activated cell. Analysis of these data suggested that a few features have patterns that differentiate between ulcerated and resolved simulations early in the time course (see S4 Fig ). Certain features of simulation outputs activated-neutrophils , total-tnf , and total-danger appeared to differentiate between outcomes, especially between 100 and 200 ticks ( Fig 4C ). In contrast, several other features ( total-oxygen , blood-flow , total-antiox , total-radical , and total-oxidase ) did not appear to differentiate between endpoints. Re-simulations starting as early as day 5 post-SCI consistently match original simulation outcomes We next sought to find a mathematical description of the decision boundary between simulations that results in higher vs. lower amounts of tissue damage. The features that figured most prominently into this description, we reasoned, would be the ones that are most predictive of damage levels. Furthermore, we expected to find good prediction accuracy with sequences the length of a full simulation (1000 ticks, or 1000 simulated hours—approximately 5 weeks). We were not only interested in prediction accuracy, but also the "minimum prediction length" or the time point before which accuracy begins to decrease. We used the minimum prediction length as an indicator of the time by which the simulation was committed to one outcome or another. The first step was to assess the predictive power of each feature individually. The classification error rate of the 1-Nearest Neighbor Classifier trained on various subsets of the features for the first 200 ticks (8.3 days) or 50 tick (2 day) prefixes is shown in S5 Fig . While classification error ranged from above 60% to just below 40%, the error never stabilized, and no single feature achieved an acceptably low error rate. We next repeated the classification, incorporating information from multiple features into each test and training sequence (a multivariate sequence). We assessed the performance of 1NN using various combinations of features: all 13 features, 8 features that were not inert across the first 200 ticks (all but total-oxygen , blood flow , total-antiox , total-radical , total-oxidase ), and three features that appeared to differ the most among the two outcomes before 200 ticks ( activated-neutrophils , total-tnf , total-danger ( Fig 4C )). We found the classifier performed best when trained on a parsimonious set of three features. While removing inert features did not appear to improve performance ( Fig 4D ), pruning to only 3 features did result in improved performance ( Fig 4D ). A possible explanation for this finding is that the remaining 5 non-inert features contributed noise rather than information that was correlated with outcomes. We expected that in this last case we would achieve a perfect classification (0% classification error) because there was complete separation between the two peaks. By 200 ticks (equal to 200 hours or roughly 8.3 days of real time), we had in fact achieved 0% error rate using 1NN with total-tnf , total-danger , and activated-neutrophils ( Fig 4D ). While 1NN using 3 features was able to separate perfectly low damage from high damage outcomes by 200 ticks (i.e. 8.3 days post-SCI), it is possible that final damage levels were determined even earlier in the simulations. We next sought to examine whether 1NN truly identified the decision time point in the model. To do so, we asked if restarting a simulation from early time points but with a different random seed would lead to the same outcome as the original simulation. For ten simulations (five resulting in ulcers and five resolving), we recorded all state variables at intervals of 25 ticks (approximately 1 day) until t = 200. Starting from each set of saved states, we initiated 20 simulations and carried them out for 500 additional ticks, or about 20 simulated days. This process yielded 100 simulations from each of eight starting time points that were originated with states from resolving simulations, as well as another 100 simulations starting from each time point in ulcerating simulations. The ulceration status of each of these new simulations was noted, and the rate of disagreement with the source simulations was plotted for each starting time point (see Fig 4E ). Contrary to the results from 1NN, the simulations seem to be determined before t = 125. In all but one source simulation, all re-simulation outcomes matched original simulation outcomes by t = 100. These findings imply that there are model configurations determining PU fate from ~4 days post-SCI. It would be worthwhile to discover these exact configurations to determine whether they might be useful in a clinical setting. Sensitivity analysis Global sensitivity analysis of the PUABM suggests that inflammation and oxygen determine to tissue damage, but reperfusion is critical to ulceration The goal of a global sensitivity analysis is to understand how changing model parameter values affect the outcomes produced in simulations. Others have used sensitivity analysis techniques to analytically determine which parameters were most highly associated with metrics of model outcomes [ 43 , 49 , 53 ]. This technique is particularly useful to determine which parameters of a model are most influential toward model outcomes. Because many of our parameters represent lumped estimates or qualitative relationships, and since we were interested in which of only three high-order mechanisms were most important, we hypothesized that we would be able to parse out parameters controlling such high order information without performing an exhaustive sweep of such a large parameter space. We first sought to explore how various model parameters might interact with each other to cause complex tissue damage outcomes. We performed a sensitivity analysis aimed at observing the relative and additive effects of each source of damage. The model includes three sources of tissue damage: lack of oxygen, presence of TNF-α, and presence of ROS ( Fig 5A ). TGF-β1 and IL-1β were not included in this analysis because they do not contribute to tissue damage directly. Thus, TNF-α, ROS, and oxygen are the three broad mechanistic sources of damage, and for each there exists a single broadly effective parameter that controls the magnitude of the effect of a given parameter on a simulation. Biologically, we view these parameters as representing an individual's overall sensitivity to each of the three damage-causing species. The first step in this analysis employed two values for each parameter, corresponding to low (zero-effect) and high (default) levels ( Fig 5B ). The zero-effect level of each parameter was defined as the value at which that source of damage no longer contributed to a simulation (an individual would be completely insensitive to presence or absence of that species). PUABM simulations in which each of three broadly effective parameters was set to either default (definitely has an effect) or zero-effect values show explicitly how each of the sources of damage affects outcomes on its own and in combination with other damage sources ( Fig 5B ). We note that when the tissue is not sensitive to oxygen (left hand column of each panel), no ancillary damage accumulates even when an ulcer appears by t = 400 h. We chose this time point because differences between parameter sets were well pronounced by this time point. Sensitivity assessed at later time points proved to be less informative, as in these extreme conditions the ulcers had already progressed significantly. In several parameterizations, time points prior to t = 400 were also less informative, due to sudden ulceration in otherwise healthy-looking tissue. 10.1371/journal.pcbi.1004309.g005 Fig 5 Sensitivity analysis reveals a unique contribution for all damage mechanisms, but simulated tissue is most sensitive to oxygen. ( A ) We partitioned parameter values according to which damage mechanism they affected. Panel ( B ) shows a global sensitivity analysis comparing the effects of the three mechanisms of damage on tissue health outcomes. Each mechanisms is simulated at two levels: with maximal or negligible effects. In panels C-E, sensitivity analysis was performed on parameters within each mechanistic module. All sensitivity analyses are shown with model snapshots at time t = 400 h (approximately 2.5 weeks). Default parameters are marked with asterisks. ( C ) Tissue sensitivity to oxygen parameters. The parameter governing oxygen production varies in each row. Each column represents a value of the parameter controlling how sensitive the tissue is to local oxygen concentrations. When oxygen production is plentiful, the simulated tissue becomes insensitive to other oxygen parameters (bottom row). Panel ( D ) illustrates tissue sensitivity to pressure intensity and period. Pressure intensity varies in each row. Each column represents a value of pressure cycle length. As the cycle length increases, the number of reperfusion events decreases for the same period of ischemia. Increased pressure leads to more damage, while fewer reperfusion events lead to lower damage at t = 400 h. Panel ( E ) illustrates tissue sensitivity to inflammatory mediators. Each column represents a value of the parameter controlling how sensitive the tissue is to local TNF-α concentrations. The parameter controlling tissue sensitivity to TGF-β1 varies in each row. Increasing sensitivity to TNF-α leads to earlier ulceration and more damage, while increased sensitivity to TGF-β1 leads to decreased tissue damage. The differences between simulations with and without sensitivity to oxygen (top row versus bottom row) are that of tissue damage severity. In all but one column, the bottom simulation (oxygen-sensitive) appears as a more severe version of the upper simulation (oxygen-insensitive). From this analysis, we hypothesize that while oxygen sensitivity contributes to tissue damage, it is not sufficient to cause significant damage or a PU, and furthermore we hypothesize that the lower the threshold of reperfusion injury, the more likely a PU is to form. Comparison of analogous simulations in each of the two panels (left panel is sensitive to ROS, right panel is ROS-insensitive) shows that often it is sensitivity to ROS that is the difference between ulcerated and damaged (but non-ulcerated) tissue in the PUABM. The simulation in the bottom right corner of the second panel shows a PU, but it is smaller and less extensive than the ulcer in the bottom right simulation of the left panel. This snapshot represents the type of PU that forms when parameters dictate that tissue is sensitive to all three damage sources. Based on these analyses, we hypothesize that tissue damage can be ascribed to inflammation, ulceration can be ascribed to ROS, and oxygen can exacerbate damage but may not be sufficient to cause significant tissue damage. Module-based exploration of the PUABM suggests that tissue is sensitive to oxygen and pro-inflammatory mediators We next sought to determine the degree to which parameters within each of the three mechanisms described above contribute to tissue damage. Within each mechanism, we identified parameters that were likely to have the greatest effect due to their mathematical form; i.e. thresholds were more likely to significantly influence outcomes than linear coefficients. Because injury in the PUABM is initiated by repeated pressure cycles, we first examined how sensitivity to pressure affected damage outcomes. To do this, we varied the maximum intensity of pressure (applied to the center of the tissue) and the length of the pressure interval ( Fig 5C ). This number indicates for how many consecutive ticks (simulated hours) pressure is applied, equal to the number of ticks during which pressure is removed. At t = 400 h (16 days) of simulated time post-SCI, very little effect of pressure interval length was observed. However, a slightly lower degree of damage was discerned in the medium-length pressure interval. Pressure intensity appeared to affect tissue health negatively. Pressure influences tissue oxygen in the PUABM. The impact of oxygen on tissue health was therefore investigated by varying in tandem parameters controlling oxygen production (how much oxygen is released from a blood vessel at each time interval) and tissue sensitivity to oxygen (a scalar factor determining the degree to which epithelial cell health is impacted by local oxygen concentrations, same as in global analysis above) ( Fig 5D ). This analysis suggested that when oxygen production from blood vessels in the PUABM is low (or zero), tissue health is affected severely, but in a manner that scales with the parameter that specifies sensitivity to oxygen. At high oxygen production levels, however, there was no apparent sensitivity of simulated tissue health to oxygen concentration. Simulated tissue sensitivities to pro- vs. anti-inflammatory/ pro-healing mediators were compared to each other ( Fig 5E ). This analysis involved comparing the effects of varying the parameter controlling the model's sensitivity to TGF-β1 (the canonical anti-inflammatory/ pro-healing mediator) and the parameter controlling sensitivity to TNF-α (the canonical early-acting pro-inflammatory mediator). Although all simulation snapshots are taken at the same time point, PU appear by t = 400 h (16 d) with higher sensitivity to TNF-α; this does not occur at lower sensitivity to TNF-α. The effect of TGF-β1 on tissue health in the PUABM is less pronounced, though there is overall less damage as the sensitivity to TGF-β1 increases. We examined some of the more detailed mechanisms contributing to the severity of inflammatory response by determining which of thresholds within that module were the most sensitive. The inflammatory response, like most biological processes, is regulated in part through signaling mechanisms that, in essence, define thresholds of activation [ 54 ]. Accordingly, various thresholds in the PUABM control state changes of inflammatory cells (e.g., when local TNF-α concentration is above threshold, macrophages convert to M1 phenotype.) There is a different threshold for each state change, as illustrated in Table 1 and Fig 1B . Sensitivity analysis indicated that the most sensitive threshold was the concentration of DAMPs necessary to activate neutrophils. We found that most thresholds had a small range over which the output was sensitive to the threshold value, and outside of this range the dynamics were stable. This was not true for thresholds controlling activation of M2 macrophages. For both IL-1β and TGF-β1, increasing the concentration necessary to induce M2 activation led to increased simulated tissue damage. In silico clinical trials for post-SCI pressure ulcers Corticosteroids reduce overall tissue damage, but not post-SCI ulceration in simulations In silico (simulated) clinical trials are an inexpensive and increasingly popular means of gleaning translational knowledge from computational models [ 55 – 59 ]. Accordingly, these methods were utilized to test both current and hypothetical or cutting-edge therapies for inflammation in the setting of post-SCI PU. The feasibility of corticosteroids as a treatment was examined for this indication using the PUABM, varying dose and timing of corticosteroids to investigate whether reducing inflammation could delay time of ulceration and/or lead to less tissue damage ( Fig 6 ). Corticosteroid administration was simulated as suppressing all inflammatory mechanisms. This effect was implemented by the following rule: any activated macrophages or neutrophils die upon encountering at least a threshold amount of steroid in their local vicinity of steroid ( S6 Fig ). We note that doses do not correlate directly to clinical doses, but instead to degree of functionality. For example, a simulated high dose (5) of corticosteroids shuts down macrophage and neutrophil activity completely, whereas lower doses all some cells to live and takes longer to see an effect. In simulations where corticosteroids were administered at the highest dose and before 250 hours, total damage was reduced. However, with simulated pressure cycling, PU still formed, suggesting that the lack of inflammation driven by macrophages and neutrophils was insufficient to ameliorate the damaging effects of ischemia and reperfusion ( Fig 6A ). By comparison, simulations with an initial acute injury but without pressure responded more favorably to steroid intervention ( Fig 6B ). These simulation results agree with the global sensitivity analysis of Fig 5B . It appeared that all doses of corticosteroids greater than zero were sufficient to achieve the full effect in both cases in the PUABM. There was, however, more damage incurred when corticosteroids were introduced at later times. While overall damage in simulations in which corticosteroids were introduced late was still less than simulations without corticosteroids, timeliness of corticosteroid application was anti-correlated with tissue damage incurred. This is likely related to the self-perpetuating nature of the pro-inflammatory mediators in this model. We hypothesize that early interruption of the positive feedback cycle was more effective in reducing inflammatory activity than the same amount of treatment later. 10.1371/journal.pcbi.1004309.g006 Fig 6 In silico clinical trials suggest little efficacy for corticosteroids or DAMP inhibitors. Simulations are shown at t = 700 h. We varied both the dose and timing of corticosteroid administration, simulated as an injection into the bloodstream, under ( A ) alternating pressure and ( B ) 40% initial injury conditions. When inflammatory cells were neutralized early enough but pressure continued, overall damage decreased, but ulceration was not prevented. Without continuous pressure cycles, the earliest dose of steroids was successful in stemming ulcer formation, but later applications did not. We then varied both the dose and timing of administration of a neutralizing antibody to HMGB1, simulated as a topical cream applied to the entire field. This targeted approach had ( C ) no apparent effect during simulations with alternating pressure, but ( D ) was able to slow ulcer formation after a 40% initial injury without pressure. Anti-DAMP antibodies have no effect on post-SCI ulceration in simulations DAMPs such as HMGB1 have become leading therapeutic targets for inflammatory indications [ 60 , 61 ]. Accordingly, dose and timing of anti-DAMP administration were varied over wide ranges using the PUABM (implemented as illustrated in S7 Fig ); however, no apparent effect was observed ( Fig 6C ). This lack of efficacy in these simulations is perhaps unsurprising, as there are multiple methods of inciting inflammation encoded in the model. Similarly, in humans and experimental animals [ 33 ] there are several ways by which the body is alerted to cell damage and potentially harmful invaders. In the PUABM, inflammation is incited secondarily when tissue cells that have accumulated damage above a threshold can activate neutrophils directly, thus compensating for the lack of DAMPs. As with corticosteroids, the same treatment was simulated on an initial acute injury without pressure cycles, yielding similar results ( Fig 6D ). Model description and work flow The PUABM represents vascularized soft tissue (skin, fat, muscle, or a composite thereof) overlaying a bony prominence. The central hypothesis of this model is that when this tissue is damaged (which can happen in a variety of ways), further complications can evolve, including inciting further tissue damage. See the Materials and Methods for a more detailed description of model rules. Pseudocode is provided in the online supplement (S1). Epithelial cell health in the PUABM depends on the continuous flow of oxygen to the field of simulated tissue. Pressure—such as that created by shifting a person's weight—is simulated as compressing the tissue and vasculature over a bony prominence, indirectly causing damage to the tissue by restricting the flow of oxygen to the tissue and secondarily by formation of oxygen species (ROS) upon reperfusion (when pressure is relieved). Injured tissue incites activation of inflammatory cells, which bring mediators to the field, some of which injure the tissue further. As simulated cycles of pressure on/off are repeated, these mechanisms are sufficient to generate pressure ulcers in silico . The molecular and cellular components of the model are enumerated in Table 1 . Model verification with literature-derived data Simulations of negative controls After tuning model parameters so that we were studying behaviors in realistic physiological ranges (see Materials and Methods ), there were a few specific mechanisms whose behavior we sought to verify. It was important to verify that the damage accumulated throughout a simulation is in fact due to the mechanisms of injury encoded in the model, and not to other spurious mechanisms. Unperturbed, the tissue remained "healthy" for thousands of simulated hours ( Fig 2A "No Pressure" data shown up to 1000 h; this behavior remained unchanged up to 5000 h). Some damage was observed when tissue was subjected to 12 h of simulated ischemia ( Fig 2A "12 h constant pressure"), but no additional injury to the tissue was observed after release of pressure. By comparison, starting with a 35% injury over a patch of tissue at time 0 h ( Fig 2A "35% Initial Injury") was sufficient to incite the inflammatory response, ultimately leading to PU formation. However, this injury was not as damaging as repeated cycles of ischemia and reperfusion carried out throughout the complete time course ( Fig 2A "Default Parameters"). Simulations of ischemia and reperfusion injury If tissue damage in the model is caused both by ischemia and by reperfusion, then, all else being equal, a period of ischemia would be expected to cause less tissue damage than the same period of ischemia followed by reperfusion. This ischemia-reperfusion (I/R) damage hypothesis was verified by varying the length of pressure cycles (turning rate) in PUABM simulations while measuring total tissue damage ( Fig 2B ). As expected, both ischemia and reperfusion caused tissue damage. Holding the amount of ischemic time constant while increasing the number of reperfusion events led to earlier ulceration and a more dramatic increase in overall tissue damage upon ulceration. With lower turning frequencies (and thus longer ischemic periods), the amount of damage increased steadily until ulceration became an inevitable outcome, whereas with shorter ischemic periods followed by short recovery, the tissue was able to maintain more health until ROS levels crossed the injury threshold, leading to an ulcer (see S1 Fig ). In both cases, the level of damage incurred long term reached similar levels ( Fig 2B , inset). These in silico experiments were concordant with in vivo studies in which rats were subjected to varying amounts of ischemia followed by reperfusion injury [ 38 ]. As in our simulations, rats with more reperfusion events incurred more damage overall. Simulations of acute inflammation yield two predominant outcomes The rules governing inflammatory mechanisms in the PUABM are based on dynamics of acute inflammation. These dynamics were initially tested in a simulated acute wound, without repeated pressure cycles (Fig 2C and 2D ). As each mediator in the simulation represents an amalgam of several mediators, the dynamics were validated at the cellular level. In a successful implementation of these mechanisms, we hypothesized that tracking the activation of neutrophils and macrophages would reveal cellular dynamics similar to those found in settings of acute inflammation. Interestingly, two characteristic patterns of inflammatory cell dynamics emerged spontaneously from the same set of default parameter values. We explore the potential causes of these two patterns below, but we initially sought to verify that both patterns agree with inflammatory dynamics reported in the literature. As observed in situations of self-sustaining inflammation [ 39 ], all cell counts in the first pattern reached a maximum and were still increasing at the end of the simulation ( Fig 2C ). Conversely, as one would expect to see in a resolving inflammation scenario [ 40 ], every cell type except M2 macrophages peaked well before the end of the simulations, displaying the second characteristic pattern ( Fig 2D ). In such simulations, neutrophil counts dropped to zero relatively early in the time course and did not rise again. Tissue cell counts from simulations corresponding to each of the characteristic patterns suggested that every simulation that had the first pattern of inflammatory dynamics resulted in an ulcer, while all simulations corresponding to the second pattern did not. This was also verified by examining visual outputs from simulations with both types of inflammatory dynamics ( Fig 3 ). 10.1371/journal.pcbi.1004309.g003 Fig 3 Simulations of the PUABM match key features of clinical images. ( A ) Simulations achieved visual appearances with characteristics similar to each stage of PU development. The first row of clinical images come from the National Pressure Ulcer Advisory Panel (images from npuap.org, used with permission) and are of different subjects. Images in the second row are from people with SCI enrolled in a prospective study of PU at each stage. Irregular shapes and increasing nearby damage are observed in both sets of clinical data. ( B ) Numbers indicate days post-injury. Simulated ulcers evolve with visual characteristics that match PU progression observed in people with SCI. Two simulation time courses are matched against one patient from our study. We match key features: irregular shapes, nearby satellite ulcers (open arrows), jagged edges (solid arrows), and decreasing tissue health across the field. Simulations of negative controls After tuning model parameters so that we were studying behaviors in realistic physiological ranges (see Materials and Methods ), there were a few specific mechanisms whose behavior we sought to verify. It was important to verify that the damage accumulated throughout a simulation is in fact due to the mechanisms of injury encoded in the model, and not to other spurious mechanisms. Unperturbed, the tissue remained "healthy" for thousands of simulated hours ( Fig 2A "No Pressure" data shown up to 1000 h; this behavior remained unchanged up to 5000 h). Some damage was observed when tissue was subjected to 12 h of simulated ischemia ( Fig 2A "12 h constant pressure"), but no additional injury to the tissue was observed after release of pressure. By comparison, starting with a 35% injury over a patch of tissue at time 0 h ( Fig 2A "35% Initial Injury") was sufficient to incite the inflammatory response, ultimately leading to PU formation. However, this injury was not as damaging as repeated cycles of ischemia and reperfusion carried out throughout the complete time course ( Fig 2A "Default Parameters"). Simulations of ischemia and reperfusion injury If tissue damage in the model is caused both by ischemia and by reperfusion, then, all else being equal, a period of ischemia would be expected to cause less tissue damage than the same period of ischemia followed by reperfusion. This ischemia-reperfusion (I/R) damage hypothesis was verified by varying the length of pressure cycles (turning rate) in PUABM simulations while measuring total tissue damage ( Fig 2B ). As expected, both ischemia and reperfusion caused tissue damage. Holding the amount of ischemic time constant while increasing the number of reperfusion events led to earlier ulceration and a more dramatic increase in overall tissue damage upon ulceration. With lower turning frequencies (and thus longer ischemic periods), the amount of damage increased steadily until ulceration became an inevitable outcome, whereas with shorter ischemic periods followed by short recovery, the tissue was able to maintain more health until ROS levels crossed the injury threshold, leading to an ulcer (see S1 Fig ). In both cases, the level of damage incurred long term reached similar levels ( Fig 2B , inset). These in silico experiments were concordant with in vivo studies in which rats were subjected to varying amounts of ischemia followed by reperfusion injury [ 38 ]. As in our simulations, rats with more reperfusion events incurred more damage overall. Simulations of acute inflammation yield two predominant outcomes The rules governing inflammatory mechanisms in the PUABM are based on dynamics of acute inflammation. These dynamics were initially tested in a simulated acute wound, without repeated pressure cycles (Fig 2C and 2D ). As each mediator in the simulation represents an amalgam of several mediators, the dynamics were validated at the cellular level. In a successful implementation of these mechanisms, we hypothesized that tracking the activation of neutrophils and macrophages would reveal cellular dynamics similar to those found in settings of acute inflammation. Interestingly, two characteristic patterns of inflammatory cell dynamics emerged spontaneously from the same set of default parameter values. We explore the potential causes of these two patterns below, but we initially sought to verify that both patterns agree with inflammatory dynamics reported in the literature. As observed in situations of self-sustaining inflammation [ 39 ], all cell counts in the first pattern reached a maximum and were still increasing at the end of the simulation ( Fig 2C ). Conversely, as one would expect to see in a resolving inflammation scenario [ 40 ], every cell type except M2 macrophages peaked well before the end of the simulations, displaying the second characteristic pattern ( Fig 2D ). In such simulations, neutrophil counts dropped to zero relatively early in the time course and did not rise again. Tissue cell counts from simulations corresponding to each of the characteristic patterns suggested that every simulation that had the first pattern of inflammatory dynamics resulted in an ulcer, while all simulations corresponding to the second pattern did not. This was also verified by examining visual outputs from simulations with both types of inflammatory dynamics ( Fig 3 ). 10.1371/journal.pcbi.1004309.g003 Fig 3 Simulations of the PUABM match key features of clinical images. ( A ) Simulations achieved visual appearances with characteristics similar to each stage of PU development. The first row of clinical images come from the National Pressure Ulcer Advisory Panel (images from npuap.org, used with permission) and are of different subjects. Images in the second row are from people with SCI enrolled in a prospective study of PU at each stage. Irregular shapes and increasing nearby damage are observed in both sets of clinical data. ( B ) Numbers indicate days post-injury. Simulated ulcers evolve with visual characteristics that match PU progression observed in people with SCI. Two simulation time courses are matched against one patient from our study. We match key features: irregular shapes, nearby satellite ulcers (open arrows), jagged edges (solid arrows), and decreasing tissue health across the field. Model validation with clinical data In simulations of the PUABM, tissue health begins to decline within the first few hours of pressure cycling, first in the area where pressure intensity is greatest, directly over a simulated bony prominence. This is manifested by a change in color from peach to red (similar to the erythematous appearance of inflamed tissue; notice the area of redness developing over time in the simulations [ Fig 3B , top 2 rows)]. While all of the patients in our cohort who had detectable ulcers at this early stage had light skin tones (and therefore simulations did as well), model parameters could be adjusted to match darker skin tones. The simulated tissue is able to recover some health in the very first rounds of pressure, during periods of reperfusion, but after a certain point, the PUABM tissue field remains red (damaged) despite reflow of oxygen and leukocytes to the region. This is comparable to a region of un-blanchable skin erythema that is the diagnostic criteria for a Stage I PU [ 14 ]. Tissue health declines further as the simulation progresses, increasing the intensity and radius of redness. Eventually, tissue cells begin to die and exit the simulation, leaving behind a white patch to indicate lack of cellular activity at that position (see 4 th panel [day 25] of the top 2 rows of Fig 3B ). The first instance of tissue cell death was considered to be the opening of a PU. From this time onward in the simulation, the PU was observed to grow outward while cells near the edge continued to decline in health. Simulations of the PUABM were initially compared to reference clinical images of PU severity (see Fig 3A ). The National Pressure Ulcer Advisory Panel (NPUAP) has issued guidelines classifying ulcer severity by depth [ 41 ], i.e. layer of tissue affected: skin, fat, muscle, or bone. The prevailing notion is that some ulcers begin as deep tissue injury before opening to the epidermal surface [ 42 ]. While the PUABM is two-dimensional, simulations were nonetheless able to capture the appearance of a variety of PU of various degrees of severity. As shown in Fig 3A , simulations achieved appearances similar to all four stages of ulcer progression. Specifically, the pattern of damaged tissue surrounding the ulcers was similar in simulations and clinical images. In both simulations and clinical ulcers, tissue immediately surrounding the ulcer showed the most damage, while tissue farther away from the pressure point generally appeared healthier (though not without some smaller areas of damage). The simulations also recapitulated evolution of PU development observed in a prospective patient cohort of 49 patients (see Materials and Methods ). The average time to ulceration post-injury was 20 ± 12 days among the RERC subjects. With default parameters in the PUABM, the time to ulceration was 405 ± 6 hours (17 ± 1 days); there was no statistically significant difference between the simulation predictions and the actual RERC ulceration times (p = 0.606 by Mann-Whitney comparison on ranks). We take this agreement to be one measure of validation of the mechanisms encoded in the PUABM and their associated parameter values. The timing of ulceration in the PUABM was not encoded explicitly; in fact, ulceration itself was not encoded in the model. Rather, the progression to ulcer occurred as a result of the accumulation of many smaller tissue-damaging events in a larger context (field of tissue). The parameter values that allowed this progression to agree with clinical data do not govern ulceration itself, but only control these smaller events (e.g. macrophage lifespan, cytokine secretion rates, etc.). When taken together across the whole simulation field, we conclude that PUABM-generated patterns are comparable to images of tissue in human patients. The reference clinical images are static snapshots of a dynamic process; moreover, the images are from different individuals. Accordingly, individual simulations of the PUABM were compared to the dynamics of PU formation in a single individual. Note that clinical days are approximate, and that we sought to determine if the PUABM would result in qualitative concordance with clinical images. Therefore, we allowed for alignment of simulation images to clinical snapshots within a several-day window, especially early in the process of ulcer formation (when it is quite difficult for caregivers to define nascent ulcers). Fig 3B shows snapshots from a single simulation in which PU reached visual appearances with striking similarities to time courses of individual patients. Key characteristics of the PU arise in both simulations and patients. Though the simulated bony prominence is circular while in reality the sacral pressure area is approximately a triangle, both simulated and actual patients develop ulcers with irregular shapes. This is especially noticeable from day 33–34 onward in Fig 3B , for both simulations and clinical images. Furthermore, in both simulations and clinical subjects, once an ulcer formed, a second nearby ulcer was more likely to develop. Secondary ulcers are marked with open arrows; they are apparent in RERC subject showing a stage IV ulcer and in the last simulation panel of Fig 3A , though the ulcer also appears in the third simulation panel in a less severe state. In our simulations, decreased oxygen and increased tissue damage surrounding the primary ulcer contribute to weakened surrounding tissue that is more vulnerable to ulceration than tissue that is farther away. The model also recapitulated irregularly shaped ulcer boundaries: compare the last panel of Fig 3A and both rows of Fig 3B to images from stage IV NPUAP and the clinical time course in Fig 3B . Jagged edges were also noted in both simulations and clinical images, which are indicated with solid arrows and appear in the rightmost panels of Fig 3A and across time courses in Fig 3B . Nearest neighbor analysis suggests that pro-inflammatory mediators drive tissue damage in the PUABM In order to gain insights that may at some point be of diagnostic utility, we sought to understand the range of behaviors that the PUABM can achieve, as well as the conditions that accompany these behaviors. To inspire new treatment strategies, we investigated the likely mechanism by which those outcomes are reached. While the structure of a given computational model can, of course, dictate its ultimate behaviors, much of the behavior of a model can be determined by the values of its parameters [ 43 – 49 ]. Accordingly, model parameters (including initial conditions, rates, thresholds, etc.) were varied alone and in combination with each other in order to gain an understanding of how variations in parameters relate to variations in output (time-evolving health of tissue). Tissue health was quantified by assessing overall damage as indicated by the life score attributed to each tissue cell, as detailed in the Materials and Methods . A second metric for tissue health was presence or absence and size of an ulcer, which was marked by the death of the first tissue cell. Taken together, total tissue damage and presence/ size of an ulcer allowed us to compare one simulation to another and qualify whether the outcome of one was better or worse than the other. Bimodal behavior in simulations with pressure cycles We observed that simulations could reach two distinct endpoints, even when initiated with the exact same parameter values. This could be a significant feature of the model, as a similar dichotomy is also observed in the clinic: two patients with otherwise similar health may have drastically different responses to the same apparent injury and similar subsequent treatment [ 50 ]. An understanding of how the model could arrive stochastically at two outcomes might lead to information that could help a clinician predict which patients might be headed for which outcomes. The dichotomous model behavior was most prominent in simulations of the acute inflammatory response in the absence of pressure, as in Fig 2C and 2D (see also S4 Fig ), but was also observed when pressure cycles were applied ( Fig 4A ). With pressure applied, and other parameters set to default ranges, all simulations led to an ulcer; however, the amount of damage to the tissue surrounding the ulcer differed from simulation to simulation. Reducing this damage would be a desirable clinical outcome, because such damage reduction may prevent the worsening of that ulcer and/or lower the risk of developing secondary ulcers in nearby tissue. Histograms of total damage at t = 1000 from 500 simulations illustrate the complexity of this problem ( Fig 4B , grey bars). Using the Expectation-Maximum algorithm, we found that the data were best approximated by a mixture of two Gaussian models, which suggested that there were in fact two levels of damage among the simulation outputs. In Fig 4B , the peaks estimated by the mixture of two Gaussians model are plotted below the tissue damage histogram for visual comparison. We notice substantial overlap between the two outcomes, and the mixing proportions of these two distributions estimated by the GMM fit are unequal. Using this parameterization of the model, the lower damage outcomes occur in only approximately 10% of the simulations. As with bimodal inflammatory responses to an acute initial injury, this mixing proportion could be tuned by using a different parameterization. However, it is difficult to calibrate this number to clinical outcomes because ulcers that resolve before they open at the surface often go unnoticed [ 51 , 52 ]. 10.1371/journal.pcbi.1004309.g004 Fig 4 Bimodal outcomes are determined by 125 ticks (5 simulated days) prior to the appearance of post-SCI ulcers. Panel ( A ) shows two simulations that were initiated with the same set of parameter values but evolved different levels of damage over time. In panel ( B ) these levels are characterized by a bimodal distribution of tissue damage at tick t = 1000 (day 42). The upper panel is a histogram of these values and the lower panel shows the bimodal distribution estimated from 1000 simulations with the same parameters and initial conditions. ( C ) Some individual features display very different trajectories between resolved and ulcerated simulation outcomes. Means and standard deviations of 1000 simulations are plotted. Upper 3 panels are the features plotted over the first 200 ticks (8.3 days); lower panels show the full time course of 1000 ticks. ( D ) 1-Nearest Neighbor performance versus length of sequences in training set was determined for various numbers of included features. ( E ) Percent of simulations resulting in ulceration after restarting from saved states at indicated restart tick. The saved states were derived from simulations that originally resulted in either ulceration or resolving inflammation. Pro-inflammatory mediators produced during simulations correlate with bimodal tissue damage Once we verified that the bimodal behavior was quantifiable, we aimed to define the variables in the PUABM most predictive of the amount of damage. Accordingly, we examined both visual and statistical patterns in model components (e.g. each cell type and mediator). Using counts of each component as a time-dependent readout, we determined which of these features was most predictive of a particular level of overall damage. At each tick, recordings were made of total amounts of tissue damage, each mediator, oxygen, and counts of each type of activated cell. Analysis of these data suggested that a few features have patterns that differentiate between ulcerated and resolved simulations early in the time course (see S4 Fig ). Certain features of simulation outputs activated-neutrophils , total-tnf , and total-danger appeared to differentiate between outcomes, especially between 100 and 200 ticks ( Fig 4C ). In contrast, several other features ( total-oxygen , blood-flow , total-antiox , total-radical , and total-oxidase ) did not appear to differentiate between endpoints. Re-simulations starting as early as day 5 post-SCI consistently match original simulation outcomes We next sought to find a mathematical description of the decision boundary between simulations that results in higher vs. lower amounts of tissue damage. The features that figured most prominently into this description, we reasoned, would be the ones that are most predictive of damage levels. Furthermore, we expected to find good prediction accuracy with sequences the length of a full simulation (1000 ticks, or 1000 simulated hours—approximately 5 weeks). We were not only interested in prediction accuracy, but also the "minimum prediction length" or the time point before which accuracy begins to decrease. We used the minimum prediction length as an indicator of the time by which the simulation was committed to one outcome or another. The first step was to assess the predictive power of each feature individually. The classification error rate of the 1-Nearest Neighbor Classifier trained on various subsets of the features for the first 200 ticks (8.3 days) or 50 tick (2 day) prefixes is shown in S5 Fig . While classification error ranged from above 60% to just below 40%, the error never stabilized, and no single feature achieved an acceptably low error rate. We next repeated the classification, incorporating information from multiple features into each test and training sequence (a multivariate sequence). We assessed the performance of 1NN using various combinations of features: all 13 features, 8 features that were not inert across the first 200 ticks (all but total-oxygen , blood flow , total-antiox , total-radical , total-oxidase ), and three features that appeared to differ the most among the two outcomes before 200 ticks ( activated-neutrophils , total-tnf , total-danger ( Fig 4C )). We found the classifier performed best when trained on a parsimonious set of three features. While removing inert features did not appear to improve performance ( Fig 4D ), pruning to only 3 features did result in improved performance ( Fig 4D ). A possible explanation for this finding is that the remaining 5 non-inert features contributed noise rather than information that was correlated with outcomes. We expected that in this last case we would achieve a perfect classification (0% classification error) because there was complete separation between the two peaks. By 200 ticks (equal to 200 hours or roughly 8.3 days of real time), we had in fact achieved 0% error rate using 1NN with total-tnf , total-danger , and activated-neutrophils ( Fig 4D ). While 1NN using 3 features was able to separate perfectly low damage from high damage outcomes by 200 ticks (i.e. 8.3 days post-SCI), it is possible that final damage levels were determined even earlier in the simulations. We next sought to examine whether 1NN truly identified the decision time point in the model. To do so, we asked if restarting a simulation from early time points but with a different random seed would lead to the same outcome as the original simulation. For ten simulations (five resulting in ulcers and five resolving), we recorded all state variables at intervals of 25 ticks (approximately 1 day) until t = 200. Starting from each set of saved states, we initiated 20 simulations and carried them out for 500 additional ticks, or about 20 simulated days. This process yielded 100 simulations from each of eight starting time points that were originated with states from resolving simulations, as well as another 100 simulations starting from each time point in ulcerating simulations. The ulceration status of each of these new simulations was noted, and the rate of disagreement with the source simulations was plotted for each starting time point (see Fig 4E ). Contrary to the results from 1NN, the simulations seem to be determined before t = 125. In all but one source simulation, all re-simulation outcomes matched original simulation outcomes by t = 100. These findings imply that there are model configurations determining PU fate from ~4 days post-SCI. It would be worthwhile to discover these exact configurations to determine whether they might be useful in a clinical setting. Bimodal behavior in simulations with pressure cycles We observed that simulations could reach two distinct endpoints, even when initiated with the exact same parameter values. This could be a significant feature of the model, as a similar dichotomy is also observed in the clinic: two patients with otherwise similar health may have drastically different responses to the same apparent injury and similar subsequent treatment [ 50 ]. An understanding of how the model could arrive stochastically at two outcomes might lead to information that could help a clinician predict which patients might be headed for which outcomes. The dichotomous model behavior was most prominent in simulations of the acute inflammatory response in the absence of pressure, as in Fig 2C and 2D (see also S4 Fig ), but was also observed when pressure cycles were applied ( Fig 4A ). With pressure applied, and other parameters set to default ranges, all simulations led to an ulcer; however, the amount of damage to the tissue surrounding the ulcer differed from simulation to simulation. Reducing this damage would be a desirable clinical outcome, because such damage reduction may prevent the worsening of that ulcer and/or lower the risk of developing secondary ulcers in nearby tissue. Histograms of total damage at t = 1000 from 500 simulations illustrate the complexity of this problem ( Fig 4B , grey bars). Using the Expectation-Maximum algorithm, we found that the data were best approximated by a mixture of two Gaussian models, which suggested that there were in fact two levels of damage among the simulation outputs. In Fig 4B , the peaks estimated by the mixture of two Gaussians model are plotted below the tissue damage histogram for visual comparison. We notice substantial overlap between the two outcomes, and the mixing proportions of these two distributions estimated by the GMM fit are unequal. Using this parameterization of the model, the lower damage outcomes occur in only approximately 10% of the simulations. As with bimodal inflammatory responses to an acute initial injury, this mixing proportion could be tuned by using a different parameterization. However, it is difficult to calibrate this number to clinical outcomes because ulcers that resolve before they open at the surface often go unnoticed [ 51 , 52 ]. 10.1371/journal.pcbi.1004309.g004 Fig 4 Bimodal outcomes are determined by 125 ticks (5 simulated days) prior to the appearance of post-SCI ulcers. Panel ( A ) shows two simulations that were initiated with the same set of parameter values but evolved different levels of damage over time. In panel ( B ) these levels are characterized by a bimodal distribution of tissue damage at tick t = 1000 (day 42). The upper panel is a histogram of these values and the lower panel shows the bimodal distribution estimated from 1000 simulations with the same parameters and initial conditions. ( C ) Some individual features display very different trajectories between resolved and ulcerated simulation outcomes. Means and standard deviations of 1000 simulations are plotted. Upper 3 panels are the features plotted over the first 200 ticks (8.3 days); lower panels show the full time course of 1000 ticks. ( D ) 1-Nearest Neighbor performance versus length of sequences in training set was determined for various numbers of included features. ( E ) Percent of simulations resulting in ulceration after restarting from saved states at indicated restart tick. The saved states were derived from simulations that originally resulted in either ulceration or resolving inflammation. Pro-inflammatory mediators produced during simulations correlate with bimodal tissue damage Once we verified that the bimodal behavior was quantifiable, we aimed to define the variables in the PUABM most predictive of the amount of damage. Accordingly, we examined both visual and statistical patterns in model components (e.g. each cell type and mediator). Using counts of each component as a time-dependent readout, we determined which of these features was most predictive of a particular level of overall damage. At each tick, recordings were made of total amounts of tissue damage, each mediator, oxygen, and counts of each type of activated cell. Analysis of these data suggested that a few features have patterns that differentiate between ulcerated and resolved simulations early in the time course (see S4 Fig ). Certain features of simulation outputs activated-neutrophils , total-tnf , and total-danger appeared to differentiate between outcomes, especially between 100 and 200 ticks ( Fig 4C ). In contrast, several other features ( total-oxygen , blood-flow , total-antiox , total-radical , and total-oxidase ) did not appear to differentiate between endpoints. Re-simulations starting as early as day 5 post-SCI consistently match original simulation outcomes We next sought to find a mathematical description of the decision boundary between simulations that results in higher vs. lower amounts of tissue damage. The features that figured most prominently into this description, we reasoned, would be the ones that are most predictive of damage levels. Furthermore, we expected to find good prediction accuracy with sequences the length of a full simulation (1000 ticks, or 1000 simulated hours—approximately 5 weeks). We were not only interested in prediction accuracy, but also the "minimum prediction length" or the time point before which accuracy begins to decrease. We used the minimum prediction length as an indicator of the time by which the simulation was committed to one outcome or another. The first step was to assess the predictive power of each feature individually. The classification error rate of the 1-Nearest Neighbor Classifier trained on various subsets of the features for the first 200 ticks (8.3 days) or 50 tick (2 day) prefixes is shown in S5 Fig . While classification error ranged from above 60% to just below 40%, the error never stabilized, and no single feature achieved an acceptably low error rate. We next repeated the classification, incorporating information from multiple features into each test and training sequence (a multivariate sequence). We assessed the performance of 1NN using various combinations of features: all 13 features, 8 features that were not inert across the first 200 ticks (all but total-oxygen , blood flow , total-antiox , total-radical , total-oxidase ), and three features that appeared to differ the most among the two outcomes before 200 ticks ( activated-neutrophils , total-tnf , total-danger ( Fig 4C )). We found the classifier performed best when trained on a parsimonious set of three features. While removing inert features did not appear to improve performance ( Fig 4D ), pruning to only 3 features did result in improved performance ( Fig 4D ). A possible explanation for this finding is that the remaining 5 non-inert features contributed noise rather than information that was correlated with outcomes. We expected that in this last case we would achieve a perfect classification (0% classification error) because there was complete separation between the two peaks. By 200 ticks (equal to 200 hours or roughly 8.3 days of real time), we had in fact achieved 0% error rate using 1NN with total-tnf , total-danger , and activated-neutrophils ( Fig 4D ). While 1NN using 3 features was able to separate perfectly low damage from high damage outcomes by 200 ticks (i.e. 8.3 days post-SCI), it is possible that final damage levels were determined even earlier in the simulations. We next sought to examine whether 1NN truly identified the decision time point in the model. To do so, we asked if restarting a simulation from early time points but with a different random seed would lead to the same outcome as the original simulation. For ten simulations (five resulting in ulcers and five resolving), we recorded all state variables at intervals of 25 ticks (approximately 1 day) until t = 200. Starting from each set of saved states, we initiated 20 simulations and carried them out for 500 additional ticks, or about 20 simulated days. This process yielded 100 simulations from each of eight starting time points that were originated with states from resolving simulations, as well as another 100 simulations starting from each time point in ulcerating simulations. The ulceration status of each of these new simulations was noted, and the rate of disagreement with the source simulations was plotted for each starting time point (see Fig 4E ). Contrary to the results from 1NN, the simulations seem to be determined before t = 125. In all but one source simulation, all re-simulation outcomes matched original simulation outcomes by t = 100. These findings imply that there are model configurations determining PU fate from ~4 days post-SCI. It would be worthwhile to discover these exact configurations to determine whether they might be useful in a clinical setting. Sensitivity analysis Global sensitivity analysis of the PUABM suggests that inflammation and oxygen determine to tissue damage, but reperfusion is critical to ulceration The goal of a global sensitivity analysis is to understand how changing model parameter values affect the outcomes produced in simulations. Others have used sensitivity analysis techniques to analytically determine which parameters were most highly associated with metrics of model outcomes [ 43 , 49 , 53 ]. This technique is particularly useful to determine which parameters of a model are most influential toward model outcomes. Because many of our parameters represent lumped estimates or qualitative relationships, and since we were interested in which of only three high-order mechanisms were most important, we hypothesized that we would be able to parse out parameters controlling such high order information without performing an exhaustive sweep of such a large parameter space. We first sought to explore how various model parameters might interact with each other to cause complex tissue damage outcomes. We performed a sensitivity analysis aimed at observing the relative and additive effects of each source of damage. The model includes three sources of tissue damage: lack of oxygen, presence of TNF-α, and presence of ROS ( Fig 5A ). TGF-β1 and IL-1β were not included in this analysis because they do not contribute to tissue damage directly. Thus, TNF-α, ROS, and oxygen are the three broad mechanistic sources of damage, and for each there exists a single broadly effective parameter that controls the magnitude of the effect of a given parameter on a simulation. Biologically, we view these parameters as representing an individual's overall sensitivity to each of the three damage-causing species. The first step in this analysis employed two values for each parameter, corresponding to low (zero-effect) and high (default) levels ( Fig 5B ). The zero-effect level of each parameter was defined as the value at which that source of damage no longer contributed to a simulation (an individual would be completely insensitive to presence or absence of that species). PUABM simulations in which each of three broadly effective parameters was set to either default (definitely has an effect) or zero-effect values show explicitly how each of the sources of damage affects outcomes on its own and in combination with other damage sources ( Fig 5B ). We note that when the tissue is not sensitive to oxygen (left hand column of each panel), no ancillary damage accumulates even when an ulcer appears by t = 400 h. We chose this time point because differences between parameter sets were well pronounced by this time point. Sensitivity assessed at later time points proved to be less informative, as in these extreme conditions the ulcers had already progressed significantly. In several parameterizations, time points prior to t = 400 were also less informative, due to sudden ulceration in otherwise healthy-looking tissue. 10.1371/journal.pcbi.1004309.g005 Fig 5 Sensitivity analysis reveals a unique contribution for all damage mechanisms, but simulated tissue is most sensitive to oxygen. ( A ) We partitioned parameter values according to which damage mechanism they affected. Panel ( B ) shows a global sensitivity analysis comparing the effects of the three mechanisms of damage on tissue health outcomes. Each mechanisms is simulated at two levels: with maximal or negligible effects. In panels C-E, sensitivity analysis was performed on parameters within each mechanistic module. All sensitivity analyses are shown with model snapshots at time t = 400 h (approximately 2.5 weeks). Default parameters are marked with asterisks. ( C ) Tissue sensitivity to oxygen parameters. The parameter governing oxygen production varies in each row. Each column represents a value of the parameter controlling how sensitive the tissue is to local oxygen concentrations. When oxygen production is plentiful, the simulated tissue becomes insensitive to other oxygen parameters (bottom row). Panel ( D ) illustrates tissue sensitivity to pressure intensity and period. Pressure intensity varies in each row. Each column represents a value of pressure cycle length. As the cycle length increases, the number of reperfusion events decreases for the same period of ischemia. Increased pressure leads to more damage, while fewer reperfusion events lead to lower damage at t = 400 h. Panel ( E ) illustrates tissue sensitivity to inflammatory mediators. Each column represents a value of the parameter controlling how sensitive the tissue is to local TNF-α concentrations. The parameter controlling tissue sensitivity to TGF-β1 varies in each row. Increasing sensitivity to TNF-α leads to earlier ulceration and more damage, while increased sensitivity to TGF-β1 leads to decreased tissue damage. The differences between simulations with and without sensitivity to oxygen (top row versus bottom row) are that of tissue damage severity. In all but one column, the bottom simulation (oxygen-sensitive) appears as a more severe version of the upper simulation (oxygen-insensitive). From this analysis, we hypothesize that while oxygen sensitivity contributes to tissue damage, it is not sufficient to cause significant damage or a PU, and furthermore we hypothesize that the lower the threshold of reperfusion injury, the more likely a PU is to form. Comparison of analogous simulations in each of the two panels (left panel is sensitive to ROS, right panel is ROS-insensitive) shows that often it is sensitivity to ROS that is the difference between ulcerated and damaged (but non-ulcerated) tissue in the PUABM. The simulation in the bottom right corner of the second panel shows a PU, but it is smaller and less extensive than the ulcer in the bottom right simulation of the left panel. This snapshot represents the type of PU that forms when parameters dictate that tissue is sensitive to all three damage sources. Based on these analyses, we hypothesize that tissue damage can be ascribed to inflammation, ulceration can be ascribed to ROS, and oxygen can exacerbate damage but may not be sufficient to cause significant tissue damage. Module-based exploration of the PUABM suggests that tissue is sensitive to oxygen and pro-inflammatory mediators We next sought to determine the degree to which parameters within each of the three mechanisms described above contribute to tissue damage. Within each mechanism, we identified parameters that were likely to have the greatest effect due to their mathematical form; i.e. thresholds were more likely to significantly influence outcomes than linear coefficients. Because injury in the PUABM is initiated by repeated pressure cycles, we first examined how sensitivity to pressure affected damage outcomes. To do this, we varied the maximum intensity of pressure (applied to the center of the tissue) and the length of the pressure interval ( Fig 5C ). This number indicates for how many consecutive ticks (simulated hours) pressure is applied, equal to the number of ticks during which pressure is removed. At t = 400 h (16 days) of simulated time post-SCI, very little effect of pressure interval length was observed. However, a slightly lower degree of damage was discerned in the medium-length pressure interval. Pressure intensity appeared to affect tissue health negatively. Pressure influences tissue oxygen in the PUABM. The impact of oxygen on tissue health was therefore investigated by varying in tandem parameters controlling oxygen production (how much oxygen is released from a blood vessel at each time interval) and tissue sensitivity to oxygen (a scalar factor determining the degree to which epithelial cell health is impacted by local oxygen concentrations, same as in global analysis above) ( Fig 5D ). This analysis suggested that when oxygen production from blood vessels in the PUABM is low (or zero), tissue health is affected severely, but in a manner that scales with the parameter that specifies sensitivity to oxygen. At high oxygen production levels, however, there was no apparent sensitivity of simulated tissue health to oxygen concentration. Simulated tissue sensitivities to pro- vs. anti-inflammatory/ pro-healing mediators were compared to each other ( Fig 5E ). This analysis involved comparing the effects of varying the parameter controlling the model's sensitivity to TGF-β1 (the canonical anti-inflammatory/ pro-healing mediator) and the parameter controlling sensitivity to TNF-α (the canonical early-acting pro-inflammatory mediator). Although all simulation snapshots are taken at the same time point, PU appear by t = 400 h (16 d) with higher sensitivity to TNF-α; this does not occur at lower sensitivity to TNF-α. The effect of TGF-β1 on tissue health in the PUABM is less pronounced, though there is overall less damage as the sensitivity to TGF-β1 increases. We examined some of the more detailed mechanisms contributing to the severity of inflammatory response by determining which of thresholds within that module were the most sensitive. The inflammatory response, like most biological processes, is regulated in part through signaling mechanisms that, in essence, define thresholds of activation [ 54 ]. Accordingly, various thresholds in the PUABM control state changes of inflammatory cells (e.g., when local TNF-α concentration is above threshold, macrophages convert to M1 phenotype.) There is a different threshold for each state change, as illustrated in Table 1 and Fig 1B . Sensitivity analysis indicated that the most sensitive threshold was the concentration of DAMPs necessary to activate neutrophils. We found that most thresholds had a small range over which the output was sensitive to the threshold value, and outside of this range the dynamics were stable. This was not true for thresholds controlling activation of M2 macrophages. For both IL-1β and TGF-β1, increasing the concentration necessary to induce M2 activation led to increased simulated tissue damage. Global sensitivity analysis of the PUABM suggests that inflammation and oxygen determine to tissue damage, but reperfusion is critical to ulceration The goal of a global sensitivity analysis is to understand how changing model parameter values affect the outcomes produced in simulations. Others have used sensitivity analysis techniques to analytically determine which parameters were most highly associated with metrics of model outcomes [ 43 , 49 , 53 ]. This technique is particularly useful to determine which parameters of a model are most influential toward model outcomes. Because many of our parameters represent lumped estimates or qualitative relationships, and since we were interested in which of only three high-order mechanisms were most important, we hypothesized that we would be able to parse out parameters controlling such high order information without performing an exhaustive sweep of such a large parameter space. We first sought to explore how various model parameters might interact with each other to cause complex tissue damage outcomes. We performed a sensitivity analysis aimed at observing the relative and additive effects of each source of damage. The model includes three sources of tissue damage: lack of oxygen, presence of TNF-α, and presence of ROS ( Fig 5A ). TGF-β1 and IL-1β were not included in this analysis because they do not contribute to tissue damage directly. Thus, TNF-α, ROS, and oxygen are the three broad mechanistic sources of damage, and for each there exists a single broadly effective parameter that controls the magnitude of the effect of a given parameter on a simulation. Biologically, we view these parameters as representing an individual's overall sensitivity to each of the three damage-causing species. The first step in this analysis employed two values for each parameter, corresponding to low (zero-effect) and high (default) levels ( Fig 5B ). The zero-effect level of each parameter was defined as the value at which that source of damage no longer contributed to a simulation (an individual would be completely insensitive to presence or absence of that species). PUABM simulations in which each of three broadly effective parameters was set to either default (definitely has an effect) or zero-effect values show explicitly how each of the sources of damage affects outcomes on its own and in combination with other damage sources ( Fig 5B ). We note that when the tissue is not sensitive to oxygen (left hand column of each panel), no ancillary damage accumulates even when an ulcer appears by t = 400 h. We chose this time point because differences between parameter sets were well pronounced by this time point. Sensitivity assessed at later time points proved to be less informative, as in these extreme conditions the ulcers had already progressed significantly. In several parameterizations, time points prior to t = 400 were also less informative, due to sudden ulceration in otherwise healthy-looking tissue. 10.1371/journal.pcbi.1004309.g005 Fig 5 Sensitivity analysis reveals a unique contribution for all damage mechanisms, but simulated tissue is most sensitive to oxygen. ( A ) We partitioned parameter values according to which damage mechanism they affected. Panel ( B ) shows a global sensitivity analysis comparing the effects of the three mechanisms of damage on tissue health outcomes. Each mechanisms is simulated at two levels: with maximal or negligible effects. In panels C-E, sensitivity analysis was performed on parameters within each mechanistic module. All sensitivity analyses are shown with model snapshots at time t = 400 h (approximately 2.5 weeks). Default parameters are marked with asterisks. ( C ) Tissue sensitivity to oxygen parameters. The parameter governing oxygen production varies in each row. Each column represents a value of the parameter controlling how sensitive the tissue is to local oxygen concentrations. When oxygen production is plentiful, the simulated tissue becomes insensitive to other oxygen parameters (bottom row). Panel ( D ) illustrates tissue sensitivity to pressure intensity and period. Pressure intensity varies in each row. Each column represents a value of pressure cycle length. As the cycle length increases, the number of reperfusion events decreases for the same period of ischemia. Increased pressure leads to more damage, while fewer reperfusion events lead to lower damage at t = 400 h. Panel ( E ) illustrates tissue sensitivity to inflammatory mediators. Each column represents a value of the parameter controlling how sensitive the tissue is to local TNF-α concentrations. The parameter controlling tissue sensitivity to TGF-β1 varies in each row. Increasing sensitivity to TNF-α leads to earlier ulceration and more damage, while increased sensitivity to TGF-β1 leads to decreased tissue damage. The differences between simulations with and without sensitivity to oxygen (top row versus bottom row) are that of tissue damage severity. In all but one column, the bottom simulation (oxygen-sensitive) appears as a more severe version of the upper simulation (oxygen-insensitive). From this analysis, we hypothesize that while oxygen sensitivity contributes to tissue damage, it is not sufficient to cause significant damage or a PU, and furthermore we hypothesize that the lower the threshold of reperfusion injury, the more likely a PU is to form. Comparison of analogous simulations in each of the two panels (left panel is sensitive to ROS, right panel is ROS-insensitive) shows that often it is sensitivity to ROS that is the difference between ulcerated and damaged (but non-ulcerated) tissue in the PUABM. The simulation in the bottom right corner of the second panel shows a PU, but it is smaller and less extensive than the ulcer in the bottom right simulation of the left panel. This snapshot represents the type of PU that forms when parameters dictate that tissue is sensitive to all three damage sources. Based on these analyses, we hypothesize that tissue damage can be ascribed to inflammation, ulceration can be ascribed to ROS, and oxygen can exacerbate damage but may not be sufficient to cause significant tissue damage. Module-based exploration of the PUABM suggests that tissue is sensitive to oxygen and pro-inflammatory mediators We next sought to determine the degree to which parameters within each of the three mechanisms described above contribute to tissue damage. Within each mechanism, we identified parameters that were likely to have the greatest effect due to their mathematical form; i.e. thresholds were more likely to significantly influence outcomes than linear coefficients. Because injury in the PUABM is initiated by repeated pressure cycles, we first examined how sensitivity to pressure affected damage outcomes. To do this, we varied the maximum intensity of pressure (applied to the center of the tissue) and the length of the pressure interval ( Fig 5C ). This number indicates for how many consecutive ticks (simulated hours) pressure is applied, equal to the number of ticks during which pressure is removed. At t = 400 h (16 days) of simulated time post-SCI, very little effect of pressure interval length was observed. However, a slightly lower degree of damage was discerned in the medium-length pressure interval. Pressure intensity appeared to affect tissue health negatively. Pressure influences tissue oxygen in the PUABM. The impact of oxygen on tissue health was therefore investigated by varying in tandem parameters controlling oxygen production (how much oxygen is released from a blood vessel at each time interval) and tissue sensitivity to oxygen (a scalar factor determining the degree to which epithelial cell health is impacted by local oxygen concentrations, same as in global analysis above) ( Fig 5D ). This analysis suggested that when oxygen production from blood vessels in the PUABM is low (or zero), tissue health is affected severely, but in a manner that scales with the parameter that specifies sensitivity to oxygen. At high oxygen production levels, however, there was no apparent sensitivity of simulated tissue health to oxygen concentration. Simulated tissue sensitivities to pro- vs. anti-inflammatory/ pro-healing mediators were compared to each other ( Fig 5E ). This analysis involved comparing the effects of varying the parameter controlling the model's sensitivity to TGF-β1 (the canonical anti-inflammatory/ pro-healing mediator) and the parameter controlling sensitivity to TNF-α (the canonical early-acting pro-inflammatory mediator). Although all simulation snapshots are taken at the same time point, PU appear by t = 400 h (16 d) with higher sensitivity to TNF-α; this does not occur at lower sensitivity to TNF-α. The effect of TGF-β1 on tissue health in the PUABM is less pronounced, though there is overall less damage as the sensitivity to TGF-β1 increases. We examined some of the more detailed mechanisms contributing to the severity of inflammatory response by determining which of thresholds within that module were the most sensitive. The inflammatory response, like most biological processes, is regulated in part through signaling mechanisms that, in essence, define thresholds of activation [ 54 ]. Accordingly, various thresholds in the PUABM control state changes of inflammatory cells (e.g., when local TNF-α concentration is above threshold, macrophages convert to M1 phenotype.) There is a different threshold for each state change, as illustrated in Table 1 and Fig 1B . Sensitivity analysis indicated that the most sensitive threshold was the concentration of DAMPs necessary to activate neutrophils. We found that most thresholds had a small range over which the output was sensitive to the threshold value, and outside of this range the dynamics were stable. This was not true for thresholds controlling activation of M2 macrophages. For both IL-1β and TGF-β1, increasing the concentration necessary to induce M2 activation led to increased simulated tissue damage. In silico clinical trials for post-SCI pressure ulcers Corticosteroids reduce overall tissue damage, but not post-SCI ulceration in simulations In silico (simulated) clinical trials are an inexpensive and increasingly popular means of gleaning translational knowledge from computational models [ 55 – 59 ]. Accordingly, these methods were utilized to test both current and hypothetical or cutting-edge therapies for inflammation in the setting of post-SCI PU. The feasibility of corticosteroids as a treatment was examined for this indication using the PUABM, varying dose and timing of corticosteroids to investigate whether reducing inflammation could delay time of ulceration and/or lead to less tissue damage ( Fig 6 ). Corticosteroid administration was simulated as suppressing all inflammatory mechanisms. This effect was implemented by the following rule: any activated macrophages or neutrophils die upon encountering at least a threshold amount of steroid in their local vicinity of steroid ( S6 Fig ). We note that doses do not correlate directly to clinical doses, but instead to degree of functionality. For example, a simulated high dose (5) of corticosteroids shuts down macrophage and neutrophil activity completely, whereas lower doses all some cells to live and takes longer to see an effect. In simulations where corticosteroids were administered at the highest dose and before 250 hours, total damage was reduced. However, with simulated pressure cycling, PU still formed, suggesting that the lack of inflammation driven by macrophages and neutrophils was insufficient to ameliorate the damaging effects of ischemia and reperfusion ( Fig 6A ). By comparison, simulations with an initial acute injury but without pressure responded more favorably to steroid intervention ( Fig 6B ). These simulation results agree with the global sensitivity analysis of Fig 5B . It appeared that all doses of corticosteroids greater than zero were sufficient to achieve the full effect in both cases in the PUABM. There was, however, more damage incurred when corticosteroids were introduced at later times. While overall damage in simulations in which corticosteroids were introduced late was still less than simulations without corticosteroids, timeliness of corticosteroid application was anti-correlated with tissue damage incurred. This is likely related to the self-perpetuating nature of the pro-inflammatory mediators in this model. We hypothesize that early interruption of the positive feedback cycle was more effective in reducing inflammatory activity than the same amount of treatment later. 10.1371/journal.pcbi.1004309.g006 Fig 6 In silico clinical trials suggest little efficacy for corticosteroids or DAMP inhibitors. Simulations are shown at t = 700 h. We varied both the dose and timing of corticosteroid administration, simulated as an injection into the bloodstream, under ( A ) alternating pressure and ( B ) 40% initial injury conditions. When inflammatory cells were neutralized early enough but pressure continued, overall damage decreased, but ulceration was not prevented. Without continuous pressure cycles, the earliest dose of steroids was successful in stemming ulcer formation, but later applications did not. We then varied both the dose and timing of administration of a neutralizing antibody to HMGB1, simulated as a topical cream applied to the entire field. This targeted approach had ( C ) no apparent effect during simulations with alternating pressure, but ( D ) was able to slow ulcer formation after a 40% initial injury without pressure. Anti-DAMP antibodies have no effect on post-SCI ulceration in simulations DAMPs such as HMGB1 have become leading therapeutic targets for inflammatory indications [ 60 , 61 ]. Accordingly, dose and timing of anti-DAMP administration were varied over wide ranges using the PUABM (implemented as illustrated in S7 Fig ); however, no apparent effect was observed ( Fig 6C ). This lack of efficacy in these simulations is perhaps unsurprising, as there are multiple methods of inciting inflammation encoded in the model. Similarly, in humans and experimental animals [ 33 ] there are several ways by which the body is alerted to cell damage and potentially harmful invaders. In the PUABM, inflammation is incited secondarily when tissue cells that have accumulated damage above a threshold can activate neutrophils directly, thus compensating for the lack of DAMPs. As with corticosteroids, the same treatment was simulated on an initial acute injury without pressure cycles, yielding similar results ( Fig 6D ). Corticosteroids reduce overall tissue damage, but not post-SCI ulceration in simulations In silico (simulated) clinical trials are an inexpensive and increasingly popular means of gleaning translational knowledge from computational models [ 55 – 59 ]. Accordingly, these methods were utilized to test both current and hypothetical or cutting-edge therapies for inflammation in the setting of post-SCI PU. The feasibility of corticosteroids as a treatment was examined for this indication using the PUABM, varying dose and timing of corticosteroids to investigate whether reducing inflammation could delay time of ulceration and/or lead to less tissue damage ( Fig 6 ). Corticosteroid administration was simulated as suppressing all inflammatory mechanisms. This effect was implemented by the following rule: any activated macrophages or neutrophils die upon encountering at least a threshold amount of steroid in their local vicinity of steroid ( S6 Fig ). We note that doses do not correlate directly to clinical doses, but instead to degree of functionality. For example, a simulated high dose (5) of corticosteroids shuts down macrophage and neutrophil activity completely, whereas lower doses all some cells to live and takes longer to see an effect. In simulations where corticosteroids were administered at the highest dose and before 250 hours, total damage was reduced. However, with simulated pressure cycling, PU still formed, suggesting that the lack of inflammation driven by macrophages and neutrophils was insufficient to ameliorate the damaging effects of ischemia and reperfusion ( Fig 6A ). By comparison, simulations with an initial acute injury but without pressure responded more favorably to steroid intervention ( Fig 6B ). These simulation results agree with the global sensitivity analysis of Fig 5B . It appeared that all doses of corticosteroids greater than zero were sufficient to achieve the full effect in both cases in the PUABM. There was, however, more damage incurred when corticosteroids were introduced at later times. While overall damage in simulations in which corticosteroids were introduced late was still less than simulations without corticosteroids, timeliness of corticosteroid application was anti-correlated with tissue damage incurred. This is likely related to the self-perpetuating nature of the pro-inflammatory mediators in this model. We hypothesize that early interruption of the positive feedback cycle was more effective in reducing inflammatory activity than the same amount of treatment later. 10.1371/journal.pcbi.1004309.g006 Fig 6 In silico clinical trials suggest little efficacy for corticosteroids or DAMP inhibitors. Simulations are shown at t = 700 h. We varied both the dose and timing of corticosteroid administration, simulated as an injection into the bloodstream, under ( A ) alternating pressure and ( B ) 40% initial injury conditions. When inflammatory cells were neutralized early enough but pressure continued, overall damage decreased, but ulceration was not prevented. Without continuous pressure cycles, the earliest dose of steroids was successful in stemming ulcer formation, but later applications did not. We then varied both the dose and timing of administration of a neutralizing antibody to HMGB1, simulated as a topical cream applied to the entire field. This targeted approach had ( C ) no apparent effect during simulations with alternating pressure, but ( D ) was able to slow ulcer formation after a 40% initial injury without pressure. Anti-DAMP antibodies have no effect on post-SCI ulceration in simulations DAMPs such as HMGB1 have become leading therapeutic targets for inflammatory indications [ 60 , 61 ]. Accordingly, dose and timing of anti-DAMP administration were varied over wide ranges using the PUABM (implemented as illustrated in S7 Fig ); however, no apparent effect was observed ( Fig 6C ). This lack of efficacy in these simulations is perhaps unsurprising, as there are multiple methods of inciting inflammation encoded in the model. Similarly, in humans and experimental animals [ 33 ] there are several ways by which the body is alerted to cell damage and potentially harmful invaders. In the PUABM, inflammation is incited secondarily when tissue cells that have accumulated damage above a threshold can activate neutrophils directly, thus compensating for the lack of DAMPs. As with corticosteroids, the same treatment was simulated on an initial acute injury without pressure cycles, yielding similar results ( Fig 6D ). Discussion We have previously put forward an approach involving extensive simulations validated against highly focused clinical data in order to streamline subsequent in vitro , in vivo , and clinical studies [ 62 ]. Herein, we demonstrate how a tissue-realistic ABM of ischemia/reperfusion injury and inflammation in epithelial tissue could recapitulate morphological features of pressure ulcers in individuals with spinal cord injury, in a manner that is highly consistent with clinical images. At the molecular level, the model was calibrated to in vivo studies of ischemia/ reperfusion injury in rat epidermis [ 38 ] as well as inflammatory dynamics reported in the literature. The model output matched clinical ulcers qualitatively, generating irregular shapes and jagged edges, as well as distinct, secondary foci of inflammation that could progress to ulcers, despite initial conditions simulating a circular area of bony protrusion. In further general agreement with clinical data, model simulations spontaneously reached endpoints of either pathogenic or resolving disease. Interrogation of this stochastic phenomenon in silico revealed that ulceration outcomes were determined before the appearance of an open ulcer—suggesting that early diagnosis and intervention are critical. From a diagnostic standpoint, our simulations suggest that the most important predictors of ulcer formation are tissue oxygen levels and the levels of pro-inflammatory mediators. Predictions of the PUABM could be compared visually with easily-obtained images of patient skin; thus, we suggest that with further calibration and validation, this model could eventually be used as a diagnostic aid to determine which patients are at higher risk for ulcer formation before ulcers progress beyond stage I. From the standpoint of therapeutic interventions, in silico clinical trials using the PUABM suggested that after a relatively early point in time, inflammation-targeting intervention are unlikely to prevent ulceration or reduce tissue damage. In our simulations, corticosteroids were incapable of preventing ulcer formation, though when applied early enough could reduce the amount of tissue damage surrounding the wound. Sensitivity analysis revealed that ulceration in the model was correlated to tissue sensitivity to oxygen. Thus, wound oxygenation may be a potential therapeutic avenue for post-SCI PU. To date, studies of hyperbaric oxygen treatment have not proven successful in treating PU [ 63 ]. This failure may be explained by the timing issues mentioned above; we hypothesize that oxygenation is key to prevent ulcers from forming, but less helpful once they do. Where data were available, model parameters were calibrated to cell-level phenomena (e.g. lifespans). Otherwise, values were chosen that would yield tissue phenotypes relevant to healthy conditions. We reasoned that for any insights derived from the PUABM about ulceration to be meaningful, the same parameterization should allow tissue to remain healthy when unperturbed by pressure. For example, diffusion rates were adjusted to ensure adequate oxygenation to tissue that was not experiencing pressure. Diffusion represents an important link between spatial and temporal scales of the model, since diffusion governs the radius of effect a mediator could have. In early tests of negative controls, after several hundred hours of simulations with no pressure, we noticed tissue health declining at the edges of the simulation grid despite toroidal boundary conditions. We concluded that oxygen concentrations over time were not sufficient to maintain tissue health in those regions, and therefore our estimate of the diffusion rate was likely too low. By changing this parameter, tissue health was maintained when unperturbed for thousands of simulated hours. Importantly, other mediator diffusion parameters were also adjusted, which did not change the dynamics of the model. All model rules take effect only in a small, local environment: the largest area a cell will search before taking an action is that encompassing its immediate neighbors. Similarly, local concentrations of mediators (rather than total amount in a simulation) are used to determine macrophage state changes, tissue damage, and all other functions that are mediator-driven. Tissue damage itself is a local concept: each cell has a health score that is factored in to its individual behavior. However, when collections of cells arranged in a tissue experience similar stimuli (e.g. pressure, local mediator concentrations, etc.) they will respond similarly and spatial patterns—like ulcers—can emerge. In addition, our model was calibrated solely to healthy tissue. Thus, it is notable that when pressure was applied, ulcers not only developed but did so in a comparable time frame to—and with morphology that mimicked—clinical outcomes: irregular shapes emerged despite being simulated as starting from a perfectly round ischemic area. The PUABM was designed to synthesize phenomenological information into a framework that would allow hypothesis testing and might lead to deeper investigation of mechanisms of interest. We chose our parameter values and ranges to yield behaviors that were deemed relevant by both clinician and biologists on our interdisciplinary team. Many of the parameters that define key mechanisms of the model represent the cumulative actions of a set of cells or molecules, some of whose functions remain ill-defined. For this reason, almost all of the parameters values for the model were estimated based on the phenomena we wished to encode, rather than taking values estimated in the literature (which could not, by definition, exist). For example, we calibrated epithelial cell parameters so that unperturbed tissue remained healthy for periods of time much longer than a simulation, as described in Materials and Methods . For inflammatory mechanisms, we chose thresholds for cell state changes such that at the population level several states could co-exist when appropriate and ulcer progression occurred on a relevant time scale. Sampling distributions were similarly unknown, and so we reasoned that we could understand the broadest range of model capabilities by sampling parameters from uniform or log-uniform distributions. Our 1NN and re-simulations yielded different estimates of the time point of bifurcation. The re-simulations had much more input information, and there is likely something in the model that was not included as a feature in 1NN that could have improved its classification time. Mediators were used in 1NN because they could conceivably be measured at the bedside, but identification of other, potentially diagnostic, features would certainly improve our understanding of this phenomenon. In addition, to address hypotheses of inside-out ulcer formation, it will be important to develop a three-dimensional version of this model that will be able to simulate wounds with depth penetrating multiple layers of tissue. The PUABM also represents an extension of our prior work, in which we created a hybrid equation- and agent-based model that simulates blood flow along with skin injury, inflammation, and ulcer formation [ 25 ]. As in the PUABM, the relationship between pressure and the course of ulcer formation was demonstrated in that prior study. The equation-based portion of that hybrid model was calibrated to data related to blood flow following experimental pressure responses in non-injured human subjects or to data from people with SCI, predicting a higher propensity to form PU in response to pressure in people with SCI vs. non-injured control subjects (both as cohorts and individual patients [ 25 ]). Thus, blood flow data could be integrated with clinical images to further improve diagnosis or treatment, in essence comprising a novel, multi-scale diagnostic platform for post-SCI ulcer formation. In conclusion, we suggest that the primary value of the PUABM lies in its ability to recapitulate a broad range of pathology with qualitative, and at times quantitative, fidelity based on a single set of parameters, as well as broadly reproducing the full range of outcomes upon variation of only a few parameters. Better measurements at the bedside and further analysis may eventually allow the model be used to identify individual wound healing phenotypes and trajectories, determine appropriate treatment course, and design and test new treatment regimens. Supporting Information S1 Text Model Pseudocode. (DOCX) Click here for additional data file. S1 Code SPARK code for PUABM. (ZIP) Click here for additional data file. S1 Table Parameters, descriptions, and default values. (DOCX) Click here for additional data file. S2 Table Information criteria outcomes for GMM Fit. (DOCX) Click here for additional data file. S1 Fig Bar graph of Tissue Damage at 10, 15, and 20 days post-SCI. For simulations in the upper plot, Pressure Cycle length was 2, 6, or 12h of ischemia followed by and equal period of reperfusion. In most simulations, ulceration occurs between day 15 (360 h) and day 20 (480 h). After ulceration, simulations with shorter pressure cycle lengths have more tissue damage initially. For comparison, the lower plot shows in vivo results [ 37 ] demonstrating that increasing the amount of ischemia increases damage (1h v 2h, 5 cycles) but for a given amount of ischemia, increasing the number of reperfusion events also increases damage (2h, 5 cycles v 1h, 10 cycles). (DOCX) Click here for additional data file. S2 Fig Histogram of ulceration versus % initial injury. A 30% initial injury leads to ulceration roughly 50% of the time. (DOCX) Click here for additional data file. S3 Fig Histogram and GMM Estimate of tissue damage at t = 1000 for 30% initial injury. Two disparate levels of tissue damage suggest two outcomes from simulations starting with the same initial conditions. Lower tissue damage is associated with no ulcer formation. (DOCX) Click here for additional data file. S4 Fig Time courses for each feature in resolved versus ulcerated simulations. Blue time courses are from simulations that led to ulcers whereas red did not. Total counts of cells, mediators, and tissue health are plotted for 1000 ticks. Time courses of some features are very similar for both outcomes, but for other the differences are stark. (DOCX) Click here for additional data file. S5 Fig 1NN over individual features. Classification accuracy (resolved versus ulcerated) for 1NN over each feature individually. Features were input as time vectors, using either all previous time points or the previous 50 ticks. (DOCX) Click here for additional data file. S6 Fig Implementation of corticosteroid treatment in model rules. Local concentrations of corticosteroid functionally disable macrophages and neutrophils in the model, causing them to no longer produce or respond to mediators. (DOCX) Click here for additional data file. S7 Fig Implementation of anti-DAMPs antibody in model rules. Anti-DAMPs antibodies are simulated as removing DAMPs from the model, lowering local concentrations in a quenching reaction. (DOCX) Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4806969/

Low-Cost National Media-Based Surveillance System for Public Health Events, Bangladesh

We assessed a media-based public health surveillance system in Bangladesh during 2010–2011. The system is a highly effective, low-cost, locally appropriate, and sustainable outbreak detection tool that could be used in other low-income, resource-poor settings to meet the capacity for surveillance outlined in the International Health Regulations 2005. The Study IEDCR contracted for a media scanning company to identify relevant news stories for this surveillance system ( Figure ). Each day, the company collected the major daily newspapers available in the capital city of Dhaka shortly after morning distribution. Trained staff members at the company read through each paper to identify any health-related article and scan them into PDF file format. Other staff members scanned television news reports and recorded relevant video clips. Figure Information flow for national media-based public health surveillance system, Bangladesh. IEDCR, Institute of Epidemiology, Disease Control and Research. The national rapid response team, consisting of key staff members from IEDCR, received a daily email containing all identified health-related newspaper articles and video clips. The team examined each news item and decided whether it warranted an outbreak response on the basis of expert clinical and epidemiologic knowledge; public health importance (e.g., number of cases and deaths reported, severity of symptoms); and verification by local health officials. For the purposes of this analysis, IEDCR retrospectively created a database of reported events sent by the media scanning company, which included the number of reported events, outbreak etiology, news source, and the outcome of each investigation. The outbreaks reported were classified by media type, etiology, and season. To assess the surveillance system's performance, we followed the guidelines for evaluating public health surveillance set by the US Centers for Disease Control and Prevention (CDC) ( 2 ).We aimed to assess all 9 recommended system qualities: simplicity, flexibility, data quality, acceptability, sensitivity, positive predictive value (PPV), representativeness, timeliness, and stability ( 3 ). However, because we did not have external data to serve as a reference standard or comparator, we were unable to assess the sensitivity and PPV of the system. We calculated the proportion of all outbreak investigations conducted by IEDCR that were first detected through the media surveillance. We interviewed key stakeholders, including the manager and staff of the contracted media scanning company, the director of IEDCR, and national rapid response team members. We also conducted a group discussion with 4 journalists representing large national newspapers to explore the processes of obtaining information for health events and timeline of reporting. A total of 36 news sources were scanned regularly in this media surveillance system: 23 (64%) were Bengali language newspapers, 6 (17%) were English language newspapers, and 7 (19%) were Bengali language television channels. From May 2010 through September 2011, the media scanning company captured and delivered 2,821 news stories to IEDCR. Of those, 2,501 (89%) were health related, 810 items included the term "outbreak" (29% of total), and 196 were the first reports of a possible outbreak (7%); 90% of all outbreak reports came from Bengali language media sources. During the same period, the national rapid response team investigated 30 outbreaks, 21 (70%) of which were first detected through this surveillance system. At the rate of US$125/year to hire the scanning company, the total cost for the 16-month period was US$167. The cost of contracting with the scanning company for each outbreak detected with this system was therefore approximately US$8 (US$167 divided by 21 outbreaks). Outbreaks of diarrhea, measles, and anthrax were reported through this system ( Table 1 ). Outbreaks were reported year-round and from 51 of the 64 administrative districts in Bangladesh ( Table 2 ). Table 1 Outbreaks first identified by event-based surveillance and investigated by IEDCR, Bangladesh, May 2010–September 2011* Outbreak no. Date Etiology reported by media Confirmed etiology 1 2010 May Anthrax Cutaneous anthrax 2 2010 Jun Anthrax Cutaneous anthrax 3 2010 Jul Unknown poisoning Unintentional pesticide poisoning 4 2010 Jul Mass psychogenic illness Mass psychogenic illness 5 2010 Jul Suspected pneumonia Bronchiolitis 6 2010 Jul Unknown animal scratch Rabies 7 2010 Jul Food poisoning Food poisoning 8 2010 Jul Diarrhea Diarrhea 9 2010 Aug Cutaneous anthrax Cutaneous anthrax 10 2010 Nov Diarrhea Cholera 11 2010 Nov Suspected high-energy biscuit poisoning Mass psychogenic illness after biscuit consumption 12 2010 Dec Suspected pneumonia Bronchiolitis 13 2010 Dec Suspected rabies Rabies 14 2011 Apr Diarrhea Cholera 15 2011 May Cutaneous anthrax Cutaneous anthrax (5 outbreaks) 16 2011 Jun Cutaneous anthrax Cutaneous anthrax (2 outbreaks) 17†2011 Jun Unusual duck and geese mortality†Avian influenza, subtype H5N1, in geese, but no human cases detected 18 2011 Jul Cutaneous anthrax Cutaneous anthrax 19 2011 Jul Cutaneous anthrax Cutaneous anthrax 20 2011 Aug Unknown disease Influenza B virus infection 21 2011 Aug Cutaneous anthrax Cutaneous anthrax *IEDCR, Institute of Epidemiology, Disease Control and Research, Bangladesh. †Non–human-related outbreak. Table 2 Case numbers for media-reported outbreaks, by cause and season, Bangladesh, May 2010–September 2011 Reported etiology No. (%) cases Pre-monsoon, Mar–May Monsoon, Jun–Sep Post-monsoon, Oct–Nov Winter, Dec–Feb Total Diarrhea 24 (50) 7 (15) 10 (21) 7 (15) 48 (100) Anthrax 3 (14) 19 (86) 0 0 22 (100) Mass psychogenic illness 9 (45) 10 (50) 0 1 (5) 20 (100) Upper respiratory infection 0 2 (33) 3 (50) 1 (17) 6 (100) Measles 3 (60) 0 0 2 (40) 5 (100) Other/unknown 31 (33) 49 (52) 5 (5) 10 (11) 95 (100) Total 70 (36) 87 (44) 18 (9) 21 (11) 196 (100) Key informant interviews consistently indicated that the system was simple, flexible, timely, and acceptable because it used existing media infrastructure and required only minimal costs to contract with a company to compile daily reports of news items. Changes to the system could be implemented effectively through frequent communications between the media scanning company and IEDCR. The system was widely acceptable by all stakeholders and was considered a valuable component of disease surveillance in Bangladesh. We were unable to quantitatively assess the coverage of remote areas by national newspapers, especially those lacking easy access to telecommunication infrastructure. The system moderately captured a representative sample of possible sources of information in the country. Both Bengali- and English-language news items were collected, although only newspapers available in Dhaka were included, so some newspapers with only local circulation were not available for review. Timeliness and stability of the system were both high. Although time from outbreak onset to reporting in the media source might vary, once media sources learned of an outbreak, publication occurred within 24 hours. Because it was low-cost, low-tech, and highly acceptable to all stakeholders, the system was highly stable. This media surveillance identified outbreaks of emerging infections that might not have been otherwise investigated, including several outbreaks that were potential public health events of international concern. In the context of global health security, international donors should support media-based surveillance to further strengthen existing traditional indicator-based approaches. The media-based surveillance system in Bangladesh fills a gap that is not covered by other global event-based surveillance systems, which collect publically available information about potential health threats mostly from Internet sources ( 4 ), such as ProMED ( 5 ), BioCaster, and HealthMap ( 6 ). Although these systems collect and analyze enormous amounts of information from the Internet regarding potential health threats, they are limited by the inability to process information in local, non-English languages or to capture information not on the Internet. In 2013, ≈341 million persons in the world (5%) spoke English as a first language ( http://www.nationsonline.org/oneworld/most_spoken_languages.htm ); clearly, most news sources in the world are written in languages other than English and, therefore, are beyond the reach of these English language event-based surveillance systems. Although all stakeholders were knowledgeable about their duties and responsibilities and the procedures which needed to be followed, no written standard operating procedures were in place by which we could evaluate process performance. Written procedures, for both the media scanning company and IEDCR staff, could enhance system performance by fostering sustainability and ensuring standardization. Our evaluation of this surveillance system was limited in 2 key ways. Because the database for the system was created retrospectively over a short period, some reports may have been missing from the database. In addition, the absence of other data sources meant that we were unable to determine the sensitivity and PPV of the system. However, neither of these limitations influences the high proportion of outbreaks detected through this system nor the low cost per outbreak detected, the most critical findings from our evaluation. Conclusions IEDCR in Bangladesh has created an innovative, low-cost, locally appropriate solution for event-based surveillance that helps to meet the purposes of the country's surveillance goals and IHR requirements. This surveillance system could serve as a model for outbreak detection in other resource-poor countries.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3384954/

Inhibitory Monoclonal Antibodies against Mouse Proteases Raised in Gene-Deficient Mice Block Proteolytic Functions in vivo

Identification of targets for cancer therapy requires the understanding of the in vivo roles of proteins, which can be derived from studies using gene-targeted mice. An alternative strategy is the administration of inhibitory monoclonal antibodies (mAbs), causing acute disruption of the target protein function(s). This approach has the advantage of being a model for therapeutic targeting. mAbs for use in mouse models can be obtained through immunization of gene-deficient mice with the autologous protein. Such mAbs react with both species-specific epitopes and epitopes conserved between species. mAbs against proteins involved in extracellular proteolysis, including plasminogen activators urokinase plasminogen activator (uPA), tissue-type plasminogen activator (tPA), their inhibitor PAI-1, the uPA receptor (uPAR), two matrix metalloproteinases (MMP9 and MMP14), as well as the collagen internalization receptor uPARAP, have been developed. The inhibitory mAbs against uPA and uPAR block plasminogen activation and thereby hepatic fibrinolysis in vivo . Wound healing, another plasmin-dependent process, is delayed by an inhibitory mAb against uPA in the adult mouse. Thromboembolism can be inhibited by anti-PAI-1 mAbs in vivo . In conclusion, function-blocking mAbs are well-suited for targeted therapy in mouse models of different diseases, including cancer. Introduction Most of our understanding of the in vivo roles of extracellular proteases, their cellular receptors and inhibitors is derived from studies using gene-targeted mice. The phenotype of a gene-deficient mouse can, however, be influenced by redundancy, when the function of the protein encoded for by the inactivated gene is compensated due to overlapping functions between proteins (Page-McCaw et al., 2007 ; Stevens et al., 2012 ). Identification of the in vivo function of a protease/receptor in an adult mouse can be obtained by administration of inhibitory monoclonal antibodies (mAbs), causing acute disruption of the target protein function(s), further offering the advantage of directly serving as a model for therapeutic targeting. In order to avoid immunogenicity effects in mice, the mAbs used should be of murine origin. Development of such mAbs requires immunization of gene-deficient mice with the autologous target protein. Generation of mAbs against murine proteases using gene-deficient mice was first described by Declerck et al. ( 1995a ). Mice lacking the gene encoding either of the two main plasminogen activators, i.e., tissue-type plasminogen activator (tPA) and urokinase plasminogen activator (uPA), were immunized with murine tPA (mtPA) and muPA, respectively. Sera from the immunized mice had high titers of anti-mtPA and anti-muPA antibodies, while similar immunizations of wild-type mice resulted in no or little reactivity toward mtPA and muPA, consistent with these proteins being self-antigens in wild-type mice (Opdenakker et al., 2003 ). In this review, the properties of mouse mAbs, all generated by use of gene-deficient mice, will be described with respect to specificity, epitope location, cross-reactivity with the homologous protein from other species (conserved epitopes), and inhibitory function in vitro and in vivo . The antibody-mediated in vivo effect will be compared with that observed in gene-deficient mice to reveal similarities or changes between acute disruption of function in the adult animal and genetic disruption in utero . Monoclonal Antibodies against Murine Plasminogen Activators To generate mAbs against mtPA, mice were immunized twice and subsequently boosted with 10 μg recombinant mtPA before spleen cell isolation and fusion to myeloma cells. Hybridomas were selected for their ability to secrete antibodies reacting with mtPA using ELISA, resulting in the isolation of 203 hybridomas from two fusions. The amount of coating antigen in the ELISA was rather high (200 ng mtPA/well), allowing antibodies with relatively low affinity to be isolated. Of the 203 mAbs, 21 were selective for mouse tPA, while the residual mAbs cross-reacted with tPA from one or more species, i.e., rat, human, and vampire-bat tPA. Importantly, no anti-mtPA mAbs recognized muPA (Declerck et al., 1995a ). The mAbs were also assayed for their inhibitory effect on the ability of tPA to activate plasminogen (Figure 1 ). Interestingly, it was demonstrated that of the 21 mAbs recognizing epitopes specific for mtPA, six were inhibitory, whereas 105 of the 182 cross-reacting mAbs were inhibitory (Declerck et al., 1995a ). Hence, based on the apparent correlation between inhibitory properties and cross-reactivity, it was suggested that epitopes conserved between species are functionally important for plasminogen activation (Declerck et al., 1995a ). Two of the anti-mtPA mAbs were used for design of a quantitative ELISA, enabling measurements of mtPA in tissue and body fluids from mice (Declerck et al., 1995b ). Figure 1 Pathways of the plasminogen activation and MMP systems targeted by mouse mAbs blocking the functions of the proteases and receptors . Degradation of various components of the extracellular matrix involves both plasmin-mediated breakdown and degradation by MMPs, such as MMP9 and MMP14/MT1-MMP. This degradation of the extracellular matrix is a prerequisite for tissue remodeling, which takes place during cancer invasion. mAbs have been developed in order to target the functions of the proteases and receptors, indicated by (a–h). tPA (a) and uPA (c) can activate plasminogen to plasmin, which in turn mediates activation of pro-uPA to uPA (b), when pro-uPA is bound to its receptor uPAR (d) at the cell surface. Both uPA and tPA are inhibited by PAI-1 (e). MMP9 (f) and MMP14 (g) primarily degrade collagen in the extracellular matrix. Collagen and cleaved fragments hereof can bind to uPARAP and are subsequently internalized with the receptor (h). Using the same overall strategy, uPA −/− mice were immunized with muPA, resulting in 38 hybridomas producing antibodies against muPA (Declerck et al., 1995a ). Specificity test showed negative reaction with mtPA for all of these antibodies and interestingly, none of the anti-muPA mAbs cross-reacted with human uPA (Declerck et al., 1995a ). A quantitative ELISA was designed using two of the anti-muPA mAbs (Declerck et al., 1995b ). Due to the low affinity of these anti-muPA mAbs, the sensitivity of this assay was 10-fold lower than that of mtPA, precluding measurements of muPA in plasma. In a study with the aim to generate high affinity anti-muPA mAbs with in vivo efficacy, uPA −/− mice were immunized six times, followed by three boosting injections with recombinant pro-muPA (Lund et al., 2008 ). To ensure isolation of high affinity mAbs, the wells were coated with low amounts of antigen (2 ng pro-muPA/well) in the hybridoma screening ELISA, resulting in identification of nine anti-muPA mAb secreting hybridomas. Especially, two mAbs (mU1 and mU3), possessing epitopes in the uPA B-chain, encompassing the catalytic site, were demonstrated to interfere with the function of mouse uPA. Specificity test demonstrated no cross-reactivity with mouse tPA, and interestingly neither antibody recognized human uPA. Both mU1 and mU3 are high affinity antibodies. mU3 displays higher affinity for both pro-muPA ( K D = 0.03 nM) and muPA ( K D = 0.2 nM), as compared to mU1 [pro-muPA ( K D = 0.2 nM) and muPA ( K D = 1.3 nM)]. These antibodies efficiently and dose-dependently block uPA-mediated plasminogen activation. mU1 prevents both plasmin-mediated pro-uPA activation and uPA-mediated plasminogen activation (Figure 1 ), while mU3 only inhibits the latter of these reactions. To test the effect of these mAbs in vitro in a multicomponent system, a cellular assay using uPA-activatable anthrax pro-toxin (Liu et al., 2001 ) was employed. In this assay, a cytotoxic effect is released as a result of cell-bound uPA activity, which serves to cleave an engineered protective antigen, PrAg-U2. Pre-incubation of murine monocyte macrophage-like P388D.1 cells with either mU1 or mU3 led to significant rescue of cells simultaneously exposed to PrAg-U2 and the recombinant toxin FP59 (Lund et al., 2008 ). Interestingly, the rescue effect of mU1 was stronger than that of mU3 in vitro . As uPA −/− mice are insensitive to anthrax pro-toxin treatment (Liu et al., 2003 ), this system was used to determine the in vivo efficacy of mU1 and mU3 in wild-type mice. Two injections of the mAbs (60 mg/kg/dose) were administered to the mice prior to anthrax pro-toxin treatment. Of the wild-type mice treated with mU1, eight out of 10 mice survived, whereas only few mice treated with either mU3 or the isotype-matched negative control mAb survived, demonstrating mU1, but not mU3, to possess in vivo efficacy by blocking uPA activity (Figure 2 ; Lund et al., 2008 , 2011a ). Figure 2 mU1-induced rescue of adult mice treated with a uPA-activatable anthrax pro-toxin. (A) FVB/n wild-type mice received intraperitoneal injections of 60 mg/kg mU1 (dotted and broken lines) or an equivalent volume of saline (solid line; day −1 and 0), followed by injections with the uPA-activatable anthrax pro-toxin (i.e., 0.6 mg/kg PrAg-U2 +  0.4 mg/kg FP59; broken and solid lines) or saline alone (dotted line; day 0). Survival of the mice, presented in percentage, was recorded for 7 days. (B) FVB/n wild-type mice were treated with 60 mg/kg mU3 (dotted and broken lines) or an equivalent volume of saline (solid line), and subsequently by 0.6 mg/kg PrAg-U2 +  0.4 mg/kg FP59 (broken and solid lines) or saline alone (dotted line), as described in A (Lund et al., 2008 ). Monoclonal Antibodies against muPAR The main function of the uPA receptor (uPAR), is to localize the proteolytic activity to the cell surface (Ellis and Danø, 1991 ). uPAR also interacts with vitronectin and various integrins, thereby possessing several biological functions independent of its role in uPA-mediated proteolysis (Høyer-Hansen and Lund, 2007 ; Smith and Marshall, 2010 ). Despite the multifunctional properties, uPAR −/− mice show no overt phenotype (Bugge et al., 1995 ). To develop mAbs against muPAR, uPAR −/− mice were immunized with recombinant soluble muPAR (Pass et al., 2007 ). The procedures to generate hybridomas were identical to those described for the muPA mAbs (Pass et al., 2007 ; Lund et al., 2008 ). One fusion resulted in 12 hybridomas, secreting antibodies reacting with muPAR, including six that cross-reacted with human uPAR in Western blotting. uPAR consist of three domains and the integrity of this structure is a prerequisite for the high affinity binding of uPA and of vitronectin (Behrendt et al., 1996 ; Høyer-Hansen et al., 1997 ). The epitopes of the anti-muPAR mAbs mR1 and mR3 are located on the N-terminal muPAR domain I, while those of mR2 and mR4 are situated on muPAR domains II-III (Pass et al., 2007 ; Rasch et al., 2008 ). They are all high affinity mAbs, with mR3 having the highest affinity ( K D = 0.5 nM). mR1, mR2, and mR4 cross-react with human uPAR and have similar affinities for uPAR from the two species (Rasch et al., 2008 ). mR1, mR2, and mR4 were all capable of blocking binding of the murine amino-terminal fragment of muPA (mATF) to muPAR (Figure 1 ; Pass et al., 2007 ; Rasch et al., 2008 ). mR1 and mR4 showed a dose-dependent inhibition of mATF binding with IC 50 values of 0.67 and 0.34 nM, respectively, being similar to that of the biological ligand (IC 50 = 0.14 nM). Intriguingly, mR3 did not interfere with the uPA-uPAR interaction, even though the complete epitope of mR1 (Gårdsvoll et al., 2011 ) is close to or overlapping with that of mR3, as these mAbs are unable to bind simultaneously to muPAR (Rasch et al., 2008 ). The binding site for vitronectin on uPAR is different from that for uPA (Wei et al., 1994 ). Some mAbs with epitopes in domain I of human uPAR block the binding of vitronectin (Høyer-Hansen et al., 1997 ), while none of the so far characterized mAbs, reacting with mouse uPAR domain I, prevents this interaction (Rasch et al., 2008 ). The anti-muPAR mAbs were tested in rescue experiments using the above-described uPA-activatable anthrax pro-toxin assay (Pass et al., 2007 ). The experiments were performed both in vitro and in vivo in mice. Pre-incubation of uPAR-positive P388D.1 cells with mR1 resulted in significant cell rescue. For the same system in vivo , it was previously demonstrated that uPAR −/− mice are insensitive to treatment with the toxins, providing evidence for the stringent need for receptor-bound uPA to release the anthrax cytotoxic effect (Liu et al., 2003 ). Administration of the anthrax pro-toxin simultaneously with two injections of mR1 (60 mg/kg/dose) to the wild-type mice, resulted in rescue of all eight mice. Only two of seven mice treated with saline or the negative control mAb survived, thus demonstrating in vivo efficacy of mR1 by blocking the uPA-uPAR binding (Pass et al., 2007 ). Monoclonal Antibodies against mPAI-1 PAI-1 is the primary endogenous inhibitor of both tPA and uPA. The first reported murine mAbs against mPAI-1 were generated using the same protocol as for the anti-mtPA mAbs (see above), except that the immunized mice were PAI-1 −/− and mPAI-1 was the antigen (Carmeliet et al., 1993 ; Declerck et al., 1995b ). Two fusions resulted in 64 anti-mPAI-1 mAb-producing hybridomas and two mAbs were successfully employed for design of a quantitative immunoassay for measurement of mPAI-1 in murine plasma. In a recent study, the aim was to generate inhibitory anti-mPAI-1 mAbs reacting with the glycosylated form of mPAI-1 bound to vitronectin (Van De Craen et al., 2011 ). PAI-1 is secreted in an active but conformationally unstable form, which rapidly loses activity unless bound to vitronectin (Declerck et al., 1988 ). The desired anti-PAI-1 mAbs should therefore also inhibit PAI-1 activity of the PAI-1-vitronectin complex. Such anti-mPAI-1 mAbs were developed as previously described (Declerck et al., 1995b ), though with an increase in antigen dose from 10 to 55 μg/injection. One fusion yielded 21 hybridomas producing anti-mPAI-1 antibodies. Two mAbs were specific for mPAI-1, 19 cross-reacted with rat PAI-1, and 14 of these also recognized human PAI-1. Five of the anti-mPAI-1 mAbs, reacting with epitopes conserved between the three species, were found to neutralize PAI-1 activity, even in the presence of a 33-fold molar excess of mouse vitronectin. In this neutralization assay, the inhibitory effect of the mAbs was determined as the ability of mPAI-1 to inhibit mtPA activity in the absence and presence of the antibodies (Figure 1 ) (Declerck et al., 1995b ; Van De Craen et al., 2011 ). The in vivo efficacy of the five most inhibitory mAbs was tested in a thromboplastin-dependent mouse thromboembolism model, designed to demonstrate the potential function of the mAbs to rescue injected mtPA from mPAI-1-mediated inactivation. Mice were injected with 10 mg/kg of the mAb via the tail vein. Thromboplastin was injected simultaneously with mtPA to evoke thromboembolism. A reduced physical activity was used as the read-out for the thromboembolic condition. Evaluation 15 min. after treatment, demonstrated all five inhibitory anti-mPAI-1 mAbs to possess in vivo efficacy, as the mAb-treated mice were more active than control mice (Van De Craen et al., 2011 ). Two of these anti-mPAI-1 mAbs have furthermore been demonstrated to decrease the level of active mPAI-1 in mouse plasma, collected from mice treated with lipopolysaccharide to induce high mPAI-1 levels in the blood (Van De Craen et al., 2012 ). Monoclonal Antibodies against mMMPs mMMP9 Matrix metalloproteinase 9 (MMP9) belongs to the subfamily of gelatinases and is active in degradation of various components in the extracellular matrix, including early collagen cleavage fragments. Anti-mMMP9 mAbs have been generated by immunization of MMP9 −/− mice with recombinant mMMP9 (Descamps et al., 2002 ; Opdenakker et al., 2003 ; Kolaczkowska et al., 2008 ). Hybridomas were identified by ELISA, in which the wells were coated with goat-anti-mouse IgG. Addition of hybridoma supernatants, followed by biotinylated mMMP9 was used to identify anti-mMMP9 producing clones. Of the 250 hybridomas secreting anti-mMMP9 antibodies, the 16 with the strongest reactions were selected (Kolaczkowska et al., 2008 ). Four of these mAbs were specific for the pro-form of mMMP9, two were specific for the active form, and the 10 remaining reacted with pro-mMMP9 as well as one or more of three degradation products (Kolaczkowska et al., 2008 ). The mAbs reacting exclusively with the active form were found to have MMP9 neutralizing activity (Figure 1 ), as measured using a fluorogenic substrate (Opdenakker et al., 2003 ). mMMP14/mMT1-MMP All of the above-mentioned gene-deficient mice have very mild, if at all comprised phenotypes and a normal life span (Carmeliet et al., 1993 , 1994 , 1995 ; Bugge et al., 1995 ; Vu et al., 1998 ). The MMP14 −/− mice have in contrast a severe phenotype and a short life span (7–13 weeks) as they develop craniofacial dysmorphism, arthritis, osteopenia, dwarfism, and fibrosis of soft tissues due to inadequate collagen turn-over (Holmbeck et al., 1999 ). To enable generation of anti-mMMP14 mAbs, immunizations were therefore initiated 4 weeks after birth of the MMP14 −/− mice, as compared with the usual age of 8–9 weeks. The mice were only immunized twice, followed by three boosting injections with 15 μg recombinant soluble mMMP14 (Ingvarsen et al., 2008 ). Due to the small size of their spleens, splenocytes isolated from five immunized MMP14 −/− mice were pooled for one fusion. Screening for hybridomas was performed using ELISA with each well coated with 12 ng mMMP14. Six hybridomas secreting anti-mMMP14 antibodies were isolated. The epitopes are located in the hemopexin domain and/or in the membrane-proximal linker region of mMMP14. The mAbs cross-react with the human form of MMP14. Interestingly, one mAb is capable of stimulating the MMP14 dimerization step on the cell surface, which is necessary for MMP14-mediated activation of the soluble gelatinase, MMP2. None of the mAbs possess any inhibitory effect on the collagenolytic activity of mMMP14 (Figure 1 ; Ingvarsen et al., 2008 ). mMMP9 Matrix metalloproteinase 9 (MMP9) belongs to the subfamily of gelatinases and is active in degradation of various components in the extracellular matrix, including early collagen cleavage fragments. Anti-mMMP9 mAbs have been generated by immunization of MMP9 −/− mice with recombinant mMMP9 (Descamps et al., 2002 ; Opdenakker et al., 2003 ; Kolaczkowska et al., 2008 ). Hybridomas were identified by ELISA, in which the wells were coated with goat-anti-mouse IgG. Addition of hybridoma supernatants, followed by biotinylated mMMP9 was used to identify anti-mMMP9 producing clones. Of the 250 hybridomas secreting anti-mMMP9 antibodies, the 16 with the strongest reactions were selected (Kolaczkowska et al., 2008 ). Four of these mAbs were specific for the pro-form of mMMP9, two were specific for the active form, and the 10 remaining reacted with pro-mMMP9 as well as one or more of three degradation products (Kolaczkowska et al., 2008 ). The mAbs reacting exclusively with the active form were found to have MMP9 neutralizing activity (Figure 1 ), as measured using a fluorogenic substrate (Opdenakker et al., 2003 ). mMMP14/mMT1-MMP All of the above-mentioned gene-deficient mice have very mild, if at all comprised phenotypes and a normal life span (Carmeliet et al., 1993 , 1994 , 1995 ; Bugge et al., 1995 ; Vu et al., 1998 ). The MMP14 −/− mice have in contrast a severe phenotype and a short life span (7–13 weeks) as they develop craniofacial dysmorphism, arthritis, osteopenia, dwarfism, and fibrosis of soft tissues due to inadequate collagen turn-over (Holmbeck et al., 1999 ). To enable generation of anti-mMMP14 mAbs, immunizations were therefore initiated 4 weeks after birth of the MMP14 −/− mice, as compared with the usual age of 8–9 weeks. The mice were only immunized twice, followed by three boosting injections with 15 μg recombinant soluble mMMP14 (Ingvarsen et al., 2008 ). Due to the small size of their spleens, splenocytes isolated from five immunized MMP14 −/− mice were pooled for one fusion. Screening for hybridomas was performed using ELISA with each well coated with 12 ng mMMP14. Six hybridomas secreting anti-mMMP14 antibodies were isolated. The epitopes are located in the hemopexin domain and/or in the membrane-proximal linker region of mMMP14. The mAbs cross-react with the human form of MMP14. Interestingly, one mAb is capable of stimulating the MMP14 dimerization step on the cell surface, which is necessary for MMP14-mediated activation of the soluble gelatinase, MMP2. None of the mAbs possess any inhibitory effect on the collagenolytic activity of mMMP14 (Figure 1 ; Ingvarsen et al., 2008 ). Monoclonal Antibodies against uPARAP/Endo180 uPARAP (also designated Endo180), a member of the mannose receptor family, is a collagen internalization receptor and a type 1 transmembrane protein (Behrendt, 2004 ; Engelholm et al., 2009 ). For the generation of mAbs, uPARAP −/− mice were immunized with recombinant soluble human uPARAP (Sulek et al., 2007 ). All procedures were identical to those described for generation of mAbs against muPA and muPAR (Pass et al., 2007 ; Lund et al., 2008 ). Interestingly, seven out of eight mAbs cross-react with mouse uPARAP. The anti-uPARAP mAb 5f4 was demonstrated to inhibit cellular collagen internalization in fibroblasts (Figure 1 ; Madsen et al., 2011 ). Due to this mechanism of action, addition of 5f4 to wild-type cultured fibroblasts leads to accumulation of collagen degradation products in the cell culture medium, being similar to the collagen fragments that occur in the medium of uPARAP −/− mouse fibroblasts, cultured without antibody addition (Jürgensen et al., 2011 ; Madsen et al., 2011 ). Although the epitope of 5f4 is located to the N-terminal domains 1–3 of uPARAP (Madsen et al., 2012 ), the antibody does not prevent binding of uPARAP to collagen in a purified system. The capacity of blocking collagen internalization may be related to another interesting characteristic as the antibody down-regulates uPARAP expression on the cell surface (Madsen et al., 2011 ). Comparison of Effects of Acute Disruption of Protein Function in the Adult Animal and Genetic Disruption in utero Studies with gene-deficient animals have demonstrated uPA to play an important role in skin wound healing and extravascular fibrin clearance in the liver as well as the presence of a strong redundancy between uPA and tPA (Bugge et al., 1996 ). Mice deficient in uPA or tPA have a marginal delay in wound healing, while uPA −/− ;tPA −/− double-deficient mice display a significant delay compared to wild-type littermates (Lund et al., 2006 ). To study the effect of acute disruption of uPA function in wound healing, tPA −/− mice were treated with mU1 (60 mg/kg), administered according to the mAb half-life (Jögi et al., 2010 ). Importantly, mU1 caused a delay in wound healing, resulting in a healing time, which was not significantly different from that of the uPA −/− ;tPA −/− mice (Jögi et al., 2010 ). Consistent with uPA and tPA being the main plasminogen activators, a reduced level of plasmin was detected in wound extracts from mU1-treated tPA −/− mice, and the plasmin level was similar to that in uPA −/− ;tPA −/− mice. This underscores that administration of mU1 blocks plasminogen activation in vivo (Jögi et al., 2010 ). Both uPA −/− ;tPA −/− and uPAR −/− ;tPA −/− mice spontaneously develop hepatic fibrin deposits (Bugge et al., 1996 ). While uPA −/− mice occasionally display fibrin accumulation in the liver, no deposits have been detected in neither tPA −/− nor uPAR −/− mice (Bugge et al., 1996 ). Hence, to evaluate the effect of mU1 and mR1 on in vivo fibrin clearance, tPA −/− mice were treated with either mAb (Jögi et al., 2007 ; Lund et al., 2008 ). Systemic treatment with mU1 as well as with mR1 resulted in hepatic fibrin accumulation in tPA −/− mice, being indistinguishable from the depositions observed in corresponding double-deficient mice (Jögi et al., 2007 ; Lund et al., 2008 ). A recent study has demonstrated uPARAP −/− mice to possess increased accumulation of both total fibrillar collagen and collagen type IV upon CCl 4 -induced liver fibrosis (Madsen et al., 2012 ). Hepatic collagen deposition in wild-type mice with liver fibrosis thus seems as a suitable model for testing the in vivo efficacy of the anti-uPARAP mAbs. In addition to the overlapping roles between uPA and tPA in certain physiological processes, it was recently demonstrated that ablation of MMP9 in combination with either plasminogen activator impairs normal gestation, resulting in a non-Mendelian distribution of the off-spring (Lund et al., 2011b ). Furthermore, combined deficiency in MMP9 and uPA, but not tPA, causes a significant delay in wound healing as compared to the single-deficient and wild-type littermate control mice. Notably, a compensatory upregulation of uPA activity is seen in wound extracts from MMP9 deficient mice (Lund et al., 2011b ). Consequently, these systems are interesting for studies using the described mAbs against mouse uPA and MMP9, employing wild-type or the relevant single-deficient mice for treatment. Conclusion Monoclonal antibodies (mAbs) raised in gene-deficient mice have the unique ability to recognize epitopes conserved between species. Additional reactivity with species-specific epitopes provides a more diverse reactivity with the antigen than mAbs raised in wild-type mice. Furthermore, some studies have suggested cross-reacting mAbs to be superior inhibitors to mAbs only reacting with species-specific epitopes. Compared to small molecular inhibitors, the mAbs possess an amazing selectivity, have a long half-life in circulation (3–6 days; Pass et al., 2007 ; Lund et al., 2008 ), and are therefore well-suited for therapy experiments in mouse models of different diseases, including cancer. Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3133267/

The Nudix Hydrolase CDP-Chase, a CDP-Choline Pyrophosphatase, Is an Asymmetric Dimer with Two Distinct Enzymatic Activities ▿

A Nudix enzyme from Bacillus cereus (NCBI RefSeq accession no. NP_831800 ) catalyzes the hydrolysis of CDP-choline to produce CMP and phosphocholine. Here, we show that in addition, the enzyme has a 3′→5′ RNA exonuclease activity. The structure of the free enzyme, determined to a 1.8-à resolution, shows that the enzyme is an asymmetric dimer. Each monomer consists of two domains, an N-terminal helical domain and a C-terminal Nudix domain. The N-terminal domain is placed relative to the C-terminal domain such as to result in an overall asymmetric arrangement with two distinct catalytic sites: one with an "enclosed" Nudix pyrophosphatase site and the other with a more open, less-defined cavity. Residues that may be important for determining the asymmetry are conserved among a group of uncharacterized Nudix enzymes from Gram-positive bacteria. Our data support a model where CDP-choline hydrolysis is catalyzed by the enclosed Nudix site and RNA exonuclease activity is catalyzed by the open site. CDP-Chase is the first identified member of a novel Nudix family in which structural asymmetry has a profound effect on the recognition of substrates.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10400928/

The global burden of neglected zoonotic diseases: Current state of evidence

The majority of emerging infectious diseases are zoonoses, most of which are classified as "neglected". By affecting both humans and animals, zoonoses pose a dual burden. The disability-adjusted life year (DALY) metric quantifies human health burden since it combines mortality and morbidity. This review aims to describe and analyze the current state of evidence on neglected zoonotic diseases (NZDs) burden and start a discussion on the current understanding of the global burden of NZDs. We identified 26 priority NZDs through consulting three international repositories for national prioritization exercises. A systematic review of global and national burden of disease (BoD) studies was conducted using pre-selected databases. Data on diseases, location and DALYs were extracted for each eligible study. A total of 1887 records were screened, resulting in 74 eligible studies. The highest number of BoD was found for non-typhoidal salmonellosis (23), whereas no estimates were found for West Nile, Marburg and Lassa fever. Geographically, the highest number of studies was performed in the Netherlands (11), China (5) and Iran (4). The number of BoD retrieved mismatched the perceived importance in national prioritization exercises. For example, anthrax was considered a priority NZD in 65 countries; however, only one national study estimating BoD was retrieved. By summing the available global estimates, the selected NZDs caused at least 21 million DALYs per year, a similar order of magnitude to (but less than) the burden due to foodborne disease (included in the Foodborne Disease Burden Epidemiology Reference Group). The global burden of disease landscape of NZDs remains scattered. There are several priority NZDs for which no burden estimates exist, and the number of BoD studies does not reflect national disease priorities. To have complete and consistent estimates of the global burden of NZDs, these diseases should be integrated in larger global burden of disease initiatives. 1 Introduction The World Health Organization (WHO) defines zoonoses as diseases and infections naturally transmitted between vertebrates and humans [ 1 ]. There are several thousand human infectious diseases documented in the literature [ 2 ], and it is estimated that 60% of them and 75% of emerging infections are zoonotic [ 3 ]. Zoonoses have high impacts on human and animal health and the ecosystem and are also responsible for enormous economic losses [ [4] , [5] , [6] , [7] ]. Zoonotic diseases disproportionately affect vulnerable groups in low-and middle-income countries (LMICs) and are often "neglected" regarding the geopolitical attention and relative funding they receive for prevention and control initiatives and research [ 8 , 9 ]. In recent years, the WHO has established a roadmap to eliminate neglected tropical diseases (NTDs), focusing on 20 diseases, including six neglected zoonotic diseases (NZDs): echinococcosis, foodborne trematodiasis, human African trypanosomiasis (HAT), leishmaniosis, rabies, and taeniasis/cysticercosis [ 10 ]. The Foodborne Diseases Burden Epidemiology Reference Group (FERG) established by the WHO and Institute for Health Metrics and Evaluation (IHME) have carried out and set standards for Global Burden of Disease (GBD) studies and the global burden of foodborne diseases, quantifying population health losses associated with diseases [ [25] , [39] ]. The metric central to these burden of disease (BoD) frameworks is the Disability-Adjusted Life Year (DALY) [ 11 ]. The DALY is a summary measure of public health that combines the effects of mortality (Years of Life Lost) and morbidity (Years Lived with Disability) into a single metric. This metric also integrates different health states defined by the disease model [ 12 ]. As a result, DALYs encompass all disease impacts of NZDs, including disabling acute and chronic outcomes, rather than just focusing on fatal outcomes [ 13 ]. Additionally, when the same methodology is used, the impacts of different diseases or injuries can be directly compared, which helps policymakers to set priorities and advocacy. There has also been development of methods to combine the human and animal health burdens caused by zoonoses that are consistent with DALYs, such as zDALY, which adds a time trade-off component for animal morbidity and mortality [ 14 ]. This review aims to describe and discuss the current state of evidence on the human burden of NZDs and the knowledge gaps that surround them. The results will help to quantify the current understanding of the public health impact of NZDs and provide a resource for more robust and comparable decision-making. The review considers only the burden that zoonoses pose on humans and does not address burden estimations for the livestock sector, as this will be covered by other areas of the Global Burden of Animal Diseases programme (GBADs), of which this review is part. 2 Methods 2.1 Selection of the neglected zoonotic diseases for the review The zoonotic diseases investigated were selected based on the national prioritization exercises from the CDC One Health Zoonotic Disease Prioritization Process Overview [ 15 ] and the WHO Joint External Evaluation Mission Reports (JEE) [ 16 ], as these are considered authoritative sources. We compared our findings with the WHO document " Ending the Neglect to Attain the Sustainable Development Goals: a Roadmap for Neglected Tropical Disease 2021-2030 " [ 10 ]. The data extraction from these exercises and reports was conducted in July 2021, when the JEE Mission Reports comprised 116 completed reports and the CDC One Health Zoonotic Disease Prioritization Process Overview page was updated on 18th May 2021. A complete list of countries considered by the CDC and the JEE can be found in the additional documents (supporting document pp. 1–4). When extracting the data, we counted the number of times each disease was included as a priority by the country. We included the first ten diseases for each country in our calculation since most countries identified ten or fewer priority pathogens. This resulted in a list of 52 zoonotic diseases. For the scope of this review, only the first 25 priority NZDs were selected. Additionally, we included foodborne trematodiasis based on the WHO roadmap [ 10 ]. Table 1 shows the diseases included in the review, and specifies the number of times each disease was mentioned in a prioritization exercise and if it was included in the WHO roadmap. Table 1 Prioritization frequencies of neglected zoonotic diseases and status of inclusion in the WHO roadmap. Table 1 Disease Frequency ⁎ Included in the WHO roadmap Rabies 94 X Brucellosis 75 Influenza a (h5n1 and h1n1) 72 Anthrax 65 Bovine tuberculosis 47 Rift valley fever 28 Non-typhoidal salmonellosis 25 Ebola virus 20 Leptospirosis 20 Crimean/congo hemorrhagic fever 17 Plague 14 Yellow fever 11 Alveolar echinococcosis 9 X Cystic echinococcosis 9 X Human african trypanosomiasis 9 X Japanese encephalitis 8 Lassa fever 8 Marburg virus disease 8 Cysticercosis 7 X Q fever 7 Toxoplasmosis 7 Campylobacteriosis 5 Glanders 5 Leishmaniosis 5 X West nile disease 5 Foodborne trematodiases / X ⁎ Number of times the disease was mentioned for a country in either the CDC one health zoonotic disease prioritization process or in the WHO joint external evaluation mission reports. The prioritization exercises and the WHO roadmap had different degrees of detail in defining diseases (e.g., leishmaniosis, covering visceral and cutaneous ones). These differences were also visible in the literature review and prioritization exercises (supporting document pp. 5–6). The results acknowledge specific sub-groups or divisions of diseases, where applicable. 2.2 Searching and eligibility criteria To identify the available evidence on the burden of the selected priority NZDs, a review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement (supporting document pp. 7–8) [ 17 ]. PubMed, Web of Science, and Embase were systematically searched for relevant articles using terms covering DALYs and pathogens. A complete list of the keywords used is presented in the supporting document (9–17). Only peer-reviewed articles published between January 1990 and November 2021 were included. Studies published before 1990 were omitted since the DALY metric was introduced in the mid-1990s [ 11 ]. Studies not focusing on humans were excluded. In addition, only studies primarily aimed to determine the burden with the DALY methodology were included. Hence, studies that presented DALY estimates, but not as the main objective, were excluded (e.g., cost-effectiveness, life-cycle assessment, and quantitative microbial risk assessment studies). Studies reporting on cause-specific subsets of GBD estimates were also excluded from the search since the most recent GBD data on these conditions were extracted regardless. Nevertheless, reference lists of the studies that used GBD data were screened to find additional peer-reviewed articles. No language restriction was applied. Publications with insufficiently detailed information, such as abstracts, editorials, or letters to editors, were excluded. 2.3 Data screening, selection and extraction The authors used Rayyan, a tool developed to manage citations, screen abstracts and apply inclusion and exclusion criteria, in order to create a database of unique titles [ 18 ]. The articles screening was conducted by one researcher (CDB). The study supervisor (BD) provided support for articles difficult to categorise initially. For each eligible paper where the full text was available, information was extracted using a data extraction grid, and the reference list was searched for additional studies. The following information was extracted: study information, reference population, DALY result and total population (supporting document p. 18). CDB performed data extraction, and NV reviewed the information extracted. The results have been displayed using the R program and Drawio [ [19] , [20] , [21] ]. 2.1 Selection of the neglected zoonotic diseases for the review The zoonotic diseases investigated were selected based on the national prioritization exercises from the CDC One Health Zoonotic Disease Prioritization Process Overview [ 15 ] and the WHO Joint External Evaluation Mission Reports (JEE) [ 16 ], as these are considered authoritative sources. We compared our findings with the WHO document " Ending the Neglect to Attain the Sustainable Development Goals: a Roadmap for Neglected Tropical Disease 2021-2030 " [ 10 ]. The data extraction from these exercises and reports was conducted in July 2021, when the JEE Mission Reports comprised 116 completed reports and the CDC One Health Zoonotic Disease Prioritization Process Overview page was updated on 18th May 2021. A complete list of countries considered by the CDC and the JEE can be found in the additional documents (supporting document pp. 1–4). When extracting the data, we counted the number of times each disease was included as a priority by the country. We included the first ten diseases for each country in our calculation since most countries identified ten or fewer priority pathogens. This resulted in a list of 52 zoonotic diseases. For the scope of this review, only the first 25 priority NZDs were selected. Additionally, we included foodborne trematodiasis based on the WHO roadmap [ 10 ]. Table 1 shows the diseases included in the review, and specifies the number of times each disease was mentioned in a prioritization exercise and if it was included in the WHO roadmap. Table 1 Prioritization frequencies of neglected zoonotic diseases and status of inclusion in the WHO roadmap. Table 1 Disease Frequency ⁎ Included in the WHO roadmap Rabies 94 X Brucellosis 75 Influenza a (h5n1 and h1n1) 72 Anthrax 65 Bovine tuberculosis 47 Rift valley fever 28 Non-typhoidal salmonellosis 25 Ebola virus 20 Leptospirosis 20 Crimean/congo hemorrhagic fever 17 Plague 14 Yellow fever 11 Alveolar echinococcosis 9 X Cystic echinococcosis 9 X Human african trypanosomiasis 9 X Japanese encephalitis 8 Lassa fever 8 Marburg virus disease 8 Cysticercosis 7 X Q fever 7 Toxoplasmosis 7 Campylobacteriosis 5 Glanders 5 Leishmaniosis 5 X West nile disease 5 Foodborne trematodiases / X ⁎ Number of times the disease was mentioned for a country in either the CDC one health zoonotic disease prioritization process or in the WHO joint external evaluation mission reports. The prioritization exercises and the WHO roadmap had different degrees of detail in defining diseases (e.g., leishmaniosis, covering visceral and cutaneous ones). These differences were also visible in the literature review and prioritization exercises (supporting document pp. 5–6). The results acknowledge specific sub-groups or divisions of diseases, where applicable. 2.2 Searching and eligibility criteria To identify the available evidence on the burden of the selected priority NZDs, a review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement (supporting document pp. 7–8) [ 17 ]. PubMed, Web of Science, and Embase were systematically searched for relevant articles using terms covering DALYs and pathogens. A complete list of the keywords used is presented in the supporting document (9–17). Only peer-reviewed articles published between January 1990 and November 2021 were included. Studies published before 1990 were omitted since the DALY metric was introduced in the mid-1990s [ 11 ]. Studies not focusing on humans were excluded. In addition, only studies primarily aimed to determine the burden with the DALY methodology were included. Hence, studies that presented DALY estimates, but not as the main objective, were excluded (e.g., cost-effectiveness, life-cycle assessment, and quantitative microbial risk assessment studies). Studies reporting on cause-specific subsets of GBD estimates were also excluded from the search since the most recent GBD data on these conditions were extracted regardless. Nevertheless, reference lists of the studies that used GBD data were screened to find additional peer-reviewed articles. No language restriction was applied. Publications with insufficiently detailed information, such as abstracts, editorials, or letters to editors, were excluded. 2.3 Data screening, selection and extraction The authors used Rayyan, a tool developed to manage citations, screen abstracts and apply inclusion and exclusion criteria, in order to create a database of unique titles [ 18 ]. The articles screening was conducted by one researcher (CDB). The study supervisor (BD) provided support for articles difficult to categorise initially. For each eligible paper where the full text was available, information was extracted using a data extraction grid, and the reference list was searched for additional studies. The following information was extracted: study information, reference population, DALY result and total population (supporting document p. 18). CDB performed data extraction, and NV reviewed the information extracted. The results have been displayed using the R program and Drawio [ [19] , [20] , [21] ]. 3 Results 3.1 Flowchart of selected studies A total of 1887 entries were retrieved from the selected databases. After removing the duplicates and applying the eligibility assessment, 73 studies were included in the review. Additionally, the most recent estimates of GBD (2019) were identified and included [ 22 ]. Thus, 74 studies were included in the review ( Fig. 1 ). Fig. 1 Flow chart of study selection. Fig. 1 3.2 Burden of disease studies by location and year Of the 74 studies included, 60 provided national or sub-national burden estimates (supporting document pp. 19–25). Nine studies were global, including the estimates from the latest GBD version (nine diseases). We excluded previous versions of the GBD to avoid overrepresentation. Two BoD assessments focused on specific regions of the world, one on the European Union and one on Asia and Africa. The highest number of single-country BoD assessments were observed in the Netherlands ( N = 11), followed by China ( N = 5) and Iran ( N = 4) (supporting document p. 26). The first BoD assessment was published in 2000 [ 26 ], excluding studies carried out by the GBD (which updates its estimates frequently). After 2008, the number of studies published for this disease set increased, reaching nine publications in 2017 alone (supporting document p.26). 3.3 Burden of disease studies by disease The number of burden estimates varied significantly between the different diseases. Out of the 73 publications, 47 focused on a single disease. The most common group investigated was foodborne diseases ( N = 14); within these publications, two are part of the FERG program [ 38 , 37 ]. Other sets of diseases examined together were grouped by the general label of infectious diseases ( N = 4) [ 25 , [40] , [41] , [42] ], arboviral diseases (N = 1) [ 43 ], and parasitic zoonoses (N = 1) [ 44 ]. One study focused on specific sequelae, namely the post-infectious irritable bowel syndrome, and quantified the DALYs caused by each disease for these specific sequelae [ 46 ]. Finally, two studies looked at the general BoD (e.g., diseases, injuries and risk factors), one from the GBD and the second in Iran [ 22 , 47 ]. The diseases that had the highest number of BoD studies were non-typhoidal salmonellosis ( N = 24), campylobacteriosis ( N = 22), toxoplasmosis ( N = 1), cystic echinococcosis ( N = 13), cysticercosis (N = 13), rabies ( N = 12), brucellosis ( N = 11), and alveolar echinococcosis ( N = 7) (supporting document p. 27). Five BoD studies were found for HAT; one focused exclusively on Rhodesian HAT, one addressed Gambian HAT, another estimated the burden for both (providing DALYs for each sub-type), and two did not distinguish between them. The same approach was found in the studies on leishmaniosis ( N = 5). One study focused only on the cutaneous type, one estimated the independent burden for each kind (cutaneous and visceral), and three provided no distinctions. Four studies investigated the burden of leptospirosis. Three publications provided estimates for bovine tuberculosis and foodborne trematodiases, with one estimate quantifying only the burden for clonorchiasis. Two estimates were identified for Q fever, influenza A H5N1, Japanese encephalitis, and yellow fever. Only one BoD assessment was found for Ebola fever, anthrax, Rift Valley fever, and Crimean/Congo hemorrhagic fever. Finally, no estimates were found for West Nile disease, Influenza A subtype H1N1, Marburg virus disease, plague, Lassa fever, and glanders. 3.1 Flowchart of selected studies A total of 1887 entries were retrieved from the selected databases. After removing the duplicates and applying the eligibility assessment, 73 studies were included in the review. Additionally, the most recent estimates of GBD (2019) were identified and included [ 22 ]. Thus, 74 studies were included in the review ( Fig. 1 ). Fig. 1 Flow chart of study selection. Fig. 1 3.2 Burden of disease studies by location and year Of the 74 studies included, 60 provided national or sub-national burden estimates (supporting document pp. 19–25). Nine studies were global, including the estimates from the latest GBD version (nine diseases). We excluded previous versions of the GBD to avoid overrepresentation. Two BoD assessments focused on specific regions of the world, one on the European Union and one on Asia and Africa. The highest number of single-country BoD assessments were observed in the Netherlands ( N = 11), followed by China ( N = 5) and Iran ( N = 4) (supporting document p. 26). The first BoD assessment was published in 2000 [ 26 ], excluding studies carried out by the GBD (which updates its estimates frequently). After 2008, the number of studies published for this disease set increased, reaching nine publications in 2017 alone (supporting document p.26). 3.3 Burden of disease studies by disease The number of burden estimates varied significantly between the different diseases. Out of the 73 publications, 47 focused on a single disease. The most common group investigated was foodborne diseases ( N = 14); within these publications, two are part of the FERG program [ 38 , 37 ]. Other sets of diseases examined together were grouped by the general label of infectious diseases ( N = 4) [ 25 , [40] , [41] , [42] ], arboviral diseases (N = 1) [ 43 ], and parasitic zoonoses (N = 1) [ 44 ]. One study focused on specific sequelae, namely the post-infectious irritable bowel syndrome, and quantified the DALYs caused by each disease for these specific sequelae [ 46 ]. Finally, two studies looked at the general BoD (e.g., diseases, injuries and risk factors), one from the GBD and the second in Iran [ 22 , 47 ]. The diseases that had the highest number of BoD studies were non-typhoidal salmonellosis ( N = 24), campylobacteriosis ( N = 22), toxoplasmosis ( N = 1), cystic echinococcosis ( N = 13), cysticercosis (N = 13), rabies ( N = 12), brucellosis ( N = 11), and alveolar echinococcosis ( N = 7) (supporting document p. 27). Five BoD studies were found for HAT; one focused exclusively on Rhodesian HAT, one addressed Gambian HAT, another estimated the burden for both (providing DALYs for each sub-type), and two did not distinguish between them. The same approach was found in the studies on leishmaniosis ( N = 5). One study focused only on the cutaneous type, one estimated the independent burden for each kind (cutaneous and visceral), and three provided no distinctions. Four studies investigated the burden of leptospirosis. Three publications provided estimates for bovine tuberculosis and foodborne trematodiases, with one estimate quantifying only the burden for clonorchiasis. Two estimates were identified for Q fever, influenza A H5N1, Japanese encephalitis, and yellow fever. Only one BoD assessment was found for Ebola fever, anthrax, Rift Valley fever, and Crimean/Congo hemorrhagic fever. Finally, no estimates were found for West Nile disease, Influenza A subtype H1N1, Marburg virus disease, plague, Lassa fever, and glanders. 4 Discussion This review aims to provide a comprehensive overview of the evidence on BoD studies for NZDs and to start a reflection on the current understanding of the global burden of NZDs. Seventy-four studies met our inclusion criteria, including the latest version of the GBD. Over two-thirds of the studies were national or sub-national. The highest number of BoD studies was found for the Netherlands, which mainly reported estimates for foodborne diseases but also on Influenza A N5H1 ( N = 2) and Q fever (N = 2) [ 41 , 48 ]. This finding was somewhat surprising given that they were not part of the CDC and WHO prioritization exercises, which focused on LMICs and thus did not contribute to our priority list. A possible explanation would be that national BoD studies are well established as input for policymaking and conducted regularly in the Netherlands [ 49 ]. Furthermore, they do not only include diseases endemic to their own country but also a subset of the global priority list. No burden studies were found in South America or South Asia. Few were retrieved from Sub-Saharan Africa, even if most of the diseases selected are considered endemic to these regions. Over the period taken into consideration, we observed an increase in the number of BoD assessments for the selected diseases. This could reflect an increasing recognition of the DALY metric in decision-making. It could also indicate a rising effort of international communities and projects to acknowledge the public health impact of zoonoses. In 2005, for example, the WHO set up a series of meetings on neglected zoonotic diseases [ 50 ] and, in 2006, launched the initiative to estimate the Global Burden of Foodborne Diseases [ 51 ], which includes many NZDs. Differences were observed between the frequency of diseases listed in country prioritization exercises and the number of BoD estimates included in the review ( Fig. 2 ). This could be caused by miscommunication between the different actors (e.g. policymakers, researchers, international organizations, etc.) due to their different aims. Some diseases frequently listed in the prioritization exercises had few BoD assessments. Anthrax, for example, was listed as a priority by 65 countries, but only one BoD study has been published. Notably, anthrax is not included in the GBD or the WHO roadmap, even though it is the only bacterial NZD included in the World Health Assembly resolution WHA 66.12 on NTDs [ 53 ]. For rabies, which had the highest frequency in the prioritization exercises ( N = 94) and is included in the GBD and the WHO road map, only twelve BoD estimates were found. Fig. 2 Differences in the prioritization exercises from the WHO and CDC and the number of burden of disease studies found. Fig. 2 In this review, we included both (sub-)national and global estimates. Local and global estimates serve different purposes, and both have strengths and weaknesses. BoD assessments at the local level strengthen the local health information systems, improve the understanding of population health and define local priorities. In contrast, global estimates aggregate the disease burden worldwide and help establish priorities at the global level. There were also diseases with a low frequency in prioritization exercises but a high number of BoD assessments, such as campylobacteriosis ( N = 22) and toxoplasmosis ( N = 16) ( Fig. 2 ). The mismatch might suggest that some countries or academic groups may not use DALYs for their BoD assessments, or indicate a lacking capacity to carry on prioritization exercises to the next step. Furthermore, countries participating in prioritization exercises and the diseases that burden them most do not necessarily overlap with the countries carrying out BoD assessments and their priorities. Finally, factors such as missing data on duration or severity make the burden estimation for specific diseases through the standard BoD framework difficult and could also contribute to the mismatch. A disease considered a disabler (e.g. low mortality rate), such as toxoplasmosis, might be overlooked in LMICs and gain more attention in high-income countries, where BoD studies are more common. Thus, some NZDs could be considered "neglected among the neglected" because they are overlooked in different contexts. Some of the prioritised NZDs are also part of other domains such as food safety, antimicrobial resistance, diarrheal disease, and maternal or neonatal health, which might contribute to their higher number of estimates. When comparing diseases for which global burden estimates with the same methodology were available, the number of prioritization studies did not always align with the severity of the burden ( Table 2 ). Brucellosis, for example, prioritised by 75 countries, has eleven estimates (national and global) and resulted in 264,073 (100540–6,187,148) DALYs globally in 2010 according to FERG. On the other hand, bovine TB caused 607,775 (458364–826,115) DALYs in 2010 but was only prioritised by 47 countries and had three burden estimates (one national, two FERG) [ 38 , 54 ]. Table 2 DALYs per the selected diseases at the global level and the number of prioritisation exercises conducted for them ⁎ Table 2 Disease Source DALYs Result from the prioritization Toxoplasmosis, acquired FERG ⁎⁎ 153,779 (772676–1,733,114) 7 Cystic echinococcosis FERG ⁎⁎ 183,573 (88082–1590 46) 9 Brucellosis FERG ⁎⁎ 264,073 (100540–6,187,148) 75 Toxoplasmosis, congenital FERG ⁎⁎ 526,515 (359756–835,537) 7 TB bovine FERG ⁎⁎ 607,775 (458364–826,115) 47 Alveolar echinococcosis FERG ⁎⁎ 687,823 (409190–1,106,320) 9 Foodborne trematodiasis FERG ⁎⁎ 2,024,592 (1652243–2,483,514) – Cysticercosis FERG ⁎⁎ 2,788,426 (213763–3,606,582) 9 Campylobacteriosis FERG ⁎⁎ 3,733,822 (2857037–5,273,652) 5 Non-typhoidal salmonellosis FERG ⁎⁎ 4,377,930 (3242020–7,175,522) 25 African trypanosomiasis GBD 82,615 (37636–155,791) 9 Cystic echinococcosis GBD 122,457 (89244–168,556) 9 Ebola GBD 195,394 (230578–160,083) 20 Yellow fever GBD 290,137 (107073–597,713) 11 Leishmaniosis GBD 696,703 (375207–1,619,382) 5 Foodborne trematodiasis GBD 780,089 (385735–1,446,031) – Rabies GBD 782,052 (320289–1,081,217) 94 Cysticercosis GBD 1,371,067 (874432–1,960,855) 7 Non-typhoidal salmonellosis (invasive) GBD 6,114,262 (3323425–9,705,738) 25 Rift Valley Fever Labeaud et al. (2011) 6156 (353–11,958) 28 Japanese encephalitis Labeaud et al. (2011) 1,062,474 (265778–1,859,170) 8 Leptospirosis Torgerson et al. (2015) 2,900,000 (1250000–4,540,000) 20 ⁎ This table only provides an overview of the different global estimates; it is important to note that they are not directly comparable since they come from different sources that applied different methodologies (as is visible, for example, from the estimates of non-typhoidal salmonella, for more information see [ 55 ]). ⁎⁎ For FERG, the estimates are reported by all transmission routes, not just foodborne. 4.1 Disability Adjusted Life Years estimates for zoonoses The most recent GBD only covered nine of the NZDs selected for this review, which accounted for approximately 10 million DALYs.24 Due to the limited number of diseases included, this number is, however, an underestimation of the total burden of NZDs. When combining the estimates for all diseases with global BoD studies, using the most recent study or, whenever available, the estimate without age or time discounting, the total burden amounts to over 21·5 million DALYs. While this result only covers seventeen diseases ( Table 2 ), it illustrates that the global burden of NZDs is substantial. To put it into perspective, the global DALY estimate due to infectious foodborne diseases calculated by FERG was 33 million in 2010, and the 2019 GBD estimate of the burden of enteric diseases was 96·8 million DALYs. Our crude global NZD DALY estimate was calculated by manually adding (the medians of) estimates based on different methodological choices and assumptions and should be interpreted with great caution. To have complete and consistent estimates of the global burden of NZDs, these diseases would need to be integrated into larger global BoD initiatives. This will not only help to understand the real burden of these diseases and help define priorities based on evidence but could also inform policies aiming to eradicate these "neglected" diseases. 4.2 Zoonosis: where does its burden lie? There is an ambiguity in what constitutes zoonoses and the role of animals as reservoirs. The WHO defines zoonoses as "diseases and infections naturally transmitted between vertebrates and humans" [ 56 ]. For this review, we interpreted this as the infection being maintained in an animal population, the reservoir, and a continuous source of human infection. Humans acquire zoonotic infections through direct contact with animals or indirect exposure routes such as vector-borne or environmental pathogens associated with the food system [ 57 ]. This first interpretation of zoonoses includes only diseases that are at stages two and three of the five-step framework for the evolution from animal to human diseases as proposed by Wolfe et al., where the animal is necessary and the pathogens can undergo only a few cycles of human-to-human transmission [ 59 ]. However, the zoonosis definition could also refer to zoonotic origins of the disease, but transmission irrespective of the animal reservoir. Examples of this would be COVID-19 or the human immunodeficiency virus (HIV), which started as an animal infection but spread to humans at some point (this phenomenon is called spillover) and later mutated into human-only strains, or stage five as described by Wolfe et al. [ 59 , 60 ]. Differentiating between diseases that may have originated in animals but independently persist in human populations and diseases that require a non-human animal host for pathogen transmission and survival enables more targeted and strategic initiatives for prevention and control. For instance, to tackle the spread of HIV, interventions do not focus on animals since the infection is mainly transmitted from human to human. On the other hand, for diseases such as rabies or brucellosis, interventions target the animal host because a permanent animal reservoir is needed to sustain the epidemic. Making this differentiation also helps determine where the disease burden lies (human or animal population). This helps understand changes in human morbidity and/or mortality as well as animal health/production and premature mortality due to disease and contributes to improving human health and animal productivity. In contrast, there are diseases like dengue where the role of the animal reservoirs is not yet clear, making it difficult to understand where the burden lies and determine the most effective initiatives [ 61 ]. Under the One Health approach, understanding where the burden lies will help assess the direct impact of the disease (on both animals and humans) and indirect ones, such as decreases in household incomes due to production losses, which may also affect health. 4.3 Disability-Adjusted Life Years as standard burden of disease metric The use of DALYs implies multiple methodological choices and assumptions [ 62 ]. Thus, direct comparison between different estimates could lead to incorrect interpretation. For example, FERG and GBD estimates are structurally different, with the first producing an incident DALY and the second a prevalent DALY; consequently, although we presented them together ( Table 2 ) to provide an overview, they should not be directly compared. Notably, there is no single or preferred way to estimate DALYs. Each methodology has its strengths and weaknesses. Qualitative or semi-quantitative prioritization exercises often reflect the notoriety or the perceived risk of a disease rather than the real threat or burden. Using DALYs to establish priorities sets up a more evidence-based and internally consistent framework for disease prioritization, limiting the participants' biases or specific interests in semi-quantitative prioritization exercises. Prioritization should consider both local concerns and DALY estimates. Diseases that do not have DALY estimates or do not appear in the GBD will not receive the proper attention and probably be included in a category such as "others". Indeed, it is interesting that some of the prioritised diseases (e.g., influenza A H1N1, plague and Lassa fever) are not acknowledged by international communities and do not have a burden estimate. This increases the chances that their importance is underestimated, especially if they mainly occur in LMICs. 4.1 Disability Adjusted Life Years estimates for zoonoses The most recent GBD only covered nine of the NZDs selected for this review, which accounted for approximately 10 million DALYs.24 Due to the limited number of diseases included, this number is, however, an underestimation of the total burden of NZDs. When combining the estimates for all diseases with global BoD studies, using the most recent study or, whenever available, the estimate without age or time discounting, the total burden amounts to over 21·5 million DALYs. While this result only covers seventeen diseases ( Table 2 ), it illustrates that the global burden of NZDs is substantial. To put it into perspective, the global DALY estimate due to infectious foodborne diseases calculated by FERG was 33 million in 2010, and the 2019 GBD estimate of the burden of enteric diseases was 96·8 million DALYs. Our crude global NZD DALY estimate was calculated by manually adding (the medians of) estimates based on different methodological choices and assumptions and should be interpreted with great caution. To have complete and consistent estimates of the global burden of NZDs, these diseases would need to be integrated into larger global BoD initiatives. This will not only help to understand the real burden of these diseases and help define priorities based on evidence but could also inform policies aiming to eradicate these "neglected" diseases. 4.2 Zoonosis: where does its burden lie? There is an ambiguity in what constitutes zoonoses and the role of animals as reservoirs. The WHO defines zoonoses as "diseases and infections naturally transmitted between vertebrates and humans" [ 56 ]. For this review, we interpreted this as the infection being maintained in an animal population, the reservoir, and a continuous source of human infection. Humans acquire zoonotic infections through direct contact with animals or indirect exposure routes such as vector-borne or environmental pathogens associated with the food system [ 57 ]. This first interpretation of zoonoses includes only diseases that are at stages two and three of the five-step framework for the evolution from animal to human diseases as proposed by Wolfe et al., where the animal is necessary and the pathogens can undergo only a few cycles of human-to-human transmission [ 59 ]. However, the zoonosis definition could also refer to zoonotic origins of the disease, but transmission irrespective of the animal reservoir. Examples of this would be COVID-19 or the human immunodeficiency virus (HIV), which started as an animal infection but spread to humans at some point (this phenomenon is called spillover) and later mutated into human-only strains, or stage five as described by Wolfe et al. [ 59 , 60 ]. Differentiating between diseases that may have originated in animals but independently persist in human populations and diseases that require a non-human animal host for pathogen transmission and survival enables more targeted and strategic initiatives for prevention and control. For instance, to tackle the spread of HIV, interventions do not focus on animals since the infection is mainly transmitted from human to human. On the other hand, for diseases such as rabies or brucellosis, interventions target the animal host because a permanent animal reservoir is needed to sustain the epidemic. Making this differentiation also helps determine where the disease burden lies (human or animal population). This helps understand changes in human morbidity and/or mortality as well as animal health/production and premature mortality due to disease and contributes to improving human health and animal productivity. In contrast, there are diseases like dengue where the role of the animal reservoirs is not yet clear, making it difficult to understand where the burden lies and determine the most effective initiatives [ 61 ]. Under the One Health approach, understanding where the burden lies will help assess the direct impact of the disease (on both animals and humans) and indirect ones, such as decreases in household incomes due to production losses, which may also affect health. 4.3 Disability-Adjusted Life Years as standard burden of disease metric The use of DALYs implies multiple methodological choices and assumptions [ 62 ]. Thus, direct comparison between different estimates could lead to incorrect interpretation. For example, FERG and GBD estimates are structurally different, with the first producing an incident DALY and the second a prevalent DALY; consequently, although we presented them together ( Table 2 ) to provide an overview, they should not be directly compared. Notably, there is no single or preferred way to estimate DALYs. Each methodology has its strengths and weaknesses. Qualitative or semi-quantitative prioritization exercises often reflect the notoriety or the perceived risk of a disease rather than the real threat or burden. Using DALYs to establish priorities sets up a more evidence-based and internally consistent framework for disease prioritization, limiting the participants' biases or specific interests in semi-quantitative prioritization exercises. Prioritization should consider both local concerns and DALY estimates. Diseases that do not have DALY estimates or do not appear in the GBD will not receive the proper attention and probably be included in a category such as "others". Indeed, it is interesting that some of the prioritised diseases (e.g., influenza A H1N1, plague and Lassa fever) are not acknowledged by international communities and do not have a burden estimate. This increases the chances that their importance is underestimated, especially if they mainly occur in LMICs. 5 Limitations The review has several limitations. First, we could have missed out on some eligible and valid BoD studies, given that we did not consider ongoing estimates or BoD studies performed but not documented in peer-reviewed articles. Second, we acknowledge that the list of priority NZDs that we used is biased to a subset of global priorities and from the countries that took part in the exercises; however, it is informative since it is derived from established policy documentation. Third, we realize that using the number of identified BoD studies as an endpoint for determining whether disease prioritization aligns with the availability of burden estimates may lead to bias in different ways, which should be kept in mind when interpreting the results. Our choice to include only the most recent GBD study to avoid skewing the results influenced the number of BoD studies available for several diseases, especially for those who have been part of the GBD for a long time [ 63 ]. 6 Future prospects This review aimed to report the current state of BoD assessments for the set of selected diseases and to reflect on the current understanding of the public health impact of zoonoses. The findings can serve to improve the knowledge and advocacy of zoonoses. This is especially important for diseases that currently do not have BoD estimates and may be neglected by governments or policymakers. Consequently, future research should focus on acquiring the data needed to quantify the burden of these diseases, such as incidence or prevalence data, mortality, duration and disease models. Our findings highlight the need for more global BoD estimates for diseases such as anthrax, given their high frequencies in the prioritization exercises and their low representation in BoD assessments on a national or international level. On the other hand, for diseases with sufficient estimates, such as brucellosis and toxoplasmosis, it could be impactful to include them in the GBD so that estimates could be compared to other diseases to provide a perspective on their relative impact. Indeed, this review suggests broadening the scope of GBD or FERG to include more diseases and increase the comparability between diseases at the global level. BoD assessments at the local level allow for a detailed look into data quality to strengthen local health information systems and better understand population health and will support local policymaking. Different efforts have been made in this field; for instance, FERG carried out different capacity-building activities and encouraged the use of information provided by BoD for evidence-informed policies. This review suggests that these efforts need to continue and stresses the importance of addressing methodological limitations inherent in the standard DALY approach. It is important to move towards a One Health approach to understand the full impact of zoonoses, develop a systematic methodology to describe the impact of animal diseases on society, including human health, and close the gap between human and animal health. This review is part of a broader initiative that aims to establish a systematic methodology for assessing the impact of animal diseases on society. The goal is to provide a comprehensive understanding of the burden of zoonotic diseases on both humans and animals. To achieve this, it is crucial to complement these estimates with similar studies focusing on the impact of these diseases on animals. 7 Conclusion This review aimed to explore the current state of evidence on the BoD estimates for selected zoonoses and to reflect on understanding these diseases and their estimates. The results showed that not all of the diseases had BoD estimates and that the numbers of BoD estimates do not reflect the frequency of the diseases in prioritization exercises. This highlights the need for further research on zoonoses in order to have a better understanding of how each disease affects humans. Declarations This research is performed in the framework of the Global Burden of Animal Diseases (GBADs) programme which is led by the University of Liverpool and the World Organization for Animal Health (OIE) ( https://animalhealthmetrics.org/ ). This research is supported through the Grant Agreement Investment ID INV-005366 with the Bill & Melinda Gates Foundation and the UK Foreign, Commonwealth and Development Office (FCDO). GBADs case studies receive additional funding from the following: European Commission, Australian Centre for International Agricultural Research (ACIAR), Brooke Foundation and the Food and Agriculture Organization of the United Nations (FAO). A full list of the GBADs collaborators can be accessed here: https://animalhealthmetrics.org/acknowledgements . Declaration of Competing Interest All authors declare that they have no conflicts of interest. Appendix A Supplementary data Supplementary material Image 1 Data availability Data will be made available on request.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3064659/

Bacillus anthracis Interacts with Plasmin(ogen) to Evade C3b-Dependent Innate Immunity

The causative agent of anthrax, Bacillus anthracis , is capable of circumventing the humoral and innate immune defense of the host and modulating the blood chemistry in circulation to initiate a productive infection. It has been shown that the pathogen employs a number of strategies against immune cells using secreted pathogenic factors such as toxins. However, interference of B. anthracis with the innate immune system through specific interaction of the spore surface with host proteins such as the complement system has heretofore attracted little attention. In order to assess the mechanisms by which B. anthracis evades the defense system, we employed a proteomic analysis to identify human serum proteins interacting with B. anthracis spores, and found that plasminogen (PLG) is a major surface-bound protein. PLG efficiently bound to spores in a lysine- and exosporium-dependent manner. We identified α-enolase and elongation factor tu as PLG receptors. PLG-bound spores were capable of exhibiting anti-opsonic properties by cleaving C3b molecules in vitro and in rabbit bronchoalveolar lavage fluid, resulting in a decrease in macrophage phagocytosis. Our findings represent a step forward in understanding the mechanisms involved in the evasion of innate immunity by B. anthracis through recruitment of PLG resulting in the enhancement of anti-complement and anti-opsonization properties of the pathogen. Introduction Bacillus anthracis , the causative organism of anthrax, is a spore-forming Gram-positive bacterium. Infection can be initiated by the spores through inhalation, ingestion, or cutaneous abrasions. In inhalational infection, alveolar macrophages phagocytose the spores deposited on the lung surface and deliver them to the regional lymph nodes. In spite of the bactericidal activity of macrophages, some of the engulfed spores survive, germinate into vegetative bacteria, kill the macrophage, and subsequently become released into the lymphatic system. Further growth of the bacteria leads to the hemorrhagic lymphadenitis, allowing the pathogen to break into the bloodstream and spread systemically through the circulation [1] , [2] . In order to initiate a productive infection, B. anthracis needs to circumvent the innate protective response of the host. It has been shown that the pathogen employs a number of strategies against the immune cells using secreted pathogenic factors. For example, lethal and edema toxins secreted along with other virulence factors such as hemolysins within the phagosomal compartment of macrophages allow the bacteria to resist being killed and to escape from the phagocytes [3] . Cleavage of mitogen-activated protein kinase kinases by lethal toxin seems to play a central role in the immunosuppressive capacity of B. anthracis to induce necrosis or apoptosis of macrophages [4] and to inhibit responses of dendritic and T cells [5] – [7] . The pore-forming hemolysin anthrolysin O is able to damage membranes of different immune cell types and to sensitize macrophages to lethal toxin [8] . However, interference of B. anthracis with the innate immune system through specific interaction of the spore surface with the host proteins such as the complement system has heretofore attracted little attention. The complement system facilitates bactericidal activity of normal human serum (NHS) in early clearance of pathogens [9] , [10] . The complement system can be activated through three different pathways: classical, lectin, and alternative. Deposition of complement C3b onto the bacterial surface is a crucial step in eliminating the pathogen. To escape complement-mediated killing, pathogens use a common evasion strategy by acquiring the fluid-phase complement factor H, complement factor H-related proteins (FHRs), complement factor H-like proteins (FHLs), the complement C4-binding protein from host serum [10] , [11] . It has also been observed that B. anthracis can directly infect non-phagocytic cells [12] and invade tissues of the nasopharynx after spore inhalation without needing to be transported by alveolar macrophages to the lymphatics [13] . A number of pathogens bind host zymogen protease plasminogen (PLG) to the bacterial surface for tissue invasion [14] . PLG is an abundant protein found in the plasma and is a central component of the fibrinolytic system. Activation of the fibrinolytic system by PLG has recently been found during B. anthracis infection in mice [15] . In the host, inactive PLG is converted to active plasmin by host-expressed tissue-type PLG activator (tPA) and urokinase (uPA). PLG activation to plasmin by invasive pathogenic bacteria such as Borrelia burgdorferi [16] or Pseudomonas aeruginosa [17] could substantially augment the organism's potential for tissue invasion and necrosis. However, B. anthracis protease InhA can accelerate the uPA-mediated plasminogen activation, thereby suggesting a mechanism of plasmin modulation in anthrax infection [15] . As a component of the exosporium [18] , this protease might be relevant to the invasive properties of the spores. The active plasmin is a broad-spectrum serine protease that dominantly degrades non-collagenous extracellular matrix (ECM) and basal membrane proteins such as laminin and fibronectin [19] . A recent study also showed that plasmin bound to the borrelial surface leads to a drastic decrease in C3b deposition, suggesting that plasmin has anti-opsonic properties [16] . Since functional complement proteins are present in the bronchoalveolar lavage fluid (BALF) [20] , complement C3-dependent opsonization is expected to play an important role in the early stages of inhalational B. anthracis infection. In fact, C3b can bind to B. anthracis spores opsonized by the normal human serum (NHS) and thus enhance phagocytosis by human macrophages [21] . Complement-deficient A/J mice are highly susceptible to the attenuated B. anthracis Sterne strain [22] , and the resistant C57BL/6 mice acquire susceptibility to challenge with the attenuated Sterne strain after depletion of complement by cobra venom injection [23] . The above considerations prompted us to investigate the proteome of NHS bound to B. anthracis spores. Here we provide evidence that PLG binds B. anthracis through surface α-enolase and elongation factor-tu, and its activation to plasmin by uPA results in a reduction in C3b/iC3b deposition in spores. Recent studies suggested that bacterial surface proteins α–enolase and glyceraldehyde-3-phasphate dehydrogenase of B. anthracis bound PLG [24] , [25] . Together with the results, our findings represent a step forward in understanding the mechanisms involved in B. anthracis resistance to complement attack and opsonization resulting in the increased ability of the pathogen to invade the host. Materials and Methods Bacterial strains and reagents B. anthracis non-encapsulated Sterne strain 34F2 [pXO1 + , pXO2 − ] was obtained from the Colorado Serum Company. Non-virulent B. subtilis 168 was purchased from the American Type Culture Collection (Manassas, VA, USA). Human PLG, plasmin, rabbit anti-GroEL polyclonal antibody, and D-Val-Leu-Lys- p -nitroanilide (VLK-pNA, Sigma V7127) were purchased from Sigma. Human complement C3b (product # 204860) and uPA (product # 672112) were purchased from Calbiochem, and rabbit anti-human C3c polyclonal antibody (product # A0062) was from Dako. NHS was obtained from Innovative and goat anti-human PLG polyclonal antibody (ab6189) and rabbit anti-mouse C3 polyclonal antibody (ab11887) from Abcam. Mouse monoclonal antibody (EF12) against an abundant B. anthracis exosporium protein BclA (1 mg/ml; used 1∶1,000 dilution for Western blot) [26] was kindly provided by J.F. Kearney (University of Alabama). Horse raddish peroxidase (HRP)-conjugated secondary antibodies sheep anti-mouse IgG and donkey anti-rabbit IgG antibody were purchased from GE Healthcare and rabbit anti-goat IgG antibody from Jackson Immuno Research. PLG binding onto B. anthracis 1.6×10 9 spores or 100 µl of vegetative cells grown to A 600 of 1.5 were incubated with 150 µl of NHS or 10 µg of purified PLG in binding buffer (50 mM Tris, pH 7.5, 100 mM NaCl, and 2 mM MgCl 2 ) for 1 h at room temperature on a rocker platform and then washed 4 times with binding buffer. The 5 th wash in 100 µl of binding buffer was saved as a wash control, and bound proteins were eluted with 100 µl of 3 M potassium thiocyanate. The eluted protein was separated by SDS-PAGE and analyzed by Western blot with anti-PLG. To examine the effects of amino acids on PLG binding to the pathogen surface, 5×10 7 Sterne spores in binding buffer were incubated with 2 µg of PLG and 50 mM amino acids for 1 h on a rocker platform. The spores were washed 2 times with 1 ml of binding buffer and incubated in 150 µl of 50 mM Tris-HCl, pH 7.5, complemented with 20 units (0.2 ug) of uPA and 150 µl of 200 µM VLK-pNA for 15 min at 37°C. Plasmin activity was measured by reading absorbance of 100 µl of reactants at 405 nm. The spores were suspended in 150 µl of binding buffer and incubated with 4 µg of PLG for 1 h. After 5 washings with 150 µl of the binding buffer, the spores were suspended in 300 µl of 50 mM Tris-HCl, pH 7.5, and incubated with 20 units of uPA and substrate. Plasmin activity was measured as described above. Two-Dimensional SDS-PAGE Spore-bound NHS proteins were eluted by a chaotropic reagent, potassium thiocyanate, as described above and dialyzed in water. Samples were subjected to two-dimensional (2D) electrophoresis as follows: the desalted proteins were dissolved in Zoom 2D protein solubilizer-1 and applied on immobilized pH 3–10 linear gradient strips according to the manufacturer's instructions (Invitrogen). Focusing started at 175 V (15 min), was ramped to 2000 V for 45 min, and finally continued at 2000 V for 30 min in an IPGrunner system (Invitrogen). After focusing, strips were equilibrated for sample buffer and then overlaid onto 4–12% SDS-PAGE. The separated proteins were silver-stained and the bands were excised from the stained gel. Mass spectrometry The potassium thiocyanate-eluted spore-bound proteins or silver-stained protein bands excised from the 2D gel were trypsinized as described [15] . Identification of the proteins was performed by LTQ-tandem MS/MS equipped with a reverse-phase liquid chromatography nanospray tandem MS using a high-resolution LTQ-Orbitrap spectrometer (ThermoFisher). The reverse-phase column was slurry-packed in house with 5 µm, 200-à pore size C 18 resin (Michrom BioResources) in a 100 µm×10 cm fused silica capillary (Polymicro Technologies) with a laser-pulled tip. After sample injection, the column was washed for 5 min at 200 nl/min with 0.1% formic acid, peptides were eluted using a 50-min linear gradient from 0 to 40% acetonitrile and an additional step of 80% acetonitrile (all in 0.1% formic acid) for 5 min. The LTQ-Orbitrap MS was operated in a data-dependent mode in which each full MS scan was followed by five MS-MS scans where the five most abundant molecular ions were dynamically selected and fragmented by collision-induced dissociation using normalized collision energy of 35%. Tandem mass spectra were matched against the National Center for Biotechnology Information mouse database by Sequest Bioworks software (ThermoFisher) using full tryptic cleavage constraints and static cysteine alkylation by iodoacetamide. For a peptide to be considered legitimately identified, it had to be the top number one matched and had to achieve cross-correlation scores of 1.9 for [M+H] 1+ , 2.2 for [M+2H] 2+ , 3.5 for [M+3H] 3+ , ΔCn>0.1, and a maximum probability of randomized identification of 0.01. The MS data were filtered to improve the quality of the data set prior to protein selection. The initial set of proteins was limited to those that could be confidently identified, and was further screened to remove proteins with few non-zero peptide hits. Preparation of recombinant proteins The target genes of B. anthracis were amplified from its chromosomal DNA by PCR with specific oligonucleotides using a Taq polymerase premix (Invitrogen). Primers used in this study were as follows: GroEL (BA0267), forward GCA AAA GAT ATT AAA TTT AGT GAA , reverse CAT CAT TCC GCC CAT ACC GCC ; enolase (BA5364), forward ATG TCA ACA ATT ATT GAT GTT , reverse TCA TCG TTT GAT GTT ATA AAA ; and EF-tu (BA0108), forward ATG GCT AAA GCT AAA TTC GAA , reverse TCA CTC AAC GAT AGT AGC AAC . The amplicons were ligated into expression plasmid pTrcHis2-TOPO (Invitrogen) and then transformed into E. coli DH5α following the manufacturer's instructions. Protein expression was induced with 1 mM isopropyl-β-D-thiogalactoside for 5 h. The 6× His-tagged fusion proteins were isolated under native conditions by Ni 2+ -NTA resin (Probond, Invitrogen) as described in the manufacturer's protocols. For binding assays, purified proteins (320 µg) were also conjugated to carboxylate-modified FluoSpheres (1.0 µm, 500 µl) in the presence of EDAC (1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide) according to the manufacturer's recommendations (Invitrogen). The resulting beads were blocked with 1% BSA and resuspended in 500 µl of PBS. Exosporium extraction and ligand blot analysis Exosporium extracts were prepared by incubating the spore suspension in 0.1M DTT, 0.05% SDS, and 0.1M NaCl, pH 10 for 2.5 h in a 37°C shaking water bath [27] , or by a sonication in 20 mM Tris-HCl, 0.5 mM EDTA, pH 7.5. Sonication was performed using a Microsonix XL ultrasonic cell disruptor (Microson) for seven 1 min bursts (output power 12 W), each separated by 2 min cooling on ice [28] . By centrifugation, spores (exosporium negative) and supernatants were separated. The spore pellets were washed twice with the spore binding buffer as described above, and were subjected to PLG binding assays. For ligand blot analysis, exosporium extracts or recombinant proteins were run on 4–12% SDS-PAGE gel and then electrophoretically transferred onto a nitrocellulose membrane. The membrane was soaked in PBS/0.05 Tween 20 (PBST) containing 1% bovine serum albumin overnight at 4°C to renature the proteins, and then it was incubated with PLG (1 µg/ml in PBST/1% BSA) for 1.5 h at room temperature. The membrane was washed 5 times with PBST and incubated with anti-PLG antibody for 1 h followed by the corresponding horseradish peroxidase (HRP)-conjugated secondary antibody. The blot was visualized by HRP reaction. PLG binding to recombinant receptors MaxiSorp 96 well plates (Nunc) were coated overnight with different concentrations of recombinant proteins (7.8–500 nM) at 4°C. Following 3 washings with PBST, wells were blocked for 1 h at room temperature with 0.1% gelatin/PBS and then washed 3 times. Afterwards, 100 µl/well of human PLG (1 µg/ml) were added and incubated for 2 h at room temperature. Unbound PLG was removed by washing 3 times with PBST. Bound PLG was incubated with anti-PLG antibody (1∶5,000) for 1 h followed by secondary antibody. PLG binding to receptors was colorimetrically measured at 450 nm after sequential addition of a HRP substrate TMB (3,3′,5,5′-tetramethylbenzidine) and sulfuric acid. C3 deposition and degradation on spore surface Spores (2×10 7 /well) were washed, resuspended in PBS, and immobilized onto MaxiSorp microplates (Nunc) overnight at 4°C. After washing with PBST, wells were blocked with PBS/0.2% gelatin for 1 h at room temperature and incubated with 10% NHS (100 µl) for 30 min at room temperature. The wells were washed 2 times with PBST and incubated with 2 µg/well of PLG for 1 h in the presence or absence of protease cocktail (100-fold dilution, Sigma). Bound PLG was activated by uPA (20 units/well) for 3 h at 37°C. Deposited C3b was then detected by incubation with anti-C3c antibody (1∶2,000) followed by HRP-conjugated secondary antibody. C3 deposition was colorimetrically measured after addition of TMB and sulfuric acid at 450 nm. Preparation of rabbit BALF and phagocytosis assays BALF was collected from New Zealand White rabbits infused with 30–40 ml of Hanks' balanced salt solution (HBSS) under the approval of the Institutional Animal Care and Use Committee of the Biocon (Rockville, MD; approval # A0900-09a). The BALF was used after centrifuging at 1,500 rpm for 20 min at 4°C. For macrophage phagocytosis assays, spores were incubated with 250 µg of BALF and/or 25 µg of NHS (as a source of C3) for 1 h at room temperature in the presence or absence of 100 µM leupeptin. Spores (8×10 6 ) were washed twice with PBS and resuspended in PBS. RAW264.7 cells were infected with the spores at MOI of 10 and centrifuged to precipitate spores for 2 min. After 30 min of incubation, the cells were washed 6 times with HBSS and lysed by 2.5% saponin, and phagocytosed spores were counted by a serial dilution method on LB agar plate. Protein staining and immunoblotting Proteins were loaded onto 10% or 4–12% NuPAGE MES gel (Invitrogen) and separated under reducing conditions (32 mM dithiothreitol). Separated proteins were then silver-stained using GelCode SilverSNAP kit (Pierce) according to manufacturer's instructions or immunoblotting. For immunoblotting, the separated proteins were electrophoretically transferred to a nitrocellulose membrane using an iBlot gel transfer system (Invitrogen). After blocking with 5% dried milk solution, the membrane was probed with the primary antibody using PBST containing 5% milk, and was incubated with the corresponding HRP-coupled secondary antibody (1∶10,000 dilution) for 1 h at room temperature. Then the membrane was washed in PBST and visualized with the most sensitive West Femto chemiluminescent substrate system (Thermo Scientific). Statistical analysis P-values were calculated by the paired student's t-test. Statistical significance was determined by analysis of variance (ANOVA) prior to Student's t-test. Significance was set at P-values less than 0.05. Error bars in all the figures indicate standard error of the mean (SEM) in a two-tailed t-test. Bacterial strains and reagents B. anthracis non-encapsulated Sterne strain 34F2 [pXO1 + , pXO2 − ] was obtained from the Colorado Serum Company. Non-virulent B. subtilis 168 was purchased from the American Type Culture Collection (Manassas, VA, USA). Human PLG, plasmin, rabbit anti-GroEL polyclonal antibody, and D-Val-Leu-Lys- p -nitroanilide (VLK-pNA, Sigma V7127) were purchased from Sigma. Human complement C3b (product # 204860) and uPA (product # 672112) were purchased from Calbiochem, and rabbit anti-human C3c polyclonal antibody (product # A0062) was from Dako. NHS was obtained from Innovative and goat anti-human PLG polyclonal antibody (ab6189) and rabbit anti-mouse C3 polyclonal antibody (ab11887) from Abcam. Mouse monoclonal antibody (EF12) against an abundant B. anthracis exosporium protein BclA (1 mg/ml; used 1∶1,000 dilution for Western blot) [26] was kindly provided by J.F. Kearney (University of Alabama). Horse raddish peroxidase (HRP)-conjugated secondary antibodies sheep anti-mouse IgG and donkey anti-rabbit IgG antibody were purchased from GE Healthcare and rabbit anti-goat IgG antibody from Jackson Immuno Research. PLG binding onto B. anthracis 1.6×10 9 spores or 100 µl of vegetative cells grown to A 600 of 1.5 were incubated with 150 µl of NHS or 10 µg of purified PLG in binding buffer (50 mM Tris, pH 7.5, 100 mM NaCl, and 2 mM MgCl 2 ) for 1 h at room temperature on a rocker platform and then washed 4 times with binding buffer. The 5 th wash in 100 µl of binding buffer was saved as a wash control, and bound proteins were eluted with 100 µl of 3 M potassium thiocyanate. The eluted protein was separated by SDS-PAGE and analyzed by Western blot with anti-PLG. To examine the effects of amino acids on PLG binding to the pathogen surface, 5×10 7 Sterne spores in binding buffer were incubated with 2 µg of PLG and 50 mM amino acids for 1 h on a rocker platform. The spores were washed 2 times with 1 ml of binding buffer and incubated in 150 µl of 50 mM Tris-HCl, pH 7.5, complemented with 20 units (0.2 ug) of uPA and 150 µl of 200 µM VLK-pNA for 15 min at 37°C. Plasmin activity was measured by reading absorbance of 100 µl of reactants at 405 nm. The spores were suspended in 150 µl of binding buffer and incubated with 4 µg of PLG for 1 h. After 5 washings with 150 µl of the binding buffer, the spores were suspended in 300 µl of 50 mM Tris-HCl, pH 7.5, and incubated with 20 units of uPA and substrate. Plasmin activity was measured as described above. Two-Dimensional SDS-PAGE Spore-bound NHS proteins were eluted by a chaotropic reagent, potassium thiocyanate, as described above and dialyzed in water. Samples were subjected to two-dimensional (2D) electrophoresis as follows: the desalted proteins were dissolved in Zoom 2D protein solubilizer-1 and applied on immobilized pH 3–10 linear gradient strips according to the manufacturer's instructions (Invitrogen). Focusing started at 175 V (15 min), was ramped to 2000 V for 45 min, and finally continued at 2000 V for 30 min in an IPGrunner system (Invitrogen). After focusing, strips were equilibrated for sample buffer and then overlaid onto 4–12% SDS-PAGE. The separated proteins were silver-stained and the bands were excised from the stained gel. Mass spectrometry The potassium thiocyanate-eluted spore-bound proteins or silver-stained protein bands excised from the 2D gel were trypsinized as described [15] . Identification of the proteins was performed by LTQ-tandem MS/MS equipped with a reverse-phase liquid chromatography nanospray tandem MS using a high-resolution LTQ-Orbitrap spectrometer (ThermoFisher). The reverse-phase column was slurry-packed in house with 5 µm, 200-à pore size C 18 resin (Michrom BioResources) in a 100 µm×10 cm fused silica capillary (Polymicro Technologies) with a laser-pulled tip. After sample injection, the column was washed for 5 min at 200 nl/min with 0.1% formic acid, peptides were eluted using a 50-min linear gradient from 0 to 40% acetonitrile and an additional step of 80% acetonitrile (all in 0.1% formic acid) for 5 min. The LTQ-Orbitrap MS was operated in a data-dependent mode in which each full MS scan was followed by five MS-MS scans where the five most abundant molecular ions were dynamically selected and fragmented by collision-induced dissociation using normalized collision energy of 35%. Tandem mass spectra were matched against the National Center for Biotechnology Information mouse database by Sequest Bioworks software (ThermoFisher) using full tryptic cleavage constraints and static cysteine alkylation by iodoacetamide. For a peptide to be considered legitimately identified, it had to be the top number one matched and had to achieve cross-correlation scores of 1.9 for [M+H] 1+ , 2.2 for [M+2H] 2+ , 3.5 for [M+3H] 3+ , ΔCn>0.1, and a maximum probability of randomized identification of 0.01. The MS data were filtered to improve the quality of the data set prior to protein selection. The initial set of proteins was limited to those that could be confidently identified, and was further screened to remove proteins with few non-zero peptide hits. Preparation of recombinant proteins The target genes of B. anthracis were amplified from its chromosomal DNA by PCR with specific oligonucleotides using a Taq polymerase premix (Invitrogen). Primers used in this study were as follows: GroEL (BA0267), forward GCA AAA GAT ATT AAA TTT AGT GAA , reverse CAT CAT TCC GCC CAT ACC GCC ; enolase (BA5364), forward ATG TCA ACA ATT ATT GAT GTT , reverse TCA TCG TTT GAT GTT ATA AAA ; and EF-tu (BA0108), forward ATG GCT AAA GCT AAA TTC GAA , reverse TCA CTC AAC GAT AGT AGC AAC . The amplicons were ligated into expression plasmid pTrcHis2-TOPO (Invitrogen) and then transformed into E. coli DH5α following the manufacturer's instructions. Protein expression was induced with 1 mM isopropyl-β-D-thiogalactoside for 5 h. The 6× His-tagged fusion proteins were isolated under native conditions by Ni 2+ -NTA resin (Probond, Invitrogen) as described in the manufacturer's protocols. For binding assays, purified proteins (320 µg) were also conjugated to carboxylate-modified FluoSpheres (1.0 µm, 500 µl) in the presence of EDAC (1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide) according to the manufacturer's recommendations (Invitrogen). The resulting beads were blocked with 1% BSA and resuspended in 500 µl of PBS. Exosporium extraction and ligand blot analysis Exosporium extracts were prepared by incubating the spore suspension in 0.1M DTT, 0.05% SDS, and 0.1M NaCl, pH 10 for 2.5 h in a 37°C shaking water bath [27] , or by a sonication in 20 mM Tris-HCl, 0.5 mM EDTA, pH 7.5. Sonication was performed using a Microsonix XL ultrasonic cell disruptor (Microson) for seven 1 min bursts (output power 12 W), each separated by 2 min cooling on ice [28] . By centrifugation, spores (exosporium negative) and supernatants were separated. The spore pellets were washed twice with the spore binding buffer as described above, and were subjected to PLG binding assays. For ligand blot analysis, exosporium extracts or recombinant proteins were run on 4–12% SDS-PAGE gel and then electrophoretically transferred onto a nitrocellulose membrane. The membrane was soaked in PBS/0.05 Tween 20 (PBST) containing 1% bovine serum albumin overnight at 4°C to renature the proteins, and then it was incubated with PLG (1 µg/ml in PBST/1% BSA) for 1.5 h at room temperature. The membrane was washed 5 times with PBST and incubated with anti-PLG antibody for 1 h followed by the corresponding horseradish peroxidase (HRP)-conjugated secondary antibody. The blot was visualized by HRP reaction. PLG binding to recombinant receptors MaxiSorp 96 well plates (Nunc) were coated overnight with different concentrations of recombinant proteins (7.8–500 nM) at 4°C. Following 3 washings with PBST, wells were blocked for 1 h at room temperature with 0.1% gelatin/PBS and then washed 3 times. Afterwards, 100 µl/well of human PLG (1 µg/ml) were added and incubated for 2 h at room temperature. Unbound PLG was removed by washing 3 times with PBST. Bound PLG was incubated with anti-PLG antibody (1∶5,000) for 1 h followed by secondary antibody. PLG binding to receptors was colorimetrically measured at 450 nm after sequential addition of a HRP substrate TMB (3,3′,5,5′-tetramethylbenzidine) and sulfuric acid. C3 deposition and degradation on spore surface Spores (2×10 7 /well) were washed, resuspended in PBS, and immobilized onto MaxiSorp microplates (Nunc) overnight at 4°C. After washing with PBST, wells were blocked with PBS/0.2% gelatin for 1 h at room temperature and incubated with 10% NHS (100 µl) for 30 min at room temperature. The wells were washed 2 times with PBST and incubated with 2 µg/well of PLG for 1 h in the presence or absence of protease cocktail (100-fold dilution, Sigma). Bound PLG was activated by uPA (20 units/well) for 3 h at 37°C. Deposited C3b was then detected by incubation with anti-C3c antibody (1∶2,000) followed by HRP-conjugated secondary antibody. C3 deposition was colorimetrically measured after addition of TMB and sulfuric acid at 450 nm. Preparation of rabbit BALF and phagocytosis assays BALF was collected from New Zealand White rabbits infused with 30–40 ml of Hanks' balanced salt solution (HBSS) under the approval of the Institutional Animal Care and Use Committee of the Biocon (Rockville, MD; approval # A0900-09a). The BALF was used after centrifuging at 1,500 rpm for 20 min at 4°C. For macrophage phagocytosis assays, spores were incubated with 250 µg of BALF and/or 25 µg of NHS (as a source of C3) for 1 h at room temperature in the presence or absence of 100 µM leupeptin. Spores (8×10 6 ) were washed twice with PBS and resuspended in PBS. RAW264.7 cells were infected with the spores at MOI of 10 and centrifuged to precipitate spores for 2 min. After 30 min of incubation, the cells were washed 6 times with HBSS and lysed by 2.5% saponin, and phagocytosed spores were counted by a serial dilution method on LB agar plate. Protein staining and immunoblotting Proteins were loaded onto 10% or 4–12% NuPAGE MES gel (Invitrogen) and separated under reducing conditions (32 mM dithiothreitol). Separated proteins were then silver-stained using GelCode SilverSNAP kit (Pierce) according to manufacturer's instructions or immunoblotting. For immunoblotting, the separated proteins were electrophoretically transferred to a nitrocellulose membrane using an iBlot gel transfer system (Invitrogen). After blocking with 5% dried milk solution, the membrane was probed with the primary antibody using PBST containing 5% milk, and was incubated with the corresponding HRP-coupled secondary antibody (1∶10,000 dilution) for 1 h at room temperature. Then the membrane was washed in PBST and visualized with the most sensitive West Femto chemiluminescent substrate system (Thermo Scientific). Statistical analysis P-values were calculated by the paired student's t-test. Statistical significance was determined by analysis of variance (ANOVA) prior to Student's t-test. Significance was set at P-values less than 0.05. Error bars in all the figures indicate standard error of the mean (SEM) in a two-tailed t-test. Results Identification of human serum proteins interacting with B. anthracis spores Spores of the Sterne strain 34F2 were incubated with NHS. After an extensive wash, the bound proteins were eluted by a chaotropic solution followed by desalting. Desalted spore-bound proteins were separated by 2D electrophoresis and visualized by silver staining ( Figure 1 ). Protein spots were excised from the gel and trypsinized to be identified by LC-MS/MS ( Table S1 ). The identified proteins included complement, acute phase proteins, and proteases. Among them, complement factor H and PLG were relatively abundant spore-bound proteins ( Figure 1B ). To support this finding, we trypsinized total eluted proteins from the spore surface and submitted them to LC-MS/MS. As shown in Table S2 , PLG and complements factor H and C3 were abundant among the spore-bound proteins, with high numbers of peptide hits and high peptide identification scores. 10.1371/journal.pone.0018119.g001 Figure 1 2D gel analysis of human serum proteins interacting with B. anthracis Sterne spores. Proteins were eluted from NHS-treated spores and separated by 2D electrophoresis. Proteins were visualized by silver staining and were identified by LTQ-MS/MS as described in Materials and Methods . (A) Control eluates from spores alone without NHS. (B) Proteins eluted from NHS-treated spores by a chaotropic reagent. Protein identifications by LTQ-tandem MS/MS were indicated in gel B. PLG binding to spores is exosporium-dependent We further examined whether PLG binds to the spore surface of B. anthracis using Western blotting. Purified PLG bound efficiently to the spore surface, while plasmin showed more efficient binding in the NHS ( Figure 2A ), suggesting less existence of PLG in the NHS. To examine whether PLG binding to the spores was exosporium-dependent, we extracted exosporium of spores using an alkaline buffer containing DTT-SDS-NaCl and sonication. To test for the removal of exosporium proteins, we carried out Western blot analysis with antibodies against BclA and GroEL, major exosporium proteins. BclA and GroEL were efficiently extracted by both alkaline solution and sonication ( Figure 2B ). The resulting exosporium-negative spores were then subjected to the PLG binding assays. Removal of exosporium revealed a significant decrease in PLG binding onto the surfaces ( Figure 2C ). No detectable amount of PLG was bound in the exosporium-negative, non-pathogenic B. subtilis as well. Heat treatment of spores (65°C for 30 min) did not show a significant change PLG binding ( Figure 2D ), suggesting heat-insensitive receptor-mediated PLG binding. These data demonstrated that exosporium was involved in PLG binding to spores. 10.1371/journal.pone.0018119.g002 Figure 2 Binding of PLG to B. anthracis spores is exosporium-dependent. (A) Spores (1.6×10 9 ) were incubated with PLG (10 µg) or NHS (150 µl) and washed, and bound proteins were eluted by 3 M potassium thiocyanate. (A) Western blot of 5 th washed ( W ) and eluted ( E ) fractions with anti-PLG antibody. Lane W demonstrates an absence of detectable protein in the last wash before elution. U indicates unbound proteins. Pla H and Pla L represent heavy chain and light chain of plasmin, respectively. (B) Exosporium was removed by alkaline DTT-SDS-NaCl buffer or sonication as described in Materials and Methods . Removal of exosporium was confirmed by Western blot analysis of major exosporium proteins BclA and GroEL. (C) Exosporium-positive (NT), exosporium-negative (by alkaline extraction or sonication), or B. subtilis 168 spores were incubated with PLG and eluted by a chaotropic reagent. Bound PLG was analyzed by Western blotting as described. U , unbound; W , 5 th wash; and E , elution with a chaotropic salt. (D) B. anthracis spores were incubated at 65°C for 30 min, and subjected to PLG binding assay as described above. No significant change of PLG binding by heat treatment (65°C) was seen compared with no heat treated control (-). B. anthracis spore-bound PLG is functional To examine whether PLG bound to pathogen receptors can be activated to plasmin, spores were incubated with PLG and its activator uPA. Spore-bound PLG was efficiently activated by uPA in a concentration-dependent manner ( Figure 3A ). This PLG binding to pathogen surface was abrogated by 50 mM of lysine analogues such as arginine and 6-amino-n-caproic acid (6-ACA), indicating that binding of PLG to spores was specific to the presence of lysine residues in the spore surface proteins ( Figure 3B ). To compare the binding specificity of PLG and plasmin, purified enzymes were incubated with the spores. The amount of bound proteins after elution and PLG activation by uPA was measured using a colorimetric assay. The assay demonstrated a similar affinity of the spores for PLG and plasmin, within less than 100 nM ( Figure 3C ). 10.1371/journal.pone.0018119.g003 Figure 3 Spore-bound PLG is activated by uPA. (A) Plasmin activity of bound PLG after uPA activation. Spores (5×10 7 ) were incubated with PLG and washed, and 20 units of uPA and plasmin substrate VLK-pNA were added. After incubation for 15 min at room temperature, the absorbance was measured at 405 nm. (B) Effect of lysine analogues on spore-bound PLG/plasmin activity. Spores (5×10 7 ) were incubated with PLG (2 µg) in the presence of 50 mM amino acids, and assayed as described in panel A . (C) Binding affinity of PLG and plasmin to spores. PLG and plasmin were incubated with spores (2.4×10 7 ) and the amount of bound proteins was assayed by the colorimetric method described in panel B. The amount of plasmin was determined using a standard activity curve. Identification of the spore PLG receptors To screen for the PLG receptors, exosporium extracts were prepared in an alkaline buffer as described above and subjected to a ligand blot analysis overlaid with PLG and probed with anti-PLG antibody. The bands of silver-stained gel ( Figure 4A ) corresponding to each PLG binding protein in a ligand blot ( Figure 4B ) were excised from the gel and subjected to tryptic digestion and peptide mass fingerprinting using a LTQ-tandem MS/MS. Bands 1–3 were identified to be chaperonin-60 kDa (GroEL), α-enolase, and translation elongation factor-tu (EF-tu), respectively ( Figure 4C ). To further confirm the putative PLG receptors in experiments with recombinant proteins, the genes of the above proteins were PCR-amplified from a bacterial chromosome using specific primers. The amplicons were inserted into pTrcHis-TOPO vector to generate proteins and all recombinant proteins were expressed in E. coli and purified ( Figure 5A ). Strong PLG binding to α-enolase and EF-tu, but weak binding to GroEL, was observed in a ligand blot assay ( Figure 5B ). Binding affinity of recombinant receptors to PLG was further determined in a microplate coated with different concentrations of receptor proteins. PLG binding to the receptors was dose-dependent and the affinity was in the order of α-enolase>EF-tu>GroEL ( Figure 5C ). To further support the differential binding, we conjugated beads (particle diameter of 1 µm) with recombinant proteins to mimic spores. PLG was able to bind strongly with α-enolase-conjugated beads and weakly with EF-tu-congugated beads ( Figure 5D ). Together, these results suggest that α-enolase and EF-tu are PLG receptors of B. anthracis spores. 10.1371/journal.pone.0018119.g004 Figure 4 Identification of PLG receptors. Spore extracts obtained from an acidic buffer were separated on a 4–12% SDS-PAGE. (A) The proteins were visualized by coomassie brilliant blue staining (CBB). (B) Far-Western (Ligand) blot with PLG. The proteins were transferred onto a nitrocellulose membrane, and the membrane was incubated with PLG. PLG-bound proteins were visualized by Western blotting with anti-PLG antibody. (C) Identification of proteins interacting with PLG. Bands (1–3) were excised from the gel and subjected to in-gel digestion with trypsin. LTQ-MS/MS was performed to identify the proteins. 10.1371/journal.pone.0018119.g005 Figure 5 PLG binding of recombinant receptors. (A) Recombinant proteins were expressed on plasmid pTrcHis2-TOPO by IPTG induction and purified by Ni + -chelate column chromatography. The purified proteins were visualized on SDS-PAGE by Coomassie blue staining. (B) Analysis of PLG binding of recombinant receptors by ligand blot. The bound PLG was detected by anti-PLG antibody. (C) Analysis of PLG binding of recombinant receptors by ELISA. Different concentrations of receptors were immobilized on a plate. After incubation with PLG (0.1 µg/well), binding activity was analyzed by anti-PLG antibody antibody. ▪, enolase; ▴, EF-tu; •, GroEL; and ○, control BSA. (D) PLG binding of receptor-conjugated beads. Receptor-conjugated beads (50 µl) were incubated with PLG (1 µg) and eluted with 100 µl of 3 M potassium thiocyanate. The eluted protein (10 µl) from the beads was analyzed by Western blotting with anti-PLG. Surface-bound PLG exhibits anti-opsonic activities It has been known that upon complement activation, C3b, along with cleavage products such as iC3b, is covalently attached to target surfaces to opsonize the pathogenic organisms for pathogenesis [29] , [30] . To assess whether spore-bound PLG is able to degrade C3b deposited on the surface, we performed a C3b deposition and degradation assay based on whole-spore ELISA. Spores were immobilized onto microplates and treated with NHS and PLG. After extensive washing, spore-bound PLG was activated by uPA, and deposition of C3b on the spore surface was monitored. Spore-bound active plasmin led to a drastic decrease in C3b molecules on the surface compared to control without PLG activation by uPA, or those pretreated with PLG in the presence of protease cocktail ( Figure 6A ). Plasmin-mediated C3b degradation was confirmed by Western blot analysis after incubation of C3b with plasmin-coated spores. Both α and β chain of C3b were degraded by spore-coated plasmin, which was inhibited by leupeptin treatment ( Figure 6B ). These suggest that after its activation by uPA, spore surface-bound PLG exhibits anti-opsonic activity by cleaving C3b molecules. 10.1371/journal.pone.0018119.g006 Figure 6 Degradation of deposited C3b by B. anthracis spores-bound PLG. (A) Spores (2×10 7 /well) were immobilized onto a microplate and incubated with 10% NHS (100 µl) for 30 min at room temperature. Washed spores were incubated with 2 µg/well PLG for 1 h in the presence or absence of protease cocktail. Bound PLG was activated by uPA (20 units/well) for 3 h at 37°C. Deposited C3b was detected by anti-C3c IgG and HRP-conjugated secondary antibody followed by TMB reaction. C3b deposition is expressed as the mean absorbance at 450 nm of quadruplicates. Error bars indicate ± standard deviation. *PEF-tu>GroEL ( Figure 5C ). To further support the differential binding, we conjugated beads (particle diameter of 1 µm) with recombinant proteins to mimic spores. PLG was able to bind strongly with α-enolase-conjugated beads and weakly with EF-tu-congugated beads ( Figure 5D ). Together, these results suggest that α-enolase and EF-tu are PLG receptors of B. anthracis spores. 10.1371/journal.pone.0018119.g004 Figure 4 Identification of PLG receptors. Spore extracts obtained from an acidic buffer were separated on a 4–12% SDS-PAGE. (A) The proteins were visualized by coomassie brilliant blue staining (CBB). (B) Far-Western (Ligand) blot with PLG. The proteins were transferred onto a nitrocellulose membrane, and the membrane was incubated with PLG. PLG-bound proteins were visualized by Western blotting with anti-PLG antibody. (C) Identification of proteins interacting with PLG. Bands (1–3) were excised from the gel and subjected to in-gel digestion with trypsin. LTQ-MS/MS was performed to identify the proteins. 10.1371/journal.pone.0018119.g005 Figure 5 PLG binding of recombinant receptors. (A) Recombinant proteins were expressed on plasmid pTrcHis2-TOPO by IPTG induction and purified by Ni + -chelate column chromatography. The purified proteins were visualized on SDS-PAGE by Coomassie blue staining. (B) Analysis of PLG binding of recombinant receptors by ligand blot. The bound PLG was detected by anti-PLG antibody. (C) Analysis of PLG binding of recombinant receptors by ELISA. Different concentrations of receptors were immobilized on a plate. After incubation with PLG (0.1 µg/well), binding activity was analyzed by anti-PLG antibody antibody. ▪, enolase; ▴, EF-tu; •, GroEL; and ○, control BSA. (D) PLG binding of receptor-conjugated beads. Receptor-conjugated beads (50 µl) were incubated with PLG (1 µg) and eluted with 100 µl of 3 M potassium thiocyanate. The eluted protein (10 µl) from the beads was analyzed by Western blotting with anti-PLG. Surface-bound PLG exhibits anti-opsonic activities It has been known that upon complement activation, C3b, along with cleavage products such as iC3b, is covalently attached to target surfaces to opsonize the pathogenic organisms for pathogenesis [29] , [30] . To assess whether spore-bound PLG is able to degrade C3b deposited on the surface, we performed a C3b deposition and degradation assay based on whole-spore ELISA. Spores were immobilized onto microplates and treated with NHS and PLG. After extensive washing, spore-bound PLG was activated by uPA, and deposition of C3b on the spore surface was monitored. Spore-bound active plasmin led to a drastic decrease in C3b molecules on the surface compared to control without PLG activation by uPA, or those pretreated with PLG in the presence of protease cocktail ( Figure 6A ). Plasmin-mediated C3b degradation was confirmed by Western blot analysis after incubation of C3b with plasmin-coated spores. Both α and β chain of C3b were degraded by spore-coated plasmin, which was inhibited by leupeptin treatment ( Figure 6B ). These suggest that after its activation by uPA, spore surface-bound PLG exhibits anti-opsonic activity by cleaving C3b molecules. 10.1371/journal.pone.0018119.g006 Figure 6 Degradation of deposited C3b by B. anthracis spores-bound PLG. (A) Spores (2×10 7 /well) were immobilized onto a microplate and incubated with 10% NHS (100 µl) for 30 min at room temperature. Washed spores were incubated with 2 µg/well PLG for 1 h in the presence or absence of protease cocktail. Bound PLG was activated by uPA (20 units/well) for 3 h at 37°C. Deposited C3b was detected by anti-C3c IgG and HRP-conjugated secondary antibody followed by TMB reaction. C3b deposition is expressed as the mean absorbance at 450 nm of quadruplicates. Error bars indicate ± standard deviation. *P<0.001 (paired Student's t-test). (B) Spores were incubated with plasmin and washed extensively as described in Figure 2A . C3b was incubated with the spores in the presence or the absence of leupeptin and C3b degradation was analyzed by Western blot with anti-C3c antibody. Rabbit BALF decreases complement-dependent phagocytosis In inhalation anthrax, the lung serves as a portal of entry. Before alveolar macrophages phagocytose deposited spores, they may interact with the BALF in the alveolar space. Therefore, we examined whether the interaction with BALF increases or decreases spore phagocytosis. When rabbit complements were incubated with the BALF, degraded C3b fragment was detected in Western blot with anti-C3 antibody, suggesting the existence of C3 degrading protein(s) ( Figure 7A ). To confirm the existence of PLG and its binding to spores, spores were incubated with BALF and the bound proteins were submitted to Western blot analysis with anti-PLG antibody. As shown in Figure 7B , PLG in the rabbit BALF bound to spores. To evaluate the role of spore-bound PLG on the phagocytosis, we incubated B. anthracis spores with BALF without activation of PLG by uPA in order to examine the intact BALF effects. Murine macrophages RAW264.7 cells were then infected with the spores in the presence of NHS. The macrophages were lysed and the number of viable bacteria was determined by counting CFU following dilution plating. Phagocytosis in the presence of NHS alone increased the CFU 3-fold per well at 30 min post-infection and decreased the CFU per well in the presence of NHS and BALF ( Figure 7C ). A decrease was observed in the NHS only treated samples when leupeptin, a plasmin inhibitor, was amended to the exposure media. Taken together, PLG in the BALF may play a role in host defense in the lung. 10.1371/journal.pone.0018119.g007 Figure 7 Rabbit BALF decreases spore phagocytosis of macrophages. (A) Rabbit complements were incubated with BALF (3 and 30 µg) in the PBS for 1 h and subjected to Western blot analysis with anti-C3b antibody. Degradation of C3b was indicated by arrow (27 kDa of α′3 chain). (B) PLG in the BALF interacts with spores. BALF was incubated with spores, eluted by a chaotropic salt and subjected to Western blot analysis with anti-rabbit PLG antibody. (C) BALF decreases NHS-mediated spore phagocytosis. RAW264.7 cells were infected with NHS-treated spores in the presence of BALF and/or leupeptin. Phagocytosed spores were determined by a serial agar plating method. Discussion Following a proteomic approach, we have found that B. anthracis spores bind a number of human serum proteins that are involved in humoral and innate immunity. Spore-bound proteins included complement proteins, blood coagulation/fibrinolysis regulators, acute phase proteins, and cell surface and extracellular proteins. Deposition of complement C3 onto B. anthracis spores has been reported to be required for opsonin-dependent phagocytosis by macrophages [21] . Inhaled pathogens such as Mycobacterium tuberculosis [31] , [32] and Pseudomonas aeruginosa [33] were phagocytosed by macrophages in a C3-dependent manner to be subsequently cleared by the lungs. Although C3 opsonization of pathogens including B. anthracis facilitates early infection steps [16] , [20] , [21] , the pathogens inhibit the host complement attack and apparently utilize diverse escape mechanisms. Our proteomic data suggest that complement regulators are involved in control of complement activation. Complement factor H, FHR-1, C1 inhibitor, and C4BP were acquired by B. anthracis spores ( Table S2 ). Complement factor H, a 150-kDa plasma glycoprotein, is the central fluid-phase regulator of the alternative complement pathway. Further study on complement factor H -mediated C3b opsonization is warranted. In the present study, we focused on whether spore-bound PLG regulates spore opsonization by C3b. The results showed that spore-bound PLG induces a significant decrease in C3b opsonization by uPA activation. Therefore, it is likely that the acquisition of host regulators masks the pathogenic surface, resulting in survival of B. anthracis spores. This might represent a novel mechanism to inhibit the host innate immune system during early B. anthracis infection. Acquisition of PLG and its subsequent conversion to active plasmin promote dissemination of bacteria in the host [34] . Several invasive bacterial pathogens utilize the PLG system to invade tissues [16] , [35] , [36] , and use of this system has been extended to viruses [37] and parasites [38] . Increasing evidence proposes a so-called "bacterial metastasis" that is facilitated by the binding and activation of PLG and by the colonization and invasion of PLG-bound bacteria into tissues [35] . As shown in this study, B. anthracis is capable of binding PLG on the outer surface of spores. The bound PLG is activated to plasmin by the addition of human activator uPA, promoting a surface-associated plasmin activity. The binding and activation are inhibited by the presence of the lysine analogue 6-ACA, suggesting a lysine-dependent binding of PLG to spores. This supports the existence of lysine binding sites in PLG, which has been shown for several pathogens to be mediated by PLG kringle domains [39] – [41] . This may account for the similar binding capacity to spore receptors in spite of the conformational change by proteolytic cleavage between the kringle and protease domains [42] . Heat-resistant spores of B. anthracis were retained in the lungs of mice challenged with aerosolized Sterne spores for all infection periods [43] . Unlike the lung, homogenates of other organs such as lymph node, liver, and spleen showed the presence only of vegetative bacilli in the same experiment [43] . Accordingly, spores are the first cells of B. anthracis which invade through lung barriers, and vegetative cells are the type that circulate in the bloodstream and invade organs. The binding of host PLG or plasmin might represent a mechanism to regulate several physiological processes, e.g., fibrinolysis, ECM degradation, cell migration, the processing of growth factors, and bacterial metastasis into several organs [44] . The Gram-positive bacteria group A streptococci interact with PLG via GAPDH and α–enolase on the surface [45] , [46] . Proteomic analysis of the PLG-binding proteins in the human pathogen M. tuberculosis identified glutamine synthase A1, HSP70 (DnaK), HSP60 (groEL), EF-tu, and other proteins that are metabolic proteins and are localized extracellularly [47] – [49] . In our study, PLG-binding proteins of B. anthracis spores were identified by ligand Western blotting of spore exosporium extracts to be GroEL, α-enolase, and EF-tu. They are extracellularly localized, consistent with those of other human pathogens. Since a non-virulent B. subtilis has no exosporium, these PLG receptors might not be localized in outmost surface of its spores resulting in no PLG binding as shown in Figure 2 . This could be a feature characteristic of a virulent B. anthracis in contrast to a non-virulent B. subtilis . The PLG binding to receptor proteins is usually mediated by a carboxyl-terminal lysine residue. In a competition assay, lysine and its analogue 6-ACA significantly inhibited PLG binding to spores, suggesting that a lysine residue is involved in PLG binding. Although spore- or cell-bound PLG is activated by the host PLG activation system, some pathogens present an endogenous PLG activation system, e.g., streptokinase [50] , staphylokinase [51] , and Pla [52] . B. anthracis secretes two major metalloproteases, Npr599 (or NprB) and InhA, that exhibit PLG-degrading activity [15] . However, proteolysis of PLG by these proteases did not display plasmin activity [15] . This activity has been shown in the NprB homologue bacillolysin MA, produced by B. megaterium , which converts PLG into angiostatin and mini-PLG [53] . Another NprB-like protease, aureolysin of S. aureus , has the ability to convert PLG into angiostatin and mini-PLG and activates pro-uPA to uPA [54] . This observation opens up the possibility that in addition to the host activation, B. anthracis -bound PLG may be activated by a bacterial activation system such as pro-uPA activation by bacterial proteases. Thus, whether PLG activation during B. anthracis infection is due to secreted proteases, the host PLG activation system, or both deserves further study. Based on the above considerations, complement regulators from bacterial pathogens could be vaccine candidates to fight bacterial infections. For instance, GNA1870, a lipoprotein of Neisseria meningitidis , was identified to be a complement factor H binding protein and meningococcal vaccine candidate [55] . In animal models, antibodies binding to GNA1870 inhibit binding of complement factor H and thus render the bacterium susceptible to the alternative complement pathway. Furthermore, the bound antibodies activate the classical pathway, thereby initiating and enhancing complement attack [55] . PLG binding proteins as regulators of complement C3b might be useful as additional vaccine targets to avoid the immune evasion by the above discussed mechanisms. A recent study showed that Streptococcus suis enolase localizes on the cell surface and facilitates bacterial adherence, and that enolase confers complete protection against infection to mice [56] . Proteomic analysis showed that B. anthracis expresses α-enolase as a dominant immunogenic antigen [57] , [58] ; however, the function of enolase has not been established. Another study demonstrated that B. anthracis glyceraldehydes-3-phosphate was predominantly interacted with plasminogen and that immunization with the recombinant protein resulted in a significant protection upon challenge with B. anthracis in the murine model [25] . Therefore, we propose that B. anthracis proteins interacting with PLG (i.e. α-enolase) function as a protective antigen and are vaccine candidates to inhibit innate immune evasion by the pathogen. In summary, we have shown that B. anthracis may utilize the host PLG system to regulate complement opsonization in order to evade innate immunity as reported elsewhere [17] , [31] , [35] , [59] , [60] . We have identified PLG as a spore-associating protein. PLG-bound spores were capable of exhibiting anti-opsonic properties by cleaving C3b molecules through the surface receptors α-enolase and EF-tu. PLG-dependent anti-oposonization was confirmed in the rabbit BALF by its C3b degradation and anti-phagocytic activity. This suggests new avenues for development of anti-opsonization agents with capacity in affecting innate immunity. Supporting Information Table S1 Identification of human serum proteins interacting with B. anthracis spores separated by 2D SDS-PAGE. Spore-bound serum proteins were separated onto 2D SDS-PAGE and the each spot excised from the gel was subjected to LTQ-MS/MS. (XLS) Click here for additional data file. Table S2 Identification of human serum proteins interacting with B. anthracis spores. Spore-bound serum proteins were subjected to LTQ-MS/MS. Percentage of wrong identification was estimated to be ∼2%, and #IDs ≥5 are listed. #Peptides and #IDs represent the number of different peptides detected from this protein and the total number of peptide identifications from this protein, respectively. (XLS) Click here for additional data file.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2897299/

A Rapid Antimicrobial Susceptibility Test for Bacillus anthracis ▿

An effective public health response to a deliberate release of Bacillus anthracis will require a rapid distribution of antimicrobial agents for postexposure prophylaxis and treatment. However, conventional antimicrobial susceptibility testing for B. anthracis requires a 16- to 20-h incubation period. To reduce this time, we have combined a modified broth microdilution (BMD) susceptibility testing method with real-time quantitative PCR (qPCR). The growth or inhibition of growth of B. anthracis cells incubated in 2-fold dilutions of ciprofloxacin (CIP) (0.015 to 16 μg/ml) or doxycycline (DOX) (0.06 to 64 μg/ml) was determined by comparing the fluorescence threshold cycle ( C T ) generated by target amplification from cells incubated with each drug concentration with the C T of the no-drug (positive growth) control. This Δ C T readily differentiated susceptible and nonsusceptible strains. Among susceptible strains, the median Δ C T values were ≥7.51 cycles for CIP and ≥7.08 cycles for DOX when drug concentrations were at or above the CLSI breakpoint for susceptibility. For CIP- and DOX-nonsusceptible strains, the Δ C T was <1.0 cycle at the breakpoint for susceptibility. When evaluated with 14 genetically and geographically diverse strains of B. anthracis , the rapid method provided the same susceptibility results as conventional methods but required less than 6 h, significantly decreasing the time required for the selection and distribution of appropriate medical countermeasures.

PMC

Anthrax

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7223921/

Progress in research and application development of surface display technology using Bacillus subtilis spores

Bacillus subtilis is a widely distributed aerobic Gram-positive species of bacteria. As a tool in the lab, it has the advantages of nonpathogenicity and limited likelihood of becoming drug resistant. It is a probiotic strain that can be directly used in humans and animals. It can be induced to produce spores under nutrient deficiency or other adverse conditions. B. subtilis spores have unique physical, chemical, and biochemical characteristics. Expression of heterologous antigens or proteins on the surface of B. subtilis spores has been successfully performed for over a decade. As an update and supplement to previously published research, this paper reviews the latest research on spore surface display technology using B. subtilis . We have mainly focused on the regulation of spore coat protein expression, display and application of exogenous proteins, and identification of developing research areas of spore surface display technology. Introduction Bacillus subtilis is an important industrial microorganism. Its genetics and physiology have been studied intensively. Among bacteria, the understanding of its genetic background and physiology is second only to Escherichia coli (Kunst et al. 1997 ; Sonenshein et al. 2002 ). Spore surface display is a method of anchoring exogenous functional proteins on the surface of spores by means of a special structure (Zhang et al. 2019 ). B. subtilis spore surface display has many advantages. First, spores are resistant to harsh environmental conditions, and this is conducive to the use and stability of exogenous proteins in complex environments (Wang et al. 2011 ). Second, spores are synthesized in the cytoplasm of bacterial cells, so any heterologous protein to be anchored on the spore surface does not need to cross any membrane (Kim and Schumann 2009 ). Third, molecular chaperone in the cytoplasm of B. subtilis can appropriately promote the secretion and expression of foreign proteins (Muller et al. 2000 ). The first spore display system was established by Isticato et al. ( 2001 ), using CotB as an anchor protein to display tetanus toxin (TTFC) on the surface of B. subtilis spores. With growing knowledge of the B. subtilis genome and proteome, spore surface display has now been successfully applied in many fields, including oral vaccine development, antibody production, biocatalysis, bioremediation, and creating of diagnostic tools (Fig. 1 ) (Georgiou et al. 1997 ; Li et al. 2019 ). Fig. 1 Applications of B. subtilis spore surface display B. subtilis spore surface display follows two main approaches: a recombinant approach and a nonrecombinant approach (Isticato and Ricca 2014 ; Ricca et al. 2014 ). The recombinant approach requires modification of the bacterial genome to express a protein of interest as a fusion with spore coat protein (Hinc et al. 2013 ; Isticato and Ricca 2014 ), and the nonrecombinant approach is based on the direct adsorption of heterologous proteins on the spore surface or anchoring exogenous proteins on the spore surface with a cross-linking agent (Isticato et al. 2019 ; Ricca et al. 2014 ). The display by recombination approach avoids the isolation and purification steps of foreign proteins, the production process is simple, and it is the mainstream of B. subtilis spore surface display technology (Chen et al. 2017b ; Kim and Schumann 2009 ). In this review, we summarize the application of genetic recombination-based spore surface display technology in many fields, discuss new and developing research, and determine the future prospects of the technology. Formation and structure of B. subtilis spores Bacteria have many strategies to cope with the challenges of their environment (Tasaki et al. 2017 ). These strategies often involve rapid changes in gene expression, which temporarily alter the phenotype of cells and allow them to survive. A more complex and persistent example of stress response is sporulation, in which the bacterial genome is isolated in a protected space (spore) until environmental conditions improve, at which point spores will germinate to form vegetative cells with reproductive capacity (Setlow 2014 ). Among Gram-positive bacteria, B. subtilis and a few similar species are the most commonly used experimental systems, and many studies have been conducted to assess the process and morphology of sporulation (Higgins and Dworkin 2012 ). Formation of B. subtilis spores It is challenging for B. subtilis to form spores; their formation is controlled by a series of regulatory and structural genes whose expressions themselves are tightly regulated (Bejerano-Sagie et al. 2006 ). When nutrients are depleted, sporulation is triggered by the activation of histidine sensor kinases, including KinA, KinB, and KinC, which shuttle phosphate through an extended phosphorelay, resulting in phosphorylation of the master regulator of sporulation,transcription factor Spo0A (Molle et al. 2003 ). KinA is the major kinase responsible for initiation of sporulation and KinA (or KinB) overexpression during exponential growth is sufficient to induce entry into sporulation (Fujita and Losick 2005 ). In fact, inducing KinA synthesis beyond a certain level leads toentry into sporulation regardless of nutrient availability (Eswaramoorthy et al. 2010 ). The effect of a kinC mutation on sporulation is weaker than that of kinA or kinB (Lopez et al. 2009 ). Phosphorylated Spo0A can directly activate or inhibit the transcription of many genes; it indirectly controls genes involved in asymmetric cell division and those involved in the activation of sporulation-specific sigma factors and ultimately promotes spore formation (Hilbert and Piggot 2004 ). The formation of spores can be roughly divided into the following steps (Eichenberger et al. 2003 ; Higgins and Dworkin 2012 ): In harsh environments, B. subtilis cells begin to form dormant spores that resist adverse environments, and the activity of σ H begins to increase. Cells then divide unequally using specific asymmetric septum to form large mother cells and a small forespore. The mother cell is necessary for spore formation, but it is eventually lysis and the prospore eventually produces a mature spore. Mother cell and prospore express different σ factors; σ E factors are expressed in mother cells, while σ F factors are expressed in prospores (Losick and Stragier 1992 ), and phosphorylated Spo0A can induce the activation of σ E and σ F factors (Wang et al. 2006 ). After unequal division is completed, the maternal plasma membrane gradually encapsulates the forespore, so the outer membrane of the forespore encapsulates two layers of membrane structure. After that, the activated or synthesized σ G and σ K begin to induce gene expression in the forespore and mother cell. Lastly, specific structures such as spore crust, cortex, and spore coat are gradually synthesized. The cortex is composed of peptidoglycan (PG), which is located between the inner and outer membrane of spore, and spore PG precursors are synthesized in the mother cell (Popham 2002 ). The spore coat is formed in the mother cell and covers the outer surface of the prospore (Henriques and Moran 2007 ; Kim et al. 2006 ). Dipicolinic acid (DPA) synthesized in the mother cell gradually fills the forespores, which could help the forespores dehydrate continuously, the mother cell lyses, and mature spore is generated (Fig. 2 ) (McKenney et al. 2013 ). Fig. 2 The sporulation and germination cycle in B. subtilis . Adapted from McKenney et al. ( 2013 ) Structure of B. subtilis spores The B. subtilis spore is a complex structure. The spore core contains the chromosomal DNA that is maintained in a compact state by small acid-soluble proteins (SASPs). The original membrane that surrounded the forespore surrounds the core and the peptidoglycan rich cortex surrounds this membrane. Surrounding the cortex, the spore coat consists of about 80 proteins deposited by the mother cell arranged in inner and outer layers (Fig. 3 ) (Liu et al. 2016 ; McKenney et al. 2013 ). Fig. 3 Spore structure Spore coat The assembling coat is synthesized in the mother cell and is targeted to the outer forespore membrane by SpoIVA (Wang et al. 2009 ). SpoIVA binds and hydrolyzes ATP, allowing it to self-assemble into cable-like structures (Setlow 2012 ; Ramamurthi and Losick 2008 ) that form a basement layer that serves as a platform for coat assembly. Other proteins involved in assembly are SpoVID that directly interacts with SpoIVA (Mullerova et al. 2009 ; Wang et al. 2009 ) and SafA, which is necessary for the encasement of the spore (Mullerova et al. 2009 ). SafA was found to affect the localization of about 16 inner coat protein fusions (McKenney et al. 2010 ) substantiating its central role in coat assembly. Three layers of the B. subtilis spore coat are observed in thin-section electron microscopy: an inner coat, an outer coat, and crust (Warth et al. 1963 ; McKenney et al. 2010 ). The outer coat is indispensable for spore formation, yet its specific functions remain unclear. Compared with the outer coat of spores, the inner coat is a selective permeability barrier that protects spore DNA from being destroyed by some chemical agents (McKenney et al. 2013 ). The spore coat makes spores resistant to chemical reagents and external lysozymes, and prevents the nuclei from being degraded or ingested by protozoa (Setlow 2006 ); however, the resistance to heat, radiation, and some other chemical reagents is poorly understood (Borch-Pedersen et al. 2016 ). Spore cortex The cortex and core of spores are the key structures in the formation and maintenance of dormant spores. The spore cortex is thick, mainly composed of PG, which can reach 10% of the total dry weight of the spore. The structure of the PG is similar to that in vegetative cell wall, but in cortex, the structure is relatively loose, which is extremely important for maintaining spore resistance and dormancy (Aguilar et al. 2007 ; Higgins and Dworkin 2012 ). Because some amino acid residues of N-acetylmuramic acid of PG in the spore cortex are replaced by short peptides, the degree of cross-linking of PG in cortex is lower than that of PG in vegetative cell wall (Popham 2002 ). In addition, B. subtilis spore-cortex PG was found to be O-acetylated, a common PG modification that reduces sensitivity to the innate immune anti-microbial lysozyme (Laaberki et al. 2011 ). However, since lysozyme is unable to penetrate the outer coat (Driks 1999 ), this modification would not appear to be useful. Spore core The innermost layer of spores is the core, it is surrounded by the inner forespore membrane, germ cell wall (McKenney et al. 2013 ). The inner forespore membrane is located in the inner of cortex, it It has extremely low permeability, and small molecular substances are difficult to penetrate, which can prevent DNA-damaging substances from penetrating the inner forespore membrane to cause damage to the spore core DNA (Setlow 2006 ). Spore core contains most of its enzymes, chromosomal DNA, ribosomes, and tRNA, it also contains the special small molecule, DPA, it is chelated with Ca 2+ in the spore core, exists as a calcium salt, CaDPA, which can promote dehydration of the spore core and increase the thermal resistance of the spore (Higgins and Dworkin 2012 ; McKenney et al. 2013 ). SASPs are tightly bound to the core DNA of the spore, which can make the spore tolerate damage from UA radiation, drying, and high temperature, and can be used as a carbon source and energy source during spore germination (Setlow 2007 ). Regulation of B. subtilis spore coat protein expression The anchoring proteins used in B. subtilis spore surface display can be linked to exogenous proteins through their C- or N-termini. The correct selection of anchoring proteins is key to successfully displaying exogenous proteins on the spore surface. A suitable anchoring protein needs to meet the following requirements: (1) it must have a strong anchoring domain to ensure that foreign proteins can be immobilized on the spore surface (Potocki et al. 2017 ); (2) they must be compatible with foreign proteins, be able to form fusion proteins, and should not be able to interact with each other (Lee et al. 2003 ); and (3) anchored proteins must be resistant to protease hydrolysis (Lee et al. 2003 ; Potocki et al. 2017 ). To date, various spore coat proteins, such as CotB, CotC, CotE, CotG, CotX, CotY, CotZ, CgeA, and OxdD, have been successfully used as the anchoring proteins to display exogenous proteins on the spore surface of B. subtilis . Spore surface display of B. subtilis using CotB as an anchoring protein CotB was the first spore coat protein to be used in B. subtilis spore surface display. Its expression and assembly require the assistance of a variety of regulatory factors and proteins (Kodama et al. 2011 ; Zilhao et al. 2004 ). The expression of cotB is regulated by the maternal cell-specific sigma factors and transcription regulators GerE and GerR (Cangiano et al. 2010 ). CotB has a strongly hydrophilic C-terminus, which is composed of three serine-rich repeats; the serine residues accounts for more than 50% of the CotB C-terminus. Some studies have shown that CotB modification requires the involvement of CotG and CotH (Zilhao et al. 2004 ), and CotG is known to interact directly with CotB. Mutation of cotG results in the accumulation of a 46-kDa CotB protein in cells, but the specific mechanisms for this remain unclear. CotH, or proteins regulated by CotH, can prevent CotG from being hydrolyzed by proteases in the cell before assembling into spores, and it has an indirect regulatory effect on CotB (Nguyen et al. 2016 ). Isticato et al. deleted the amino acid residue in position 105 of the CotB C-terminus (CotBΔ105), used CotBΔ105 as an anchoring protein, and integrated the tetanus toxin gene into the amylase gene locus of the cotB -deleted B. subtilis genome. It was found that exogenous protein could not be expressed in spores, which proved that the fusion protein could not be assembled on the surface of spores in the absence of the original cotB gene (Henriques and Moran 2007 ). Therefore, the cotB , cotG , and cotH genes of B. subtilis should be retained when CotB is being used as an anchor to display exogenous proteins. Spore surface display of B. subtilis using CotC as an anchoring protein CotC is an abundant, 66-amino-acid protein known to assemble in the outer coat in various forms: a monomer of 12 kDa, a homodimer of 21 kDa, and two less abundant forms of 12.5 and 30 kDa, probably due to posttranslational modifications of CotC (Isticato et al. 2010 ; Isticato et al. 2008 ). Assembly of CotC requires expression of both cotH and cotE , but CotC does not accumulate in the mother cell compartment when its assembly is prevented by mutation of CotH (Isticato et al. 2004 ). In contrast, overexpression of cotH allows the accumulation of CotC in the mother cell compartment, suggesting that CotH, or a CotH-dependent factor, acts to prevent degradation of CotC in the mother cell and then allows its assembly within the coat (Baccigalupi et al. 2004 ). The mechanism of assembly of CotC is of interest, as the abundant CotC protein has been used as a vehicle for the display of heterologous proteins at the spore surface (Isticato et al. 2007 ). At present, heat-labile enterotoxin B subunit, urea, ethanol dehydrogenase, β-galactosidase, proinsulin, enolase, and trehalose synthase have all been successfully displayed on the spore surface using CotC as molecular carrier, which improves their tolerance to harsh environments (Hinc et al. 2010b ; Romero et al. 2007 ). Spore surface display of B. subtilis using CotG as an anchoring protein CotG is a 24-kDa protein regulated by mother cell RNA polymerase σ K and transcription regulator GerE. Like CotC, the expression of CotG is also indirectly regulated by GerR because GerR can activate SpoVIF, which plays an active role in GerE and GerE-dependent genes (Cangiano et al. 2010 ). The assembly of CotG on spore surfaces is mainly as 32- and 36-kDa proteins. Thirty-two kilodalton CotG may be formed by abnormal migration of unmodified initial proteins. Thirty-six kilodalton CotG may be produced by extensive cross-linking of proteins when proteins are assembled into the spore coat (Eichenberger et al. 2004 ). Like CotB and CotC, CotG assembly also requires cotH expression. CotH protects CotG from protease hydrolysis before sporulation, which is essential for the formation and assembly of CotG (Naclerio et al. 1996 ; Zilhao et al. 2004 ). Therefore, cotH should be retained when CotG is used as anchoring protein to display exogenous proteins (Saggese et al. 2014 ). Spore surface display of B. subtilis using other anchoring proteins OxdD is a secondary component of the spore shell and has oxalate decarboxylase activity. It can catalyze the conversion of oxalate into formate and CO 2 . Its molecular weight is approximately 43 kDa (Garcia-Ramon et al. 2017 ). oxdD gene is transcribed by a σ K -recognized promoter during sporulation and is negatively regulated by GerE. Therefore, OxdD is produced in the mother cell chamber of sporangia and depends on SafA assembly in the coat. Genetic and cytobiological analyses have shown that OxdD is located in the outer layer of the spore. As an anchoring protein, OxdD could encapsulate the exogenous proteins under the spore surface, providing more effective protection for the exogenous proteins and reducing effects on spore formation (Romero et al. 2007 ). CotH is an intermediate morphogenetic protein that plays a role in the assembly of the spore shell, but differs from CotG. CotH, as an inner layer protein of 42.8 kDa, has a strong correlation with CotB and CotG. The expression of cotH is regulated by σ K . As mentioned earlier, the assembly of spore coat proteins CotB, CotC, and CotG in CotH-mutant strains also has multiple validity defect, indicating that the inner and outer layers of the spore coat require CotH function (Isticato et al. 2015 ). CotZ, a key component of the crust belongs to the last encasement class and is more abundant at the mother cell proximal pole of the forespore (Imamura et al. 2010 ). It is dependent on σ E , σ K and the transcription factor GerE for expression (McKenney et al. 2013 ). CotZ is a 16-kDa protein, and it has been found to act as a new anchoring motif for the efficient display of UreA of Helicobacter acinonychis on the spores (Imamura et al. 2011 ; Hinc et al. 2013 ). In the case of the CotZ-UreA fusion protein, the calculated number of recombinant protein molecules is 2.5 × 10 2 from a single spore. This fusion protein is more effective in stimulating immunological response than other antigens in mice. Similar to CotZ, CgeA is another 14 kDa crust protein. CgeA is dependent on σ K and the transcription factor GerE for expression (Imamura et al. 2011 ). Iwanicki et al. ( 2014 ) described an example application of presented vector system to display CagA protein of Helicobacter pylori in fusion with CgeA spore coat protein. Applications of B. subtilis spore surface display B. subtilis has a well-established fermentation and production technology, and spores are resistant to harsh environmental conditions (Wang et al. 2011 ), so the application of spore surface display technology is very extensive. To date, this technology has been used in the production of multimeric proteins, oral vaccine preparations, and industrial enzyme production (Guoyan et al. 2019 ). Table 1 summarizes the related applications of spore surface display of foreign proteins based on recombinant approach in previous studies. Table 1 List of fusion and target proteins, used vectors, and application of Bacillus subtilis spore surface-displayed proteins based on recombinant approach Fusion protein Bacterial strain Target protein Used vector Substrate/antibody Product Application Reference CotB B. subtilis PY79 TTFC pGEM Anti-TTFC ― * Oral vaccination (Isticato et al. 2001 ) B. subtilis PY79 and RH103 TTFC pET28b Anti-TTFC ― Oral vaccination (Duc et al. 2003 ) B. subtilis PY79 and PP108 TcdA ― Anti-TcdA ― Oral vaccination (Hong et al. 2017 ) B. subtilis DB403 Tm1350 pHS p- Nitrophenyl butyrate p -Nitrophenyl Industrial biocatalysis (Chen et al. 2015a ) B. subtilis DB403 DSM pHS p -Nitrophenyl butyrate p -Nitrophenyl Industrial biocatalysis (Chen et al. 2015b ) B. subtilis PY79 VP28 pDG364 White Spot Syndrome virus ― Vaccine for shrimps (Nguyen et al. 2014 ; Pham et al. 2016 ) B. subtilis PY79 UreA pGEM Anti-UreA ― Anti-Helicobacter vaccine (Hinc et al. 2010b ) B. subtilis HU58 MPT64 pcotVac Anti-MPT64 ― Vaccine against tuberculosis (Sibley et al. 2014 ) B. subtilis PY79 RSM2e3 pDG1664 Anti-RSM2e3 ― Influenza vaccine (Zhao et al. 2014 ) B. subtilis 168 FliD pDL Anti-FliD ― C. difficile oral vaccines (Negri et al. 2013 ) B. subtilis PY79 GST-Cpa247-370 pDG1664 Anti-Cpa247–370 ― Vaccine against necrotic enteritis (Hoang et al. 2008 ) B. subtilis PY79 PA pDG364 Anti-PA ― Anthrax vaccine (Le et al. 2007 ) B. subtilis PY79 and RH201 pDHAFB Anti-His ― Bioremediation (Hinc et al. 2010a ) CotC B. subtilis PY79 (Spo + ) TTFC and LTB pRH22 and pIM51 Anti-TTFC and anti-LTB ― Clostridium tetani and E. coli vaccine (Mauriello et al. 2004 ) B subtilis 168 (trp − ) BmADH pJS700 Ethanol and NAD + Acetaldehyde and NADH Industrial biocatalysis (Wang et al. 2011 ) B. subtilis PY79 and PP108 TcdA ― Anti-TcdA ― Oral vaccination (Hong et al. 2017 ) B subtilis 168 (trp − ) OmpC pDG364 ― ― Vaccine against Salmonella (Dai et al. 2018 ) B. subtilis DB431 and BB80 VP28 and VP26 pDG1662 Anti-Vp28 and anti-Vp26 ― Oral vaccination (Valdez et al. 2014 ) B. subtilis WB600 Urease B and CTB pUS186 Rat anti UreB serum ― Oral vaccine for H. pylori (Zhou et al. 2017) B. subtilis WB600 CsCP pEB03 Rat anti-rCsCP serum ― Vaccine against Clonorchis sinensis (Tang et al. 2016 ; Tang et al. 2017 ) B. subtilis WB600 TP20.8 pGEX TP20.8-specific antibody ― Vaccine against Clonorchis sinensis (Zhou et al. 2008b ) B. subtilis WB600 CsPmy PEB03 Rat anti-rCsPmy serum ― Vaccine against Clonorchis sinensis (Sun et al. 2018 ) B. subtilis WB600 CsTP22.3 pGEX Rat anti-TP22.3 sera ― Vaccine against Clonorchis sinensis (Zhou et al. 2008a ) B. subtilis WB600 CsLAP2 PEB03 Rat anti-CsLAP2 serum ― Vaccine against Clonorchis sinensis (Qu et al. 2014 ) B. subtilis WB800N TreS pDG1730 D-maltosee D-trehalose Industrial biocatalysis (Liu et al. 2019 ) B. subtilis WB600 VP4 pEB03 Rabbit anti-rVP4 serum ― Vaccine against grass carp reovirus (Jiang et al. 2018 ) B subtilis 168 (trp − ) hGH pJS700 Anti-hGH ― Oral vaccination (Lian et al. 2014 ) B. subtilis PY79 UreA pGEM Anti-UreA ― Anti-Helicobacter vaccine (Hinc et al. 2010b ) B subtilis 168 β-galactosidase pKH40 ONPG ONP Industrial biocatalysis (Tavassoli et al. 2013 ) B. subtilis 168 (trp − ) GP64 pJS700 GP64-specific antibody ― Vaccine against Bombyx mori Nucleopolyhedrovirus (Li et al. 2011 ) B. subtilis 168 (trp − ) HSA pJS700 HSA-specific antibody ― Oral vaccination (Mao et al. 2012 ) CotE B. subtilis DB104 Tyrosinase pCSK1 L-tyrosine ― Industrial, medical, and environmental applications (Hosseini-Abari et al. 2016 ) B. subtilis DB104 β-galactosidase pDG1728 Anti β-galactosidase, antibody mouse IgM ― Industrial biocatalysis (Hwang et al. 2013 ) B. subtilis DB104 Lipase A and Lipase B pHPS9 pNPP ― Industrial biocatalysis (Kim 2017 ) CotG B.subtilis DB403 Nitrilase pHS Tomalononitrile, Succinonitrile, Glutaronitrile 2-cyanoacetic acid, 3-cyanopropionic acid, 4-cyanobutyric acid Industrial biocatalysis (Chen et al. 2015c ) B. subtilis DB104 DhaA pHY300PLK 2-CEES Chloride Bioremediation (Wang et al. 2019 ) B. subtilis DB104 β-galactosidase pDG1728 Anti β-galactosidase, antibody mouse IgM ― Industrial biocatalysis (Hwang et al. 2013 ) B. subtilis DB104 ω-transaminase pHPS9 (S)-α-methylbenzylamine and pyruvate Acetophenone Industrial biocatalysis (Bum-Yeol et al. 2011 ) B. subtilis DB104 GFP UV pCSK1 ― ― Diagnosis (Kim et al. 2007 ) B. subtilis MI111 Phytase pHT304 Sodium phytate Inorganic phosphate Industrial biocatalysis and animal probiosis (Mingmongkolchai and Panbangred 2018 ) B. subtilis WB800N TreS pDG1730 D-Maltosee D-Trehalose Industrial biocatalysis (Liu et al. 2019 ) B. subtilis 168 c-trp ChiS pDHAFB Chitin 3,5-dinitrosalicylic acids and N-acetyl glucosamine Biopesticide (Rostami et al. 2017 ) B. subtilis DB403 Nitrilase pHS. 3-Cyanopyridine 3-Carboxypyridine Industrial biocatalysis (Chen et al. 2017a ) B. subtilis DB403 L-arabinose isomerase pHS D-galactose D-tagatose Industrial biocatalysis (Qi et al. 2018 ) B. subtilis WB600 NanA pEASY Pyruvate Neu5Ac Industrial biocatalysis (Xu et al. 2011 ) B. subtilis DB104 Streptavidin pHPS9 Anti-streptavidin, Antibody ― Biological diagnosis (Kim et al. 2005 ) CotX B. subtilis 168 (trpC2) β-galactosidase pJS700a ONPG ONP Industrial biocatalysis (He et al. 2015 ; Wang et al. 2016 ) CotY B. subtilis 168(trp−) β-galactosidase pJS700a ONPG ONP Industrial biocatalysis (He et al. 2015 ) CotZ B. subtilis 168(trp−) β-galactosidase pJS700a ONPG ONP Industrial biocatalysis (He et al. 2015 ) B. subtilis WB800(trp-) DPEase pET22b(+) D-fructose D-allulose Industrial biocatalysis (He et al. 2016 ) CgeA B. subtilis 168 CagA pMUTIN4 Mouse anti-CagA antibody ― Vaccine formulation (Iwanicki et al. 2014 ) OxdD B. subtilis PY79 Phytase pDG364 Sodium phytate Inorganic phosphate Industrial biocatalysis and animal probiosis (Potot et al. 2010 ) * Not available Application in polyprotein production B. subtilis can spontaneously form spores in harsh or nutrient poor environments. Spores have strong resistance to adverse environments, such as high temperature, chemical reagents, ultraviolet rays, and lysozymes. The spore coat is a complex structure comprising at least 70 different proteins. Spore surface display requires the expression of exogenous proteins fused with coat proteins, so that the exogenous proteins are assembled on the spore surface directly without transmembrane localization after synthesis in the mother cell. The fusion proteins can be immobilized by spore surface display, which improves the stability of the protein and makes isolation and purification easier. Liu H et al. fused trehalose synthase with spore-anchoring proteins CotC and CotG for display on the surface of B. subtilis spores, and immunofluorescence, Western blot analysis, and enzyme activity assays showed that trehalose synthase was indeed present on the spore surface. The trehalose synthase on the surface of the recombinant spore can react with maltose as a substrate to form trehalose, after reused four times, the recombinant spore retained most of the enzymatic activity. (Liu et al. 2019 ). β-Galactosidase is a high molecular weight protein (116 kDa). It is active in a tetramer state and can affect the structure of host cells in general surface display systems. To date, this protein has been displayed on the spore surface using the B. subtilis spore coat proteins CotC, CotE, CotG, CotX, CotY, and CotZ, the enzyme expressed on the surface of the spore still retains its activity (He et al. 2015 ). The active polyproteins were anchored on the spore surface by spore surface display technology, which demonstrates that this technology represents a new method for the production of polyproteins. Application in preparation of oral vaccine CotB and CotC were selected as anchoring proteins to display antigens on the surface of B. subtilis spores. Since the first successful display of surface antigens, the list of displayed antigens has grown steadily (Amuguni and Tzipori 2012 ; Rosales-Mendoza and Angulo 2015 ). Spores have good resistance to stress; therefore, vaccines developed with this method can tolerate the acidic environment of the gastrointestinal tract and have a long shelf life (Zhou et al. 2008a ). They can pass through the gastrointestinal mucosa smoothly and quickly induce the body to produce a protective immune response. In addition, the use of spores as vaccine carriers can improve the efficiency of the immune response (Batista et al. 2014 ; Vogt et al. 2016 ). In recent years, Clonorchiasis sinensis, caused by Clonorchis sinensis , has become increasingly prevalent. Effective prevention strategies are urgently needed to control this food-borne infectious disease. Previous studies have shown that C. sinensis paramyosin (CsPmy) functions as a preferred vaccine. Sun et al. ( 2018 ) displayed CsPmy on the spore surface using CotC as anchoring protein. The expression of CsPmy on the spore surface was analyzed by SDS-PAGE, Western blot analysis, and immunofluorescence assay, and the results showed that CsPmy was successfully expressed on spore surfaces and the fusion protein had good thermostability. Specific IgGs in sera and intestinal mucosa were increased after intraperitoneal and intragastrical immunization. Oral immunization with B. subtilis spore expressing CsPmy on the surface was a promising, safe, and needle-free vaccination strategy against clonorchiasis (Mingmongkolchai and Panbangred 2018 ). In addition, CsPmy, CsCP, TP20.8, CsTP22.3, and CsLAP2 have also been successfully displayed on spore surfaces for immunization against clonorchiasis sinensis (Tang et al. 2017 ; Zhou et al. 2008a ). Salmonellosis is a major public health problem throughout the world. Dai et al. have assessed the potential use of B. subtilis spores for the expression of a major protective antigen of Salmonella serovar pullorum, OmpC. Mice immunized with recombinant spores expressing the OmpC antigen presented significant higher levels of OmpC-specific serum IgG and mucosal SIgA antibodies than mice immunized with nonrecombinant spores ( p < 0.01) (Dai et al. 2018 ). These results indicate that B. subtilis spores have broad applicability in vaccine development. Application in the production of industrial enzymes Industrial enzymes are at the core of the biocatalysis and biotransformation industries. They are characterized by high catalytic efficiency, high specificity, and low pollution in the production process. It can be difficult to separate enzymes from substrates, and the reaction conditions are usually strictly controlled. This leads the enzymes to be easily inactivated and makes their reuse difficult. However, enzymes can be easily separated from their substrates by displaying them on the surface of spores. The excellent stress resistance of spores can enhance the stability of enzymes in complex environments and promote the reuseabilty of enzymes. He et al. produced D-allulose by using D-psicose 3-epimerase (DPEase) expressed and displayed on the surface of B. subtilis spores. DPEase was fused at the C-terminus of the anchoring protein, CotZ, via a peptide linker, and trophic genes were used as selection markers during chromosomal integration. The optimal temperature and pH of the fusion protein CotZ-DPEase were 55 °C and pH 7.5–8.0, respectively, and the anchored DPEase exhibited high thermostability. Under optimal conditions, 60% of the yield was maintained after five cycles of utilization. Therefore, this biocatalyst system, capable of expressing and immobilizing DPEase on the spore surface of B. subtilis , was an appropriate alternative for D-allulose production (He et al. 2016 ). Lipases expressed in microbial hosts have great commercial value, but their applications are restricted by the high costs of production and harsh conditions used in industrial processes. Chen et al. successfully displayed the thermophilic lipase Tm1350 on the B. subtilis spore surface. The results showed that spore surface-displayed Tm1350 had more stable enzyme activity than free enzyme. Meanwhile, recycling experiments showed that the recombinant spores could be used for up to three reaction cycles without a significant decrease in catalytic rate (84%) (Chen et al. 2015a ). These studies have played a positive role in development of the application of spore surface-displayed enzymes in the industrial field. Application in the field of biological control of environmental pollution Enzymatic technology has been applied to the treatment of environmental pollution due to its advantages of stability against environmental stress and high catalytic efficiency. Tyrosinases, which are copper-containing monooxygenases, could be used for bioremediation of phenol-polluted environments and production of L-DOPA and melanin from L-tyrosine, are widely used for environmental applications (Sok and Fragoso 2018 ). Hosseini-Abari et al. displayed tyrosinase on spore surfaces using CotE as a molecular carrier. Tyrosinase activity on spores was monitored in the presence of L-tyrosine and CuSO 4 . Recombinant spores could be used repeatedly, with 62% of enzymatic activity remaining after washing six times with Tris-HCl buffer (Hosseini-Abari et al. 2016 ). Chitinase is a hydrolytic enzyme that has the specific function of hydrolyzing chitin into chitosan or N-acetylglucosamine. Chitinase is mainly used to control pests in agriculture. It can be used alone as an insecticide or used in conjunction with other microorganisms to control pests (Rishad et al. 2016 ). Rostami et al. fused chitinase with CotG and successfully displayed it on the surface of B. subtilis spores. Enzyme activity assays showed that the surface-displayed chitinase was active and was also able to inhibit the growth of Rhizoctonia solani and Trichoderma harzianum fungi (Rostami et al. 2017 ) This suggests a new bioremediation method to treat the problem of residual organophosphorus pesticides in the environment. Application in animal feed preparation Feed enzymes must remain active under the harsh conditions of feed preparation and the gastrointestinal tract. The strong stress resistance of spores enables them to be used as new tools for improving bioactive molecular preparations. E. coli phytase (AppA) has been widely used as an exogenous feed enzyme for monogastric animals. Sirima et al. displayed AppA on the spore surface of B. subtilis using spore coat protein CotG as an anchoring protein. AppA was successfully produced on the spore surface as verified by Western blot analysis and phytase activity assays. The highest enzyme activity was observed at 55 °C and thermal stability measurements demonstrated that more than 30% activity remained after 30 min incubated at 60 °C (Mingmongkolchai and Panbangred 2018 ). Research hotspots on surface display of B. subtilis spores As mentioned above, many B. subtilis spore surface display systems have been developed. However, up to now, most of these studies have been confined to the laboratory. Therefore, research on how to scale up the production of target proteins has become an active area of research. Strategies include introducing linker peptide chains (Huang et al. 2015 ), using multiple anchoring proteins to display exogenous proteins at the same time (Iwanicki et al. 2014 ), and increasing the number of copies of exogenous genes (Xu et al. 2011 ). Research on improving the sporulation efficiency of B. subtilis is another recent approach to optimizing spore surface display (Devi et al. 2015 ; Tojo et al. 2013 ). An appropriate intermediate ligand can improve the folding efficiency of foreign target proteins and anchoring proteins. It can also change the interactions between foreign target proteins and anchoring proteins, as well as between target foreign proteins and the cell surface. Strauss et al. found that the activity of lipase on the spore surface was positively correlated with the length of the intermediate. Lipase activity increased from 0.8 to 83 U/mg when the length of the intermediate increased from 10 to 92 amino acids (Strauss and Götz 1996 ). Hinc et al. found that the binding mode of anchoring proteins and foreign target proteins was the key factor for the success of spore display. The conformation of linker peptides could affect the results of spore surface display, and alpha helices have shown to be most effective under some conditions (Hinc et al. 2013 ). Using multiple anchoring proteins to display exogenous proteins at the same time can also improve spore display efficiency. The structure of B. subtilis spores is complex and contains dozens of different proteins. The number of potential anchoring proteins in spores is an important factor that restricts display efficiency. Therefore, the simultaneous display of various exogenous proteins by multiple anchoring proteins has become a hot area of research (Liu et al. 2019 ). At present, the chromosome insertion sites selected by the researchers are all the growth non-essential amyE gene. Iwanicki et al. constructed spore surface display integrative vectors using the non-essential genes lacA and pyrD as insertion sites, and using CotC and CotG as anchoring proteins, thus creating a multi-anchoring protein display system (Iwanicki et al. 2014 ). Conclusions and future perspectives B. subtilis spore surface display technology has developed rapidly over the past decade, and many coat proteins, including CotB, CotC, CotE, CotG, CotZ, CgeA, and OxdD, have been successfully used to display exogenous proteins or polypeptides on the spore surface. The nonpathogenicity of B. subtilis make this technology applicable to food and biological industries. The resistance of spores to stress makes industrial enzymes displayed on their surface more stable, and also provides the advantages of easy purification and recycling of immobilized enzyme, which can greatly reduce the cost of industrialization. B. subtilis spore surface display provides feasible avenues to improve industrial production efficiency, while providing for food and biological safety. At the same time, there are still some problems in spore surface display, such as the limited number of anchoring proteins on the spore surface, which is not conducive to a large number of exogenous proteins. Further, the success of surface display on spores depends on the fusion of anchored and target proteins, so it is critical to choose the correct fusion and anchor partner (Hinc et al. 2010b ). B. subtilis spore surface display technology has shown great promise for use in vaccine and drug preparation, enzymatic catalysis, biological detection, and other areas because of its unique advantages. It is believed that with further research on surface display using B. subtilis , this technology will play an important role in even more fields in the future.

PMC