Datasets:

health360

/

Healix-Shot

like 5

Follow

Health 360 7

potential to create new opportunities for rural economic enterprise, usually lags behind that in urban centres (Shucksmith, 2018). Seasonality may act as a constraint on tourism as a source of permanent employment in scenic areas (Schmallegger and Carson, 2010). An idyllic vision of low density settlement amid green fields can be accompanied by low levels of public transport provision and poor accessibility to essential services (OECD, 2010). These features impact across the socio-economic spectrum stimulating out-migration. As well as being 'pushed' in this way, opportunities in an external destination can exert a 'pull' on the potential migrant, although the relationships between these factors are complex (Castles et al., 2014: 28). The desire to attain self-actualization through contact with more liberal attitudes and lifestyles and access to a wider range of social and economic experiences than are present in a rural area of origin is recognised also as being part of the attraction of the urban (Arnett, 2000). The rural out-migrant imaginary is likely to be a counter-idyllic one. The role of an idyllic imaginary has received extensive attention in literature relating to counterurbanisation in which new migrants to the countryside dominate (e.g., Boyle and Halfacree, 1998;Halfacree, 2012). Recognised tropes include open landscapes, fresh air, benefits of a countryside environment for children, low levels of crime and a perceived sense of community, some of which are recognised as being highly problematic (Matthews et al., 2000;Cloke, 2003;Rye, 2006;Woods, 2011;Christiaanse and Haartsen, 2017;McHugh-Power et al., 2017). By comparison with documentation of the quest for an idyll by newcomers, its influence on return migration has received relatively less attention, although elements of an idyll are known to influence return to a place of origin (e.g. Ní Laoire, 2007;Haartsen and Thissen, 2014;Morse and Mudgett, 2017). The evidence suggests that the imaginary is imbricated with links to family and place and may change over the life-course. Ní Laoire's research (2007: 332) among a group of thirty returnees to rural Ireland pointed to inter-relationships between "classic counterurbanisation discourses" and notions of family and kinship. Morse and Mudgett's (2017) respondents in rural Vermont in the United States of America (USA) reported homesickness and a longing for landscape as reasons for return. Haartsen and Thissen's (2014) research among young returnees in the Netherlands identified three inter-linked dimensions of relations between self, others and environment, following Gustafson (2001a: 9). Gustafson (2001b: 650) uses the metaphor of 'roots' to signify such "emotional bonds with the physical environment (which) often also contain notions of local community, shared culture, and so forth", associations described by Pascual-de-Sans (2004) as ideotopes. Motivations for return among returnees in rural Northern Ireland included childhood memories, life-course stage, lifetime experiences and future life-course expectations, in which a rural idyll may feature (Stockdale et al., 2013). The negative influences of anonymity and absence of a feeling of belonging, in a migration destination, may also influence a quest to recapture the familiarity of home (Relph, 1976;Hay, 1998;Ahmed, 1999). Thus, migration may signify freedom, opportunities and new experiences but can be associated with uprootedness and the perceived loss of previously meaningful places (Relph, 1976;Fielding, 1992). This article seeks to add to the literature by exploring the proposition, among first generation returnees, that counter-idyllic elements which propel out-migration and idyllic elements may be present in the imaginaries of out-migrants and returnees to varying extents and that one may over-ride the other at particular points in time. An idyllic imaginary, however conceived, is not sufficient to attract international emigrants back to rural areas. Economic resources in the form of employ ment, investment capital to permit a business to be established, inheritance of a property or business or transferrable pension benefits, among retirees, are usually required. Nevertheless, positive recollections of a rural childhood and continued transnational contacts with family and friends, through various media of communication and return visits, are known to be influential (Conway and Potter, 2006;Barcus and Brunn, 2010). Return can permit certain aspects of a remembered idyll to be attained, but re-establishment in an area of origin, as long-recognised, can be problematic (Gmelch, 1980). The physical place may differ from when the migrant left, social relationships may need to be re-negotiated and service facilities are likely to compare unfavourably with those in a large overseas city (Gustafson, 2001a;Barrett and Mosca, 2013). Tensions may therefore be present. Methodology and sample characteristics The reported study is based on a convenience sample of thirty-four returned migrants, who were interviewed in 2014 by university geography students, usually a family member of the interviewee, using a schedule prepared by the author and following recognised protocols (Drooglever Fortuijn and van der Meer, 2006). Statistical representativeness was not sought. Following an ethnographic methodology, the study sought instead to capture a range of experiences of migration and return over time and, thereby, contribute to greater understanding of the phenomenon. The interviewee was required to have migrated from the Republic of Ireland (RoI) as an adult, lived overseas for one year or more, and returned either permanently or for one year or more before re-migrating (and, possibly, returning again). The definition of migration is long-established and currently used by Eurostat (2018). All left from and returned to a place of less than 1500 population, usually the same place, permitting the migrant to compare recollections of that place whilst absent with the reality on return. The interview schedule included closed questions profiling the social and economic attributes of the interviewee at the various migration stages. Transnational contacts with home were queried with reference to the methods used. The fora where contact took place with Irish people in the destination area(s) were recorded. Opportunities were provided in open questions to report (in the respondent's perceived order of priority) the influences on migration, experiences overseas, the reasons for return and experiences associated with rural residence again. The presence of an idyllic imaginary, which involved proximity to family and friends in a rural setting, emerged unprompted from the narratives. The data were entered in Excel and the qualitative statements were analysed in an iterative way to identify key themes and sub-themes (Saldăna, 2013). Detailed insights emerged relating to varying periods of absence from the early 1950s until recently. Most returned from a major city. The sample comprises twenty-four males and ten females, an imbalance among returnees that contrasts with the traditional national RoI pattern (Punch and Finneran, 1999). This imbalance arises, in part, from higher levels of international, rural male, emigration (and later return) during the recessionary years of 2008-2010. Also, some rural female emigrants were excluded from the analysis because of returning to larger centres of population. The average period of time spent overseas was 6.4 years with a maximum of 30 years, a minimum of 1 year and a mode of 2 years. Absences of one or two years were reported in particular by migrants on short-term work permits in Australia. The majority left in their late teens and early twenties when freedom to move is greatest (Rogers et al., 1978). Twenty-eight respondents reported their marital/ conjugal status as being 'single' at migration, five had a spouse or partner and one migrant had a spouse and children. Most had at least a secondary level (high school) education and over one-third of both males and females held a third-level qualification (Table 1). Four males, who completed their education before free access to secondary schooling became available (in 1967), held a primary education only. Unemployment and under-employment were high: some 65% overall, including more males than females, were unemployed or under-employed, especially during the economic recession of 2008-2010. Employment in an occupation below the qualification level was particularly marked among those with a third-level qualification. More generally, unemployment was widespread among professionals, skilled and unskilled workers in the construction industry. Female nurses and a newly-qualified teacher were unable to find employment in their professions also. In order to maintain confidentiality, the respondents are referred to by a number and specific locations are not named (e.g. rural#1). Migration and life overseas Most respondents reported several reasons for migrating in order of importance. Personal factors (and supporting family in one case), such as finding employment or better employment, were ranked first by all occupational groups, following Sjaastad (1962). Linked aspirations to gain experience/ skills and improve longer-term employment prospects were cited by both males and females over time. A 21-year-old male civil engineer who migrated to London, in 2010, because of "no available jobs, no income, wishing to gain experience and to experience travel" (rural#14), echoed a qualified nurse who migrated, age 22 in 1985, to work in a London hospital, because of being "unemployed, to gain experience in the workforce and to experience living in another country" (rural#31). Individualism in seeking freedom and adventure featured (Thissen et al., 2010), especially for recent long-distance migrants to Australia. A graduate, employed as a barman, who migrated to Sydney in 2011 for one year and worked in construction, reported that he sought "adventure", his "friends had migrated", he sought "a more lucrative income" and he was "bored with home life" (rural#21). The main destinations were major international cities with established Irish communities (notably, the Greater London Area [GLA], Perth, Sydney, New York and San Francisco in order of importance), reflecting path-dependency (Walter, 1980). Reciprocal labour market access for British and Irish citizens helps to explain a dominance of British destinations. Short-term work permits in engineering and certain skilled trades were available in Australia, in the late 2000s, and some recent graduates combined travel with work (Glynn et al., 2013). The sample respondents were legally resident in the USA. 'Strong ties' with family and friends were used by many skilled and unskilled migrants in finding employment in the Britain and even in Australia and the USA (Granovetter, 1973). Graduates and longer-distant migrants to Australia usually found employment through the 'weaker ties' of agencies and job advertisements. Regular contact with Irish people at work and socially in pubs, playing Gaelic games and attending matches was the norm and a potential source of social and psychological support. Migration acted as an 'escalator' from under-employment and unemployment to work that aligned more closely with the qualifications held (Newbold, 2001). Additional formal qualifications were gained: for example, a degree in civil engineering, a nursing qualification, various certificates and diplomas, specialised driving licences and training in trades. Valuable experience was gained also in the vocational area, even during a year's contract, as described by the nurse mentioned above: "Definitely get more responsibility in work leading to better chances of promotion, more professional qualifications" (rural#31). Positive comparisons were made by both men and women with the area of origin, in terms of the wider range of shops, goods, prices and transport available in the destination. More diverse lifestyle options became accessible. An under-employed typing teacher who moved to the GLA in 1998 and became a secretary said: "There were so many amenities close at hand, within walking distance. There was always somewhere to go every night of the week, cinema, theatre, clubs, etc. Buses/trains were so plentiful" (rural#16). Returnees from Australia valued the superior weather conditions and the outdoor lifestyle. Moving from a rural area or small town to a large city required personal adaptation in ways that suggest disruption. Contrasts with the area of origin were recalled in references to the physical scale of the destination area and longer travel distances to work. The larger and ethnically diverse populations represented a major change for migrants in the 1980s. A female shop assistant who moved to the GLA, age 21 in 1986, said that "There was a far more diverse population in England … in (place of origin) everyone was Irish, mainly from (place of origin)" (rural#17). A 25 year old builder, who migrated to San Francisco in 1988, from a small town, commented that "America was more urbanised, modern and people (were) very different" (rural#26). There were regrets for the countryside in expressions such as "missing the countryside", "life was much simpler in the countryside in Ireland" and missing a slower pace of life in comparison with the "faster pace of life" in a city. A sense of anomie was reported also (Ahmed, 1999), as described by the secretary above, who valued all of the material advantages of city life: "It was hard to get used to being anonymous in London, no

one knew you, cared about you or where you were going/doing/wearing etc. No one even said 'hello', so it was a bit daunting but you get used to it and soon stop eye contact with people and avoid saying 'hello'" (rural#16). Respondents who migrated at various periods of time to different cities referred also to missing family and friends and feeling homesick. A woman who worked in an Irish bank in the GLA had difficulty in having her accent understood initially (rural#2). There was therefore a sense of loss of being part of a familiar social field (Gustafson, 2001a). All respondents maintained transnational contact with family and friends in Ireland through media which changed over time from letters to telephone, email, Skype and social networking sites, illustrating the 'elasticity of place' (Barcus and Brunn, 2010). Twenty-two migrants visited at least once annually. Short-term migrants in Australia and New Zealand did not visit whilst away but maintained regular contact. Reasons for return Features of an idyllic countryside, involving membership of a social field and perceived environmental qualities that contrasted with those of a city, together with the availability of employment, dominated the reasons for return, following Gustafson (2001a). Only one respondent referred to establishing a small business (as a carpenter) and he migrated later to work on a more lucrative project; entrepreneurship was therefore limited among the sample, although noted in some other Irish research (Cawley, 2015). Proximity to family and friends was ranked first by many males and females over time, followed by access to employment. For those with children, proximity to family may have been linked to anticipation of support with childcare, although this was not mentioned specifically. Younger returnees were most likely to rank the availability of employment in first place. Whilst access to employment facilitated return, the motivations for return were inter-linked with life-course stage, in terms of expecting the birth of a child, planning a family, concerns about the welfare of children and retirement, and were often related to coming 'home' or establishing a 'home' in Ireland, as noted by Ní Laoire (2007). Fourteen migrants returned with a spouse/partner and a child or children whose welfare they cited. The female bank employee above returned with her spouse to her small town, in 2002, when work was available, because she "was pregnant, wanted child to be born and raised in Ireland and missed home" (rural#2). The trope of establishing a 'home' and family in Ireland was an aspiration also for a 25-year-old engineer who spent two years in Perth and returned with his partner, in 2012, to "settle back home in Ireland", "find work, marry and start a family" (rural#23). Welfare of children was framed with reference to features of an idyllic countryside being "a better lifestyle for them" than a city, as expressed by a construction worker who returned in 1996 after thirteen years in Britain (rural#11). Lack of safety in a city was referred to. Recapturing features of rurality was present as a motivation in references to missing country life. Family responsibilities to care for ageing or ill parents and personal difficulties influenced return in different decades, at times when employment was not necessarily easily accessible in Ireland. A young man returned earlier than expected, in 1959, when the Irish economy was underdeveloped, after five years working as a delivery man in London, to care for his dying mother and help his father on the farm. His experience of life in London related to the positive features of a regular income; recollections of the Irish countryside did not feature (rural#13). A 22-year-old carpenter returned after one year in Sydney, in 2012, to part-time employment, because of parental illness and his own homesickness. He had close contact with Irish people and appreciated: "(the) higher wages, more to do socially, greater sense of optimism in Australia, fast-paced lifestyle". He missed playing Gaelic sports but did not refer to missing other aspects of life in the countryside, although these may have been inherent in his homesickness (rural#21). Other young people returned after brief periods of absence on obtaining employment in their home area. Their recollections of the area of origin when overseas related more to the absence of employment or appropriate employment than positive features of country life. They include a male teacher who returned from Italy with his Italian wife, in 2011, where he had enjoyed living, on obtaining employment and because of owning a house in Ireland, which saved on rent in Italy (rural#1). A 24-year-old man spent a year working in Australia and travelling in South East Asia and returned on the expiry of a visa, in 2013, to return to college and did not report any sense of loss of an idyll whilst away (rural#32). A 23-year-old skilled worker who found employment in Perth for a year and returned to employment, in 2008, enjoyed the climate and culture in Australia and made no reference to missing rural Ireland (rural#33). Remembered idyllic features of proximity to family and membership of a familiar social field exerted an influence, in particular, on the return of emigrants with children or planning a family, who were able to find employment in the area of return. In other cases, usually among single people in their twenties, when obligations to parents, the termination of a work permit and the availability of employment were instrumental in their return, an idyllic imaginary was absent in the reasons cited. The shorter periods of absence, when the more negative features that led to emigration may have been recollected more clearly, together with the foreshortening of the potential to gain the experience and self-fulfilment sought, seemed to override any idyllic recollections of the countryside of origin. Life after return The experience following return was queried in terms of: (i) perceived differences (changes) in the area of origin, (ii) the features enjoyed most, and (iii) any difficulties experienced. The differences arose primarily from changes in the Irish economy: especially, improved employment and housing developments during the 1997-2007 years of growth and unemployment and emigration of friends during the recession of 2008-2011. Increased ethnic diversity was noted, associated with international immigration from the late 1990s. The bank employee who returned with her spouse, in 2002, because of wishing her child to be born in Ireland, appreciated "being home and catching up with friends" but the physical and social environment had changed: "(place) had grown, massive housing developments around what was once a small village. More nationalities had arrived in Ireland" (rural#2). These differences between expectations and experience point to the tensions that exist between the recollected idyll and a changed reality. The features enjoyed most on return and the difficulties experienced also revealed tensions between motivations and the experience of living in the countryside again. Thus, the attainment of the idyll by those who sought it, or appreciated (or not) by those whose return was motivated by other factors, may be identified. The features enjoyed most by returnees highlighted relations with others and being part of a wider social field, irrespective of the period of absence, pointing to the endurance of social relationships over time. Reuniting with family and friends and being known, primary reasons for return, were also most enjoyed: "familiarity of home", "being closer to loved ones", "feeling at home". Features of rural life that were missed when overseas were recaptured by many on return, in terms of "a more relaxed atmosphere", "a slower pace of life" and "peace and quiet". However, many of those who sought the perceived advantages of rural life for their children, as opposed to a city life, reported tensions arising from other aspects of rurality. A newly-qualified teacher who moved to London to gain employment in 1989 and returned in 2001, with her spouse and children, to provide the latter with a better lifestyle, appreciated: "Space-less densely populated; cleaner and greener-less graffiti, better air quality; food tasted better; less danger". Nevertheless, she found that "anything official or bureaucratic often took longer than expected" (rural#25). A nurse who returned from London with her family, after seventeen years' absence, in 1990, to raise their children and "settle down" and "to be near to family and friends", found that the countryside was: "Too quiet and basic, the children found it very difficult to find things to do" (rural#3). The construction worker, who returned with his family in 1996, after thirteen years in Britain, to provide "a better lifestyle" for his children adverted to "The nosiness of people" and his wife missing her family and friends whom she left behind (rural#11). Low levels of basic retail and transport services were restrictive for many in the countryside. In the words of a secretary, who cited schooling and a better environment for their children as influential in the family returning from the GLA, there were: "Less shops, therefore less choice and value for money, as there was less competition" (rural#27). A skilled tradesman who gained a degree in civil engineering whilst living in New York, and returned with his spouse and children, in 1991, because of having a house in Ireland, the growing economy and preferring "to raise kids in the country rather than the city", said that "the slower pace of life got taking used to again" (rural#8). Unlike Morse and Mudgett's (2017) respondents in Vermont, specific landscape features were not cited as reasons for return, except by an early retiree who referred to the contrasts between a Scottish industrial city where he lived and his home area. He returned to make a better life for his son and to return 'home'. After twenty-seven years' absence, he enjoyed "walking on the beach and day trips to (my) island home" which was then unpopulated. He did not experience any problems associated with returning (rural#12). Other respondents who, usually, had been absent for shorter periods of time and prioritised access to employment in an improved economy as a motivation for return, also valued renewed contact with family and friends. In general, broader features of an idyllic countryside received less attention from them. They reported difficulties in finding suitable employment, reduced incomes, poor rural services, loss of friendships and loss of independence. Thus, the teacher who returned to employment with his Italian wife, in 2011, enjoyed reuniting with his family but regretted leaving friends made abroad, had a reduced income and found that many of his friends had emigrated (rural#1). The nurse, referred to earlier, who returned at the end of a one-year contract in 1986 to take up training as a midwife, moved in with her parents and saved money, but she referred to a "loss of privacy/anonymity -'nosey' neighbour syndrome!". The lack of public transport was also a constraint on her social life (rural#31). Loss of friendships made overseas echoes Ralph's (2014) findings for Irish returnees from the USA. Returnees from Australia missed the pleasant climate and the associated lifestyle. Features of a remembered idyllic countryside were recaptured by returnees with children in particular who had been absent for periods of ten years or more. They experienced tensions on return, notably relating to deficiencies in service provision. Many of this group adapted, as De Bree et al. (2010) and Ralph (2014) noted elsewhere. Others migrated again to support their families. An air traffic controller, who returned from South Africa in 2006, because of missing 'home' and the availability of employment, valued being closer to family again. In 2011, he moved to the Middle East with his wife and two young children in order to earn income to meet their mortgage payments (rural#30). In response to a lack of suitable employment and income, in the area of return, a construction worker (rural#11) and a managerial employee (spouse of rural#16) commuted weekly to Scotland and England, respectively, whilst their families remained in Ireland. Younger returnees who had spent shorter periods of time abroad and returned out of family obligation, on the termination of work permits and, or, the availability of employment also experienced tensions relating to the reality that influenced emigration initially, notably in terms of reduced incomes. Their social circle was also reduced through emigration. A number migrated again, using the experience and transnational links that they had established. Conclusions The migration of young working age people from rural areas involves substantial losses of human and

social capital (Jentch and Shucksmith, 2004). Their return contributes human, social and economic benefits (Stockdale, 2006;Dustmann et al., 2011). A rural idyllic imaginary that includes aspects of landscape, a more relaxed pace of life and close contact with family in a community environment, features among the factors that influence return and further research is recommended on its role (Stockdale et al., 2013;Haartsen and Thissen, 2014). This article was designed to contribute to better understanding of the factors that influence rural return migration by seeking to capture key features of the economic migrant experience from departure, living in a city environment, to return and post-return, using Irish evidence. The sample is relatively small but it includes rich qualitative information relating to the rural migrant life-course. Males outnumber females in the sample but information is provided on family return which includes a spouse and a child or children. Unemployment and constraints on career and personal fulfilment in a rural area of origin emerge as causes for out-migration in the late teens and early twenties, as noted by Stockdale (2006). The advantages of city life are appreciated on arrival, particularly access to employment, higher incomes and superior services and the transition from the area of origin may be facilitated by social and, or, work contact with compatriots. Tensions arise, however, during the reality of adapting to life in a major city, from the larger scale of the place, a sense of anomie in an unfamiliar environment and the loss of close contact with family and friends (Ahmed, 1999). Regret for a loss of features of a rural idyll emerged unprompted from some of the narratives. Reuniting with family and friends was a major reason for return and became possible when employment opportunities improved in Ireland. Urban life was also re-evaluated at certain stages in the life-course and, notably, in the context of marriage, the birth of a child or children attending school, following Ní Laoire (2007). References to the countryside being a better and safer environment for children and proximity to family suggest that migrants wished their children to grow up in contact with remembered positive features of rural society and proximity to relatives. Help with childcare from grandparents may have been anticipated, although not stated specifically. For those motivated by recapturing a remembered rural idyll, being part of a familiar social field and a peaceful countryside were attained but the reality of rural residence, notably poor service provision, compared unfavourably with city living. Places had changed also as a result of housing development and new residential immigration. Tensions were therefore present between expectations and reality. Most families readapted but some spouses commuted weekly to Britain, for employment and income reasons, to support their families who remained in Ireland. Returnees who were influenced more by obligations to parents, the termination of a work permit or improvements in the Irish economy, than by an idyllic imaginary, recaptured features of 'home' in terms of proximity to family and friends. They experienced tensions relating to employment and income, poor services, the loss of privacy associated with moving back to live with parents, and the emigration of friends. Some of them migrated again using the transnational links that they had established and some returned again. Out-migration from rural areas is inevitable for many younger people because of the limited availability of higher education, training and desired employment opportunities within commuting distance, and a desire for personal fulfilment (Haartsen and Thissen, 2014). Depopulation, social and economic decline are continuing features of some rural areas in Western Europe. Unless wider scale abandonment is to be accepted, measures are required to retain larger numbers of younger people, following completion of higher education or training and to attract others back, as part of the pursuit of the 'good countryside' (Shucksmith, 2018). The reported research illustrates clearly that many overseas emigrants, who were the focus of this paper, retain close links with their family and friends whilst absent and some begin to re-value more idyllic features of their areas of origin, particularly when they consider the upbringing of their children. If economic and employment circumstances permit, they will consider returning to the area of origin. Such returnees may recapture the social idyll that they seek but tensions arise, particularly from poor levels of provision of everyday services. In the case of other returnees who may have spent shorter periods of absence and whose return may be involuntary, reuniting with family and friends is important but an idyllic rural imaginary does not appear to compensate for negative features of rurality. Some of them will migrate again although, possibly, return at a later stage. The retention of working age populations and the attraction of emigrants back are objectives of rural development policies in many countries (Shucksmith, 2018). As a contribution to policy in Ireland, the results illustrate that access to employment is necessary to facilitate return for most emigrants. So also is improved access to basic transport and other services to provide an acceptable quality of life for them. The motivations and requirements of different groups of returnees to the countryside differ, however, and these differences need to be recognised. Features of a remembered idyll may compensate for the reality of rural life for some returnees but not for others who may migrate again using their transnational knowledge and links. Using Automatic Speech Recognition to Optimize Hearing-Aid Time Constants Automatic speech recognition (ASR), when combined with hearing-aid (HA) and hearing-loss (HL) simulations, can predict aided speech-identification performances of persons with age-related hearing loss. ASR can thus be used to evaluate different HA configurations, such as combinations of insertion-gain functions and compression thresholds, in order to optimize HA fitting for a given person. The present study investigated whether, after fixing compression thresholds and insertion gains, a random-search algorithm could be used to optimize time constants (i.e., attack and release times) for 12 audiometric profiles. The insertion gains were either those recommended by the CAM2 prescription rule or those optimized using ASR, while compression thresholds were always optimized using ASR. For each audiometric profile, the random-search algorithm was used to vary time constants with the aim to maximize ASR performance. A HA simulator and a HL simulator simulator were used, respectively, to amplify and to degrade speech stimuli according to the input audiogram. The resulting speech signals were fed to an ASR system for recognition. For each audiogram, 1,000 iterations of the random-search algorithm were used to find the time-constant configuration yielding the highest ASR score. To assess the reproducibility of the results, the random search algorithm was run twice. Optimizing the time constants significantly improved the ASR scores when CAM2 insertion gains were used, but not when using ASR-based gains. Repeating the random search yielded similar ASR scores, but different time-constant configurations. INTRODUCTION Recent studies have shown that automatic speech recognition (ASR) can be used, in combination with signal-processing algorithms mimicking the effects of hearing loss (HL), to predict the speech-identification performances of older hearing-impaired (OHI) listeners [see Schädler et al. (2015) and Fontan et al. (2020a) for a discussion of the advantages of ASR-based metrics by comparison to other objective measures of speech intelligibility]. This was demonstrated for unaided (Kollmeier et al., 2016;Schädler et al., 2018;Fontan et al., 2020a) and aided (using simulated or real hearing aids; Fontan et al., 2020b;Schädler et al., 2020) speech perception. Based on these findings, it has been speculated that ASR-based prediction systems could also be used to assess speech-intelligibility benefits resulting from various hearing-aid (HA) configurations. Recently, Fontan et al. (2020c) used an ASR system to evaluate and to improve the insertion gains recommended by the CAM2 HA fitting rule (Moore et al., 2010b). For each of their hearing-impaired (HI) participants, 625 gain functions (corresponding to systematic variations of CAM2 gains by 0, ± 3, or ± 6 dB) were assessed. Each gain function was applied to speech stimuli using an HA simulator. The amplified speech material was then degraded using the HL simulator developed by Nejime and Moore (1997). Based on each participant's audiogram, both the elevation of hearing thresholds and loudness recruitment were mimicked. Spectral smearing, which is also implemented in the original HL simulator to mimic the loss of frequency selectivity, was not used used by Fontan et al. (2020c), since its simulation resulted in weaker correlations between ASR scores and human speech intelligibility (Fontan et al., 2020a). Finally, the amplified and degraded stimuli were fed to the ASR system for computing recognition scores. Fontan et al. (2020c) compared the benefits associated with the insertiongain function yielding the highest ASR scores (the "optimized" gains yielding a mean improvement of 13 percentage points) to those obtained with CAM2 gains in a group of OHI participants. Significantly higher human speech-identification scores were observed for speech amplified with optimized gains than for speech amplified according to the gains recommended by CAM2. These significant improvements were observed both for word and sentence materials. Gonçalves Braz et al. (2022) extended this work and combined ASR with several random-search (RS) algorithms to optimize not only insertion gains but also compression thresholds. This approach is referred to as OPRA-RS, which stands for "Objective Prescription Rule based on ASR and Random Search." Using slow time constants for the compressor of the simulated HA, optimized insertion gains and compression thresholds were determined for 12 audiometric profiles corresponding to different levels of HL severity. ASR scores yielded by the optimized parameters were significantly higher than those obtained with CAM2 (mean improvements ranged from 2 to 10 percentage points for the different RS algorithms). Significant differences were observed between RS algorithms in terms of ASR score and convergence speed. A limitation of Gonçalves Braz et al.'s (2022) study is that only one set of time constants was used. However, aided speech intelligibility depends on the attack and release times of the HA compressor (Moore et al., 2011;Hopkins et al., 2012). Small time constants (i.e., "fast" compression) help perceiving rapid changes in loudness, such as those occurring when a weak speech sound (e.g., a consonant) precedes or follows a speech sound with higher energy (e.g., a vowel; Souza, 2002;Hopkins et al., 2012). At the same time, when fast compression is implemented in a multi-channel HA that processes each frequency channel independently, it tends to reduce spectral contrasts (i.e., by flattening the speech spectrum) and may thus have a deleterious effect on the perception of speech formants, which are crucial for the identification of vowels (Bor et al., 2008). By causing rapid variations of the signal amplitude at the onset and offset of speech sounds, fast compression speeds can also distort the signal envelope and therefore negatively impact speech intelligibility Moore, 1992, 2008;Stone et al., 2009). These distortions are more likely to happen when high compression ratios are used (Verschuure et al., 1996). Despite their impact on speech intelligibility, there is currently no consensus as to the best time constants that should be used: time constants used clinically and commercially in hearing aids vary broadly, with attack and release times ranging from 0.5 to 2,000 ms and 10 to 5,000 ms, respectively (Moore and Sȩk, 2016). The present study extends the work of Gonçalves Braz et al. (2022) by investigating whether OPRA-RS can also be used to optimize time constants. Attack and release times were optimized for HA configurations that corresponded to the compression thresholds and/or the insertion gains recommended either by OPRA-RS or by CAM2 for the same 12 audiometric profiles as used in Gonçalves Braz et al. (2022). As these HA configurations sometimes involved high compression ratios (>3), and that in such cases, fast compression can distort the signal envelope and thus affect speech intelligibility (Souza, 2002), only "slow" compression speeds were used. To assess ASR performance, speech stimuli were first amplified using an HA simulator and then degraded to mimic the perceptual consequences of the elevation of hearing thresholds and loudness recruitment. The resulting speech signals were eventually fed to an ASR system for recognition. The optimization of compression speed was carried out twice in order to assess the reproducibility of the outcomes in terms of ASR scores and optimized time constants. Figure 1 describes the processing chain used to optimize time constants for a given input audiogram. At initialization, the RS algorithm randomly selects attack and release times within two ranges of

possible values. These time constants, as well as the compression thresholds and the insertion gains prescribed by OPRA-RS or CAM2 for a 65-and 85-dB-SPL speech input level (IGSP65s and IGSP85s, respectively; see Supplementary Datasheet for more details), are transmitted to an HA simulator. The HA simulator amplifies 50 speech stimuli corresponding to five 10-word lists of the speech intelligibility test of Fournier (1951), which is the test most often used by French audiologists for speech audiometry (Rembaud et al., 2017). The amplified speech signals are then degraded by the HL simulator, according to the input audiogram. The resulting speech signals are finally processed by an ASR system developed for the French language. Overview of the Optimization Chain A total of N iterations are used to assess N time-constant configurations. After each iteration, the ASR score and average log-likelihood of recognized words yielded by the current time constants are compared to those obtained with the best timeconstant configuration found up to the current iteration. If the FIGURE 1 | Components of the OPRA-RS optimization chain, with associated input data (in italics) and output data (right panel). The parameters randomized by the RS algorithm are highlighted in red. current configuration yields a higher ASR score (or the same ASR score but with a higher log-likelihood) than the previous best configuration, the current configuration is used as a baseline for the next iteration. Otherwise, the best previous configuration serves as a baseline for the next iteration. Based on the current number of iterations i, the search ranges are reduced around the baseline time constants for the next iteration, following the equation: where stepsize corresponds to the step (in ms) used to define possible values within the search range. Simulation of Hearing-Aid Processing A 5-channel HA simulator implemented in MATLAB TM (Moore et al., 2010a) was used to amplify the speech signals. The frequency ranges of the five HA channels were 0.1-0.7, 0.7-1.4, 1.4-2.8, 2.8-5.6, and 5.6-8 kHz. In each channel, the simulator used two dynamic range compressors placed in series: the wide dynamic range compression function was applied in the first compressor, while the second compressor was used as a limiter. For further details about the implementation of the HA simulator, see Fontan et al. (2020c). Simulation of Hearing Loss The functioning of the HL simulator, also implemented in MATLAB TM , is detailed in Nejime and Moore (1997). As done in Gonçalves Braz et al. (2022), the simulator was used to mimic two of the perceptual consequences of age-related HL: Based on the input audiogram, a linear filter simulated the elevation of hearing thresholds, while loudness recruitment was simulated by raising the signal envelope (Moore and Glasberg, 1993). Automatic Speech Recognition System The ASR system used in the study consisted of Hidden Markov Models and Gaussian Mixture Models. It was implemented using the Julius ASR engine (Lee and Kawahara, 2009). The acoustic models were trained on approximately 100 h of French radio broadcast news. These speech recordings were not processed to mimic HA amplification or HL and did not include the 50 word recordings used in the study to evaluate time constants. The lexicon used by the ASR system only comprised the 50 target words. A more detailed description of the ASR system is given in Gonçalves Braz et al. (2022). Test Procedure The processing chain was used to optimize time constants for 12 audiograms, using the compression thresholds selected by OPRA-RS and the insertion gains prescribed either by OPRA-RS or by CAM2. The audiograms, shown in Figure 2, represented mean or individual audiometric thresholds falling into levels 4-7 of the Wisconsin Age-Related Hearing Impairment Classification Scale (WARHICS; Cruickshanks et al., 2020). The audiograms corresponded to mild-to-moderately severe losses, with thresholds generally increasing as a function of frequency, as is typical of age-related HL. The mean audiograms were based on the data collected by Humes (2021). Some of the hearing thresholds required by the HL simulator (corresponding to the frequencies 0.125, 0.25, 0.75, and 1.5 kHz) were not included in the mean audiograms reported by Humes (2021). Those missing thresholds were intra-or extrapolated using third-least-squares polynomial regressions. The individual audiograms corresponded to older patients (mean age: 70 years; age range: 63-78 years) with sensorineural HL. For each of the four WARHICS levels, one mean audiogram and two individual audiograms were used. The RS algorithm that yielded the highest ASR performance in Gonçalves Braz et al. (2022) was used in the present study. This algorithm tunes all parameters (here, time constants) in all HA channels simultaneously. As in Gonçalves Braz et al. (2022), four independent RS threads were run in parallel. Each thread consisted of 1,000 iterations, during which time constants were randomly varied within predefined search ranges, using 10-ms steps. At the start of the RS, the search ranges were 10-500 ms for attack times, and 300-2,000 ms for release times. These ranges correspond to those generally associated with a slow compression system (Moore, 2008a,b;Moore et al., 2010a;Moore and Sȩk, 2013). For each audiogram, the final time-constant configuration yielding the highest ASR performance across the four search threads was selected. In what follows, unless explicitly mentioned, only data from the first repetition of the RS algorithm are used. Figure 3 compares the ASR scores achieved with default and optimized time constants, using either the insertion gains recommended by CAM2 (left panel) or those calculated by OPRA-RS (right panel). The default time constants correspond to the fixed compression speeds used by Fontan et al. (2020c) and Gonçalves Braz et al. (2022). Those were 200, 100, 100, 100, and 100 ms for attack times, and 2,000, 1,500, 1,200, 1,000, and 1,000 ms for release times for HA channels 1-5, respectively. RESULTS With the insertion gains recommended by CAM2, it can be noticed that ASR scores tended to be higher after the optimization of time constants (median ASR score: 92%) than with default time constants (median ASR score: 88%). As Kolmogorov-Smirnov tests indicated that the ASR scores were not normally distributed (p ≤ 0.044 in both conditions), a Wilcoxon signed-rank test was used to assess the significance of the observed difference. The results show that ASR scores are significantly higher after the optimization of time constants (Z = 2.8; p = 0.005). In contrast with this general trend, for two out of the 12 audiograms, all time-constant configurations tested during the RS yielded lower ASR scores than those obtained with the default constants. The improvements due to the optimization of time constants seem to be larger for the most severe HLs than for milder HLs. For example, for audiograms corresponding to level 7 of the WARHICS scale, the ASR score improved by 10 percentage points on average, whereas an average improvement of 1.3 percentage point is observed for audiograms corresponding to level 4 of the WARHICS scale. A Spearman correlation was computed to assess the existence of a significant association between HL severity, represented by the pure-tone average (PTA) for frequencies of 0.5, 0.75, 1, 1.5, 2, 3, and 4 kHz, and the improvement in terms of ASR score due to the optimization of time constants. The results indicate a significant positive relationship between the two variables (ρ = 0.62; p = 0.03), that is, the higher the PTA, the larger the benefit due to the optimization of time constants. Contrary to the ASR scores obtained with CAM2 gains, no improvement was observed after the optimization of time constants when using the gains recommended by OPRA-RS. For six out of the 12 audiograms, all time-constant configurations tested during the RS yielded lower ASR scores than those obtained with default time constants. The reproducibility of the ASR scores was assessed by comparing the outcomes of the two repetitions of the RS algorithm. For CAM2, the median ASR score achieved during the second repetition of the algorithm (91%) was very close to the score achieved during the first repetition (92%); a Wilcoxon test revealed that no significant difference existed between the ASR scores yielded by each of the repetitions (Z = −1.7; p = 0.10). For OPRA-RS, all ASR scores remained equal across repetitions of the RS algorithm. shows the distribution of attack and release times yielding the highest ASR performances for the 12 audiograms with the insertion gains recommended by CAM2 (left panel) or by OPRA-RS (right panel). In the cases for which better ASR scores were achieved with the default time constants used by Gonçalves Braz et al. (2022), these default values were retained as best configurations. Median attack and release times across channels are 135 and 1,190 ms, respectively, for CAM2 gains, and 100 and 1,200 ms, respectively, for OPRA-RS gains. Contrary to attack times, optimized release times span the entire possible range of values. Optimized time constants seem less variable for OPRA-RS than for CAM2. This is at least partially due to the fact that, for OPRA-RS, a larger proportion of the optimized time constants correspond to the default constants used by Gonçalves Braz et al. (2022). Finally, the time constants obtained during the two repetitions of the RS algorithm were compared. As default time constants corresponded to fixed values, the HA configurations for which the best ASR scores were achieved with default time constants were excluded from this analysis. For the remaining HA configurations (N = 16), the median absolute differences across repetitions were 85 and 355 ms for attack and release times, respectively. The minimum and maximum absolute differences were 0 and 420 ms for attack times, and 20 and 1,640 ms for release times. DISCUSSION This study provides proof of concept that RS can be used for the optimization of HA time constants for a given audiometric profile. This approach might prove particularly useful since there is currently no consensus as to the time constants that should be used to maximize speech intelligibility for a HI individual (Moore and Sȩk, 2016). It has been shown that knowledge of the HA user's cognitive abilities might help to choose slow or fast compression (Gatehouse et al., 2003;Souza and Sirow, 2014), but the results of studies addressing the relationship between hearing abilities and optimal time constants are heterogeneous (Hopkins et al., 2012;Moore and Sȩk, 2016). Within this context, OPRA-RS represents a novel approach that, given the audiometric profile of the HA user, can be used to systematically explore a large number of timeconstant configurations and assess their impact in terms of speech intelligibility. ASR scores for the optimized time constants were reproducible across repetitions of the RS algorithm, but were associated with different combinations of time constants. This is possibly due to an interaction between attack and release times, as the two parameters were optimized simultaneously. Future studies should optimize each parameter independently to assess their reproducibility. It might also be interesting to extend in future studies the search ranges used for attack and release times, which were limited in the present study to values generally associated with slow compression. For the mild-to-moderately-severe HLs used in the this study, the optimization of time constants yielded significant improvements in ASR scores for CAM2, but not for OPRA-RS. In addition, the improvements observed for CAM2 were small (4 percentage points, corresponding to 2 out of the 50 words used in the study). These observations are likely due to ceiling effects in the two test conditions, even before the optimization of time constants. Indeed, it was observed that more severe HLs, yielding the lowest ASR scores with CAM2 and default time constants, were associated with higher improvements after the optimization of time constants. To limit such ceiling effects and thus to assess if clinically significant benefits can be obtained, future studies should use more challenging experimental conditions (e.g., speech materials that are shorter and/or presented in noise). Finally, it should be determined if, as shown by Fontan et al. (2020c) for the fine-tuning of insertion gains, the benefits observed in ASR performance due to the fine-tuning of time constants translate into speech-intelligibility benefits for actual listeners with age-related HL, and if these benefits are clinically relevant. Also, in the present study, CAM2 as a baseline prescription since it was used in the previous experiments on ASR-based optimization of HA parameters (Fontan et al., 2020c;Gonçalves Braz et al., 2022). It should be determined if significant improvements are also observed for those prescription rules that are

more widely used in clinical practice, such as NAL-NL2 (Keidser et al., 2011). DATA AVAILABILITY STATEMENT The datasets generated and analyzed for this study can be obtained from the corresponding authors for any research purpose. AUTHOR CONTRIBUTIONS LF initiated the idea. LG designed and implemented the randomsearch algorithms, under the supervision of JP. MS provided scientific advice about fitting algorithms, and the hearing-aid and hearing-loss simulations. LF, LG, MS, and CF analyzed and interpreted the data. LF and CF wrote the manuscript. All authors approved the final version of the manuscript. Molecular typing of methicillin-resistant and methicillin-susceptible Staphylococcus aureus isolates from Shiraz teaching hospitals by PCR-RFLP of coagulase gene. Background and objectives To investigate coagulase gene polymorphism of MRSA and MSSA isolates from Shiraz teaching hospitals from 2011 to 2012. Materials and Methods A total of 302 isolates of Staphylococcus aureus were collected from clinical specimens in three major teaching hospitals and confirmed on the basis of morphological characteristics and biochemical tests. The isolates were subjected to molecular typing on the basis of coagulase enzyme gene polymorphism by PCR-RFLP. Results There were 27 and 28 different RFLP patterns for AluI and HaeIII restriction enzymes respectively. This study showed that the discriminatory power of coagulase gene typing by Hae III enzyme was more than that of Alu I enzyme. Conclusion PCR-RFLP method is rapid, reproducible, simple and efficient for typing Staphylococcus aureus isolated from clinical specimens. This study showed that Hae III discriminatory power is better than Alu I for typing Staphylococcus aureus isolates. INTRODUCTION Staphylococcus aureus is the most common cause of nosocomial infections that causes skin and soft tissue infections such as boils, carbuncle, cellulitis and abscesses (1,2). Drug-resistant strains of the bacteria are rapidly being developed; thus, the treatment of this organism is difficult. Methicillin resistant Staphylococcus aureus (MRSA) infection is a major cause of increasing morbidity and mortality (1). The prevention and control of Staphylococcus aureus infections depends initially on detection of the risk factors in individuals exposed to S. aureus, analysis of the isolates by discriminatory DNA typing methods, and understanding the transmission of the bacterial infection (1). Molecular typing can play an important role in the epidemiological study of nosocomial infections, such as methicillin-resistant Staphylococcus aureus (MRSA) infection (3). In many countries, Staphylococcus aureus genotyping methods have become a part of the upcoming health care system and also the study of the strain origin, clonal relatedness and the epidemiology of the infection (2). In the past two decades, a variety of molecular typing methods including pulsed-field gel electrophoresis (PFGE)(4), multilocus enzyme electrophoresis (MLEE), multilocus sequence typing (MLST) (5), spa typing (6), and coagulase genotyping (7) have developed to differentiate the related strains from unrelated ones (8) and to identify and compare the genotypes of S. aureus (9). Genotyping methods are relatively stable in natural conditions and produce reproducible results. These methods are rapid, do not require laboratory culture. The ability for typing by these methods is nearly 100% (2). Pulsed-field gel electrophoresis and multilocus sequence typing are the most reliable and the highest discriminative typing methods, but are very hard, difficult, timeconsuming and expensive to be used in a clinical microbiology laboratory (1). Coagulase gene (coa) typing is a simple, accurate, reproducible enough, easy to interpret and discriminatory method for typing Staphylococcus aureus isolates from various sources (2,(9)(10)(11). It is well-known that coagulase enzyme is produced by most of the strains of S. aureus. Today, the ability to produce the coagulase in the clinical microbiology laboratory is used to detect S. aureus in human infections. It has been demonstrated that coagulase is an important virulence factor during the infection process (1,11,12). There are many different allelic forms of the S. aureus coa gene, each isolate produces one or more than one of these forms (11,13,14). The discriminatory power of coagulase gene typing depends on the variability of the region containing the 81 bp tandem repeats at the 3' coding region of the coagulase gene that differs both in the number of tandem repeats and the location of AluI and HaeΙΙΙ restriction sites among different isolates (1)(2)(3)11,12,(15)(16)(17). So, different S. aureus isolates could be discriminated using the coagulase gene typing method (11,18). In Iran, there is a little information about the genetic differences between aureus isolates from various hospitals and in particular, no information about coagulase gene polymorphisms. In present study, we have used coagulase gene typing method on the basis of the PCR-RFLP to discriminate S. aureus strains obtained from different specimens in teaching hospitals in Shiraz, Iran. Also, the ability of AluI and HaeIII restriction enzymes, for differentiating S. aureus isolates in PCR-RFLP method was also evaluated (1). Cultivation and identification of bacteria. From August 2011 to July 2012, 302 clinical isolates of Staphylococcus aureus were collected from three major teaching hospitals in Shiraz, Iran. S. aureus strains were identified based on morphological characteristics and biochemical tests. Of all bacterial isolates, 39% were from female and 61% from male subjects. The percentage of the isolates from different wards of hospitals and various samples are presented in Tables 1 and 2. Antibiotic susceptibility test. Disks containing cefoxitin (30 µg , Mast, UK) was used to discriminate the MRSA from MSSA isolates according to CLSI guideline recommendations. MHA (Muller Hinton Agar) plates were cultured with 0.5 McFarland standard of the bacterial broth culture and antibiotic disks were placed on the plates. Then this plate was incubated for 18 h at 37°C (1). There are several studies reporting that cefoxitin disk is preferable to oxacillin for detection of MRSA strains (19)(20)(21)(22)(23). Extraction of Genomic DNA. Bacterial whole DNA was extracted from isolates by using the smallscale phenol-chloroform extraction method and used as template in PCR (24). Polymerase chain reaction. This method was carried out with a little difference from the method described by Himabindu et al. The forward primer sequence for coa gene was 5'CGAGACCAAGATTCAACAAG and the reverse primer sequence was 5'AAAGAAAACCACTCACATCA-3' (1). PCR conditions were as follows: 94°C for 5 min, followed by 30 cycles of 95°C for 30 sec, 55°C for 45 sec and 72°C for 1.5 min, followed by a final extension of 72°C for 7 min. The PCR products were separated by 1.5% agarose gel electrophoresis. Restriction Fragment Length Polymorphism (RFLP) analysis. Alu I (Fermentas, Lithuania) digestion of the products was performed by adding 15 µL of PCR product to 15 µL of the mixture that contained 2 U of Alu I, 3 µL of buffer enzyme and 8.11 µL of distilled injection water. Then, the reaction mixture was incubated at 37°C for 16 h. Hae III (Fermentas, Lithuania) digestion of the product was performed by adding 15 µL of PCR product to 15 µL of the mixture that contained 5 U of Hae III, 3 µL of buffer enzyme and 5.11 µL of distilled injection water. Then, the reaction mixture was incubated at 37°C for 16 h. To prevent evaporation of reaction mixture, 20 µL of sterile PCR oil was added. The restriction digest fragments were detected by 3% agarose gel electrophoresis. In this study, S. aureus strain 25923 was used as positive control. Discriminatory power. The ability of a typing method to discriminate different types of the unrelated strains sampled from the hospital was assessed according to the Hunter-Gaston formula (1,25). This is also called discriminatory index (D). D= Numerical index of discrimination, N= The total number of isolates in the sample Population, s= The total number of types obtained, nj= The number of isolates belonging to the jth type contains (1,25). Coagulase gene amplification: The size of PCR products produced after coagulase gene amplification was separated into 6 different banding patterns in electrophoresis (Fig. 1, Table 3). The majority of isolates grouped in pattern 3 (729bp). We pairs) consisting of multiples of 81bp tandem repeat units. After RFLP analysis of all the isolates, based on the number and size of produced bands, 26 different patterns were observed ( Table 5). The pattern 5 "324-405" was found to be the predominant pattern. The isolates that belonged to pattern 26 produced only 1 band which indicates that this PCR product is not digested by HaeIII. This probably indicates the lack of HaeIII restriction sites amongst them. The discriminatory power was detected as 0.90. DISCUSSION Typing can be used to prevent or reduce an epidemic infection, reducing costs and hospital infection rates in hospitals (1). In this study, the polymorphism of coagulase gene among MRSA and MSSA isolates were investigated using PCR-RFLP analysis. The results of typing can be used to discriminate the strains within a given species (26). If strains from two patients have the same fingerprint, it indicates that both of them are infected from the one source. Discriminatory power of coagulase gene typing is high. It is an appropriate method for epidemiologic investigation of Staphylococcus aureus infection since it is applicable for typing of large number of strains in a short time. he discriminatory power of this method is lower than PFGE and MLST but it is quicker and less costly (27). RFLP method can help to trace the source of infection and transmission; thus, this typing method can be used to prevent and control the spread of infection (1). The purpose of this study was to investigate genetic variation in coagulase gene of aureus strains isolated from teaching hospitals in Shiraz, Iran. Although, MRSA strains are important cause of the nosocomial infections, no data are available on molecular typing of MRSA or even MSSA isolates by PCR-RFLP of coagulase gene in hospital isolates from Iran. There are few studies on this subject in Iran and one of them was in veterinary practice (9). However, in a study in Urmia, Iran, 26 S. aureus isolates from urine and skin in two hospitals were analyzed by PCR-RFLP using HaeIII enzyme. PCR products ranged from 490-790 bp in size and 6 distinct RFLP banding patterns were observed after the digestion of PCR products (2). The discriminatory index of coagulase gene typing by PCR-RFLP on the basis of HaeIII enzyme is more than of this method based AluI. Thus, HaeIII enzyme is better than AluI enzyme for typing aureus. In conclusion, our study proved that, of PCR-RFLP of coagulase gene is a rapid, simple and efficient method for typing S. aureus strains isolated from different clinical specimens in Shiraz teaching hospitals. This typing method can be used for tracing the source and transmission route of S. aureus infections and helps to prevent and control those related infections in our hospitals. Restriction fragment length polymorphism genotyping of human Staphylococcus aureus isolates from two hospitals in urmia region of iran using the Image guided dose escalated prostate radiotherapy: still room to improve Background Prostate radiotherapy (RT) dose escalation has been reported to result in improved biochemical control at the cost of greater late toxicity. We report on the application of 79.8 Gy in 42 fractions of prostate image guided RT (IGRT). The primary objective was to assess 5-year biochemical control and potential prognostic factors by the Phoenix definition. Secondary endpoints included acute and late toxicity by the Radiotherapy Oncology Group (RTOG) scoring scales. Methods From October/2001 and June/2003, 259 men were treated with at least 2-years follow-up. 59 patients had low, 163 intermediate and 37 high risk disease. 43 had adjuvant hormonal therapy (HT), mostly for high- or multiple risk factor intermediate-risk disease (n = 25). They received either 3-dimensional conformal RT (3DCRT, n = 226) or intensity modulated RT (IMRT) including daily on-line IGRT with intraprostatic fiducial markers. Results Median follow-up was 67.8 months (range 24.4-84.7). There was no severe (grade 3-4) acute toxicity, and grade 2 acute gastrointestinal (GI) toxicity was unusual (10.1%). The 5-year incidence of grade 2-3 late GI and genitourinary (GU) toxicity was 13.7% and 12.1%, with corresponding grade 3 figures of 3.5% and 2.0% respectively. HT had an association with an increased risk of grade 2-3 late GI toxicity (11% v 21%, p = 0.018). Using the Phoenix definition for biochemical failure, the 5 year-bNED is 88.4%, 76.5% and 77.9% for low, intermediate and high risk patients respectively. On univariate analysis, T-category and Gleason grade correlated with Phoenix bNED (p = 0.006 and 0.039 respectively). Hormonal therapy was not a significant prognostic factor on uni- or multi-variate analysis. Men with positive prostate biopsies following RT had a

lower chance of bNED at 5 years (34.4% v 64.3%; p = 0.147). Conclusion IGRT to 79.8 Gy results in favourable rates of late toxicity compared with published non-IGRT treated cohorts. Future avenues of investigation for toxicity reduction include IMRT, margin reduction, and dose modulation targeted to sites of disease burden. Further work is required to maximize efficacy beyond that achieved through radiation dose escalation alone. Background Prostate cancer is the most commonly cancer diagnosis in Canadian men. External beam radiation therapy (EBRT) has an established role in the management of localized prostate cancer, although historical series have shown relatively poor outcomes and high toxicity [1]. This led to the investigation of various adjuncts to EBRT to improve the therapeutic ratio, with adjuvant hormonal deprivation, in particular, proving successful for men with high risk disease [2]. A steep radiation dose-response curve has been postulated, and investigated in 3 reported randomized studies to date using doses of between 78-79.2 Gy in the dose escalated arm [3][4][5][6]. All of these studies have shown an improvement in biochemical outcomes, although often only in particular risk stratification subgroups. The study with the most mature follow-up is also beginning to show an advantage in clinical endpoints [4]. All of the dose escalation studies have shown an increase in late toxicity, and various strategies have been implemented to help minimize this. Firstly, treatment based on 2-dimensional planning has been proven more toxic than with 3-dimensional conformal radiotherapy (3DCRT), with the latter widely practiced alongside intensity modulated radiotherapy (IMRT) [7][8][9]. Another is the recognition of the dose sensitivity of the rectum, and the routine use of dose constraints for this and other critical structures [10]. Soft tissue image guided radiotherapy (IGRT) has also been widely introduced to help deal with interfraction organ motion [11]. Although a variety of approaches are in use, intraprostatic gold fiducials are the most widespread in clinical practice [12]. After investigating the efficacy of 75.6 Gy in 42 fractions, the Princess Margaret Hospital continued to utilize all of the above approaches prior to escalating prostate dose to 79.8 Gy in 42 fractions [13]. Here we present 5-year efficacy and toxicity outcome data for this treatment practice. Study Design Retrospective analysis of a prospectively maintained institutional prostate cancer RT database. The first 302 consecutive men treated with curative intent with RT at PMH following the implementation of 79.8 Gy in October 2001 as the standard treatment approach for localized disease formed the study cohort. This project has University Health Network institutional human research ethics committee approval (REB 08-0473-CE). Study population Eligible patients had biopsy confirmed adenocarcinoma of the prostate with clinical stage T1-3N0 M0. All patient details were independently verified, corrected and updated by two different radiation oncologists. Patient enrolled on a concurrent randomized trial receiving 5 months of bicalutamide in the experimental arm were excluded from the study cohort. Staging CT and bone scans were not routinely performed for those with low and intermediate risk disease. Patients with less than 2 years of follow-up data available were excluded to reduce bias in under-reporting of toxicity due to an insufficient period of observation. Treatment Planning The radiotherapy planning and treatment delivery has been previously reported [13]. All patients had 3 gold fiducial markers (1 mm diameter by 5 mm long) inserted transrectally under ultrasound guidance into the prostate at base, apex and midgland. Patients were immobilized in the supine position with a VacLoc cradle extending from hip to thigh and leg stocks to immobilize knees. Patients were given specific instructions regarding daily preparation to include a comfortably full bladder via consumption of 500 mL of water within 1 hour of planning and an empty rectum via the use of milk of magnesia for one week beforehand. This approach was continued throughout treatment as well. If the rectum was distended with gas or feces on the scout CT, the patient was asked to evacuate prior to proceeding with the full scan. Axial images were obtained at 5 mm through the pelvis and at 2 mm intervals through the prostate, using a helical scanner. Contouring structures were delineated using ICRU-62 convention. The GTV was the prostate. The CTV was the same as GTV, except for men with a predicted risk of seminal vesicle (SV) involvement of greater than 15%, in which case the proximal 10 mm of the SV were included in the CTV [14]. The PTV was CTV plus a uniform 10 mm margin except posteriorly where 7 mm was used. No one received elective pelvic lymph node irradiation. Femoral heads, bladder walls and rectal walls were contoured throughout the treatment volume (18 mm superior and/ or inferior from limits of the CTV as appropriate). Radiation prescription The radiation prescription dose was 79.8 Gy in 42 fractions given in 5 fractions per week. Doses were prescribed to the ICRU reference point, with the PTV contained within the 95% isodose line. IGRT verification dose was incorporated into the plan. Critical structure dose constraints used are shown in table 1. The standard treatment approach incorporated a six-field class solution with gantry angles of 60, 90, 120, 240, 270 and 300 degrees, although these angles could be modified to optimize treatment plans. An anterior field (0 degrees) was used for orthogonal imaging of the fiducial markers. If dose constraints were exceeded, an IMRT inverse plan was used. In all cases a coplanar beam arrangement was applied with 6 MV for IMRT plans, and 6 or 18 MV photons for 3DCRT plans. Helios inverse treatment planning module within CADplan 6.27 was used for IMRT planning. Treatment delivery Field placement verification was done by means of daily electronic portal imaging of an antero-posterior and a lateral field using an amorphous Silicon array. Fiducial marker centre of mass was matched electronically to reference images by the treating technologist, and deviations of more than 3 mm in any orthogonal plane corrected online prior to treatment being delivered. Follow-Up Weekly review of patients was conducted by the radiation oncologist to manage any acute reactions. Acute symptoms were prospectively scored by treating Therapists and recorded in the electronic patient record using RTOG criteria [15]. There was no formal follow-up procedure, but follow-up policy included appointments between 4-12 weeks following the completion of treatment to assess toxicity resolution, then every six months for assessment of late toxicity, clinical and biochemical control. Endpoints The primary endpoint was 5-year biochemical no evidence of disease (bNED) according to the nadir + 2 definition [16]. The bNED using the previous ASTRO definition of three consecutive rises backdated is also reported to allow intercomparison with older literature. Instigation of salvage therapies and evidence of clinical disease progression prior to a PSA rise where also counted as a failure. For the ASTRO definition, hormone use lead to patients being excluded from bNED analysis. Peak physician-assessed acute and late toxicity was graded according to the RTOG criteria for actuarial reporting. Clinical relapse in lymph-nodes or bones was also recorded, as was instigation of salvage interventions (local, hormonal or chemotherapeutic). Any metachronous malignancies were recorded, as were any deaths and their causes. Statistics Patients were censored either at the time of an event, or the time of last review, whichever occurred first. Kaplan Meier curves for biochemical control using the two failure definitions were generated using SPSS v17.0. Univariate analyses for potential prognostic factors (age, PSA, Gleason score, T-category, risk stratification, hormonal use) were performed for the nadir + 2 biochemical failure definition. Multi-variate analyses were performed using a Cox-Regression model. Chi-squared tests were performed to assess for independence in categorical data. A log-rank p-value of less than 0.05 was considered significant. Results 22 men were excluded with less than 24 months of follow-up, and a further 21 men were excluded as they received hormonal therapy on the experimental arm of a concurrent clinical trial. Overall 259 patients were identified who fitted the inclusion criteria. They commenced radiotherapy between October 2001 and June 2003. Baseline characteristics of patients treated according to protocol, including risk stratification as per Canadian Consensus Guidelines, are shown in table 2 [17]. Median follow-up for all patients is 67.8 months (range 24.4-84.7). Acute toxicity The worst RTOG acute toxicity scores are shown in table 3. No instances of grade 3-4 toxicity were observed during treatment. Bladder wall 70% volume receives less than 70 Gy. 50% volume receives less than 55 Gy. Right/Left femoral head Maximum dose 55 Gy Late toxicity No RTOG grade 4-5 late toxicity was recorded. Table 4 shows late toxicity at the time of last assessment. The actuarial rate of late gastrointestinal (GI) toxicity grade 2-3 was 13.7% at 5 years, and the corresponding figure for genitourinary (GU) toxicity was 12.1%. The comparable figures for grade 3 toxicity were 3.5% (GI) and 2.0% (GU) respectively. These data are summarized in Figures 1 Photodynamic therapy (PDT) was delivered for two men with presumed isolated local relapse with negative imaging and positive prostate biopsies. A further two men had either cryotherapy, or a radical prostatectomy for local salvage. All but one of these four men had a rising PSA on subsequent follow-up. Distant Disease and Survival Six men have developed positive bone scans, four pelvic lymphadenopathy, and one both boney and cerebral metastatic disease. 39 patients have being commenced on salvage hormonal therapy, generally at a relatively early state of biochemical relapse (median PSA = 7.1 at initiation), usually in a continuous rather than intermittent fashion, and all but 2 on an LHRH agonist in the first instance. Four have progressed to become hormone refractory, 3 have received docetaxol, and the other mitoxantrone. 20 men have experienced a metachronous malignancy, with the most common primary sites being colorectal (n = 8), bladder (n = 3), lung (n = 2) and stomach (n = 2). One man progressed through hormonal therapy and chemo-therapy, and died with brain metastases, presumed due to prostate cancer, 25 months after commencing RT. Nine further men have died, however none of these were suffering biochemical relapse at this time of last follow-up. Discussion Our experience shows that dose escalated RT (DERT) to a dose of 79.8 Gy can be safely delivered using 3DCRT and IGRT. Our 5-year bNED by the nadir + 2 definition was 79.4% (95% CI 74.1 -84.6), which approximates the corresponding figure of 85% reported from the 78 Gy arm of the MD Anderson Cancer Centre (MDACC) randomized trial [4]. Our ASTRO outcomes are an improvement over our previously reported results of treating to 75.6 Gy (5 yr bNED 55% v 67.9%) [13], and the Memorial Sloan Kettering results using the same dose for high/unfavourable risk patients (5 yr bNED 63% v 43%) [18]. It is difficult to know if this is a function of radiation dose escalation, heterogeneity within the high risk patient groups, differential hormone administration, or a combination of such factors. The MDACC randomized trial had a higher proportion of high risk patients than in our series, no image guidance, and CT-planning only following the 4-field phase 1 RT. Hence, the slightly lower bNED in our series is unexpected. Although larger field RT including irradiation of the seminal vesicles have yet to be prospectively validated as improving disease control outcomes, there is some potential that these larger RT volumes may have had a beneficial impact in the MDACC series. Our series used a nomogram cut-off value of 15% prior to inclusion of the seminal vesicles in the CTV. Potentially, similar methods may be helpful in determining the risk of extra-capsular extension and lymph node involvement, although the selection of cut-off values in each of these respects will be empirical in the absence of prospective data. In the era of CTV-PTV margin reduction to reflect the improved understanding and management of organ motion, it remains important to reflect on the Prostate-CTV expansion. The risk of extracapsular disease, subclinical seminal vesicle and pelvic nodal involvement should be considered when the CTV is delineated. Local relapses were identified on post-treatment biopsy, suggesting that despite more accurate radiation delivery of a higher dose some tumour clonogens will survive. One option is to continue escalating the radiation dose, as has been explored at Memorial Sloan Kettering Cancer Center [19]. Another promising avenue is to exploit the likely fraction sensitivity of prostate cancer [20]. Several centres have recently been investigating the feasibility

of hypofractionation [21][22][23][24][25]. Our own experience with 60 Gy delivered in 3 Gy fractions over 4 weeks utilizing IGRT and IMRT is now being explored in a randomized trial compared with a 78 Gy in 2 Gy fraction standard arm (PROFIT -Prostate Fractionated Irradiation Trial) [26] A further strategy is to identify and direct additional treatment at the dominant prostate lesion [27]. Additional follow-up will be necessary to determine the effect of DERT and IGRT on important clinical endpoints such as local, nodal or boney relapse as well as prostate cancer specific survival (PCSS). Only after a median follow-up of 8.7 years did any advantage for such failure endpoints appear for the MDACC RCT, which is now beginning to suggest a PCSS advantage [4]. The advent of effective systemic therapy for hormone refractory prostate cancer, with the potential of others under investigation, may dilute survival impacts of DERT compared with historical cohorts [28,29]. Severe late rectal toxicity was unusual in our cohort, with grade 3 reactions reported in 3.5% of patients. The Cleveland Clinic used ultrasound based IGRT with intensity modulated RT (IMRT) in a hypofractionated manner to a dose of 70 Gy in 28 fractions, and reported a corresponding figure of 2% [22]. PMH has a low threshold for intervention for men with rectal bleeding, which may contribute in differences between series. Cleveland clinic has also shown that IGRT removes any prognostic impact of rectal filling at the time of treatment planning on efficacy and rectal toxicity [30,31]. On the basis of interfraction motion studies, IGRT has become the standard of care, and it is pleasing to see our results supporting such theoretical advantages [11]. The dose escalation randomized trials reported rate of grade 2 or greater late toxicity in some detail [3,4]. For the Actuarial RTOG grade 2-3 late genitourinary toxicity for 259 men treated for localized prostate cancer with IG-3DCRT to 79.8 Gy Figure 2 Actuarial RTOG grade 2-3 late genitourinary toxicity for 259 men treated for localized prostate cancer with IG-3DCRT to 79.8 Gy. Minimum followup is 2 years. Dutch trial at the 7 year time-point, there was no difference in late grade 2+ GU late toxicity between the 68 Gy and 78 Gy arms (41% v 40%, p = NS), but a significant worsening of grade 2+ GI late toxicity (25% v 35%, p = 0.04) [3] Similarly, the MDACC trial comparing 70 Gy with 78 Gy showed similar trends in 10 year GU (8% v 13%, p = NS) and GI (13% v 26%, p = 0.013) grade 2+ late toxicity [4]. Allowing for differences in duration of followup, the corresponding GI and GU figures from the current series of 12.6% and 12.1% compare favourably with the standard arms of these trials. Although there is potential for under-reporting of late toxicity when data is collected retrospectively, moderate to severe toxicity requiring interventions is less likely to be missed than grade 1 toxicities. It is therefore plausible that newer technologies have allowed the delivery of an extra 8-10 Gy without an increase in toxicity in moderate to severe late rectal toxicity. Both of these randomized studies incorporated some 2-dimensional planning, as well as no IGRT. It would appear that IGRT using either ultrasound or gold fiducials plus standardized 3DCRT planning with relatively conservative DVH constraints can reduce toxicity markedly. Although there is some evidence that the superior dosedistributions achievable with IMRT offers further a reduction in rectal toxicity compared with 3DCRT, such series usually do not utilise IGRT [7,32]. IMRT was used infre-quently in the current series, and given the approximate halving in the rate of Grade 2-3 rectal toxicity in the Cleveland clinic IGRT cohort compared with our own, it is possible that IMRT used in the Cleveland experience could be responsible for some of this improvement. However, the use of a smaller 4-8 mm non-uniform CTV-PTV expansion at that institution may also have contributed to this result. Although there is a suggestion from our series of an association between the use of hormonal therapy and late GI toxicity, this may be due to confounding factors such as larger prostate volumes in men managed with cytoreductive hormonal therapy. Further work in this area is needed to clarify whether such a relationship between hormonal therapy and toxicity exists. Moderate grade 2 bladder toxicity in the form of irritation or slight hematuria occurred in approximately 12% of our patients during follow-up. One must note that in general urinary toxicity was not enhanced by the delivery of higher doses of radiation in the randomized trials. However, the use of IGRT would ensure that the prostatic urethra receives full dose throughout the treatment course, and therefore one might expect similar toxicity levels as reported in the randomized trials of between 17-41%. This may be part of the reason why bladder DVH constraints have generally correlated relatively poorly with late toxicity [33]. Biochemical control by nadir + 2 definition and risk stratification for 259 men treated for localized prostate cancer with IG-3DCRT to 79.8 Gy Although toxicities can be self limiting, a focus of further work should be to try to minimize late moderate urinary toxicity. Objective scoring and management of lower urinary tract symptoms prior to treatment may be helpful, with cytoscopic assessment and intervention warranting consideration for men with poor premorbid function. Urethral sparing approaches may be technically feasible, but raises the concern of underdosing malignant tissue which can occur in a peri-urethral distribution as well as difficulties in visualizing the urethral on a daily basis. Greater education regarding bladder base anatomy on CT and MRI may assist with dose reduction to the bladder trigone. It is likely that a combination of such approaches will be required. There is little consensus on the optimal margin expansion from CTV to PTV. The main contributors to this are interfraction motion, intrafraction motion, and contouring variation. The use of IGRT with a zero mm action threshold should address the first issue. Recent work to quantify intrafraction motion suggests that 2 mm will account for this in the majority of patients for the majority of fractions [34][35][36]. The use of real time tracking could potentially reduce this further. Contouring variation can be minimized with education and MRI fusion to some extent [37]. Overall the current trend would be to reduce margins to between 5-7 mm, which would be expected to further reduce rates of late rectal toxicity, although mature follow-up of current cohorts will be needed to confirm this. We were not able to show any advantage for the small number of selected men with intermediate and high risk disease who received adjuvant hormone therapy in the current series, and one area of controversy remains the optimal manner to integrate hormonal therapies (HT) with DERT for men with intermediate and high risk disease [24]. The picture becomes less clear in the era of PSA screening and Gleason grade migration. Although trials are currently underway to address the additive advantage of HT with DERT as standard treatment, such results are pending. In the meantime, selection of men with highintermediate risk prostate cancer for a short course of HT will be empirical, with a number of suggested approaches including nomograms, Gleason 4+3 disease, PSA>15, and 2-3 intermediate risk factors [38,39]. What is clear is that in the absence of definitive evidence of benefit in this patient subgroup, that toxicities must be discussed with prospective patients, and their involvement is crucial in the decision making process. Our cohort was predominantly made up of men with lowintermediate risk disease, and did not have a large number of biochemical failures. Hence it is difficult to define risk factors in a relatively homogeneous cohort with good out-comes. The predictive power of positive biopsies has been previously reported, [13] although the current series suggests a systematic bias in clinical practice to biopsy men with a rising PSA. The lack of prognostic significance for the different dominant Gleason 7 patterns may reflect the impact of Gleason grade migration, redefinition of overall Gleason scoring from biopsies and the subsequent 'Will Rogers Effect' [40]. Conclusion Dose escalated radiotherapy using predominately 3D conformal approaches and IGRT to a dose of 79.8 Gy is feasible and leads to good rates of biochemical control. Improved CTV definition to address extracapsular extension, seminal vesicle and pelvic nodal occult disease remains an avenue for investigation. Significant rectal and bladder toxicity is unusual, but could potentially be reduced with the routine use of IMRT and a smaller CTV-PTV margin expansion. Further work is needed to optimize total dose, dose per fraction, integration with hormonal therapy, and intra-prostate tumour targeting. Multipotent Mesenchymal Stromal Cells Interact and Support Islet of Langerhans Viability and Function Type 1 diabetes (T1D) is a widespread disease, affecting approximately 41.5 million people worldwide. It is generally treated with exogenous insulin, maintaining physiological blood glucose levels but also leading to long-term therapeutic complications. Pancreatic islet cell transplantation offers a potential alternative treatment to insulin injections. Shortage of human organ donors has raised the interest for porcine islet xenotransplantation. Neonatal porcine islets are highly available, can proliferate and mature in vitro as well as after transplantation in vivo. Despite promising preclinical results, delayed insulin secretion caused by immaturity and immunogenicity of the neonatal porcine islets remains a challenge for their clinical application. Multipotent mesenchymal stromal cells (MSCs) are known to have pro-angiogenic, anti-inflammatory and immunomodulatory effects. The current state of research emphasizes the great potential of co-culture and co-transplantation of islet cells with MSCs. Studies have shown enhanced islet proliferation and maturation, insulin secretion and graft survival, resulting in an improved graft outcome. This review summarizes the immunomodulatory and anti-inflammatory properties of MSC in the context of islet transplantation. INTRODUCTION Human pancreatic islet transplantation through portal vein infusion, is a current clinical beta-cell replacement therapy to treat patients with advanced Type I Diabetes (T1D). However, live-long immunosuppression, difficulties to achieve long-term islet graft function and insulin independence as well as the shortage of suitable pancreata from heart-beating brain-dead donors for islet isolation, are still important limitations for ongoing allo-transplantation programs. Pig islet xenotransplantation is a promising alternative to overcome the bottleneck of islet availability for the treatment of T1D. However, clinical application of pig to human islet transplantation will depend on genetic engineering of pigs to overcome immune barriers and to reduce risks of pathogen infection of porcine viruses (1). Recently, significant progress has been achieved with the transplantation of pig organs presenting several genomic modifications to prevent hyperacute rejection (2,3) and cellular immune responses (4,5). In immunosuppressed nonhuman primates, long-term control of diabetes by the transplantation of adult porcine islets had been successfully achieved (6,7). Other strategies to protect porcine islets from the host immune system include islet encapsulation in semi-permeable hydrogel, such as alginate (8) functionalized by bioactive ligands or by poly(ethylene glycol) (PEG) derivatives (9,10). Mesenchymal stem cells (MSCs) are multipotent cells and play an important role in tissue repair, angiogenesis and their immunomodulatory action on immune cells have been widely studied (11,12). In the field of islet transplantation MSC are investigated for the improvement of islet function and graft survival after transplantation. Numerous studies of co-culture and co-transplantation with MSCs indicate a functional support. However, due to variable transplantation settings and origins of MSCs the immunomodulatory role, as well as their ability to reduce inflammatory processes in vivo remains controversial. This review summarizes the immunomodulatory and anti-inflammatory properties of MSC in the context of islet transplantation and evokes some of the current challenges of islet xenotransplantation. MULTIPOTENT MESENCHYMAL STROMAL CELLS (MSCs), ALSO CALLED MESENCHYMAL STEM CELLS Multipotent mesenchymal stromal cells (MSCs) are selfrenewing multipotential progenitor cells, differentiating along the osteogenic, chondrogenic and adipogenic lineages (13). MSCs have first been isolated from the bone marrow over 50 years ago and bone marrow-derived MSCs (BM-MSCs) still represent the most conventional source. A variety of other tissues also contain MSCs, including adipose tissue, umbilical cord blood, Wharton's jelly, amniotic fluid, endometrium, skin and skeletal muscle (14)(15)(16)(17)(18)(19)(20)(21)(22). It remains unknown which source is most suitable for the clinical use in the context of islet cell transplantation and further research is needed concerning this matter. To facilitate isolation and expansion as well as to standardize characterization, the International Society for Cellular Therapy (ISCT) proposed three minimal criteria for defining mesenchymal stem cells. First, MSCs must

be plastic-adherent when maintained in standard culture conditions. Second, MSCs must express CD105, CD73 and CD90, and lack expression of CD45, CD34, CD14 or CD11b, CD79alpha or CD19 and HLA-DR surface molecules. Third, MSCs must differentiate to osteoblasts, adipocytes and chondroblasts in vitro (23). MSCs have been shown to perform various beneficial functions, making them highly interesting for application in cell-based therapy, especially also for islet transplantation. MSCs SUSTAIN ANGIOGENESIS One major limitation of islet graft survival is a delayed revascularization after transplantation. After isolation, islet cells are cut off from their oxygenation via micro-vascularization and are temporarily dependent after transplantation on diffusion of nutrients and oxygen in order to ensure survival. Neovascularization is finalized after approximately two weeks, however with a lower capillary density and a significantly reduced perfusion compared to islets prior to transplantation (24). Further remodeling takes up to another three months (16). Several studies have shown a pro-angiogenic potential of MSCs. MSCs promote angiogenesis through expression and release of different pro-angiogenic cytokines, including vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF), transforming growth factor beta (TGF-b), as well as annexin 1 (ANXA1), matrix metalloproteinase (MMP) and Angiopoietin-1 (Ang-1) (16). The impact of VEGF is contentious, showing not only beneficial proangiogenic but also damaging proinflammatory effects (16). MSCs seem to keep a balance through its valuable anti-inflammatory property, discussed later. Kinnaird et al. demonstrated further that co-culturing islets with MSCs enhanced neovascularization of islets through promotion of proliferation and migration of endothelial and smooth muscle cells (25). Also, other studies showed that co-transplantation of islets with MSCs improves graft survival and function by increased neovascularization, shortening the post-transplantation ischemia period (26)(27)(28). IMMUNOMODULATORY PROPERTIES OF MSCs MSCs do not express co-stimulatory molecules that activate the immune system, such as CD40, CD80 or CD86 (29). Originally, it was thought that MSCs express only low or no human leukocyte antigen (HLA) class I and II molecules. It has since then been demonstrated that MSCs, like all somatic tissues, express MHC class I molecules constitutively and have the ability to express MHC class II when exposed to inflammatory cues such as interferon-g (30). In vitro studies showed that, attracted by a number of complement proteins, growth factors, proinflammatory cytokines and chemokines, MSCs migrate towards sites of inflammation supporting the hypothesis that MSCs possess anti-inflammatory properties (17,31,32). MSCs can express potent inhibitory molecules of both, innate and adaptive immune effectors (33), however, after transplantation, this may not allow to circumvent acquired alloimmunization, as observed in human trials (30). Nevertheless, immediate events such as acute toxicity associated with the administration of MSCs have not been described (34,35). MSCs have also shown to exert an immunomodulatory effect through phenotype-alteration of different immune cells, including dendritic cells (DC), T-and B-cells, as well as natural killer cells (NK cells) (36). Several authors have described an inhibitory effect of MSCs on immune cell proliferation, generating an immunosuppressive local milieu (16,(37)(38)(39)(40). Research has further shown that MSCs induce modifications of the adaptive immune system, notably T-cells, entailing T-cell anergy. MSCs act on T-cells through physically hindering contact with antigen-presenting cells (APCs) (41) or by an indirect suppression of T-cell activation via MSCs by hindering the maturation of DCs through cell-to-cell contact. These semi-mature DCs possess a tolerogenic phenotype, thus restraining T-cell activation (42). Also, MSCs are able to inhibit T-cell reactivity through the downregulation of proinflammatory cytokines (37,43) and to escape cytotoxic T-cell-mediated apoptosis (44,45). Importantly, MSCs inhibit T-lymphocyte proliferation through soluble factors, such as TGF-b1 and HGF (41,46) or nitric oxide (47). TGF-b1 plays a well-documented role in MSCs immunomodulation, including a role in regulatory T cell (Treg) induction and/or expansion (48)(49)(50). MSCs promote the expression of regulatory T-cells (Treg) (43,51). Early studies showed that stable islet allograft function in cynomolgus monkey was associated with increased numbers of regulatory T-cells in peripheral blood (43). Further, when co-transplanted with allogeneic islets in diabetic cynomolgus monkeys, MSCs derived from islet recipient were more efficient to prolong islet survival, when compared with 3rd party MSCs or islet derived MSCs from the donor. Using recipient-derived MSCs, they observed decreased number of memory T cells, reduced anti-donor T cell proliferation and higher Treg:T cell ratios (52). CO-TRANSPLANTATION OF ISLETS WITH MSCs FROM DIFFERENT SOURCES Various possible sources of MSCs have been tested for cotransplantation with islet cells so far. In murine models, several studies showed improved and prolonged graft survival, function, morphology and revascularization, as well as induction of beta cell proliferation following transplantation of murine islets with autologous (53), syngeneic (27,54,55), allogeneic (27,(56)(57)(58)(59) or xenogeneic MSCs (60). In mice, co-transplantation of autologous MSCs delayed islet allograft rejection and generated a local immuneprivileged site in mice (53). Rackham et al. studied the effects of co-transplantation of syngeneic murine MSCs and islet cells, and observed an improved graft outcome (54). In a subsequent study they examined the underlying factors, suggesting Annexin A1 to be a key contributor to the improved graft function through direct and indirect mechanisms (61). The exact mechanisms remain unclear. Co-transplantation of islets with MSCs in syngeneic rodent models showed better outcomes of islet survival and function than islets transplanted alone (26,27,62). Karaoz et al. described an improved islet function after co-culturing allogenic rat MSCs and islet cells, suggesting paracrine actions through IL-6, TGF-b1, osteopontin and fibronectin (59). Further, allogeneic MSCs resulted in improved islet xenograft survival and function in immune-competent diabetic mice (63). In cynomolgus monkeys, intraportal co-infusion of allogenic MSCs and islets, increased islet engraftment and function, shown by a reduced number of islets necessary to reach normoglycemia (43). Co-encapsulation studies using islets versus islets and MSC also showed beneficial effects on islet function (64)(65)(66). Intraperitoneally syngeneic transplantation of co-encapsulated islets and MSCs showed significantly lower glycaemia compared to islets encapsulated alone. By week 6, 71% of mice transplanted with islets and MSCs were cured, whereas only 16% of the islets-alone group was cured at that point. Interestingly, islet area in recovered capsules was significantly higher when co-encapsulated with MSCs suggesting that MSCs promote survival of islet cells independently from its effects on revascularization. In this study co-encapsulation of islets with MSC did not inhibit pericapsular fibrotic overgrowth (PFO), suggesting that MSCs have no influence on the inflammatory process that causes fibrotic overgrowth (64). PFO is an inflammatory host reaction, induced through the leakage of antigens from semi-permeable microcapsules, that severely impairs islet viability and graft function. However, in a mouse model of islet allotransplantation, co-encapsulation of MSCs (stimulated or not with a cocktail of pro-inflammatory cytokines) with islets in alginate microcapsules, prevented pericapsular fibrotic overgrowth (PFO) compared to islets encapsulated alone (66). Further mice receiving islets coencapsulated with stimulated and unstimulated MSC achieved higher percentages of normoglycemic mice (100% versus 71.4%, respectively) compared to mice transplanted with islets encapsulated alone (9.1%). Similarly, in vitro rat MSCs and rat islet cells when co-encapsulated in a ligandfunctionalized polyethylene glycol (PEG) hydrogel (67) led to a doubling of the stimulation index compared to islets encapsulated alone. Co-encapsulation of islet cells and MSCs in addition with cell adhesion peptides led to a significant sevenfold increase of the stimulation index compared to islets encapsulated alone (67). Human islets co-cultured in direct cell contact with human MSCs compared to islets co-cultured with MSCs but without cell-to-cell contact, displayed significantly enhanced insulin secretion in the presence of cell-to-cell contact. This effect was identified to be dependent on N-cadherin interaction, since impeding N-cadherin interaction with antibodies led to a reversal of the enhanced insulin secretion. Additionally, mice transplanted intraperitoneally with human islets coencapsulated with MSCs in hydrogel microspheres, composed of calcium alginate and covalently crosslinked to polyethylene glycol showed significantly lower blood glucose levels and prolonged islet graft survival (57). Others have shown that improved graft function correlates with enhanced revascularization of islets transplanted under the kidney capsule (68)(69)(70). Accordingly, research findings revealed significantly higher apoptosis rates in islet cells cultured without MSCs (16). Taken altogether, these findings support the hypothesis that co-transplantation of MSCs and islet cells is beneficial and that MSCs are useful for future therapeutic applications. IMPROVED NEONATAL PORCINE ISLET FUNCTION, SURVIVAL AND GRAFT OUTCOME The main disadvantage of neonatal or juvenile porcine islets, also called porcine pancreatic islet cell clusters (ICCs), is their lack of integrity and maturity. ICCs are obtained by in vitro digestion of neonatal or juvenile pig pancreas with subsequent short time culture in a specific maturation media to increase the beta cell mass for transplantation ( Figure 1) (71). Also, porcine pancreatic ICC co-transplanted with human MSC into immune deficient diabetic mice reached normoglycemia significantly earlier than mice transplanted with ICC alone (60). He et al. demonstrated an improved and accelerated development of ICCs in diabetic rhesus monkeys after cotransplantation with allogeneic simian MSCs into diabetic rhesus monkeys (72). Additionally, the group described an enhanced expression of genes implicated in the development of endocrine cells and insulin and further demonstrated enhanced expression and activation of PDGFR-a in neonatal islets through MSCs confirming earlier studies demonstrating the capability of PDGFR-a to stimulate beta-cell proliferation (73). Further, He et al. suggest an inhibition of the Notch1 signaling provoked by PDGFR-a, leading to an improved islet development and maturation. It is known that Notch1 downregulates the expression of several genes and transcription factors implicated in the development of endocrine cells and insulin (72,74). Juvenile porcine exocrine pancreas-derived MSCs (pMSCs) cocultured with direct cell to cell contact of juvenile porcine ICCs significantly enhanced beta-cell function, suggesting that cell signaling via adhesion molecules are important (57,65). However, co-encapsulation of such ICCs with pMSCs do not effectively prevent PFO and graft survival was rapidly impaired after transplantation of capsules in immunocompetent mice. Therefore, further research is required to enable efficient longterm survival of encapsulated juvenile porcine islets. Possible approaches being evaluation of modified alginate chemical composition (75) or the use of different anti-fibrotic polymers (65). IMMUNOMODULATION STRATEGIES TO INCREASE XENOGRAFT SURVIVAL To overcome the immunological barrier between pig and humans, genetic modifications have been performed in pig strains to reduce immunogenicity of organs and tissue. The first genetically modified pig, i.e. with a single human transgene for a complement regulating gene (hDAF), allowed survival of pig organs in immunosuppressed non-human primates for several months. Since then, genetic engineering, using CRISPR-CAS9, allowed cloning of animals with additional genetic modifications. Today, pigs with over 10 genetic modifications, both, deletions of pig antigens and inclusions of human transgenes are under investigation for transplantation purposes (76). Immunosuppressive regimens are still necessary but recently heart transplantation from genetically modified pigs [a1,3-galactosyltransferase-knockout and knock in human CD46 (77,78) and thrombomodulin (79)] to a non-human primate (baboon) reached long term survival of 195 days (80). Furthermore, immunoregulatory therapies (tolerance induction) using Treg-based therapeutic approaches are under investigation. Regulatory T cells (Tregs) are immune-suppressive T cells that are critical for the maintenance of tolerance in vivo (81). Chimeric antigen receptors (CARs) are synthetic fusion proteins that have been developed to genetically modify T cells in order to create a specificity toward designated antigens. The application of the CAR technology to Tregs, may allow to reduce immune responses for solid organ and cell transplantation. CAR Treg therapies are currently developed using genetic modifications for xenogenic pig antigens with the aim to improve graft acceptance of xenotransplanted tissue i.e. porcine islets. This might be achieved through infusion of ex vivo expansion of donor-specific Tregs (55,82). CAR-Tregs technology started with a study of MacDonald and colleagues which successfully transduced human Tregs with a CAR targeting the human leukocyte antigen (HLA) class I-A2 (A2-CAR) (83). In a human skin xenograft transplant model, HLA class I-A2 specific CAR-Tregs alleviated rejection of skin transplants (84). Since co-transplantation of autologous MSC delayed islet allograft rejection, it is possible that genetically modified, MSC, could be exploited as a target cell in porcine ICC xenografts to foster islet function and to increase trafficking and activation of adoptive transferred CAR-Treg cells to increase tolerance toward pig ICC xenografts. An additional challenge for islet transplantation is the precise quantification of beta cell mass (BCM) or endocrine cell mass (ECM) in vivo. Imaging the progressive loss of beta cells following islet transplantation should allow the development of individualized therapies for the management of patients post-transplant (85). Recently, a suitable biomarker for beta cell quantification, the dipeptidyl aminopeptidase-like protein 6 (DPP6) has been identified as a promising target for human BCM

imaging in healthy individuals as well as diabetic patients (86,87). First imaging and biodistribution studies using SPECT/CT and radiolabeled high-affinity camelid single-domain antibody (nanobody) directed specifically against human DPP6, allowed to visualize transplanted DPP6-expressing Kelly neuroblastoma cells or insulinproducing human EndoC-bH1 cells in immunodeficient mice. Importantly, neonatal pig islets expressing near-infrared fluorescent protein (iRFP) were non-invasively monitored through multispectral optoacoustic tomography (MSOT). MSOT signals, obtained after islet transplantation under the kidney capsule in mice, and obtained after subcutaneous and intramuscular islet transplantation in pigs, allowed to distinguish graft mass changes (88). Such reporter geneexpressing islets are also promising tools to evaluate the efficacy of newly developed biomaterials for encapsulation and transplantation of porcine islets. CONCLUSION Diabetes is a worldwide disease, affecting over 40 million people and putting an important burden on the healthcare system. Exogenous insulin represents the predominant treatment modality for type 1 diabetes, but is associated with long-term complications. Islet cell transplantation is a highly promising approach for treating type 1 diabetes aiming at reestablishing a physiological insulin secretion through replacement of the endocrine tissue. Despite improving preclinical and clinical results over the past decades, the need for immunosuppression and donor shortage limits the clinical application of this procedure. The implementation of porcine pancreatic ICCs with porcine MSC might represent a promising alternative to help to overcome the problem of donor shortage; especially neonatal or juvenile pigs providing high islet yields. Encapsulation techniques could resolve the need for immunosuppression, shielding the islets from immune attacks while still enabling the exchange of oxygen, insulin and nutrients. Yet, delayed and impaired graft outcome due to immature islet cells and the formation of pericapsular fibrosis continue to severely limit the clinical application of encapsulated islet transplantation. AUTHOR CONTRIBUTIONS NK: designed and wrote the manuscript; LB contributed to manuscript writing; LB and BE: critically revised the manuscript; CGG: designed, contributed to manuscript writing; supervised the writing of the manuscript. All authors contributed to the article and approved the submitted version. ACKNOWLEDGMENTS This review article has been supported by the Insuleman Foundation, the Child Foundation and the Foundation La Colombe. Bacteria drug resistance profile affects knee and hip periprosthetic joint infection outcome with debridement, antibiotics and implant retention Background Evaluate the effect of bacteria drug resistance profile on the success rates of debridement, antibiotics and implant retention. Methods All early acute periprosthetic infections in hip and knee arthroplasties treated with DAIR at our institution over the period from 2011 to 2015 were retrospectively analyzed. The success rate was evaluated according to the type of organism identified in culture: multidrug-sensitive (MSB), methicillin-resistant Staphylococcus aureus (MRSA), multidrug-resistant Gram-negative bacteria (MRB) and according to other risk factors for treatment failure. The data were analyzed using univariate and multivariate statistics. Results Fifty-seven patients were analyzed; there were 37 in the multidrug-sensitive bacteria (MSB) group, 11 in the methicillin-resistant Staphylococcus aureus (MRSA) group and 9 in the other multidrug-resistant Gram-negative bacteria (MRB) group. There was a statistically significant difference (p < 0.05) in the treatment failure rate among the three groups: 8.3% for the MSB group, 18.2% for the MRSA group and 55.6% for the MRB group (p = 0.005). Among the other risk factors for treatment failure, the presence of inflammatory arthritis presented a failure rate of 45.1 (p < 0.05). Conclusion DAIR showed a good success rate in cases of early acute infection by multidrug-sensitive bacteria. In the presence of infection by multidrug-resistant bacteria or association with rheumatic diseases the treatment failure rate was higher and other surgical options should be considered in this specific population. The MRSA group showed intermediate results between MSB and MRB and should be carefully evaluated. Background Periprosthetic joint infection (PJI) is one of the most demanding and challenging complications for orthopedic surgeons [1]. It has a low incidence in total hip (THA) and total knee (TKA) arthroplasty, ranging between 0.2 and 2% of primary procedures [2,3]. Despite its low incidence, PJI is the third leading cause of revision THA in the United States, representing 14.8% of all revision surgeries [4]. The ageing population has led to an increase number of arthroplasties performed worldwide, and the amount of cases of PJI is estimated to grow significantly in the coming decades [3]. Kurtz et al. [5] estimated that by the year 2030, there will be a 613% increase in the demand for TKA and a 174% increase for THA as a result of degenerative processes. The most appropriate treatment for PJI is still debated. Among the many options for treatment; Debridement, Antibiotics and Implant Retention is most indicated in cases of early acute an hematogenic acute infections. It has the advance of being little aggressive while maintaining the implants, though its success rate is debatable in the literature [6,7]. In acute cases where a mature biofilm has not yet been formed on the surface of the prosthetic material, the most used treatment is DAIR with exchange of the modular components [6,8]. The factors that contribute to poor prognosis in the treatment of PJI include age, gender, ethnicity, diabetes, rheumatic disease, obesity, malnutrition and multidrug-resistant infections [2,[9][10][11]. The emergence of multidrug-resistant bacteria in hospital environments is a phenomenon that accompanies the development of new antimicrobials and their indiscriminate use. Some studies have demonstrated that infections by these agents represent an important risk factor for PJI treatment failure [2]. The universe of drug resistant bacteria is very complex with a miscellaneous of pathogens with resistance profiles ranging to one or few commonly used antibiotics, in particular methicillinresistant Staphylococcus aureus (MRSA) and vancomycinresistant Enterococcus spp (VRE), to multi-drug resistant bacteria, such as multidrug-resistant Pseudomonas and Acinetobacter [2,11,12]. It is important to understand the difference between such organisms, as the success rates deteriorate with the increase of drug resistance profile [2]. Some authors [12] have suggested that infections in TKA by multidrug-resistant microorganisms other than MRSA have worse results in the two-stage exchange of modular components, with an increase incidence of prosthesis failure and even an increase in limb amputation. However, to the best of our knowledge there are no studies that compare the success rate of DAIR for acute infections caused by resistant Gram-negative microorganisms or other microorganisms. The objective of the present study is to analyze the influence of the infecting organism in the success rate of treatment in acute periprosthetic knee and hip infections treated with DAIR. Our hypothesis is that both MRSA infections and infections by multidrug-resistant Gramnegative bacteria will have poorer results than infections caused by multi-sensitive bacteria in the matter of prosthesis retention. Methods A retrospective study was carried out using data collected prospectively for all acute PJI of total knee and hip arthroplasty from our institution over the period from 2011 to 2015. We included all cases of acute infection (up to 3 weeks after the index surgery) that meet the criteria of PJI defined by the Musculoskeletal Infection Society [13] (Table 1) and had identification of the infecting organism in cultures from the initial diagnostic aspiration or intraoperative samples. Cases of partial arthroplasty, sub-acute or chronic infections (more than 3 weeks after arthroplasty), cases with negative culture results or previous infections in the affected joint were not included in this study. Cases where the treatment used was not DAIR were also excluded from the analysis. All specimens collected for culture were sent to the microbiology laboratory in sterile containers with thioglycolate broth, and the joint fluid samples collected by aspiration were sent for analysis in blood culture bottles. The following data were collected and evaluated: gender, age, joint affected, comorbidities (smoking, diabetes mellitus and inflammatory diseases), time of post-arthroplasty infection, microorganism responsible for the infection and its sensitivity profile and the success or failure of the initial treatment. Success was defined as implant retention without the need for new surgical procedures after the first DAIR procedure. Patients were divided into three groups according to the bacteria that caused the infection: multidrug-sensitive bacteria (Group 1), methicillin-resistant Staphylococcus aureus (MRSA) (Group 2) and Gram negative multidrugresistant bacteria (Group 3), as to compare success rates among them. All patients received antibiotic prophylaxis 1 h previously to skin incision. Teicoplanin (400 mg) and amikacin (500 mg) were the antibiotics of choice when preoperative culture was not available, which occurred in 52 of the 57 cases (between 2011 and 2015 it was not a routine in our institution to perform pre-operative aspiration for all cases of suspected PJI). In the 5 cases that had a positive preoperative culture obtained by joint aspiration, the antibiotics were target to the specific antibiotic sensitivity profile of the pathogen. None of these 5 cases were multi-drug resistant bacteria. After surgery patients received at least 2 weeks of intravenous antibiotic (teicoplanin 400 mg 12/12 h in the first 48 h, then 400 mg once a dayfor gram positive bacteriaand amikacin 500 mg once a dayfor gram negative bacteria) followed by 4 weeks of oral antibiotics, if it was possible based on bacteria resistance profile. Extensive debridement was performed always in the presence of the assistant surgeon supervision. The irrigation was carried out with at least 10 l of sodium chloride 0,9% solution, no local antimicrobial agents were utilized during or after irrigation. The modular components were exchanged (polyethylene liner in total knee arthroplasty and polyethylene liner and femoral head in total hip arthroplasty) and the implant fixation stability were tested. After the debridement an irrigation, and prior to implantation of the new liner and femoral head, the surgical team exchanged gloves and re-draping was performed. At least five samples among periprosthetic tissue, bone and joint fluid were collected during the procedure and prior to debridement and irrigation. In all cases we used drains that were removed 24 to 48 h after surgery depending upon the drain out. The mean surgical time was 75 min (ranging between 50 to 90 min). Normally distributed variables were described as mean ± standard deviation, and not normally distributed ones as median (interquartile range). For univariate analyses, the Kruskal-Wallis test for comparison of independent samples was used for continuous variables, and Fisher's exact test was used for categorical variables. To evaluate the independent determinants of the outcome 'treatment success or failure', multivariate logistic regression analysis was used. P < 0.05 was considered statistically significant. The sample size was not calculated since all available patients were included in the study. The statistical software SPSS 22 (IBM Corp.) was used for the analyses. Multivariate logistic regression was performed to investigate the effects of gender, joint affected, diabetes, smoking, inflammatory disease and type of bacteria in the outcome 'treatment success' ( Table 2). The model was statistically significant, with χ2 = 024.938 and p < 0.001, explaining 67.7% (Nagelkerke R 2 ) of the treatment success variance, and correctly classified 89.5% of the cases. Results Data were available from 57 patients subjected to DAIR and exchange of modular components for acute hip (31/ 57) or knee (26/57) periprosthetic infection: 37 in the multidrug-sensitive bacteria group, 11 in the methicillinresistant Staphylococcus aureus group and 9 in the Gram negative multidrug-resistant bacteria group. The characteristics of the groups are summarized in Table 3. There was no statistically significant difference in the distribution of patient's risk factors for PJI between the groups. All patients were submitted to DAIR between 10 and 21 days after the index surgery. There was no statistical difference in outcome regarding the time between the index surgery and DAIR procedure (p = 0,108). The median follow-up time was 5.0 (range 3.0-7.0) years. The multidrug-sensitive bacteria group (37 patients) contained 70% of Gram positive (26/37), which was composed most frequently by Staphylococcus aureus (19/26) followed by Staphylococcus epidermides (7/26). Gram negative multidrug-sensitive bacteria (11/37) was composed by Escherichia coli (4/11), Acinetobacter (2/ 11), Enterobacter cloacae (2/11) and others. There was no case of polymicrobial infection diagnosed during the first DAIR procedure. It was not necessary to adjust initial antibiotic therapy (teicoplamin and amikacin) based in the results of intra operative cultures in the multidrug-sensitive bacteria and MRSA groups. In the multidrug-resistant bacteria group, it was necessary to adjust the antibiotic in 6/9 (66%) cases regarding bacteria resistance. Of these 6 patients, 3 had failure treatment as outcome (2 Klebsiella pneumoniae and 1 Escherichia coli) which had to be changed accordingly (meropenem and tigecycline, respective). Therapeutic success after DAIR with the exchange of modular components

was different between groups: treatment failure was 8.3% in the multidrug-sensitive bacteria group, 18.2% in the methicillin-resistant Staphylococcus aureus group and 55.6% in the Gram negative multidrug-resistant bacteria group (p = 0.005) ( Table 4). Of the 10 cases that failed initial treatment, 5 had a successful second DAIR procedure without the need for revision or replacement of metal implants (4 by multidrug-resistant bacteria and 1 multidrug-sensitive), 3 underwent two-stage revision with the placement of a cement spacer between surgeries (one by MRSA, one by multidrug-resistant Escherichia coli and the other by methicillin-sensitive Staphylococcus aureus), and 2 cases of TKA progressed to transfemoral amputation due to infection after a failed two-stage revision. About the two patients that evolved to amputation, one had an infection by MRSA and the other initially had an infection by a multidrug-sensitive Escherichia coli, which evolved to a polymicrobial infection after many attempts of debridement. None of these cases presented new recurrences of the infection until our last control. The significant variables of the model were rheumatic disease (p = 0.04), infection by multidrug-sensitive bacteria (p = 0.034), infection by methicillin-resistant Staphylococcus aureus (p = 0.01) and infection by multidrug-resistant Gram-negative bacteria (p = 0.048). The odds ratio for infection failure due to rheumatic disease was 45.1, related to methicillin-resistant Staphylococcus aureus was 58.5 and related to multidrug-resistant gram-negative bacteria was 49.9. Discussion Our study demonstrated that the general incidence of infection recurrence in the treatment with DAIR was 18%. Controlling for other variables, only two risk factors among all evaluated were significant: multidrugresistant bacterial infection (MRSA or Gram negative) and patients with previous rheumatic diseases. The treatment of periprosthetic hip and knee infections is a challenge for both the orthopedist and the patient and is one of the most feared complications after arthroplasty [2]. In acute infections, the main objective is to avoid biofilm formation around the prosthetic material through a careful procedure that includes irrigation, debridement, the exchange of modular components and broad-spectrum antibiotic therapy [8]. The prognosis in these cases can be discouraging when associated with risk factors such as multidrugresistant bacteria, rheumatic diseases, diabetes and smoking [2,9,10,12,14,15]. The recurrence of infection after the initial treatment of a periprosthetic joint infection with DAIR varies in the literature between 14 and 69% [10,16,17]. The highest number of cases study, as described by Cochran et al. [10], presented a reinfection rate of 28,2% in the first year and 43,2% after 6 years of the surgical procedure with DAIR. The reinfection rate of 18% (10/57) found in our study agrees with the values described in the literature and is acceptable for this type of treatment. Regarding the type of microorganism causing the infection, our study showed a treatment failure rate of only 8.3% (3/37) when the bacterium found in the culture was multidrug-sensitive. In the MRSA group, we observed a failure rate of 18.2% (2/11). The odds ratio for recurrence of infection in the MRSA group, in comparison with the other groups, was 58.5 (p = 0.01). More outstanding, the group of patients with Gram negative multidrug-resistant bacteria had a failure rate of 55.6% (5/9), with an odds ratio of 49.9 (p = 0.048). Most studies in the literature corroborate our findings pointing to multidrug-resistant bacteria as a factor of poor prognosis in the success of PJI treatment [18,19]. The reasons for worst outcomes in resistant organism are multifactorial, such as limited choose of effective antibiotics, more aggressive biological behavior related to biofilm production and greater virulence [2,20]. Bradbury et al. [18] found an infection recurrence rate of 84% after irrigation and debridement treatment when the bacteria found in the cultures was MRSA compared to the overall failure rate of 18%. Based on these findings, some authors have gone so far as to recommend a two-stage revision in acute infection caused by MRSA [18,19]. In contrast, our findings demonstrate a reasonable chance of success in MRSA compared to multidrug-resistant Gram-negative bacteria, which raise a question regarding the role DAIR in these patients. Hsieh et al. [21] compared the infection recurrence rate after irrigation and debridement of Gram-positive bacteria with Gram-negative bacteria and obtained a success rate of 47% vs. 27%, respectively. Even when compared to other types of initial surgical procedures, non-MRSA multidrug-resistant bacteria tend to have worse outcomes than MRSA. Vaso et al. showed a failure rate of 33% for non-MRSA multidrug-resistant bacteria versus 10% for MRSA in two-stage revision knee arthroplasty. However, some authors did not find a significant difference in results when comparing multidrugsensitive and multidrug-resistant bacteria [22,23]. Most studies, however, have used small sample sizes, which hinders further conclusions. Regarding antibiotic therapy management, in only 6/ 57 cases (all multidrug resistance bacteria) the initial antibiotic protocol was not suitable based on bacteria resistance profile. Half of those cases (3/6) had failed DAIR procedures. We analyze the risk and benefit of expanding the spectrum of initial antibiotic therapy protocol for PJI, but due to the concern in increasing bacteria resistance and the fact that nowadays all patients with suspected PJI are submitted to joint aspiration previously to any surgical procedure (with very few incidence of negative cultures and dry aspiration), we find it unsuitable to modify our existent protocol. Another risk factor that was statistically significant in our study was the presence of associated rheumatic disease with an odds ratio for treatment failure of 45.1. It is already well established in the literature that rheumatic diseases is an independent risk factor for PJI [24]. Bongartz et al. [25] demonstrated an increased risk of PJI in patients with rheumatic disease, with an odds ratio of 4.08 (95% CI 1.35-12.33). The present study has limitations regarding its retrospective design and sample size. Despite the increase in bacterial drug profile resistance in the last decade, the low incidence of PJI, particularly related to multidrugresistant organism, results in the absence of large studies involving these types of microorganisms in medical literature. The prevention and treatment of acute prosthetic infection is extremely important today because of its severity, difficulty to treat, and its large contribution to the total number of failed joint replacements. As the number of joint replacements is expected to rise, this problem will only become more relevant in the next years. It is important to identify multi-drug resistant bacteria and inflammatory disease as factors of poor prognosis in the treatment of acute infection with DAIR. These data may warrant future clinical trials comparing DAIR and revision arthroplasty in this high-risk situation, and at present allow for a more informed treatment decision of surgeons, infection specialists and patients. Conclusion Infection by MRB, MRSA and inflammatory arthritis were independent risk factors for failure after treatment with DAIR for acute hip and knee periprosthetic infections. In the presence of these organisms, DAIR should be consider with caution. Cognition and Its Shaping Effect on Sexual Conflict: Integrating Biology and Psychology While genetic variation is of crucial importance for organisms to be able to adapt to their ever-changing environments over generations, cognitive processes can serve the same purpose by acting at shorter time scales. Cognition, and its resulting behaviour, allows animals to display flexible, fast and reversible responses that, without implying a genetic change, are crucial for adaptation and survival. In the research field on sexual conflict, where studies focus on male and female mating strategies that increase the individual’s reproductive fitness while forcing a cost on the partner, the role that cognition may play in how such strategies can be optimised has been widely overlooked. However, a careful analysis of behavioural studies shows that animals can develop and change their responses depending on what they perceive as well as on what they can predict from their experience, which can be of prime importance for optimising their reproductive fitness. As will be reviewed here, largely psychological processes, such as perception, memory, learning and decision-making, can not only modulate sexual conflict, but can also have a big impact on the reproductive success of a given individual. This review highlights the need for a more integrative view of sexual conflict where cognitive processes are also considered as a fundamental part of an animal’s adaptive mating response. INTRODUCTION In the case of sexual reproduction,where both males and females are required for it to be successful, optimal fitness is generally not reached simultaneously by both sexes (Parker, 2006). This has been argued to be due to the differential investment in gamete production that is faced by males and females, as egg production is more costly than sperm production (e.g., Bateman, 1948;Schärer et al., 2012, but see also e.g., Dewsbury, 1982;Janicke et al., 2016). As a consequence of such initial differential investment, reproductive fitness for females is said to rely on their capacity to produce eggs, whereas for males it is argued to depend on their access to females (e.g., see Arnold, 1994 for a review). It is such differences in gamete production and subsequent interests on the two sides that are taken as the starting point of what is referred to as sexual conflict. Before addressing the main topic-the modulating role of cognition on responses that are related to sexual conflict-we would like to briefly point out some of the problems with the definition of sexual conflict. On the one hand, sexual conflict has been defined as "a conflict between the evolutionary interests of individuals of the two sexes" (Parker, 1979, p. 124). However, as we have argued previously (Alvarez and Koene, 2018), we think we cannot state that males and females have evolutionary interests as such, and that it is more accurate to say that they have individual (albeit not necessarily conscious) interests, i.e., increasing one's benefits with the lowest possible costs, with evolutionary consequences. Just to mention one extreme example, male seed beetles (Coleoptera: Bruchidae) possess spiny genitalia that allow them to anchor themselves to the female while mating (Stutt and Siva-Jothy, 2001) and to transfer proteins, via the wounds that are caused, that increase their male fertilisation success (Hotzy et al., 2012). This is beneficial for males as it helps them securing paternity, but the injuries inflicted by the spines to the female copulatory tract pose a high cost for the females, including a reduced longevity and reproductive success (Stutt and Siva-Jothy, 2001). The conflict arises because such strategies entail a cost for their partner and, as a consequence (and not as an aim), an evolutionary arms race can emerge between males and females as they try to adapt and to counter-adapt to the strategies employed by the other sex (e.g., Arnqvist and Rowe, 2005;Koene, 2012;Rice and Gavrilets, 2014;Alvarez and Koene, 2018). In the case of seed beetles, females have evolved thicker tracts so they can resist to the harm induced by their mating partner (Rönn et al., 2007). On the other hand, if we stick to the definition of sexual conflict in terms of evolutionary interests, sexual conflict can then be seen as a cooperation between males and females to increase their net reproductive fitness in the long run instead of as a conflict (e.g., Cordero and Eberhard, 2003). For example, female seed beetles mating with males who are able to induce larger injuries (i.e., that further increase their paternity success) will sire sons with that same capacity, increasing the long-term fitness of that female, and males mating with females with thicker copulatory tracks will sire daughters with increased survival likelihood and thus also potentially increased maternal success. Thus, depending on the perspective that is taken, whether we should talk about conflict or collaboration to reach common goals becomes debatable. In fact, a clear and consensual definition of sexual conflict is still lacking (see Tregenza et al., 2006) and many studies on the topic do not provide a clear definition of sexual conflict. Having all this in mind, and being aware that our definition may still need to be refined, we have defined sexual conflict, at an individual level, as the disagreement over investment that ensues because males and females adopt or develop strategies that are only aimed to increase their own fitness but that impose a cost on the mating partner (see Alvarez and Koene, 2018). As stated by Kokko and Jennions (2014), "sexual conflict can occur over every facet of breeding". At the pre-copulatory level, it starts with the investment (or not) in searching for a

mate, deciding to accept (or reject) a potential mate, the number of matings that take place and how many gametes are transferred. At the postcopulatory level, conflict can exist over additional matings (with the same or different individuals), induced physiological effects or physical harm, the number of and investment in offspring, and the amount of parental investment. When we focus on the costs, i.e., what we consider the actual root of the conflict, that are faced by both males and females when mating, there is indeed a great variety of studies that provide clear evidence of both sexual conflict and of the sophistication of the mechanisms involved (Arnqvist and Rowe, 2005;Koene, 2012;Rice and Gavrilets, 2014;Alvarez and Koene, 2018). What is notable is that, regardless of the particular mechanism and species under study, most of the research conducted on sexual conflict has largely focussed on the physiological responses that are involved, as well as on the genetic variations underlying such physiological processes (Chapman et al., 2003;Rice and Gavrilets, 2014). For instance, on the male side, studies analyse how the production and transfer of seminal fluid proteins increase male fitness but decrease female fitness (see e.g., Poiani, 2006;Chapman, 2008;Koene, 2012;Perry et al., 2013). An example of this can be seen in fruit flies, where it has been shown that, among other effects, seminal fluid proteins increase females' investment in egg production (Wigby and Chapman, 2005) and decrease their life expectancy because of toxic effects that these proteins have (Lung et al., 2002). An example from the female side are studies that reveal how (cryptic) female choice affect male fitness as determined by the degree of sperm acceptance (Firman et al., 2017). After all, such "sperm rejection" (via sperm ejection or sperm digestion), is a female driven process that is costly for the male since he has invested in the production of spermatozoa, accessory gland products, mate searching, courtship, and copulation, possibly after investing in competition against other males (see e.g., Dewsbury, 1982;Janicke et al., 2016). However, as it has been increasingly pointed out by researchers of fields like that of evolutionary biology, physiological and genetic responses cannot be fully understood in the absence of the social environment and the cognitive processes that are constantly regulating animals' activity (sensu organic selection by Baldwin, 1896(see e.g., West-Eberhard, 2003Diogo, 2017; for general reviews). Within the particular domain of mating, the modulating role of cognition has become increasingly acknowledged (e.g., Bateson and Healy, 2005;Prum, 2017;Ryan, 2021). Animals need to be able to identify sexual partners as such, to distinguish between receptive and non-receptive conspecifics, to identify their own sexual arousal, to assess the quality of a potential mate and to pursue him or her to achieve successful mating (e.g., Pfaus et al., 2001). Moreover, this evaluation needs to be orchestrated with the constant monitoring of the ever-changing environmental conditions of the given time and space in which mating is about to take place. In other words, mating is importantly affected by cognitive processes that include motivation, perception, learning, memory and decision-making (see Pfaus et al., 2001 for a review on how learning shapes mating in rats) and thus, they cannot be overlooked if we aim to understand it fully (see Bateson and Healy, 2005;Ryan, 2021 for a more general review). Likewise, as we have previously pointed out (Alvarez and Koene, 2018), we think that the current available data on sexual conflict cannot lead to a comprehensive understanding of this topic since it leaves out an important aspect of an animal's life, i.e., its ontogeny, that, as will be further argued, can crucially shape the mating response and, therefore, the outcome of the sexual conflict (i.e., the costs) that is faced at each mating encounter. With the term ontogeny we refer to the individual's experience and how this affects the development of behavioural patterns as well as the display of physiological responses over time. It is thus an adaptive response, to the particular scenario of mating, that is not accounted for by genetic aspects but by flexible and fast responses that are mediated by cognitive processes (Baldwin, 1896;West-Eberhard, 2003;Diogo, 2017). In this regard, it is interesting to note that many authors do refer to essentially psychological processes such as anticipation, which is an animal's prediction based on previously acquired knowledge, as determinants of reproductive investment (see e.g., Cattelan and Pilastro, 2018;Dore et al., 2018;Fuss, 2021), which is directly linked to sexual conflict (Kokko and Jennions, 2014). However, in the research field of sexual conflict, to our understanding, and to the best of our knowledge, a conceptual gap still exists, and we think it is limiting the way in which it is being analysed. Interestingly, although sexual conflict has been suggested as a source for cognitive variation (Cummings, 2018), the way in which cognitive processes may be affecting sexual conflict remains largely unexplored, and at both theoretical and experimental levels, the explanation of mating traits that are related to sexual conflict is mostly given in purely physiological or genetic terms (Arnqvist and Rowe, 2005;Rice and Gavrilets, 2014;Chapman, 2015). We think that the reason for not taking into account cognitive processes and their modulatory role on sexual conflict largely stems from an implicitly-assumed definition of mating as an automatic or innate response that is solely driven by physiological mechanisms that are in turn determined by genetic factors. This implicit theoretical framework, which seems to be shared by the majority of researchers, as inferred from the main reviews on the topic (Arnqvist and Rowe, 2005;Rice and Gavrilets, 2014;Chapman, 2015), leaves out cognitive processes that are fundamental for mating to occur and that, necessarily, have the potential to modulate sexual conflict. We think that there is already evidence showing that cognitive processes exert a big influence on the way sexual conflict mechanisms work, but that they have not been analysed in such terms nor have they been conceptually taken together to widen the frame from which sexual conflict is understood. The aim of this review is therefore, without being exhaustive, to put together available examples that highlight the need for a more integrative view and approach. Most of the examples will highlight how different cognitive processes are intertwined with sexual conflict responses that allow males to increase their chances of fertilising eggs, compared to baseline conditions, which implies a higher investment on the female side, or that determine the degree of sperm acceptance by females (the less they accept, the higher the costs for the male; the more they accept, the higher the costs females may face). As a whole, they illustrate that cognition has a clear impact on sexual conflict and they show that individual ontogeny matters not only when it comes to mating (Pfaus et al., 2001) but also when talking about sexual conflict. Mate Choice and Sexual Conflict: Perception, Memory, Comparison, and Decision-Making Mate choice could be regarded as more related to sexual selection than to sexual conflict, since preferences for a relatively betterquality partner or the rejection of a non-preferred partner constitute just an attempt to maximise one's reproductive fitness. However, as argued by Kokko and Jennions (2014), whenever there is sexual selection, sexual conflict is also present, and mate choice has clear consequences in terms of sexual conflict. For example, both males and females have been found to invest significantly less in their own offspring if they were mated with a non-preferred partner, that they showed preference for in a choice test. Such smaller investment is costly for the partner because it can imply lower survival for the offspring, reducing overall reproductive fitness, and, as argued by Kokko and Jennions (2014), it can be seen as a conflict between the mating partners (over e.g., provisioning of the young). This can be observed in young mice, that have been shown to have a decreased survival rate when their fathers were mated with nonpreferred, compared to preferred, females (Gowaty et al., 2003). Likewise, female canaries that have been exposed to unattractive male songs decrease the allocation of testosterone to their own eggs, which is known to compromise the survival rate of their own offspring (Gil et al., 2004). Mate choice is thus highly related to sexual conflict. Although mate choice has been defined as a cognitive process that starts off with the perception and assessment of a conspecific (e.g., Bateson and Healy, 2005;Ryan et al., 2009;Dougherty, 2020;Ryan, 2021), as pointed out previously, it often seems to be understood as driven by an output from a nervous system that receives a specific amount of stimulation that triggers a sexually receptive response (i.e., a physiological response more than a cognitive one; e.g., Bakker, 1999;Iwasa and Pomiankowski, 1999;Andersson and Simmons, 2006;Kopp et al., 2018). It is true that, within any given species, sensory organs are responsive to specific ranges of stimulation and, as such, it is natural to observe predispositions for particular stimuli that are common to the majority of the individuals of the opposite sex (e.g., natural preference for red bellies in female guppies, Kodric-Brown, 1993; see also Ryan, 2021 for a review). However, this does not imply that perceptual preferences are purely innate nor fixed (e.g., see Weary et al., 1993). Experience with less or more similar phenotypes is important for subjects to learn to discriminate among them. For example, although pheromones can be seen as chemicals that trigger sexual arousal in a mechanistic way, studies in male rats have shown that these animals need to learn about them. Sexually naïve male rats do not display a preferential distinction between the odours from receptive or non-receptive females; even when they are experienced, they also need to learn not to attempt mating with a non-receptive female, despite the fact that she is not producing pheromones indicative of sexual receptivity (reviewed in Pfaus et al., 2001; see also e.g., Dukas, 2005 for another example in fruit flies). Just as experience (i.e., memory) can shape the response to chemicals, it can also shape perceptual preferences for certain phenotypes (see Witte and Nöbel, 2006 for a review). An example of this can be found in butterflies, where sexually naïve females prefer mating with wild-type males that have two eyespots on their wings rather than with males with what is called "enhanced ornamentation" (i.e., four eyespots), unless they had been exposed to the ornamented males before. In the latter case they showed a higher preference for enhanced ornamentation to the detriment of wild type males (Westerman et al., 2012). Simple exposure to a phenotype does not only alter mate preferences, but also has clear consequences in terms of sexual conflict. For example, in the Pacific field cricket (Teleogryllus oceanicus), females that have experience with an unattractive male (defined by the type of songs he produces) will show a higher predisposition to mate with a second unattractive male and, what is crucial in terms of sexual conflict, to retain the spermatophore of the latter for longer (Rebar et al., 2011). In other words, mere exposure to a type of (unattractive) male results into acceptance of a higher amount of sperm from that type of male, increasing females' costs, thus shaping sexual conflict. In wolf spiders (Schizocosa uetzi), the consequences of experience-based mate preferences can be even more drastic. In an experiment where the tibia of males' forelegs was painted either brown or black, females that had never been exposed to any of the two types did not show any preference for either phenotype. On the contrary, if they had been exposed to just one of them before reaching maturity, not only did they show an increased likelihood to mate with the already known phenotype, but they also displayed an increased probability to cannibalise the male with the unfamiliar phenotype, clearly illustrating the conflict caused by sexual cannibalism (Hebets, 2003). As argued above, this situation could be understood just from the perspective of sexual selection (i.e., the female is minimising her reproductive costs as she is avoiding mating with an unpreferred male), but it is also an example of extreme sexual conflict as costs for the male reach maximum values. More evidence showing that cognitive processes affect sexual conflict at the stage of mate choice comes from the fact that the attractiveness of a mate is not an absolute and that sexual conflict varies accordingly. In the aforementioned female guppies of the species Poecilia

reticulata, mate choice is not solely regulated by the exact amount of colouration of the male partner (i.e., the objective physical qualities, determined by the presence of carotenoid -yellow, orange or red-skin spots). Indeed, females' willingness to accept more or less sperm from a given male was shown to depend on the relative quality of that male: males of intermediate levels of attractiveness inseminated about three times the sperm (1521.3 × 10 3 spermatozoa) when the female had to choose between them and a "less attractive" male than when the female had to choose between the same intermediate attractive male and a more attractive one (534.4 × 10 3 spermatozoa) (Pilastro et al., 2004). Although we do not know the exact female physiological response that is involved (e.g., sperm digestion or sperm ejection), these results also show that females' assessment of the attractiveness of one male relative to that of others, leads to a lower or higher cost for the males under assessment, since they invested in sperm, accessory gland products, courtship and copulation behaviour in each mating. The rejection of sperm in each reproductive event, which implies a cost for the male, and thus, is a source of sexual conflict, would also be expected to change according to other variables that are known to affect mating. For example, in guppies, females alter their initial mate preferences after observing an older, but not a younger, female close to a male that they would not have chosen (Dugatkin and Godin, 1992;Godin et al., 2005). In these mate-choice-copy studies, the quantity of sperm that females accepted from males that were preferred or non-preferred by older or younger females was not quantified. Thus, we cannot know to what extent it may affect sexual conflict. However, there is evidence coming from the feral fowl (Gallus gallus) showing that the assessment of the social environment in a particular temporal moment is also shaping sexual conflict. In the feral fowl, females have been observed to eject, at least, half the sperm of non-dominant males but not that of dominant males. Importantly, the rejection of ejaculates can be altered when the hierarchy is changed after removing dominant and subdominant males (Pizzari and Birkhead, 2000). In both species, sperm acceptance/rejection depends on the notion of relative quality, which importantly implies not only perceiving but also comparing the different available options and deciding the optimal response to each situation. Whether females accept less, more or none of the sperm will depend on the outcome of those cognitive processes and so will the costs faced by the males, in terms of sperm and accessory gland proteins production, courtship and copulation in each mating encounter. The examples just mentioned are instances of how the cognitive processes that affect sexual selection necessarily also affect sexual conflict. Although the boundaries between sexual selection and sexual conflict could be argued not to be clear enough in some of them (see Kokko and Jennions, 2014; see also Arnqvist, 2004 for a discussion on the difficulty of establishing boundaries between the two concepts), we use these examples to illustrate that, with time, individuals accrue knowledge about their environment, including possible sexual partners (and possibly also the costs that mating with these entail), through (direct or not) experience, and that this knowledge determines the willingness to mate with, i.e., to accept/reject more or less sperm from, a particular partner. In that sense, we can say that cognitive processes regulate the activation of physiological or behavioural responses that are central to sexual conflict such as those related to sperm acceptance or to resource allocation to the offspring. It is the constant execution of such cognitive processes that allows, both males and females, to assess the quality of their partner and therefore, to flexibly adjust their response to the particular situation. Thus, cognitive processes such as perception, memory, comparison and decision-making can be said to play a modulatory role in sexual conflict that affects mating at all stages, from mate recognition to mate acceptance or to offspring resource allocation. Preparedness: A Special Type of Learning Experience As stated above, anticipation to a mating encounter has been argued to be a determinant of reproductive investment. As a matter of fact, being able to foresee a mating encounter allows animals to adopt different strategies to the specific scenario where mating is about to take place (see Hollis, 1982 for a discussion on the concept of preparedness). It is thus of great interest for animals to learn to identify and to pay attention to specific cues that signal the availability or receptivity of a potential partner so that the likelihood of being accepted increases, which in turn affects the level of sexual conflict. A special type of learning experience that allows animals to anticipate relevant events is that of classical conditioning (also referred to as Pavlovian conditioning). Classical conditioning occurs when an initially neutral stimulus (e.g., an acoustic stimulus such as the sound of a metronome) is presented together with another stimulus of biological relevance, such as food, that is referred to as the unconditioned stimulus (US). After repeated presentations, the neutral stimulus becomes a conditioned one (CS) that signals the presence of the US, and animals typically increase their rate of responding to the CS without needing to wait for the US to be present (Pavlov, 1927(Pavlov, /2003. Within the reproduction scenario, classical conditioning is important for the development of a correct mating response since it helps animals to learn about specific cues (CSs) that are consistently paired with successful mating (US). An already cited example is that of male rats (see Pfaus et al., 2001 for a review on the role of learning in rats' mating responses), that show an increased preference for odours (CSs) that are indicative of female sexual receptivity because they have been associated with successful mating (US). Indeed, Pavlovian conditioning has been shown to be of great importance for the development of preferences for pheromones, for the correct discrimination between receptive and non-receptive females and for a conditioned ejaculatory preference for certain females (see Pfaus et al., 2001 for rats; see also e.g., Dukas, 2005 for fruit flies or Domjan and Gutiérrez, 2019 for Japanese quail). Classical conditioning is, thus, crucial for the successful identification of a sexual partner that leads to successful mating. Importantly, accumulated successful mating experiences determine physiological responses that are involved in sexual conflict mechanisms. For example, in male rats, the more sexual experience they gain, the larger testes, the heavier penises and the greater the secretions from male accessory glands (reviewed in Pfaus et al., 2001). As already mentioned, these gland secretions are one of the most widespread sexual conflict mechanisms that induce high costs for females such as a decrease in sexual receptivity, an increase in egg investment, or even a decrease in their survival rate (see Poiani, 2006;Koene, 2012; for reviews on the costs of accessory sex gland secretions; see also Ramm and Stockley, 2016 for a review in rodents). Such a link between classical conditioning, more mating experience and increased male accessory gland products, indicates that classical conditioning can play a modulating role in sexual conflict. Pavlovian conditioning has been shown to be a useful tool for animals of different species to increase their reproductive success. For example, experiments with blue gourami (Trichogaster trichopterus), in which males were exposed to a light (CS) followed by the presence of a female, showed that being able to use the light as a predictor for the arrival of a female resulted in less aggressive behaviour toward her (see Hollis, 1999 for a review of the different studies she led on associative learning and mating in blue gourami). Such behavioural change in the male induced the females to spawn faster than the females that were being courted by males who had not been subjected to that learning experience. More importantly, the number of offspring sired by classically conditioned males was significantly higher than that sired by the control males for whom the CS had been unreliably paired with the presence of the female. When looking at the results obtained in this experiment (Hollis et al., 1997), the difference in the number of offspring produced by conditioned and by control males is in the magnitude of hundreds of fry. If we take the control group, i.e., females who mated to males that were not able to anticipate their availability, as the baseline of females' investment, assuming that males were of equal quality across groups, we can say that, in this particular scenario, classical conditioning prepared males for the mating encounter in such a way that they increased their reproductive fitness by inducing females to increase their investment, i.e., the costs, by producing or releasing a significantly higher number of ova. Another example of how classical conditioning induces the sexual partner to increase its reproductive investment is observed in the Japanese quail Coturnix japonica. When males are exposed to cues that signal a mating encounter, they produce larger ejaculates and larger numbers of spermatozoa (Domjan et al., 1998). With a differential conditioning procedure, in which males were exposed to a female in one context but not in a different one, mating in the context that had been paired with access to a female increased males' mating success, as measured by the number of fertilised eggs (78 vs. 39%; Adkins- Regan and MacKillop, 2003). It is important to note that the females used during training were different to the ones used in the test (i.e., higher fertilisation rate cannot be explained by a higher male preference for a particular female); that the females used in the test were sexually naïve, and that they had been randomly assigned to mate in either one context or the other (i.e., female preferences for males was not taken into account, so this variable can either explain the differences observed in the number of fertilised eggs). Thus, males that were able to rely on external cues were also able to impose a higher reproductive cost to the female they mated with. This ability to anticipate a mating encounter was shown to be of special importance in situations where one female mated with two different males. When this occurs, paternity is usually shared equally between the two males (i.e., 50-50%). However, when two males are competing for a single female and one of them has been subjected to Pavlovian conditioning, the one who is able to rely on a cue that signals the availability of the female is quicker at mating and, most relevant, sires a higher proportion of the offspring (72 vs. 28%). Such higher paternity success was shown to be explained by that learning experience alone, independent from whether they mated first or second with the female (Matthews et al., 2007). As argued by the authors, males of both conditions had the same mating experience during the experiment and prior to the test, so differences in fertilisation were not likely to be due to differences in sperm production but in sperm release, and/or, as argued by Adkins-Regan and MacKillop (2003), the higher paternity rates could also be due to higher production of foam (a substance that is transferred to the female during copulation and that is known to increase fertilisation success). Although research on the exact physiological mating response affected by classical conditioning is still lacking, these results show that classical conditioning allowed males to increase females' investment toward their own sperm, from the baseline of 50%. The experiments conducted with blue gourami and Japanese quail show that using cues as indicators of a mating opportunity helps individuals to prepare better for the mating encounter but that much research still needs to be done to analyse the effects of classical conditioning in terms of sexual conflict. Nonetheless, they also show that classical conditioning is a special instance of learning with a potentially important modulatory role in sexual conflict: males who can predict the availability of a (fertile) female adjust their mating response to obtain a higher reproductive fitness, independently of female preferences (so to the female's costs), by optimising the mating process in terms of copulation time and/or amount of sperm transfer, and possibly also by modulating the amount of accessory gland secretions that are transferred along with sperm (as happens e.g., in D. melanogaster; Mohorianu et

al., 2018). A Yet To-Be-Developed Experimental and Theoretical Field As pointed out above, studies on the cognitive processes that shape mating responses and on the physiological mechanisms that are involved in sexual conflict are largely performed independently from one another. The evidence here gathered shows that, just as cognitive processes play an important role in sexual selection, there exists a modulatory relationship between cognition and sexual conflict that has not been fully explored and we are convinced that properly considering the interplay between cognitive processes and physiological mechanisms in the context of reproduction will be a fruitful direction for this field of research. For example, as pointed out earlier, studies in guppies have not measured the extent to which social variables that are already known to affect the attractiveness of a potential mate alter the amount of sperm that a female is willing to accept, which would allow to measure the cost suffered by the male. Likewise, the studies with blue gourami and Japanese quails also reveal that much research needs to be done to understand the modulatory effect of classical conditioning. The experiments reported show beyond doubt that Pavlovian conditioning can serve to increase the partner's investment (as measured e.g., by the larger number of eggs laid by blue gourami females), but they also show that learning results in an increased investment by the individual that is able to anticipate the mating encounter (as shown by the production of larger ejaculates by learning Japanese male quails). The extent to which the extra gains obtained from being able to predict a mating encounter outweigh the costs remains to be explored. Just like learning males, females also increase their own investment when they can predict a mating encounter. In an experiment conducted with Japanese quail, it was observed that sexually naïve females who had been subjected to Pavlovian conditioning in which a context was paired with just the presence of a male (copulation was prevented) laid a higher proportion of fertilised eggs when they mated in the context that had been consistently paired with the presence of males, compared to the females that mated in a context in which males were never present (Adkins-Regan and MacKillop, 2003). In this particular situation, the males that the females encountered at the test were novel to them, as they had not been exposed to them during training. Just knowing that the cage in which they were introduced was a good predictor for the presence of a male was enough to affect females' reproductive success, also increasing the reproductive benefits of the male. The physiological responses that were modulated by this learning are still unknown. Pavlovian conditioning may alter sperm rejection or it might also affect the composition of the female reproductive tract fluid, leading, in this particular case, to increased chances of successful insemination and lower costs for the male. In this regard, female reproductive tract fluid is much understudied, but there is very recent evidence showing that its composition changes after mating (McDonough-Goldstein et al., 2021). It would thus be interesting to test whether Pavlovian conditioning could also alter the composition of females' fluid, just as it changes seminal fluid in males. Finally, when taken together, most of the studies seem to indicate that experience with a given partner or phenotype enhances animals' motivation to mate, thus it seems that learning is mostly playing a facilitating role. However, we prefer using the term "modulatory" instead of facilitating because it might also serve to hamper mating under specific circumstances. For example, in mosquitofish (Gambusia holbrooki), where coercive mating occurs, it has been observed that male sexual harassment results in a decreased foraging efficiency, and that females can reduce such costs by aggregating with other females (Pilastro et al., 2003;Dadda et al., 2005). It could thus be possible that if, after classical conditioning training, females are able to anticipate an unwanted mating encounter, they might be able to avoid mating entirely or, if not possible, to display sexually antagonistic strategies to a maximum level. DISCUSSION As we have argued, the research that has been reviewed here shows that sexual conflict strategies are not solely determined by genetic or physiological responses that are independent of the knowledge that animals acquire throughout their life. On the contrary, it shows that experience, i.e., ontogeny, affects the way in which those sexual conflict strategies are displayed, altering the amount of reproductive investment a mating partner will face. Experience, whether interpreted in terms of Pavlovian conditioning or not, affects the mating sequence and the way in which sexual conflict responses are displayed. These responses range from motivation to mate to the number of resources that are allocated to the offspring sired by a particular partner. Genetic variability is without question a driving force for evolution. The differences in both the genotype and phenotype provide a source of trait variability that enables organisms to adapt to the surrounding conditions of their environment. On the one hand, these adaptations take considerable (evolutionary) time to occur, which would render an individual in a vulnerable state: if the genetic variation has not been inherited or no beneficial mutation has taken place, there is little room for the organism to adapt successfully. On the other hand, cognitive processes are constantly ongoing and allow animals to monitor their surroundings as well as their internal states, to accumulate and update their knowledge about their Umwelt and to regulate and/or adjust their behaviour accordingly. As a consequence, cognitive processes and behaviour offer temporal, flexible and quick responses that are shorter-or longer-term adaptations that can be as crucial for survival as evolutionary changes. Indeed, learning and behaviour have been argued to be one of the main driving forces of evolution (Baldwin, 1896;Roe and Simpson, 1958;Piaget, 1976;West-Eberhard, 2003;Ginsburg and Jablonka, 2010;Diogo, 2017) because they allow animals to successfully regulate their activities on a daily basis. Here it is important to highlight that although cognitive abilities (e.g., attention, memory span, processing speed or the ability to learn) may become common, and variable, due to genetic variations (Plomin et al., 2013, but see also Nisbett et al., 2012), the adaptive response an animal develops to a particular situation is not genetically determined. In that sense, the acquired behavioural response can be reversible as it can change according to e.g., learning contingencies (e.g., counterconditioning), and most importantly, it does not involve a genetic change that is inherited by the offspring. In that regard, cognitive processes have been considered as sources of new and fast adaptations in research areas such as that of comparative psychology (Shettleworth, 2009) or eco-evolutionary dynamics (Svensson, 2019), but they have been overlooked in the field of sexual conflict (e.g., Chapman et al., 2003;Shackelford and Goetz, 2012). As we have already discussed, most of the sexual conflict strategies that animals employ are still largely examined under a very mechanistic approach in which only physiological mechanisms and genetic traits seem to be considered (Alvarez and Koene, 2018). The current review shows that there is still a conceptual gap that needs to be addressed in the domain of sexual conflict as there is ample evidence that sexual conflict is not solely regulated by genetic traits but also by each individuals' assessment of different aspects within a mating situation. Experience affects the way in which a mating partner is perceived, the motivation to mate or the capability to prepare for a mating encounter altering the physiological response that will be displayed. As pointed out by Tinbergen (1963), we cannot reach a comprehensive view of any biological trait without taking into consideration both ultimate and proximal causes. Importantly, ontogeny of behaviour cannot be understood as directly determined by genetic factors (reviewed by Sánchez and Loredo, 2007), but rather as the result of cognitive processes (motivation, attention, and perception), experience (learning and memory) and the subsequent decisions. Cognitive processes are fundamental for displaying an adaptive mating response and they modulate the physiological responses an animal experiences in a particular moment, including those related to sexual conflict. The fact that ejaculate size, sperm acceptance or paternity success are enhanced under certain cognitive conditions is already highlighting the need to take ontogeny into account for a good understanding of sexual conflict. We sincerely hope that this review will instigate a field of research that will focus on the interplay of the two factors, closing the existing gap. AUTHOR CONTRIBUTIONS BÁ and JK: conceptualization. BÁ: writing-original draft. JK: co-writing original draft, and writing-review and editing. Both authors contributed to the article and approved the submitted version. Spinal intradural extramedullary bizarre parosteal osteochondromatous proliferation of bone (Nora's Lesion): First case report Bizarre parosteal proliferation of bone (Nora's lesion) is a known bony lesion that affects mainly hands and feet. In this article, we present the first case of spinal intradural extramedullary Nora's lesion along with the management. Radiologically, the tumor was initially diagnosed as a meningioma. However, histopathological analysis confirmed bizarre parosteal proliferation of bone. It was successfully managed by surgical resection followed by physical rehabilitation. Introduction Nora's lesion or bizarre parosteal osteochondromatous proliferation (BPOP) of bone is one of several described types of osteochondromas. It is an exophytic bony outgrowth consisting of bone, cartilage, and fibrous tissue that typically has continuation with the underlying periosteum. [1,2] In addition to Nora's lesion, osteochondromas include solitary osteochondromas, postirradiation osteochondroma, multiple osteochondroma, epiphyseal osteochondroma, subungual exostosis, and para-articular osteochondroma. [3] It was first described in small bones of hands and feet by Nora et al. in 1983 followed by several cases reports. [2,4,5] There are more than 160 reported cases of BPOP mostly in the long bone and other parts of the bony skeleton. [2,4] In this report, we present the first case of spinal intradural extramedullary Nora's lesion. Case Report A 61-year-old female, with an unremarkable past medical history, presented to our emergency department complaining of progressive right lower limb weakness started 15 days before presentation associated with back pain around the right mid-thoracic region. The back pain started 4 months before presentation. No lower limb pain was reported. There was a negative history of fever, trauma, corticoid use, or malignancy. There was no urinary/ fecal incontinence or retention. On physical examination of the lower limbs, she was found to have a mild hypertonic muscular tone bilaterally. The power of the right and left lower limbs was 3/5 and 5/5, respectively. Patellar and Achilles tendon reflexes were brisk 3/4 bilaterally, Babinski sign was equivocal, and gait was unsteady. Sensory examination showed diminished light touch and pinprick sensations on and below the T4 level. Diminished proprioception marked mainly at the right side. Rectal examination showed preserved anal tone. The upper limbs' neurologic examination was normal. Investigations Laboratory analyses for complete blood count, electrolytes, erythrocyte sedimentation rate, and C-reactive protein were within normal range. Gadolinium-enhanced magnetic resonance imaging (MRI) of the spine showed homogeneously enhanced oval intradural extramedullary lesion at the thoracic level measuring 12 mm × 14 mm located at the T4-T5 level. The lesion is exerting a significant local mass effect, causing marked cord compression. The lesion was hypointense on T2-weighted image [ Figure 1]. On computed tomography scan, the lesion showed calcifications associated with osteosclerosis of the right lamina and pedicle [ Figure 2]. Differential diagnosis of the lesion included mainly a calcified meningioma followed by other primary or secondary spinal tumors. Less likely differentials were infectious or inflammatory processes. Operative management The patient was managed operatively through a laminectomy at the level of T4 and T5. The paraspinal muscles and spinal ligaments were intact. However, the spinous processes and the right T4 lamina were looking abnormal. The outer dural layer was attached to the lamina. However, the dura was easily detached from the lamina and found intact after the laminectomy was done. The dura was opened in a midline longitudinal fashion. Once opened, a white, nonhypervascular, hard tumor was seen. It had stony consistency with calcifications forming a hard shell-like structure. The tumor was attached to the dura and nerve root [ Figure 3]. With careful dissection under the microscope, we were able to remove the tumor in a piecemeal fashion [ Figure 4]. After complete removal of the tumor, dural, subcutaneous tissue, and skin closures were done in the routine fashion. Histopathology Histopathological analysis of the tumor showed a disorganized proliferation of fibrous, cartilaginous, and

bony tissues. At some locations, the cartilaginous component showed irregular groups of enlarged binucleated and atypical chondrocytes. Immunodetections with anti-MDM and Ki67 antibodies showed no immunolabeling. The Ki67 was about 1%. The histopathological features of the tumor were consistent with a BPOP of bone (Nora's lesion). Postoperative course, outcome, and follow-up Corticosteroid was given. No central nervous system infection or cerebrospinal fluid leakage occurred. Rehabilitation started 2 weeks after the operation. The patient was allowed to do walking exercises. At the 6-month follow-up, her right lower limb showed a significant improvement in power, grade of about + 4/5. She was able to walk without any difficulty. No clinical or radiological signs of recurrence were noted. Follow-up MRI of the spine showed completely resected tumor without signs of recurrence [ Figure 5]. Discussion Nora's lesion is a formed by heterotopic ossification. [1] It is an exophytic bony outgrowth consisting of bone, cartilage, and fibrous tissue that typically has continuation with the underlying periosteum. [1,2] In our case, although the tumor was attached to the dura, which in turn was attached to the lamina, the tumor itself was completely intradural with no apparent attachment to the lamina. In the literature, there is one reported case of spinal intradural extramedullary chondroma and around 20 cases of cranial intradural chondroma. [6,7] Histopathologically, Nora's lesion -in contrast to osteochondroma -has less organized microscopic appearance and nuclear atypia with atypical chondrocytes. [8] It may present hypercellularity and mitotic activity. [8] To our knowledge, this is the first reported case of spinal intradural extramedullary BPOP of bone (Nora's lesion). In bony lesions, Nora's disease has a tendency to recur locally with no metastases and almost no malignant transformation was reported. [1,2,4] The initial recurrence rate is up to 51% after the first resection and 22% after the second one. [5,9] Recurrence is usually within the first 2 years. [5,9] We are managing the patient postoperatively as if the tumor has similar behavior. However, whether this hypothesis is true or not will be determined with long-term follow-up of the patient and will be reported in future communications. Conclusion This is a first case report of spinal intradural extramedullary bizarre parosteal proliferation of bone (Nora's lesion). The patient was successfully managed with surgical resection and physical rehabilitation. Close follow-up is being done for possible recurrence based on known tumor behavior in the bony skeleton. Declaration of patient consent The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest. Anti‑cholinesterase activity of the standardized extract of Syzygium aromaticum L. Background: Clove ( Syzygium aromaticum ) is a well-known culinary spice with strong aroma; contains a high amount of oil known as clove oil. The major phyto-constituent of the clove oil is eugenol. Clove and its oil possess various medicinal uses in indigenous medicine as an antiseptic, anti-oxidant, analgesic and neuroprotective properties. Thus, it draws much attention among researchers from pharmaceutical, food and cosmetic industries. Objective: The aim of the present study was to determine the anti-cholinesterase activity of the methanol extract of clove, its oil and eugenol. Materials and Methods: In vitro anti-cholinesterase activity of S. aromaticum was performed by a thin layer chromatography bio autography, 96 well micro titer plate and kinetic methods. Reverse phase high performance liquid chromatography (RP-HPLC) analysis was carried out to identify the biomarker compound eugenol in clove oil. Results: Acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibition study revealed that eugenol possess better inhibition of the enzymes than extract and oil. Clove extract, its oil and eugenol showed better inhibition of AChE than BChE. Polyphenolic compound eugenol was detected through RP-HPLC analysis. The content of eugenol in essential oil was found to be 0.5 µ g/ml. Kinetic analysis of the cholinesterase inhibition study of the extract; clove oil and eugenol have shown that they possess mixed type of inhibition for AChE and non-competitive type of inhibition for BChE. Conclusion: These results might be useful in explaining the effect of clove as anti-cholinesterase agent for the management of cognitive ailments like Alzheimer’s disease. INTRODUCTION The decrease in the brain acetylcholine (ACh) amount is the hallmark of Alzheimer's disease (AD), which is largely manifested by impairment of cognitive function leading to changes in the behavioral pattern in an older individual. [1] The mechanistic approach related to AD revealed that the minimization of acetylcholinesterase (AChE) as well as butyrylcholinesterase (BChE) activity can compensate decrease label of ACh. [2] The present cholinesterase inhibitors (ChIs) synthetic origins have been used primarily for treatment of such diseases condition. These agents are often associated with common side-effects in AD patients. [3] A safer approach to prevent the progression of such diseased condition would be use of natural anti-cholinesterases. Several reports have been published on effectiveness and some of are approved for treating cognitive disorders of AD type. [4] Syzygium aromaticum (Myrtaceae), a tropical evergreen plant, widely used spice since time immemorial as human nutrition and flavoring agent. [5] Clove has a strong bonding with various ethnic cultures of Indian subcontinent and it has been used in rituals to medicinal practices. In Ayurveda as well as Iranian traditional medicinal formulations; clove oil is used as nervous stimulant and cognitive enhancer. [6] Phenolic compound, eugenol is reported to be the major chemical constituent present in the clove oil, which is also responsible for its biological activities. The oil also contains eugenol acetate, β-caryophyllene, chavicol, humulenes in lesser amounts. [7] Eugenol is reported for its wide range of pharmacological proper ties including analg esic, anti-oxidant, anti-inflammatory, anti-allergic, anti-carcinogenic and anti-mutagenic activities. [5,8] Previous studies on eugenol A B S T R A C T with an animal model showed significance beneficial effect on nerve function. It was also reported that eugenol content of clove oil is capable of improving learning and memory process of a mice model under elevated oxidative stress condition. [6] Other experimental study shows that treatment of diabetic mice with eugenol can alter vascular and neuronal complication in brain. [9] The above mentioned publications of S. aromaticum, its oil and eugenol provided the basis of the present study to examine their anti-cholinesterase activity. Further, kinetic analysis of the anti-cholinesterase data also revealed the types of inhibition potential of the studied plant materials. MATERIALS AND METHODS Dried flower buds of S. aromaticum were collected from local market of Kolkata, West Bengal, India and authenticated as well. A voucher specimen (SNPS-1479) of the same has been preserved in the herbarium of the School of Natural Product Studies, Jadavpur University, Kolkata, India for further use. Preparation of plant extract Fresh flower buds of S. aromaticum were shade dried and pulverized to coarse power. The plant extract was prepared by cold maceration. A hydroalcholic mixture was added with 25 g powder with occasional shaking for 72 h to obtain effective extraction. [10] The extract was collected after filtration in a beaker. The above procedure was repeated thrice for exhaustive extraction. All the extracts were pooled together and distilled under reduced pressure using rotary vacuum evaporator to collect the final semisolid extract which dried in a lyophilizer to remove residual moisture, yield was found to be 8% w/w. Isolation of clove oil Clove oil was isolated by Clevenger apparatus adapting the reported method. [11] 10 g of powdered clove was soaked in water and boiled at 100°C for 5 h to collect the buff colored oil on water. The collected oil was dried with anhydrous sodium sulfate to remove moisture and reserved in a refrigerator at 4°C until further study, yield of the oil was found to be 6% v/w. Standardization of the extract through HPLC Reverse-phase HPLC (RP-HPLC) analysis of clove oil and standard eugenol were performed on HPLC Shimadzu Prominence, Kyoto, Japan. Clove oil (10 mg/ml) and standard eugenol [ Figure 1] (1 mg/ml) were dissolved in methanol for the chromatography analysis. Isocratic mobile phase (methanol:water: 75:25) condition was maintained throughout the process. The samples were eluted from the column at room temperature with flow rate of 1 ml/min which was detected at 254 nm. LC solution software was used for the study of chromatogram. Identification and quantification of eugenol was verified by peak areas obtained in HPLC analysis. ChIs activity Bio-autography method Thin layer chromatography (TLC) analysis of AChE and BChE inhibition was carried out according to the method. [12] 2.5 mm silica gel TLC plate, F 254 (Merck, Darmstadt, Germany) was spotted with methanol extract (10 mg/ml) and it was developed with an optimized mobile phase methanol:water (75:25). The plate was dried at room temperature. Freshly prepared 1 mM DTNB or Ellman's reagent [13] and 1 mM ATCI solution in phosphate buffer was sprayed over the plate for AChE inhibiton and 1 mM BTCI solution instead of ATCI was used for studying BChE inhibition. The plates were dried for 25 s and then 3 U/ml of AChE/BChE enzyme solutions were sprayed over the respective plates. The plates were kept aside for 5 min to develop colorless spots over a yellow background as AChE/BChE inhibition zones. The plates were immediately photographed, as the white spots gets disappear fast. To confirm the true inhibition, another TLC plate was developed in a similar manner without spotting the extract, called as false positive. Appearance of white spots at the similar place on the TLC plate with extract considered as false inhibition. well micro titer plate assay The inhibition activity of AChE/BChE was determined by modified method [13] by using Bio Rad 96-well microplate reader (680 XR, USA). [14] Extract, oil, eugenol and galantamine as a positive control were dissolved in methanol for the study. In each well of separate 96-well micro titer plate, 125 µL of 3 mM, DTNB (Ellman's reagent), 25 µL of 15 mM ATCI/BTCI in phosphate buffer, 50 µL of phosphate buffer (pH: 7.2) and extract/oil/eugenol at various constrations (5-100 µg/ml) and galantamine (1-20 µg/ml) were added. The plates was measured at 405 nm at every 13 s for 65 s. AChE/BChE (25 µL of 0.22 U/ml) in phosphate buffer was added to all the wells and reading measured at 405 nm at every 13 s for 104 s. [14] This procedure was repeated for 6 times. IC 50 value was calculated by plotting the inhibition percentages against the logarithm of the concentration. Kinetic study Method applied for the kinetic study of enzyme inhibition was similar to the 96-well microplate assay with the main difference being the fixed inhibitor concentration and varied substrate (ATCI/BTCI) concentration (0.75-14 mM). The kinetic parameter K m (Michaelis-Menten constant), V max (maximal velocity) and dissociation constant/inhibition constant (K i ) was calculated over through the Lineweaver-Burk and their secondary replots. The experiment was performed in triplicate. [13] Statistical analysis All the results of the data were expressed as mean ± standard error of the mean. The linear regression parameters were used to determine the IC 50 value of enzyme activity. Kinetic parameters (K m , V max and K i ) were obtained by using non-linear regression analysis. Comparison of enzyme kinetic data was performed by using one-way analysis of variance followed by Newman-keuls comparison test. P < 0.05 was considered to be significant. Graphs were plotted using Graph Pad Prism 5.01. Standardization of the extract through HPLC In the present study, eugenol content in S. aromaticum oil was accessed through RP-HPLC analysis. The retention time of eugenol in the extract [ Figure 2a] was found at 13.633 min which was similar to the retention time of standard eugenol [ Figure 2b]. The calibration curve of eugenol has shown linear relationship between peak area and standard eugenol concentration with a correlation coefficient, r 2 = 0.9958. The amount of eugenol in oil was found to be 0.5 µg/ml. Bio-autography method AChE/BChE inhibitory activity of the standardized methanol extracts (10 mg/ml) of S. aromaticum was evaluated by using TLC bioautographic assay. Biochemical detection of plant extract was screened through an optimized mobile phase. As a result, variable white spot with yellow staining was observed

in a TLC plate [ Figure 3a- DISCUSSION It has been hypothesized that cognition and behavior of AD patients closely associated with a decrease in concentration of brain ACh level. The hypothesis was supported by the reports suggested that inhibition of AChE along with BChE remains important target for the treatment of neurodegenerative disorders. Therefore, substrate specific AChE and BChE inhibitors are important options for treatment of AD. [15,16] Until recently, large numbers of naturally-occurring compounds have been found to inhibit AChE/BChE. [17,18] Most of the prescribed ChIs such as huperzine, rivastigmine and galantamine are obtained from plant sources. This provides an opportunity to explore new ChIs from plant sources. [19] ChE inhibition potential of S. aromaticum was assessed by TLC bioautography method and found to be positive. For further confirmation of the inhibitory potential of clove extract, oil, eugenol and galantamine 96-well microplate method was applied. Enzyme inhibition efficiency of clove oil was better than the extract. Eugenol demonstrated highest AChE inhibitory among the extract and oil, but less active than the reference compound galantamine. All the tested compounds exhibited weak affinity toward BChE. Enzyme kinetics of AChE inhibitory activities of methanol extract of S. aromaticum extract, clove oil, eugenol and galantamine was screened. The V max value of clove extract, oil and eugenol were decreased by comparing with the control group, while the K m values of above constituents were increased, which indicated a reversible and mixed type of inhibition. [20] However, V max value of reference compound galantamine did not showed any significant change with respect to control group, but K m values decreased, which suggested the type of inhibition would be reversible and competitive type. [21] In the case of BChE, extract, oil and eugenol were showed reversible and non-competitive type of inhibition, [22] while galantaine possessed mixed type of inhibition. Results also indicated that galantamine have a significant (P < 0.05) value than the control group. The phytoconstituents of the S. aromaticum extract and its oil may be explored as potential lead for the treatment of AD. CONCLUSION The present study suggested that S. aromaticum extract and its oil, along with eugenol has potential anti-cholinesterase property. Therefore, this culinary spice may be explored further for its use in the management of AD and related cognitive disorders. The power of the web in cancer drug discovery and clinical trial design: research without a laboratory? The discovery of effective cancer treatments is a key goal for pharmaceutical companies. However, the current costs of bringing a cancer drug to the market in the USA is now estimated at $1 billion per FDA approved drug, with many months of research at the bench and costly clinical trials. A growing number of papers highlight the use of data mining tools to determine associations between drugs, genes or protein targets, and possible mechanism of actions or therapeutic efficacy which could be harnessed to provide information that can refine or direct new clinical cancer studies and lower costs. This report reviews the paper by R.J. Epstein, which illustrates the potential of text mining using Boolean parameters in cancer drug discovery, and other studies which use alternative data mining approaches to aid cancer research. The discovery of effective treatments for cancer represents a key goal for pharmaceutical companies who wish to identify drugs that can prolong survival time and even reverse cancers, while having an acceptable toxicity profile. However, the average cost of bringing a drug to the commercial market in the USA is now estimated at $1 billion per FDA approved drug, and many factors have compounded the expense of these developments such that cancer drug discovery is now both extremely slow and costly even for a potential blockbuster. Among the many factors contributing to the cost are the high price of clinical trial organisation and the bench research hours required to validate the efficacies and toxicities associated with a drug despite the use of time saving technologies such as high throughput screening to determine efficacies and genomic analyses of drug effects. A growing number of papers highlight the use of data mining tools to determine associations between drugs, genes or protein targets, and possible mechanism of actions or therapeutic efficacy which could be harnessed to provide information that can refine or direct new clinical cancer studies. One common method of data mining is referred to as text mining. Richard Epstein 1 provides a number of examples of how text mining using Boolean terms can be used to determine associations between a cancer type or drug and the symptoms or efficacies observed. For example he describes how phenotypes and environmental factors associated with either squamous cell carcinoma or adenocarcinoma (e.g. smoking and lymph node metastasis for squamous cell carcinoma vs. hormone and liver metastasis for adenocarcinoma) can be deciphered. Mechanistic associations of different drugs such as tyrosine kinase inhibitors and metalloprotease inhibitors can also be calculated: Growth or replication is more strongly associated with tyrosine kinase inhibitors and invasion and metastasis inhibition is more strongly associated with metalloprotease inhibitors. Epstein also provides examples of how text mining can determine associations between types of cancers and a particular gene for example, AKT. 2 The gene for AKT encodes a retroviral protein which is a pivotal cell signalling protein which when activated leads to inhibition of cellular apoptosis and activation of its downstream target (mammalian target of rapamycin (mTOR)), which increases mRNA translation through combination with its protein RAPTOR (regulatoryassociated protein of mTOR). By text mining associations, AKT is associated with a number of cancers of which the most prominent is prostate cancer. When the association is then compared with cancers in which the mTOR inhibitor temsirolimus has been used, prostate cancer also gives the strongest correlation as the disease in which this drug has been most commonly used. Therefore, text mining can detect correlations between specific cancers and their associated gene defects and the drugs that are used for that cancer. A number of papers have shown how text mining has contributed greatly to identifying critical genes and drugs in a number of cancers. For example Pospisil et al have used a combined textual-structural mining approach to identify potential enzyme targets in the extracellular space of cancerous cells for six common, lethal human tumors, by searching databases such as PubMed abstracts, NCBI Entrez, UniProt, (a universal gene/protein database) and Interpro, a conserved protein domains database. By using keyword and gene ontology terms and by clustering these terms to specific cell locations, a list of cancer-related hydrolases for each tumor type have been identified as therapeutic targets including prostatic acid phosphatase (ACPP also known as PAP), prostatespecific antigen (PSA) and sulfatase 1 (SULF1). 3,4 Another study by Turk et al have used text mining of National Cancer Institute's DTP drug repository to search for compounds showing increased toxicity in MDR cells and discovered 22 compounds with MDR specific toxicity, and a further 15 drugs showing increased cytotoxicity in cells with P-glycoprotein. Analysis of these compounds has led to the formulation of structure activity relationships linking mechanism of action with metal chelation, and shows that p-glycoprotein is not the only target of compounds that are effective in MDR cells. 5 However, there are a few known limitations to text mining. Associations are based on the use of constant terminology for a drug or gene, whereas this may not be the case for example, where drug names are changed by companies. Also there is restricted access to full text journals and also restriction to abstract publications and some chemical and physical science journals in databases such as pubmed so that data associations need to be retrieved from a more limited source of overall citations. Text mining may therefore be better utilized by combining it with other data mining tools such as microarray database mining. This uses microarray gene data from experiments which have analysed genomes or sets of genes of particular cells or tissues. 6 This allows the discovery of drug sensitive and disease specific genes which can be used to identify targets for cancer therapy. Such analyses yield vast amounts of gene data as even a whole genome can be available on a chip. When text mining and microarray data mining are combined, powerful analyses of data can be applied to decipher cancer therapy targets. For example, Ho and colleagues have identified a set of 64 genes that are specifically expressed in endothelial cells compared with non endothelial cell types from combined text mining and microarray analyses. 7 There are also caveats that can be applied to microarray mining. 8 Microarray analysis results in a vast amount of gene data from a dataset of samples that is normally at least 100 fold less than the gene data generated. There is also the problem of noise where artifactual expression may be seen with platforms that are not stringent enough to filter outlier data, and the requirement for repeat array analysis which sometimes shows variance in the fold change produced in single genes. 9 This type of analysis always requires a post test to confirm the gene changes observed, either quantitative pcr or western blotting of the associated proteins. Microarray gene data can be complemented by proteomic data analysis such as mass spectrometric analysis, SELDI-TOF (Surface-enhanced laser desorption/ionization-time of flight) and bio-plex technologies which allow analysis of very large numbers of proteins on an array format and combined analyses of proteins and genes (often referred to as pathway analyses). These techniques both complement and strengthen the observations seen with the genes alone. For example, this approach has been used to determine biomarkers which can give a very early prognosis of ovarian cancer. 10 Such an approach has also recently helped to lead to discovery of biomarkers which can predict a favourable response to prostate cancer vaccines (Bodman-Smith et al, paper in preparation). The power of data mining has now been harnessed by a growing industry specialising in the production of databases which can utilise text or gene or protein data. A selection of databases with direct application to cancer drug or target molecule discovery are presented in Table 1. The tools that these databases provide for both the drug industry and academia can maximise the mining process compared to manual mining techniques. Whereas data mining has not yet resulted in blockbuster discovery on its own merit, the use of this technology harnessed with the power of dedicated databases and bench top research, has already contributed to deciphering mechanisms of action of genes and drugs and should allow a much more rapid progress toward discovery of effective cancer therapies in the future. Computer assisted histomorphologic comparision and the expression of AgNORs in the central and peripheral giant cell lesions of the oral cavity and giant cell tumor of the long bone Objective: Computer-assisted image analysis was attempted to ascertain, if any of the previously described histologic features along with argyrophilic nucleolar organizer regions (AgNORs) could be used to determine the aggressiveness of the central giant cell granuloma of the jaws (CGCG), peripheral giant cell granuloma of the oral cavity (PGCG) and giant cell tumor of the long bones (GCT). Study Design: The study consisted of 20 cases of CGCG, 20 cases of PGCG and 5 cases of GCT. The histological features included were number of giant cells, number of nuclei in each giant cell, number of blood vessels, fractional surface area (FSA) and relative size index (RSI) of giant cells. The histologic parameters were measured using Motic image plus analyzer and AgNORs were evaluated using silver stain. Results: The statistical analysis showed significant differences among various histological parameters between CGCG, PGCG and GCT. A statistically significant difference was noted for the mean number of nuclei, FSA and RSI when GCT was compared with CGCG and PGCG. FSA of histologically aggressive central giant cell granuloma (HA-CGCG) was more compared to histologically non-aggressive central giant cell granuloma (HNA-CGCG). No statistical correlation was observed for AgNORs of multinucleated giant cells and mononuclear cells among CGCG, PGCG and GCT. Conclusion: Based on the present study findings, CGCG and GCT are distinct and separate entities and not a continuum of a single disease process. Histological parameters alone have a little implication on predicting clinical behavior of CGCG. AgNORs alone as a proliferative marker has a limited value in assessing the proliferation

potential of giant cell lesions. INTRODUCTION Giant cell lesions of the maxillofacial skeleton and other bones are a controversial matter and uncertainty still exists regarding their basic pathology and biologic behavior. Varying clinical and histological parameters have attracted the interest of many researchers towards understanding this diverse group of lesions which sometimes behave neoplastically. [1][2][3][4][5] Before 1953, workers generally did not distinguish between giant cell lesions of the jaws and giant cell tumors of the jaws. Jaffe [6] introduced the term giant cell reparative granuloma of the jaws and was the first to distinguish this lesion from giant cell tumors that usually involve the epiphyseal regions of long bones. Giant cell granulomas may occur within the bone (central giant cell granuloma, CGCG) or on the gingival or edentulous alveolar process (peripheral giant cell granuloma, PGCG). CGCG is an intrabony, non-neoplastic, slow-growing lesion affecting women more often than men. CGCG is a relatively uncommon, locally aggressive bone lesion, with variable clinical behavior. It accounts for less than 7% of all benign jaw lesions, with more than 60% of all cases occurring before the age of 30 years. [7] The radiographic appearance of CGCG is not pathognomic and may be confused with other lesions, such as ameloblastoma, odontogenic keratocyst, aneurysmal bone cyst and hyperparathyroidism. Histologically, it is characterized by the presence of few to many multinucleated giant cells and mononuclear cells within a fibrous stroma. In contrast, PGCG is a relatively common lesion thought to arise in reaction to local stimulatory factors and runs an indolent course. [8,9] Giant cell tumor of long bones (GCT) is a locally aggressive, neoplastic lesion with a high recurrence rate. Malignant transformation occurs in 15-30% of the cases. Histologically, GCT consists of evenly distributed numerous giant cells lying in a cellular matrix composed of spindle-shaped cells and scanty collagen. The giant cells measure about 100 microns in diameter and contain numerous vesicular nuclei up to 50 or more, which are situated towards the center of the cell, leaving a clear area of cytoplasm around the periphery. [4] The literature of the last few years contains several references to so-called giant cell lesions of jaws and pathologists have attempted to identify histopathologic parameters in order to predict clinical behavior and prognosis of giant cell lesions. [1][2][3][4][5] Nucleolar organizer regions (NOR) represent the loops of DNA actively transcribing to ribosomal RNA and thus to ribosome and ultimately to protein. NORs are associated with acidic argyrophilic non-histone proteins, which are visualized with the use of silver staining technique, the argyrophilic region (AgNOR). Some authors have applied AgNOR technology and found a positive correlation between AgNOR and recurrences and/or aggressiveness of the giant cell lesions, whereas others have found no correlation between PGCG and CGCG. [10,11] So far, not a single study has been performed to compare AgNOR count between CGCG, PGCG and GCT. Therefore, the visualization of NOR distribution appears to be a promising method to diagnose and assess the prognosis of giant cell lesions. As giant cells are an interesting and promising area of research, the present study used image cytometry to examine various histologic parameters along with AgNOR count to assess the aggressiveness of the giant cell lesions of the jaws. MATERIALS AND METHODS The study comprised of 20 previously diagnosed cases of CGCG, 20 cases of PGCG and 5 cases of GCT retrieved from the archives of Department of Oral Pathology. Complete clinical information and follow-up data was obtained for all 45 cases. Serum calcium or phosphorous, alkaline phosphatase and parathormone levels were within normal limits. Clinical data for CGCG was reviewed for clinical presentation, radiological appearance, treatment and follow up, without knowledge of the histopathologic findings. Based on the previously established by Chuong et al., [2] we divided the 20 cases of CGCG into two groups (a) clinically non-aggressive lesions (CNA-CGCG), characterized by minimal or no symptoms, slow growth, absence of root resorption or cortical perforation and no recurrences; (b) clinically aggressive lesions (CA-CGCG) characterized by pain, rapid growth, root resorption, cortical perforation and recurrences. Hematoxylin and eosin staining Formalin-fixed, paraffin-embedded hematoxylin and eosin stained sections of 4-µm thickness were assessed under light microscope by two observers so as to divide them into (a) histologically aggressive (HA-CGCG) and (b) histologically non-aggressive (HNA-CGCG) lesions based on the established criteria by various investigators, [1][2][3][4][5] which included size, shape, characteristics of nuclei and mononuclear cells, presence or absence of osteoid and vascularity of the lesion. AgNOR staining and interpretation AgNOR staining and counting was performed according to Ploton et al. method. [11,12] The final working solution was freshly prepared by mixing of one volume of 2% gelatin in 1% formic acid solution and two volumes of 50% aqueous silver nitrate solution. Slides were incubated with this silver solution for 30 minutes at 45°C in the dark. The silver reaction product was observed as discrete black dots under routine light microscopy (×1000 magnification). One hundred nuclei from multinucleated giant cells and 100 nuclei from mononuclear cells were randomly selected and the AgNOR dots were counted for each case. Randomness was accomplished by moving the slides in a sequenced fashion where adjacent fields did not overlap. Statistical analysis All statistical analysis were performed by using SPSS 15.0 (SPSS, Chicago, IL). Data were expressed as mean ± SD. ANOVA test and Student's t-test were used for analysis. The criterion for significance was P < 0.005. RESULTS A total of 45 cases of giant cell lesions consisted of 20 cases of CGCG, 20 cases of PGCG and 5 GCT cases. We observed 10 CA-CGCG and 10 cases of CNA-CGCG lesions depending on their clinical behavior. Age and gender The median age at the time of initial diagnosis was 35, 25, 33 and 30 years for PGCG, CA-CGCG, CNA-CGCG and GCT, respectively. The ratio of males to females was 6:14, 4:6, 3:7 and 4:1 for PGCG, CA-CGCG, CAN-CGCG and GCT, respectively. A distinct gender predilection for females was noted as 62% of patients were females [ Table 1]. Location In the PGCG group, there were 16 cases in the mandible and 4 cases in the maxilla. In the CGCG group, there were 14 cases in the mandibular posterior region, posterior to the permanent mandibular first molar and 6 in the maxilla. There were 6 cases in the mandible and 4 in the maxilla in the CA-CGCG group, while the CNA-CGCG group comprised of 10 mandibular lesions. All 5 cases of GCT were of the long bones [ Table 1]. Histologic features We observed 5 HA-CGCG and 15 HNA-CGCG lesions, irrespective of their clinical features. Their cytometric parameters are summarized in Tables 2-4. Number of giant cells The mean number of giant cells per 25 HPF in PGCG, HA-CGCG, HNA-CGCG and GCT was 3.19, 4.14, 1.03 and 4.56, respectively. The difference was not statistically significant when CGCG and PGCG were compared, but when PGCG and CGCG were compared with GCT, a significant difference was observed. There was no statistically significant differences were observed between HA-CGCG and HNA-CGCG and between HA-CGCG and GCT for number of giant cells. Number of nuclei per giant cell The mean number of nuclei per giant cells in PGCG, HA-CGCG, HNA-CGCG and GCT were 26.97, 27.07, 9.85 and 150.25, respectively. No significant difference was observed between CGCG and PGCG, but a significant difference was observed when PGCG and CGCG were compared with GCT. No significant difference was noted between HA-CGCG and HNA-CGCG, but a significant difference was observed between HA-CGCG and GCT for mean number of nuclei per giant cell. Number of blood vessels The mean number of blood vessels in PGCG, HA-CGCG, HNA-CGCG and GCT was 2.19, 2.47, 0.69 and 1.10, respectively. There was no significant differences were observed between PGCG, CGCG and GCT and no difference was noted when HA-CGCG and HNA-CGCG were compared. FSA The mean FSA for PGCG, HA-CGCG, HNA-CGCG and GCT was 0.0031, 0.0033, 0.0009 and 0.0111, respectively. The FSA occupied by the giant cells was significantly greater in PGCG than CGCG. The statistically significant difference was observed when PGCG and CGCG were compared with GCT. We observed a statistically significant difference for FSA between HA-CGCG and HNA-CGCG and also between HA-CGCG and GCT [ Figures 1-3]. RSI The mean RSI for PGCG, HA-CGCG, HNA-CGCG and GCT was 0.091, 0.072, 0.020 and 0.216, respectively. The RSI was significantly greater for PGCG than CGCG. A significant difference was observed for RSI when PGCG, CGCG and GCT were compared, but no significant difference was observed between HA-CGCG and HNA-CGCG. AgNOR staining No statistically significant correlation was observed for AgNOR counts of multinucleated giant cells and mononuclear cells when PGCG, HA-CGCG, HNA-CGCG and GCT were compared with each other [Figures 4 and 5] [ Table 5]. DISCUSSION The subject of giant cell lesions of the jaws is little understood and a matter of debate because the lesions are not pathognomic and may be confused, both radiologically as well as histologically, with other lesions of the jaws. The question of whether or not "true" giant cell tumors exist in the jaws has been argued for many years and is still unresolved. Controversies surrounding the relationship between CGCG of the jaws and GCT of the long bones have revolved around their biological behavior, histopathologic features and clinical response to conservative therapy. In the present study, we have attempted to ascertain which, if any, previously described histologic features could be used to determine the aggressiveness of giant cell lesions of the jaws. Age, gender and location The mean age for PGCG, C-NA, CA and GCT appeared to be in agreement with the previous reports. [5,[13][14][15] Also, female predilection is observed in PGCG and CGCG, showing accordance with other studies. Although studies have shown that the mandibular anterior region is the most common location for CGCG, the present study is not in agreement with the same as the most of our case (70%) were in the mandibular posterior region, posterior to the permanent mandibular first molar. [1][2][3][4][5] Histology and image analysis The comparison of CGCG and PGCG in the present study showed no significant difference in the number of giant cells, nuclei in each giant cell and number of blood vessels present in the lesions. As per our findings, since PGCG and CGCG cannot be differentiated merely on routine histopathological examination, all PGCG cases should be subjected to through clinical and radiological examination to rule out possible central bone involvement. Surprisingly, we found a greater FSA and RSI for PGCG than CGCG, thus making PGCG more aggressive than CGCG. These findings are against the commonly observed surgical findings where CGCG is found to be more aggressive than PGCG. This needs further clarification, although possible role of inflammation cannot be ruled out. Using histologic criteria established by various investigators [1][2][3][4][5] , we divided 20 cases of CGCG into 5 HA-CGCG (25%) and 15 HNA-CGCG (75%) lesions. The present study observed no significant differences except a greater FSA in HA-CGCG lesions as compared to HNA-CGCG lesions, which is in agreement with other finding in the literature. [3] However, it was noted that this histologic difference was not as readily apparent as the differences in their biological behavior. It was surprising to note that although 5 HA-CGCG lesions were clinically aggressive, the remaining 5 clinically aggressive lesions were not HA-CGCG lesions. Thus, based on the present study findings, no histologic features could be used to separate aggressive lesions from those that are clinically indolent. In other words, the lesions that show clinically aggressive course may not always exhibit concurrent histological aggressiveness. To emphasize the discrepancy between histologic appearance and biological behavior, we also prefer the non-committal term "giant cell lesion" for CGCG. [5] Because of the variations in clinical behavior, histology and prognosis, it is better to designate CGCG as either potentially aggressive or nonaggressive collectively on the basis of their clinical, radiological and histological features, which will be of more help to the surgeon than to designate all of these lesions as giant cell granulomas. [5] GCT of long bone is a well-recognized neoplasm with distinctive clinical and histopathologic features. When HA-CGCG were compared with GCT, no statistically significant differences were observed in relation to number of giant cells and number of blood vessels, but we observed a striking difference in the number of nuclei per giant cells, FSA and RSI differentiating

the aggressive neoplastic nature of GCT from CGCG. [17] Thus, our findings support the viewpoint of Abrams et al. [1] and Jaffe [6] that CGCG and GCT are distinct lesions. Although cases of GCT were few, we do not support the previously proposed concept that CGCG and GCT represents a continuum of a single disease process. It is well-recognized that endothelial cells represent a phenotypically diverse group of cells that vary morphologically and functionally from site to site. Various markers have been used to assess the vascularity in giant cell granuloma and suggested more number of blood vessels at the periphery of giant cell granuloma, which may be dependent on several factors. [15] Vered et al. has demonstrated low mean microvascular volume of VEGF and b-FGF positive blood vessels in CGCG implying low angiogenic activity, thus does not support the designation of CGCG as a true proliferative vascular lesion. [16] In the present study, no statistically significant difference was found in vascularity of CGCG, PGCG and GCT; thus, all the lesions were equally vascularly dependent. The present study also examined AgNOR counts in giant cells and stromal cells of the giant cell lesions in an attempt to assess whether it could delineate the lesions of varying clinical behavior. Our results showed no statistically significant difference among the giant cell lesions of varying behavior. The differences in AgNOR counts exclusively among peripheral and central giant cell lesions of varying behavior was studied by Souza et al., [11] who observed no difference in AgNOR counts between PGCG and CGCG in their study. On the contrary, Whitaker et al. [10] observed a significant increase in the number of AgNORs in recurrent CGCG compared to non-recurrent/nonaggressive CGCG. These differences could be attributed to the technique and sample size of various studies. Further investigations are necessary to clarify the significance of AgNOR histochemical expression in the histological behavior of giant cell lesions of the jaws. In conclusion, based on the present study findings, we suggest that GCT and CGCG are distinct lesions and not a continuum of a single disease process. We also prefer the noncommittal designation of "giant cell lesion" for CGCG. [5] The observation of this study depicts that histological parameters of CGCG may not be correlating in all the cases of giant cell lesions in predicting their clinical behavior and prognosis. AgNORs alone as a proliferative marker has a limited value in assessing the proliferation potential of giant cell lesions of the jaws. Stapled intestinal anastomosis is a simple and reliable method for management of intestinal caliber discrepancy in children Purpose Popularity of minimally invasive surgeries has led to the development of stapled intestinal anastomosis for adults. The advanced instruments used in this technique are getting suitable with the small intestinal lumens of neonates and infants. We reviewed and compared the intraoperative and postoperative results of stapled and hand-sewn anastomoses in children. Methods The operative data of children who underwent stapled and hand-sewn anastomoses between March 2005 and December 2011 were collected and analyzed retrospectively. Furthermore, we compared patients who underwent anastomoses for colostomy closure of anorectal malformation (4 stapled, 9 hand-sewn) and those who underwent anastomoses for treatment of ileal atresia (3 stapled, 11 hand-sewn). Results In the 47 patients who underwent stapled anastomosis, no intraoperative complications were observed and postoperative complications included wound infection (n = 3), delayed gastric emptying (n = 1), and ileus (n = 1). No complications suggesting anastomotic dilatation were identified. It was observed that patients who underwent stapled anastomosis for colostomy takedown with caliber discrepancy had significantly shorter surgery time than those who underwent hand-sewn anastomosis. Conclusion Our results suggest that stapled anastomosis is safe and effective for various surgical diseases in neonates, infants, and children. Introduction The safety and efficacy of stapled gastrointestinal tract anastomosis in adults have been extensively documented [1]. The recent enthusiasm regarding minimally invasive techniques has led to the development of stapled intestinal anastomosis [2,3]. The advanced instruments used in this technique may be suitable for the smaller intestinal lumens of neonates and infants. Hand-sewn techniques have traditionally been used to perform intestinal anastomosis in pediatric patients in many cases. When treating small intestinal atresia and stoma closure, great discrepancy between diameters of the proximal and distal intestine caused by disuse atrophy are often observed, which may cause difficulties and complications. To overcome size discrepancy, proficiency in performing anastomosis is required when using hand-sewn techniques [4,5]. In theory, stapled functional end-to-end anastomosis does not require a special technique and does not impair the passage of intestinal contents immediately after completion because the side-to-side nature of the procedure retains the unique diameter of the target intestine and preserves patency. Stapled side-to-side functional end-to-end intestinal anastomosis is a potentially useful technique that is not affected by intestinal size discrepancy and does not require specialized surgical experience. Stapled anastomosis can be used to perform intestinal anastomosis in children easily and reproducibly. In 2009, we began performing consecutive stapled intestinal anastomosis in newborns, infants, and children. The aim of this study was to assess the feasibility and outcome of stapled intestinal anastomosis in neonates and infants, especially in those who underwent anastomosis for caliber discrepancy, such as for ileostomy/ colostomy closure and treatment for small intestinal atresia. Patients and methods All patients who underwent intestinal anastomosis with stapling instruments between April 2009 and December 2011 were retrospectively analyzed, and demographic data, intraoperative results, and outcomes were recorded. Data collected included sex, primary diagnosis, age at surgery, weight, type of anastomosis, surgery time, estimated blood loss, anastomotic leakage, intestinal obstruction, time until initial oral feeding, time of discharge, and postoperative complications. End-to-end and end-to-side stapled intestinal anastomosis methods were used. Functional end-to-end anastomosis was used for intestinal lesions, congenital intestinal atresia, stoma closure after intestinal preparation, and primary bowel resections with subsequent stapled anastomosis. End-to-side anastomosis, also called Roux-en-Y anastomosis, was performed for portoenterostomy, hepatocholangiojejunostomy, and esophagogastric dissociation. An Endocutter ETS 35 or ETS Flex 45 stapler with 1.0 or 1.5-mm staples (Johnson & Johnson K.K., Tokyo, Japan) was used to perform functional end-to-end and Roux-en-Y anastomoses. For very small intestines, an Endocutter stapler was used to perform stapled anastomosis if the intestinal lumens could admit a 22-Fr soft catheter. Stapled anastomosis was contraindicated when the intestinal lumen could not admit a 22-Fr soft catheter or when stapling would significantly compromise the total intestinal length or the ileocecal valve. An important consideration of stapled anastomosis is that the suture line of side-to-side anastomosis does not overlap when the stapler is fired across the jointed limbs, and the staple lines are only oversewn at points of bleeding for hemostasis or for reinforcement of double-stapled areas (Fig. 1). The subgroups analyzed were patients who underwent intestinal anastomoses in hand-sewn manner, undergoing colostomy closure for anorectal malformation, and undergoing primary anastomosis for congenital ileal atresia from 2005 to 2011. An absorbable suture material was used to perform hand-sewn anastomosis in an end-to-end manner. In general, the method of intestinal anastomosis was based on the surgeon's preference. Some surgeons did not perform stapled anastomosis, and others always performed stapled anastomosis, when permitted by the intestinal size and surgical circumstances. We compared patients in the hand-sewn and stapled groups who underwent either colostomy closure for anorectal malformation or treatment for ileal atresia. This comparison was performed to assess the potential benefits of stapled side-to-side functional end-to-end anastomosis for the management of significant size discrepancy between diameters of the proximal and distal intestines. Student's t test and x 2 test were used to evaluate differences between the groups. P \ 0.05 was considered to be statistically significant. In majority of the patients, staplers with 1.0-mm staples (n = 41) were used to perform stapled intestinal anastomosis. Staplers with 1.5-mm staples were used in the procedures for the remaining six children. No intraoperative complications were observed in any of the 47 cases. Postoperative complications included wound infection (n = 3), delayed gastric emptying (n = 1), and ileus (n = 1). Every complication was improved by conservative treatment. Until date, we have noted no postoperative intestinal obstruction because of strictures or anastomotic dilatation with subsequent stasis/overgrowth related to anastomosis. The follow-up period ranged from 3 months to 3 years after surgery. Patients who underwent stapled functional or hand-sewn end-to-end anastomosis with significant size discrepancy were compared. In 13 infants who underwent colostomy closure for anorectal malformation without any other additional surgeries (4 stapled and 9 hand-sewn), no differences in infant size or age was observed between those who underwent stapled and hand-sewn anastomoses; however, the mean surgery time in infants who underwent stapled anastomosis was significantly lower than that in infants who underwent hand-sewn anastomosis. The length of postoperative hospital stay and length until initial oral feeding were shorter in patients who underwent stapled anastomosis, but were not significantly different from those who underwent hand-sewn anastomosis ( Table 2). Among the 14 newborns who were treated for ileal atresia (3 stapled and 11 hand-sewn), there were no significant differences in any of the parameters; however, surgery time, length of postoperative hospital stay, and length until initial oral feeding were shorter in those who underwent stapled anastomosis than those who underwent hand-sewn anastomosis. Two patients who underwent hand-sewn anastomoses suffered from anastomotic strictures and delayed oral feeding, but none of the patients who underwent stapled functional end-to-end anastomosis had anastomotic strictures (Table 3). Discussion Stapling devices have a long history in surgery. The safety, efficacy, and technique of stapled gastrointestinal tract anastomosis in adults have been extensively documented since 1978. Several studies, including a Cochrane review of 6 trials that involved 955 ileocolic anastomoses, have suggested that stapled intestinal anastomosis has fewer leaks with no differences in surgery time or the incidence of stricture or wound infection compared with those of hand-sewn anastomosis [6]. On the other hand, hand-sewn techniques have traditionally been used to perform intestinal anastomosis in children. In 1974, Talbert et al. [7] described a modification of the Duhamel procedure that used a linear stapling-dividing instrument. The mechanical stapler for intestinal anastomosis in children had been used in surgeries for a limited number of diseases, such as Hirschsprung's disease. Few studies have documented stapled anastomosis for other diseases in children. In 1995, Powell published a series of seven successful intestinal anastomoses that used mechanical staplers in infants less than 4 months old [8]. In 2008, Wrighton et al. reported an experience with stapled intestinal anastomosis in infants less than 1 year old, and compared surgical data and outcome with those of infants who underwent hand-sewn anastomosis [9]. In the present study, stapled intestinal anastomosis was shown to have shorter surgery time than [10] documented 64 consecutive stapled intestinal anastomoses. It concluded that stapled anastomosis was an effective approach applicable to various surgical diseases in newborns and infants. In the present study, we found that stapled anastomotic complications were not more severe than those associated with hand-sewn anastomosis. The recent enthusiasm for minimally invasive surgeries and standardization of surgical procedures may present a need for a simple and reproducible technique of anastomosis in children. In particular, size discrepancy between diameters of the proximal and distal intestines can complicate the surgical anastomotic technique. Several geometrical methods have been used to manage this discrepancy in order to reduce the risk of anastomotic leakage and stenosis [4,5]; however, there is no ideal technique for the management of size discrepancy. Almost all of the methods should be performed according to the surgeon's experience and skills. In addition, the nature of hand-sewn techniques makes the site of anastomosis edematous to a greater or lesser extent immediately after completion; however, stapled side-to-side functional endto-end intestinal anastomosis does not require direct access to the entrance of a disused and distal intestine to maintain patency. The side-to-side stapled anastomosis yields theoretically wider anastomosis than those resulting from the hand-sewn technique. This may account for the lower incidence of strictures found in some adults, although no significant differences were found in obstructive outcomes in our study. We found that surgery time was reduced during colostomy takedown, and anastomotic complications were consistently not more severe than those of the hand-sewn technique. For treatment of congenital ileal atresia, no significant differences were observed in surgery time, length until initial oral feeding, and length of postoperative stay, but our results may

indicate the low number of cases. Further studies involving more patients are needed to confirm that surgery time, length until initial oral feeding, and length of postoperative stay are significantly reduced in the stapled group for the patients with caliber discrepancy. Our data may indicate a strong tendency toward good patency for anastomosis with size discrepancy. Functional end-to-end anastomosis may be a superior procedure that anyone can perform simply for intestinal anastomosis with caliber discrepancy compared with hand-sewn anatomosis that are dependent on surgical skills. Stapled anastomosis could not be performed when the intestinal lumen could not admit a 22-Fr soft catheter. The size limitation of the stapler is a major concern; however, in the review of Wrighon et al. [9], 25 anastomoses were performed in infants between 600 and 1000 g. There may be infrequent cases in which the stapler may not be suitable. There is only one report of adverse outcomes after stapled intestinal anastomosis in 2-month-old infants and 3-year-old children with partial obstructions after stapled anastomosis [11]. Anastomosis was noted to have dilated significantly, which caused subsequent volvulus. The authors concluded that a functional end-to-end anastomosis may be susceptible to massive dilatation because of the pouch that can be created in the anastomotic region. We must note that functional end-to-end anastomosis should not be excessively long to prevent pouch formation, stasis, and subsequent dilatation. Our findings showed that in infants who underwent stapled intestinal anastomosis, especially those who underwent colostomy closures with caliber discrepancy, had shorter surgery time than those who underwent handsewn anastomosis. There were no differences in adverse outcomes. Stapled functional end-to-end anastomosis does not require a special technique, even if caliber discrepancy is apparent. In conclusion, stapled anastomosis, when permitted by intestinal size, is one of the most simple and reliable methods for intestinal anastomosis, even in newborns and infants with caliber discrepancy. This technique may become an alternative to traditional hand-sewn techniques in small children. In the future, the technique's impact on long-term outcome should be determined to verify these findings. Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited. RNA-Binding Proteins Implicated in Mitochondrial Damage and Mitophagy The mitochondrial lifecycle comprises biogenesis, fusion and cristae remodeling, fission, and breakdown by the autophagosome. This cycle is essential for maintaining proper cellular function, and inhibition of any of these processes results in deterioration of bioenergetics and swift induction of apoptosis, particularly in energy-craving cells such as myocytes and neurons. Regulation of gene expression is a fundamental step in maintaining mitochondrial plasticity, mediated by (1) transcription factors that control the expression of mitochondrial mRNAs and (2) RNA-binding proteins (RBPs) that regulate mRNA splicing, stability, targeting to mitochondria, and translation. More recently, RBPs have been also shown to interact with proteins modulating the mitochondrial lifecycle. Importantly, misexpression or mutations in RBPs give rise to mitochondrial dysfunctions, and there is strong evidence to support that these mitochondrial impairments occur early in disease development, constituting leading causes of pathogenesis. This review presents key aspects of the molecular network of the disease-relevant RBPs, including transactive response DNA-binding protein 43 (TDP43), fused in sarcoma (FUS), T-cell intracellular antigen 1 (TIA1), TIA-related protein (TIAR), and pumilio (PUM) that drive mitochondrial dysfunction in the nervous system. INTRODUCTION Adenosine triphosphate (ATP) production by mitochondria is essential for most cellular activities. In addition to ATP generation, however, mitochondria are heavily involved in calcium homeostasis, production and modulation of reactive oxygen species (ROS), and in the execution of apoptosis. Mitochondria are highly dynamic organelles characterized by rapid movement and undergo some five fusion-fission cycles every hour to properly maintain their function (Twig et al., 2008;Pernas and Scorrano, 2016). Mitochondrial fusion is the process in which mitochondria fuse together to spread metabolites, proteins, and DNA throughout the network to maintain mitochondrial (mt) DNA replication and oxidative phosphorylation (OXPHOS) capacity (Chen et al., 2005(Chen et al., , 2010Silva Ramos et al., 2019). It is mediated by optic atrophy 1 (OPA1), and mitofusin-1 and 2 (MFN1/2) (Chen et al., 2003;Olichon et al., 2003). Mitochondrial fission, on the other hand, is the process in which mitochondria divide to separate dysfunctional/depolarized mitochondrial sections in a daughter mitochondrion that will be targeted by autophagy, otherwise known as mitophagy (Twig et al., 2008). It is primarily regulated by dynamin-related protein 1 (DRP1) and dynamin-2 (DYN2) with the aid of adaptor proteins mitochondrial fission 1 (FIS1), mitochondrial fission factor (MFF), and mitochondrial dynamics proteins 49 and 51 (MiD49/51) (Smirnova et al., 1998;Yoon et al., 2003;Gandre-Babbe and van der Bliek, 2008;Otera et al., 2010;Palmer et al., 2011;Lee et al., 2016). Additionally, folds of the inner membrane of the mitochondrion (known as cristae) that are formed to increase the surface area for housing the electron transport chain (ETC) complexes and ATP synthase continuously remodel to improve mitochondrial function (Enriquez, 2016). Collectively, these mitochondrial morphology events comprise the mitochondrial life cycle. Mitochondrial dynamics are altered according to the energy requirements of the cell, nutrient availability, stress, and aging, and depend on transcriptional and post-transcriptional mechanisms. While transcription factors mediate the expression of nuclear and mitochondrial genes, RNA-binding proteins (RBPs) regulate splicing, stability, localization, and translation events. More recently, RBPs have been shown to interact directly with proteins on mitochondrial surface, too. In this review, we present findings that implicate RBPs misregulation in mitochondrial damage. We focus on transactive response DNAbinding protein 43 (TDP43), fused in sarcoma (FUS), T-cell intracellular antigen 1 (TIA1), TIA-related protein (TIAR), and pumilio (PUM), as there is substantial experimental data that show their involvement in mitochondrial pathology. General features, such as the neurological symptoms associated with their perturbation, molecular and cellular function, target mRNAs and subcellular localization have been described in our previous review, and are thus not described here (Ravanidis et al., 2018). TDP43 Mutations or deregulation of transactive response DNA binding protein 43 (TDP43 or TARDBP) expression have been associated with a spectrum of neurodegenerative diseases including frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS) (Ravanidis et al., 2018). Electron microscopy (EM) analyses of patient brain samples as well as cellular and animal models of TDP43 proteinopathy revealed prominent mitochondrial impairment, including abnormal cristae architecture and diminished cristae surface area (Wang et al., 2019). Further, increased TDP43 expression induced mitochondrial dysfunction, including decreased mitochondrial membrane potential and elevated production of ROS (Wang et al., 2019; Figure 1). Alzheimer's disease (AD) pathology includes mitochondrial perturbations such as alterations in respiratory function, mitochondrial biogenesis, and mitophagy (Cai and Tammineni, 2017;Chakravorty et al., 2019). Using the APP/PS1 transgenic mouse model co-expressing the familial AD Swedish mutations (APP K 595N,M 596L ) and mutant human presenilin 1 (PSEN1-E9) under stress conditions, Davis et al. (2018), found increased accumulation of the N-terminal (27 kDa, N27) and C-terminal (30 kDa, C30) fragments of TDP43 in mitochondria. Immunoprecipitation from cortex lysates, to reveal the interacting partners of TDP43, showed enrichment for mitochondrial proteins, including prohibitin-2 (PHB2) and voltage-dependent anion channel 1 (VDAC1). PHB2 is a scaffold protein and a mitophagy receptor located in the inner mitochondrial membrane. It is involved in targeting mitochondria for autophagic degradation by interacting with microtubule-associated protein 1A/1B-light chain 3 (LC3) conjugated to phosphatidylethanolamine (LC3-II), which is found in autophagosomal membranes (Lahiri and Klionsky, 2017;Wei et al., 2017). Accordingly, PHB2 knockdown was shown to drastically reduce mitochondrial clearance (Wei et al., 2017). In addition, PHB2 is involved in mitochondrial membranes' fusion by stabilizing indirectly the long forms of dynamin-like GTPase OPA1, which mediates mitochondrial inner membrane fusion and cristae morphogenesis. Loss of PHB2 impairs the stability of OPA1, affects mitochondrial ultrastructure, and induces the perinuclear clustering of mitochondria (Merkwirth et al., 2012). Overexpression of TDP43 was found to increase PHB2 levels, whereas TDP43 knockdown reduced PHB2 and LC3-II expression in HEK293T cells treated with carbonyl cyanide m-chlorophenylhydrazone (CCCP), an inducer of mitophagy (Davis et al., 2018). Accordingly, an increase in the E3 ubiquitin ligase parkin (PRKN)-positive punctate staining (indicative of mitophagy) in cells treated with CCCP was observed, which was enhanced with TDP43 overexpression and reduced when TDP43 levels were knocked down (Davis et al., 2018). In parallel with these findings, in NSC34 cells that exhibit motor neuron features, overexpression of full length or C-terminal fragments of TDP43 (TDP25 and TDP35) led to increased levels of LC3-II and decreased levels of autophagy receptor p62 (SQSTM1) (Hong et al., 2012). Collectively, these results suggest that TDP43 overexpression is linked to enhanced mitophagic flux. TDP43 expression also affects mitochondrial dynamics. Using transgenic mice expressing full-length human TDP43, Xu et al. (2010) observed aggregates of mitochondria, with decreased cristae and vacuoles within the mitochondrial matrix, adjacent to the nucleus, accompanied by enhanced levels of FIS1 and profission phosphorylation of DRP1 at Ser616, both key mediators of the mitochondrial fission machinery (Taguchi et al., 2007). Conversely, a marked reduction in MFN1 expression, which plays an essential role in mitochondrial fusion, was observed (Xu et al., 2010). Corroborating evidence came from Wang et al. (2013), showing that overexpression of wild-type TDP43 in primary motor neurons reduced mitochondrial length and density in neurites. Further, transgenic mice overexpressing wildtype or mutant TDP43 displayed significantly shorter, smaller, and damaged mitochondria (Wang et al., 2013). In contrast, artificial miRNA-mediated suppression of TDP43 in primary motor neurons resulted in significantly increased mitochondrial length and density in dendrites (Wang et al., 2013). In addition, co-expression of MFN2 with mutant TDP43 completely prevented all TDP43-induced mitochondrial abnormalities (Wang et al., 2013). Informative findings have also arisen from work in Drosophila. Khalil et al. (2017) found that overexpression of human wild-type TDP43 in neurons resulted in abnormally small mitochondria. The mitochondrial fragmentation was correlated with a specific FIGURE 1 | Mitochondrial perturbations induced by TDP43. PRKN in complex with HDAC6, ubiquitinates nuclear TDP43 promoting its cytoplasmic localization and proteasomal degradation. However, as revealed from research in aging or neurodegenerative diseases, TDP43 often persists in the cytosol and forms aggregates. Excess cytosolic TDP43 interacts with VDAC1, located in the outer mitochondrial membrane, but it is still unclear if interferes with its functions. Polyubiquitination of VDAC1 by PRKN is essential for driving mitophagy. Moreover, cytosolic TDP43, translocated to the outer mitochondrial membrane, directly interacts with PHB2 and, in parallel, increases its protein levels. PHB2 is known to interact with LC3-II to induce mitophagy. PHB2 is also involved in mitochondrial membranes fusion by stabilizing indirectly the long forms of OPA1. Additionally, TDP43 directly interacts with MFN2, a mitochondrial membrane protein regulating mitochondrial fusion, and possibly stabilizes its expression. Concurrently, TDP43 leads to reduced levels of another fusion protein, MFN1, and increases levels of FIS1 and DRP1 phosphorylated at Ser616, proteins promoting mitochondrial fission. Finally, TDP43 downregulates PRKN mRNA and protein levels, and impairs the proteasome, leading to the accumulation of cleaved PINK1 (cPINK1) in the cytosol. During stress conditions cPINK1 aggregates recruit PRKN to the mitochondria launching mitophagy in otherwise healthy mitochondria (non-selective mitophagy). decrease in the levels of Marf, the MFN ortholog in Drosophila. Importantly, overexpression of Marf or inactivation of profission Drp1 ameliorated the defects (Khalil et al., 2017). Similar mitochondrial dysfunctions were observed in another Drosophila study, and likewise the mitochondrial fission defects were rescued by co-expression of mitochondrial pro-fusion genes Marf, Opa1, and the dominant negative mutant form of Drp1 (Altanbyek et al., 2016). Using immunoprecipitation from cortical human brain tissue, TDP43 was found to also interact directly with pro-fusion factor MFN2 (Davis et al., 2018). Knocking down TDP43 in HEK293T cells led to a reduction in MFN2 expression levels, whereas TDP43 overexpression marginally increased MFN2 levels (Davis et al., 2018). Previously, MFN2 repression was shown to inhibit mitophagy and result in the accumulation of damaged mitochondria in muscles during aging (Sebastian et al., 2016), indicating that changes in the balance of mitochondrial fission/fusion machinery affect not only architecture dynamics but mitophagy as well. Under steady-state conditions, PTEN-induced kinase 1 (PINK1), a mitochondrial serine/threonine kinase, is imported in the inner mitochondrial membrane where it is cleaved by the serine protease presenilin-associated rhomboid-like (PARL) (Yamano and Youle, 2013). Following cleavage, PINK1 is released into the cytosol where it is recognized by the N-end rule E3 enzymes, ubiquitin protein ligase E3 component N-Recognin 1 (UBR1), UBR2, and UBR4 for constitutive and rapid proteasome-mediated degradation (Yamano and Youle, 2013). When

mitochondria are damaged, PINK1 is not cleaved and is subsequently anchored to the outer mitochondrial membrane where it recruits and activates, via phosphorylation, the E3 ubiquitin ligase PRKN to trigger selective mitophagy (Pickrell and Youle, 2015). Both PINK1 and PRKN exhibit mutations that have been linked to autosomal recessive early-onset Parkinson's disease (PD) (Kitada et al., 1998;Hatano et al., 2004;Rohe et al., 2004;Valente et al., 2004). Using human TDP43 knock-in flies, TDP43-infected mouse primary neurons, TDP43-transfected HEK293T cells, and TDP43 Q331K transgenic mice, Sun et al. (2018), showed that TDP43 downregulates PRKN mRNA and protein levels via mechanisms requiring both the RNA-binding and the protein-protein interaction functions of TDP43. Unlike PRKN, TDP43 did not regulate PINK1 at the mRNA level. Instead, overexpression of TDP43 lead to cytosolic aggregates of cleaved PINK1 due to impaired proteasomal activity, and compromised mitochondrial respiration (Sun et al., 2018). Upregulation of PRKN expression or RNAi-mediated downregulation of PINK1 levels suppressed TDP43-induced degenerative phenotype in Drosophila, indicating that PRKN and PINK1 are important components of TDP43-induced proteinopathy (Sun et al., 2018). Additionally, it has been reported that accumulation of cleaved PINK1 induces non-selective mitophagy and non-apoptotic cell death (Lim et al., 2015). In this article, it is shown that cleaved PINK1 cytosolic aggregates trigger PRKN translocation to healthy mitochondria, leading to non-selective mitophagy (Lim et al., 2015). In another study, PRKN was shown to ubiquitinate nuclear TDP43, and together with HDAC6, promote cytosolic TDP43 accumulation reminiscent of ubiquitinated wild-type or mutant TDP43 found in the cytosol in several neurodegenerative diseases (Hebron et al., 2013). Moreover, Prkn knockout mice exhibited high levels of TDP43, underscoring an indispensable role for PRKN in mediating TDP43 clearance and cytosolic localization (Wenqiang et al., 2014). A dual regulation of mitophagy and apoptosis by PRKN via VDAC1, a direct partner of TDP43 in mitochondria (Davis et al., 2018), has also been revealed. Previously, VDACs have been shown to mediate mitophagy via recruitment of PRKN in the mitochondria (Geisler et al., 2010;Sun et al., 2012;Li et al., 2014). More recently, PRKN was shown to monoor poly-ubiquitinate VDAC1. Polyubiquitination was required for PRKN-mediated mitophagy, whereas mono-ubiquitination was required for mitochondrial calcium influx and apoptosis (Ham et al., 2020). The role of TDP43 in the mono-or polyubiquitination of VDAC1 by PRKN has yet not determined. FUS Mutations in the FUS or translocated in liposarcoma (FUS/TLS) gene give rise to familial ALS and occasionally FTLD-FUS, both displaying FUS-positive inclusions (Ravanidis et al., 2018). Interestingly, however, in the majority of FTLD-FUS cases, no FUS mutations have been identified, but rather an increase in wild-type FUS expression highlighting a dose-dependent role in neurodegeneration (Sabatelli et al., 2013;Deng et al., 2015). Several systems have been used to model FUS-proteinopathies, in all of which wild-type or ALS-mutant FUS overexpression led to progressive neurodegeneration reiterating findings in patients (Huang et al., 2011;Ravanidis et al., 2018). Several studies implicate mitochondrial damage as an early event that precedes cell death in FUS proteinopathies (Deng et al., 2015So et al., 2018; Figure 2). Deng et al. (2015) showed that overexpression of wild-type or ALS-associated mutant FUS in HEK293 cells reduced the mitochondrial membrane potential and increased the production of mitochondrial ROS. Increased levels of ROS drive mitochondrial translocation of the pro-fission protein DRP1 in ASTCa1 cells, leading to mitochondrial fragmentation (Wu et al., 2011). Likewise, Deng et al. (2015) observed mitochondrial fragmentation in wild-type or mutant FUS-overexpressing HT22 cells, cultured neurons, and transgenic fly motor neurons. They then performed EM to compare healthy control and FTLD-FUS brain mitochondria. While in controls most mitochondria appeared healthy with well-organized cristae as packed-stacks of membrane sheets and with only a few FUS-immunostaining signals, in FTLD patients mitochondria displayed a marked loss or disruption of cristae with frequent detection of "onion-like" deformed shapes and FUS-immuno-positive signals, in close association with the mitochondria (Deng et al., 2015). Similarly, So et al. (2018), using transgenic hFUS mice, revealed that FUS, which is abundant at the pre-synaptic terminal of the neuromuscular junction (NMJ), caused a significant decrease in the number of mitochondria, while many of those that remained had pronounced abnormalities including disorganized cristae and large vacuoles as early as postnatal day 15. Interestingly, mitochondria in the post-synaptic muscle endplate were abundant and of normal appearance, consistent with other studies demonstrating that mitochondria at distal axon terminals undergo the earliest damage in the course of ALS disease (Magrane et al., 2012;Ruffoli et al., 2015). Deng et al. (2015) moved on to demonstrate that heat shock protein 60 kDa (HSP60), an ATP-dependent mitochondrial chaperone, interacted with FUS and mediated FUS mitochondrial localization. siRNA-based downregulation of HSP60 levels reduced mitochondrially localized FUS without altering its overall cellular levels; in fact, levels of nuclear and cytoplasmic FUS increased as a result. Accordingly, HSP60 downregulation increased the size of mitochondria and partially rescued mitochondrial defects as well as neurodegenerative phenotypes caused by wild-type or mutant FUS overexpression in transgenic fly photoreceptors. Finally, they found that HSP60 protein levels were elevated in the brains of FTLD-FUS patients (Deng et al., 2015). These observations indicate that HSP60 plays an important role in mediating the translocation of excess FUS in mitochondria, a critical early step in mitochondrial impairment and thereafter neurodegeneration. Additional mechanisms by which FUS induces mitochondrial damage have been brought forward. Wild-type or mutant FUS were found to interact with the mitochondrial ATP synthase β-subunit (ATP5B) , which is the essential catalytic subunit of mitochondrial ATP synthase FIGURE 2 | Mitochondrial perturbations induced by FUS. HSP60 mediates FUS translocation to the outer mitochondrial membrane. Mitochondrial-localized FUS binds to the β subunit of the F1 catalytic domain of ATP synthase (Complex V). The binding leads to disassembly of the F1 domain and accumulation of unassembled ATP synthase subunits, including ATP5B, which activates the UPR mt response leading to non-apoptotic cell death. Additionally, disruption of the F1 domain of the ATP synthase complex results in impaired ATP production and thereafter, deformed cristae. FUS induces mitochondrial perturbations in several other manners while being in excess in the cytoplasm. Mutant FUS binds to mature mRNAs coding for important mitochondrial proteins including Kif5b, Dnm1l, and Csde1, inhibiting their translation. This inhibition progressively leads to mitochondrial fission. Excess FUS drives the accumulation of PINK1 and PRKN proteins. As a consequence, RHOT1, also known as Miro1, a component of the primary motor/adaptor complex that anchors kinesin to the mitochondrial surface and a direct target of PRKN, is ubiquitinated leading to disruption in axonal motility and retrograde transport of mitochondria. Additionally, FUS has an impact on the OXPHOS process by deregulating the expression of the subunits NDUFS3 and UQCRC2 of Complexes I and III, respectively. OXPHOS deregulation leads to respiratory impairment and subsequent ATP production deterioration and deformed cristae. Finally, FUS decreases the levels of ser9 phosphorylation in GSK-3β, leading to increased GSK-3β activity. Activated GSK-3β deregulates the interaction of mitochondrial tethered membrane protein PTPIP51 and the inner protein of the ER, VAPB, disrupting mitochondria-ER associations. The ER-mitochondria disruption decreased Ca 2+ uptake by mitochondria following release from ER stores, resulting in reduced ATP production and deformed mitochondria. (Wang and Oster, 1998). FUS binding to ATP5B disrupted the assembly of ATP synthase super-complex, suppressing ATP synthesis . Previously, ATP synthase complex assembly has been closely associated with mitochondrial cristae formation (Paumard et al., 2002). ATP synthase mutants show disorganized cristae in yeast (Paumard et al., 2002;Strauss et al., 2008), which could explain the disruption or loss of cristae observed following FUS overexpression (Deng et al., 2015So et al., 2018). On top of that, whereas ATP synthase complex activities and formation were decreased, mitochondrial ATP5B protein levels were increased in FUS-overexpressing HEK293 cells and flies . This has given rise to an accumulation of unassembled ATP synthase subunits, including ATP5B, which activated the mitochondrial unfolded protein response (UPR mt ) . UPR mt is an adaptive mechanism to ensure mitochondrial proteostasis and quality control. However, excessive activation of UPR mt following severe or extended mitochondrial stresses can induce non-apoptotic neurodegeneration (Martinez et al., 2017). That is likely the case here, as downregulation of UPR mt genes ameliorated wild-type or mutant FUS-induced retinal degeneration in flies . A different perspective was brought forward by Stoica et al. (2016). They found that wild-type or ALS-associated mutant FUS decreased the endoplasmic reticulum (ER)-mitochondria associations in NSC34 motor neuron cells and in spinal cord motor neurons from FUS transgenic mice (Stoica et al., 2016). Specifically, they showed that FUS disrupted the interaction between the integral ER protein, vesicle-associated membrane protein-associated protein B (VAPB), and the outer mitochondrial membrane protein, protein tyrosine phosphatase interacting protein 51 (PTPIP51) that serve as scaffolds to tether the two organelles (De Vos et al., 2012). This disruption was accompanied by a decrease in Ca 2+ uptake by the mitochondria following its release from ER stores. Since mitochondrial ATP production is linked to Ca 2+ levels (De Vos et al., 2012), uncoupling of ER-mitochondria by FUS resulted in impaired ATP production (Stoica et al., 2016). Immunoprecipitation revealed that FUS did not bind either VAPB or PTPIP51. Instead, FUS reduced the inhibitory phosphorylation of ser9 in GSK-3β, resulting in its activation (Stoica et al., 2016). Previously, the same group has shown that GSK-3β inhibition increases the VAPBPTPIP51 interaction; however, the precise mechanism is not yet determined (Stoica et al., 2014). Hence, using the GSK-3β inhibitors AR-A014418 and CT99021H, they showed that FUS-induced defects in ER-mitochondria association as well as mitochondrial Ca 2+ levels were restored (Stoica et al., 2016). Considering that damaged ER-mitochondria associations have also been described in AD and PD (Zampese et al., 2011;Cali et al., 2012;Hedskog et al., 2013;Ottolini et al., 2013;Guardia-Laguarta et al., 2014), this indicates that perturbation of the ER-mitochondrial axis may be a general feature in neurodegeneration. Another way by which disease-causing FUS mutations induce mitochondrial impairment and neurotoxicity was deciphered by Nakaya and Maragkakis (2018). They showed that unlike wild-type FUS that predominantly binds pre-mRNAs, the ALS-associated R495X FUS mutant binds mature mRNAs in the cytoplasm (Nakaya and Maragkakis, 2018). Although R495X binding had only a moderate effect on mRNA levels, it significantly reduced the translation of mRNAs that are associated with mitochondrial function such as Kif5b, Dnm1l, and Csde1 (Nakaya and Maragkakis, 2018). These alterations were accompanied by a reduction in mitochondrial size, as previously reported (Deng et al., 2015So et al., 2018). Importantly, by introducing multiple mutations in the RRM RNA-binding domain of R495X FUS, to reduce its RNA-binding ability (Daigle et al., 2013), they partially abrogated R495Xinduced effects on mRNA translation, mitochondrial size, and neurotoxicity, uncovering a novel RNA-mediated pathway of FUS proteinopathy (Nakaya and Maragkakis, 2018). Insights into the role of PRKN in FUS-mediated mitochondrial dysfunction were revealed by Cha et al. (2020). Using Drosophila flies, they showed that when PRKN was cooverexpressed with FUS, it was able to rescue locomotive defects (Cha et al., 2020). At the cellular level, PRKN co-overexpression did not lead to any significant mitochondrial morphological improvements compared to the flies only overexpressing FUS; in fact, PRKN overexpressed alone also exhibited fragmented mitochondria (Cha et al., 2020). Instead, they found that PRKN restored the expression of mitochondrial subunits I (NDUFS3) and III (UQCRC2), which are significantly decreased in FUSinduced ALS flies. As a result, flies overexpressing both FUS and PRKN had partially restored ATP levels (Cha et al., 2020). Interestingly, complex III is one of the five mitochondrial distinct multi-subunit complexes (I-V) whose activity is reported to be dampened in spinal cord tissues of ALS patients (Sasaki and Iwata, 2007). Taken together, these observations demonstrated a protective role of PRKN in FUS-induced mitochondrial dysfunction. Contradictory findings concerning the role of PRKN in FUS-mediated defects have also been reported (Chen et al., 2016). Overexpression of wild-type or mutant FUS in HEK293 cells lead to the accumulation of PINK1 and PRKN proteins (Chen et al., 2016). As a consequence, the Ras homolog family member T1 (RHOT1, also known as Miro1), a component of the primary motor/adaptor complex that anchors kinesin to the mitochondrial surface and a direct target of PRKN, was ubiquitinated leading to the disruption in axonal motility and retrograde transport of mitochondria (Chen et al., 2016). Previously, Miro1 was shown to be phosphorylated by PINK1, which promoted its proteasomal degradation by PRKN Liu et al., 2012). RNAi-mediated downregulation of both PINK1 and PRKN restored locomotive defects in FUS transgenic

flies (Chen et al., 2016). As the PINK1/PRKN pathway also promotes mitochondrial fission (Poole et al., 2008;Yu et al., 2011), Chen et al. (2016 proposed that the upregulation of PINK1 and PRKN is partly responsible for mitochondrial fragmentation induced by wild-type and mutant FUS overexpression. TIA1 AND TIAR T-cell intracellular antigen 1 and TIA-related/like protein share an extended identity in the amino acid sequence, and like other RBPs, they translocate to the cytoplasm following cellular stress conditions forming stress granules (SG) (Ravanidis et al., 2018). Missense mutations in the TIA1 gene cause both Welander distal myopathy (WDM) (Hackman et al., 2013) and ALS, characterized by delayed SG disassembly and accumulation of non-dynamic SGs that harbor TDP43 (Mackenzie et al., 2017). Early in the analysis of TIA1 cell models, it became evident that TIA1 and TIAR affect mitochondrial dynamics (Figure 3). Using EM, Carrascoso et al. (2017) found that TIA1 or TIAR overexpression in HEK293 cells promoted mitochondrial clustering and fission. Closer inspection of mitochondria revealed changes in cristae organization, with many cristae having a slightly wider and more loosely organized intermembrane space than those of control cells (Carrascoso et al., 2017). Further, the mtDNA/nDNA ratio was similar between control and TIA1-or TIAR-overexpressing cells, suggesting that the changes in mitochondria were linked to reorganization dynamics rather than de novo mitochondrial FIGURE 3 | Mitochondrial perturbations induced by TIA1 and TIAR. TIA1 and TIAR mediate exon 4b inclusion in the pre-mRNA of OPA1 generating the OPA1 variant 5, which is associated with a smaller mitochondria, mitochondrial clustering, and remodeling around the perinuclear region. Further, cytosolic TIA1 enhances the translation of MFF mRNA and promotes DRP1 translocation to mitochondria leading to mitochondrial fragmentation. In parallel, TIA1 and TIAR induce modest downregulation of the pro-fusion proteases OMA1, YMEL1, and MFN2, further contributing to the pro-fission phenotype and mitophagy. TIA1 also has pro-apoptotic properties inhibited by FAST. FAST is released from its mitochondrial tether during stress, a process mediated by tBID and BAX. Following its release, FAST binds to TIA1 and prevents TIA1-mediated silencing of mRNAs encoding inhibitors of apoptosis, such cIAP-1 and XIAP. When TIA1 is in excess, it binds FAST and obstruct its anti-apoptotic events. Finally, TIA1 and TIAR increase LC3-II levels, yet the mechanism is unknown, leading to increased mitophagic events. biogenesis. Mitochondrial respiration and ATP production were impaired as a result (Carrascoso et al., 2017). When switched from glucose to galactose or fatty acids as cell culture substrates, to promote a switch from glycolysis to OXPHOS and determine the degree of mitochondrial dependency in cell growth, TIA1-or TIAR-overexpressing cells showed reduced proliferation rates (Carrascoso et al., 2017). Additionally, they displayed increased mitophagy rates and ROS production. Enhanced cleaved poly (ADP-ribose) polymerase 1 (PARP1) levels and delay in G1/S cycle phase transition, phenomena of early apoptosis, correlated with increased mitophagy (Carrascoso et al., 2017). Increased mitochondrial DNA damage were also observed in TIA1-or TIAR-overexpressing cells following H 2 O 2 treatment suggestive of impaired antioxidant defense (Carrascoso et al., 2017). Collectively, these results indicate that TIA1 or TIAR provoke respiratory deficiency and compromised mitochondrial function (Carrascoso et al., 2017). Mechanistically, TIA1 and TIAR mediated exon 4b inclusion in the pre-mRNA of OPA1 generating the OPA1 variant 5. OPA1 is a dynamin-like GTPase that regulates cristae junction numbers and stability, and the different OPA1 protein isoforms (eight in humans) relay instructions that help determine fusion, build cristae, and tune the morphology of mitochondria (Olichon et al., 2007;Song et al., 2007;Glytsou et al., 2016). OPA1v5, specifically, promotes mitochondrial clustering and remodeling around the perinuclear region (Song et al., 2007;Carrascoso et al., 2017). Ablation of TIA1 or TIAR in mouse embryonic fibroblasts (MEFs) favored the expression of short forms of OPA1, and the appearance of elongated mitochondria indicative of fusion phenotypes (Carrascoso et al., 2017). Furthermore, knockdown of OPA1 or overexpression of OPA1v5 triggered mitochondrial clustering mimicking TIA1 or TIAR effects (Carrascoso et al., 2017). In addition, proteases associated with fusion (OMA1, YMEL1, and MFN2) were modestly downregulated in TIA1-or TIAR-overexpressing cells, whereas the fission-associated protein MFF was slightly upregulated, further contributing to the profission phenotype (Carrascoso et al., 2017). Tak et al. (2017) independently reported similar mitochondrial phenotypes following TIA1 modulation, but provided different mechanistic insights. Likewise, they showed that TIA1 overexpression in CHANG liver cells enhanced mitochondrial fission, while downregulation enhanced mitochondrial elongation. In addition, TIA1 downregulation increased mitochondrial activity, including the rate of ATP synthesis and oxygen consumption (Tak et al., 2017). Further, they identified MFF 3'UTR as a direct target of TIA1 and showed that TIA1 promoted mitochondrial fragmentation by enhancing MFF translation. Accordingly, Tia1-null MEF cells had decreased levels of MFF and mitochondrial DRP1, thereby leading to mitochondrial elongation (Tak et al., 2017). Studies investigating the p.E384K mutant form of TIA1 (TIA1 WDM ) responsible for WDM revealed similar findings (Carrascoso et al., 2019). TIA1 WDM overexpression in HEK293 cells resulted in mitochondrial fission and mitochondrial swelling with an abnormal distribution of cristae. This led to decreased mitochondrial membrane potential and enhanced redox status (Carrascoso et al., 2019). Additionally, there was an increase in the formation of autophagosomes and autolysosomes, as well as mitophagic and apoptotic rates (Carrascoso et al., 2019). Taken together, these results revealed that similar to wild-type TIA1, disease-associated mutant TIA1 overexpression has a negative impact on mitochondrial dynamics, leading to mitochondrial dysfunction and cell death. Sanchez-Jimenez and Izquierdo (2013) used Tia1 and Tiar knock-out MEFs to study the molecular and cellular consequences. They found that TIA1 and TIAR knockout cells had two to threefold more mitochondria, six to sevenfold higher mitochondrial membrane potential, and twofold higher ROS levels. Mitochondria had atypical morphology, with some being enlarged and others being fragmented (Sanchez-Jimenez and Izquierdo, 2013). These alterations were associated with nuclear DNA damage, revealed by 8-hydroxy-2 -deoxyguanosine (8-oxo-dG) staining, and high rates of autophagy, possibly as a compensatory mechanism toward survival. Consequently, TIA1 and TIAR knockout MEFs displayed defects in cell proliferation and increased cell death (Sanchez-Jimenez and Izquierdo, 2013). A different perspective by which TIA1 is promoting apoptosis was brought forward by Li et al. (2004b). They proposed that during stress, TIA1 silences (Kedersha et al., 2000;Anderson and Kedersha, 2002), among others, the translation of mRNAs encoding inhibitors of apoptosis, and that the Fas-activated serine/threonine kinase (FAST) phosphoprotein is counteracting this function (Li et al., 2004b). They showed that FAST, which is tethered to the outer mitochondrial membrane in association with BCL-X L (Li et al., 2004a), is a constitutive prosurvival protein (Li et al., 2004b). RNAi-mediated silencing of endogenous FAST in HeLa cells resulted in apoptosis, whereas overexpression of FAST inhibited both Fas-and UV-induced apoptosis (Li et al., 2004b). Mechanistically, they found that a FAST mutant lacking its TIA1-binding domain did not inhibit apoptosis, and overexpressed TIA1 inhibited the antiapoptotic effects of FAST. They proposed that in response to stress, tBID and BAX move to the outer mitochondrial membrane, where they sequester BCL-X L , releasing FAST from its mitochondrial tether. FAST then binds to TIA1 and prevents TIA1-mediated silencing of mRNAs, including those encoding inhibitors of apoptosis, such as cIAP-1 and XIAP (Li et al., 2004b). Hence, FAST serves as a cellular sensor of mitochondrial stress, that in response to stress, modulates TIA1-regulated posttranscriptional silencing responses. PUMILIO Pumilio belongs to the evolutionary conserved Pumilio and FBF (PUF) family of RBPs comprised two paralogous members in vertebrates (Pum1 and 2), and one in Drosophila (Pum). It is an important mediator of neurological processes, including olfactory learning and motor function (Ravanidis et al., 2018). In humans, a PUM1 mutation is associated with adult-onset ataxia, whereas haploinsufficiency due to deletions or missense variants cause developmental delay and seizures (Gennarino et al., 2018). Studies in Drosophila backed yeast findings. Work by Gehrke et al. (2015), revealed that nuclear mRNAs encoding respiratory chain complexes (nRCC) are localized in a PINK1/Tom20-dependent manner to the mitochondrial outer membrane, where they are de-repressed and translated by PINK1/PRKN pathway through the displacement of translation repressors, including PUM and hnRNPF; in this case, PINK1 displayed an RNA-binding capacity competing with PUM for mRNA-binding, while PRKN mono-ubiquitinated the RBPs lowering their affinity for nRCC mRNAs (Gehrke et al., 2015). Accordingly, inhibiting PUM via RNAi was found to increase, whereas PUM overexpression decreased nRCC mRNAs abundance (Gehrke et al., 2015). In addition, FIGURE 4 | Mitochondrial perturbations induced by PUM2. PUM2 reduces cytochrome c oxidase complex (Complex IV) activity, leading to impaired respiration and deformed cristae. Interestingly, the long non-coding RNA NORAD inhibits PUM2 function by sequestering PUM2 from binding to mitochondrial mRNA targets. Further, PINK1 in association with Tom20 promote the expression of nuclear-encoding mitochondrial (nRCC) mRNAs in the outer mitochondrial membrane by competing with PUM and other translation repressors. PINK1 competes with PUM for mRNA-binding, while PRKN mono-ubiquitinates PUM and HNRNPF lowering their affinity for nRCC mRNAs and possibly leading to their proteasomal degradation. However, when PUM is in excess, it binds to the nRCC mRNAs and represses their translation. Finally, PUM2 binds to MFF mRNA and represses its translation, leading to reduced fission and mitophagy. PUM inhibition rescued ATP production, mitochondrial morphology, and neuromuscular-degeneration phenotypes of PINK1, but not PRKN mutant flies (Gehrke et al., 2015). Collectively, these findings revealed that besides its wellknown serine/threonine kinase activity, PINK1 is also an RBP competing with PUM to control mitochondria biogenesis (Gehrke et al., 2015). Electron microscopy of skeletal muscles from PUM2overexpressing mice revealed the accumulation of subsarcolemmal, irregularly shaped and abnormally enlarged mitochondria lacking normal cristae (Kopp et al., 2019). Furthermore, a global reduction in cytochrome c oxidase (complex IV, COX) activity was observed. In addition, transient expression of PUM2 in MEFs or stable expression of either PUM1 or PUM2 in HCT116 cells significantly impaired respiration, providing compelling evidence that PUM hyperactivity results in mitochondrial dysfunction (Kopp et al., 2019). Interestingly, a non-coding RNA called NORAD acts as a guardian of the transcriptome by being a preferred target of PUM2, thereby inhibiting its translation suppressive functions (Kopp et al., 2019). Research findings from D'Amico et al. (2019) associated PUM2 with aging defects via impaired mitochondrial dynamics. They reported that elevated levels of PUM2 are found in muscle and neocortex of aged mice (Edwards et al., 2007;Oberdoerffer et al., 2008;D'Amico et al., 2019) as well as muscle biopsies of aged humans . Additionally, Pum2 levels in the brains of mice strains BXD and LXS are inversely correlated with longevity (Gelman et al., 1988;Liao et al., 2010). To experimentally validate this suggestive effect on lifespan, they used Caenorhabditis elegans to show that PUF8, ortholog of PUM2, knockdown was associated with increased lifespan (D'Amico et al., 2019). Consistently, knock-down of puf8 and Pum2 improved both basal and maximal respiration in old worms and mouse myoblasts, respectively (D'Amico et al., 2019). Like in previous studies, using multi-tissue gene set enrichment analysis (GSEA) in the human GTEx cohort, they found that PUM2 expression levels were inversely correlated with clusters of genes responsible for mitochondrial function, including genes important for OXPHOS and cellular respiration (D'Amico et al., 2019). Furthermore, from CLIP-Seq data (Hafner et al., 2010), they identified a perfect PUM2 site in the 3 UTR of MFF mRNA and showed that PUM2 negatively regulated MFF translation in C2C12 and HeLa cells (D'Amico et al., 2019). Consequently, Pum2 silencing increased the percentage of C2C12 cells undergoing fission and mitophagy, and this was reversed by simultaneously performing Mff RNAi. Similarly, puf8 depletion improved mitochondrial homeostasis during nematode aging and canceled by mff1 co-depletion (D'Amico et al., 2019). Lastly, Pum2 depletion using CRISPR-Cas9 in the muscle of old mice increased MFF levels and mitophagy, and improved respiration. Collectively, these data suggest that PUM2 regulates mitochondrial dynamics and mitophagy via MFF. CONCLUSION Over the years, several lines of evidence have implicated mitochondrial dysfunctions in many diseases, particularly those associated with neurodegenerative disorders and aging. Following recent findings that mutations or misexpression of RBPs can cause neurological impairments, there has been tremendous interest in identifying their molecular pathogenetic mechanisms. Interestingly, it turned out that mitochondria are also direct and early targets of RBP misregulation reiterating their importance for cellular homeostasis. These findings suggest that pharmaceutical agents improving mitochondrial life cycle can be attractive therapeutics for easing mitochondrial dysfunction in these diseases. AUTHOR CONTRIBUTIONS SR and ED wrote and edited the manuscript. Targeted Expression of Human Vitamin D Receptor in Adipocytes Decreases Energy

Expenditure and Induces Obesity in Mice Our previous studies demonstrated a high fat diet-resistant lean phenotype of vitamin D receptor (VDR)-null mutant mice mainly due to increased energy expenditure, suggesting an involvement of the VDR in energy metabolism. Here we took a transgenic approach to further define the role of VDR in adipocyte biology. We used the aP2 gene promoter to target the expression of the human (h) VDR in adipocytes in mice. In contrast to the VDR-null mice, the aP2-hVDR Tg mice developed obesity compared to the wild-type counterparts without changes in food intake. The increase in fat mass was mainly due to markedly reduced energy expenditure, which was correlated with decreased locomotive activity and reduced fatty acid β -oxidation and lipolysis in the adipose tissue in the transgenic mice. Consistently, the expression of genes involved in the regulation of fatty acid transport, thermogenesis and lipolysis were suppressed in the transgenic mice. Taken together these data confirm an important role of the VDR in the regulation of energy metabolism. Introduction The maintenance of body weight depends on the balance between energy intake and energy utilization. Obesity results when energy consumed exceeds energy utilized. Energy is acquired through diets and can be stored in adipose tissue or utilized by the body to maintain basic cellular functions and physical activities. Energy can also be used for adaptive thermogenesis in response to a cold environment (1). The adipose tissue is unique in that it represents both arms of energy balance. The white adipose tissue (WAT) has the ability to sense the energy state of the body. When energy availability is high, the WAT stores the excess energy as triglyceride in lipid droplets. When energy is needed, triglyceride is broken down to free fatty acids to release into the circulation. This process, known as lipolysis, is regulated by two enzymes, adipose triglyceride lipase (ATGL) and hormone sensitive lipase (HSL). ATGL initiates the first, rate-limiting step of lipolysis by hydrolyzing triglyceride to diacylglyceride (2,3), which is further broken down to monoglyceride by HSL (4). Monoglyceride lipase cleaves the final ester bond of monoglyceride to release glycerol, and this step is not rate-limiting (5). Lipolysis is activated by catacholamines through the cAMP signaling pathway, leading to protein kinase A (PKA) activation. PKA phosphorylates HSL, which promotes HSL translocation to the lipid droplet and access to triglyceride stores (6). HSL and ATGL activity is suppressed by insulin during feeding, as insulin increases the amount of perilipin around the lipid droplets to prevent their access to triglycerides (7). The principal role of the brown adipose tissue (BAT) is to regulate adaptive thermogenesis through the expression of uncoupling proteins (UCPs). UCP1 separates oxidative phosphorylation from ATP production to release energy as heat (8)(9)(10). Studies have shown that increased expression of UCP1 in the BAT or its ectopic expression in the WAT results in increased metabolism and resistance to diet-induced obesity (11). The primary roles of UCP2 and UCP3 are less well characterized; however, both proteins have uncoupling capabilities (9,10). The VDR is a member of the nuclear receptor superfamily (12). Its high affinity ligand is 1,25-dihydroxyvitamin D (1,25(OH) 2 D 3 ), the hormonal form of vitamin D. In addition to its classic role in the regulation of calcium homeostasis, the vitamin D hormone has numerous non-calcemic activities (13), including regulation of adipocyte biology. 1,25(OH) 2 D 3 inhibits adipocyte differentiation in the NIH3T3-L1 preadipocyte model (14,15). Previous work from this and other laboratories has implicated a role for VDR in the regulation of global energy metabolism in vivo (16,17). VDR-null mutant mice are lean and resistant to high fat dietinduced obesity, in part due to the up-regulation of UCPs in adipose tissues. Data obtained from primary BAT confirmed the suppression of UCP1 and UCP3 by 1,25(OH) 2 D 3 (17). Consistently, VDR-null mice exhibited increased β-oxidation in the WAT. Given the broad range of regulatory activities of the VDR (13), however, global inactivation of the VDR in the VDR-null model affects the function of multiple tissues, which complicates the interpretation of the metabolic data. As such, the contribution of the VDR signaling in adipocytes to energy metabolism remains uncertain. To further define the role of the VDR in adipocyte biology, here we generated a transgenic (Tg) mouse model that specifically expresses the human (h) VDR in the adipose tissue. In contrast to the VDR-null mice, the Tg mice showed increased adipose mass and reduced energy metabolism compared to the wild-type (Wt) counterparts. Together these data provide further evidence confirming an important role of the adipocyte VDR in the regulation of energy metabolism. Materials and Methods Generation of aP2-hVDR transgenic mice. A 2 kb cDNA fragment containing the full-length coding sequence of hVDR was placed behind the 5.4 kb aP2 promoter/enhancer (18) in a plasmid vector (Fig. 1A). The aP2-hVDR-polyA cassette was released from the vector by Not1 and Sal1 restriction digestion. The fragment was purified and microinjected into CD-1 mouse embryos following standard procedures (19). Potential founders were screened using primers specific to the hVDR cDNA: 5'-CGTGTGAATGATGGTGGAGGGAGCC-3' (forward), and 5'-GTCTTGGTTGCCACAGGTCCAGGAC-3' (reverse). Founder Tg lines were confirmed by Southern blotting. The expression of hVDR in the adipose tissues was confirmed by Northern and Western blottings, using a 32 P-labeled hVDR cDNA probe or an anti-VDR antibody (Santa Cruz Biotechnology, Santa Cruz, CA), respectively. Animal treatment. Wt mice in CD1 background and aP2-hVDR Tg mice overexpressing the hVDR in the adipose tissue were usually fed a standard chow (SC) diet (Teklad Global 18% Rodent Diet 2018; Harlan Teklad), and maintained in a standard 12 light/12 dark cycle. To induce obesity, the mice were weaned on the SC diet and then at two months of age were switched to a high-fat (HF) diet containing 42% fat (TD.88137; Harlan Teklad) for 5 weeks. Body weight was monitored weekly. Total body fat content was quantified using dual energy X-ray absorptiometry (DEXA) scans (Lunar PIXImus II, Madison, WI) under anesthesia. Fat mass in different parts of the body was quantified by directly weighing the dissected fat pads. Indirect calorimetric measurement was carried out using the LabMaster System (TSE Systems, Midland, MI), and oxygen consumption, CO 2 production, energy expenditure, food and water intake, and locomotive movement were recorded for 4 days following 3-day acclimation. At the end of the experiment, the animals were sacrificed, and plasma, WAT, BAT and skeletal muscle were harvested. All animal studies were approved by the Institutional Animal Care and Use Committee at the University of Chicago. Plasma parameters. Total plasma cholesterol and triglyceride levels were determined using the Infinity cholesterol and triglyceride reagents from Thermo Scientific (Waltham, MA). Plasma non-esterified fatty acid (NEFA) levels were determined using NEFA C test kit from Wako (Richmond, VA). Plasma leptin levels were measured using a commercial ELISA kit (Diagnostic Systems Lab, Inc., Webster, TX). Glucose tolerance test. Two-month old male mice were fasted 6 hours before receiving an i.p. injection of glucose (2 mg/kg body weight). Blood samples were taken from tail bleeding at 0, 15, 30, 60 and 120 min after the glucose injection and blood glucose was measured using a Contour glucometer (Bayer HealthCare LLC, Mishawaka, IN). Random blood glucose levels were measured in 3-month old mice in the morning on three nonconsecutive days with the Contour glucometer. Northern blot. Total cellular RNAs were extracted using TRIzol reagents (Invitrogen, Grand Island, NY). RNAs were isolated from the gonadal WAT depot. Northern blot analysis was carried out as described previously (20). Western Blot. Gonadal WAT and BAT were homogenized in Laemmli buffer containing a protease inhibitor cocktail (Promega, Madison, WI). After centrifugation, the lower layer was collected and boiled for five minutes. Protein concentrations were determined using a BioRad Protein Assay kit (BioRad, Hercules, CA). Proteins were separated by SDS-PAGE and transferred onto Immobilon membranes. Western blotting was carried out as described previously (21). Quantitative real-time RT-PCR. Firststrand cDNAs were synthesized from 2 µg total RNAs in a 20 µl reaction using Moloney murine reverse transcriptase (Invitrogen, Carlsbad, CA) and hexanucleotide random primers. The firststrand cDNAs served as the template for PCR amplification, which was carried out using the LightCycler® 480 Real-Time PCR System (Roche Applies Science, Indianapolis, IN). Beta-2 microglobulin was used as an internal control. The primer sequences are available upon request. In vitro lipolysis assays. Freshly dissected WAT (100 mg) was incubated in Krebs-Ringer bicarbonate-HEPES (KRBH) buffer supplemented with 3% fatty acid free BSA and 5mM glucose. The isolated white fat was then incubated at 37°C for one hour with gentle shaking and treated with or without isoproterenol (10 µM, Sigma, St Luis, MO). Following the incubation, the media was collected and glycerol content was assessed using a commercial kit (Sigma, St. Louis, MO) according to the manufacturer's instruction. In vitro fatty acid β-oxidation assay. The rate of fatty acid β-oxidation in BAT was measured in the presence of 500 µM 3 Hpalmitate and 500 µM carnitine as described previously (17). NIH3T3-L1 adipocytes. Culture and differentiation of NIH3T3-L1 cells to mature adipocytes were performed as described previously (15). Differentiated cells were treated with ethanol or 20 nM 1,25(OH) 2 D 3 for 24 hours and the mRNA level of ATGL and HSL was quantified by real time RT-PCR. Statistical analysis. Data values were presented as means ± SEM. Statistical comparisons were made using Student's t-test, with P ≤ 0.05 being considered significant. Results Increased fat mass in aP2-hVDR Tg mice. We used the aP2 promoter/enhancer, an adipocyte-specific gene promoter (18,22), to target hVDR expression specifically in the adipose tissue. We chose hVDR as the transgene to distinguish it from the endogenous mouse VDR in genotyping, as the function of hVDR and mouse VDR is exchangeable. We identified five positive Tg lines by Southern blot (not shown) and Northern blot analyses of the WAT (Fig. 1B). In this study we used Tg line 21. Unless specifically indicated, all data were obtained from male mice. Northern and Western blot analyses confirmed a high level expression of hVDR at mRNA and protein levels in both WAT and BAT in this Tg line ( Fig. 1C and D). The expression of the hVDR transgene was undetectable in other tissues (not shown). On the SC diet, the Tg mice exhibited significantly increased body weight at two and four months of age compared to the Wt mice ( Fig. 2A). The body weight difference was even larger at 6 months of age between these two genotypes (not shown). The body fat content, estimated by weighing the sum of dissected subcutaneous, peri-renal and gonadal fat pads, was significantly higher in the Tg mice than in the Wt mice at 4 months of age (Fig. 2B). We then used DEXA scan to assess more accurately the body fat content. DEXA scan showed a trend toward increased body fat percentage in the Tg mice at two months of age (p=0.065) (Fig. 2C); at three months of age, the increase in body fat percentage was significant in female Tg mice (Fig. 2C). When placed on the HF diet for 5 weeks, the Tg mice displayed markedly higher body weight and body fat content compared to the Wt counterparts, the latter assessed by DEXA scan ( Fig. 2A and C). Thus the hVDR transgene increased body fat mass in both male and female mice. As expected, all mice on the HF diet showed higher body weight and total fat mass in both genotypes compared to mice on the SC diet ( Fig. 2A and C). On both the SC and HF diets, various fat depots of the body, including subcutaneous, gonadal and peri-renal fats, were significantly larger in the Tg mice ( Fig. 2D and E). The Tg mice also had greater BAT mass, although the difference did not reach significance ( Fig. 2D and E). The HF diet widened the difference in total body fat content and in the size of subcutaneous and gonadal fat pads between the Wt and Tg mice compared to the SC diet (Fig. 2C, D and E). Altered adipokine levels in aP2-hVDR Tg mice. As aggregate adipose mass affects adipokine production, we examined the expression of leptin and adiponectin in the WAT. The level of leptin mRNA was slightly increased in the Tg mice compared to Wt mice; however this difference was not significant (Fig. 3A). In contrast, the level

of adiponectin was significantly reduced (by 65%) in the Tg mice compared to the Wt mice (Fig. 3B). Serum leptin levels were not significantly different between the Wt and Tg mice on the SC diet; however, the levels were significantly elevated in the Tg mice on the HF diet (Fig. 3C), probably due to the greater increase in fat mass under this dietary condition. Increased cholesterol levels in aP2-hVDR Tg mice. We next examined the circulating lipid levels. There was no difference in plasma triglyceride or NEFA levels between the Wt and Tg mice ( Fig. 4A and B). Interestingly, the Tg mice exhibited an increase in the level of plasma cholesterol compared to the Wt mice (Fig. 4C). This is in contrast to the VDR-null mutant mice that had decreased plasma cholesterol levels relative to the Wt mice (17). Together these observations suggest a possible involvement of the adipocyte VDR in the regulation of plasma cholesterol levels. Reduced global metabolism in aP2-hVDR Tg mice. We assessed mouse global metabolism by indirect calorimetry. Total energy expenditure, oxygen consumption (VO2) and carbon dioxide production (VCO2) (not shown) were markedly reduced in the Tg mice compared to the Wt mice on the SC diet ( Fig. 5A and B); however, there was no difference in the respiratory exchange rate between these two genotypes (Fig. 5C), indicating no change in fuel substrate selection in the Tg mice. Food intake was the same for the Wt and Tg mice during the light and dark cycle of the day (Fig. 5D), indicating that the difference in body weight or fat mass between these mice was not due to a difference in energy intake. Interestingly, the Tg mice moved less than the Wt mice during the active dark cycle, as measured by the number of beam breaks. During the light cycle, however, when all mice are presumably less active, there was no difference in the number of beam breaks between these mice (Fig. 5E). Less physical activity may contribute to the increase in fat mass seen in the Tg mice. It is unlikely, however, that the extra weight that the Tg mice are carrying physically prevents these mice from moving. Similar results, including the metabolic data, food intake and physical activity, were seen between Wt and Tg mice when they were fed the HF diet (Fig. 6A-E). Lower β-oxidation and thermogenesis in aP2-hVDR Tg mice. The decrease in metabolism seen in the Tg mice suggests possible suppression of total fuel utilization. Since the respiratory exchange rate of these mice was not altered, it is likely that both carbohydrate and lipid oxidation are decreased. We thus quantified the expression of key genes involved in β-oxidation and glycolysis in the BAT, WAT and skeletal muscle. Skeletal muscle had no expression of the hVDR transgene. Carnitine palmitoyl transferase (CPT)1 and CPT2 are responsible for the transport of long chain fatty acids into the mitochondrial matrix for βoxidation (23). In the Tg mice, the expression of CPT1 and CPT2 in both BAT and WAT was significantly suppressed compared to the Wt mice ( Fig. 7A and B). CPT2, not CPT1, was also significantly lowered in the skeletal muscle of the Tg mice (Fig. 7C). Consistently, the rate of β-oxidation, measured in vitro using adipocytes isolated from the BAT, was reduced in the Tg mice compared to the Wt counterparts (Fig. 8). Hexokinase (HK) and pyruvate kinase (PK) are regulatory enzymes involved in glycolysis. HK phosphorylates glucose to glucose-6-phosphate in the first step of glycolysis. PK catalyzes the production of pyruvate in the final step of glycolysis. In the Tg mice, both HK and PK mRNA levels were significantly reduced in the BAT (Fig. 7A), only HK was reduced in the WAT (Fig. 7B), but no changes in HK and PK were observed in the skeletal muscle (Fig. 7C). The UCPs are involved in the regulation of adaptive thermogenesis. Consistently with the obese phenotype, the transcripts of UCP1, UCP2 and UCP3 were significantly suppressed in the BAT of the Tg mice compared to the Wt mice (Fig. 7A). Similarly, the mRNA levels of both UCP2 and UCP3 were lower in WAT of the Tg mice (Fig. 7B), but they were not significantly changed in the skeletal muscle (Fig. 7C). UCP1 is not normally expressed in WAT or skeletal muscle and thus was not examined. Decreased lipolysis in aP2-hVDR Tg mice. Although the metabolism of the Tg mice was decreased compared to the Wt mice, they maintained similar levels of NEFAs, the major fuel source for the body. NEFAs are generated from WAT through the breakdown of triglycerides in the process of lipolysis, which is regulated by ATGL and HSL. Both ATGL and HSL transcript levels in WAT were significantly suppressed in the Tg mice compared to the Wt counterparts (Fig. 9A). We further determined lipolysis in the WAT isolated from these mice by measuring glycerol release into the media. Under the basal condition, the Tg WAT released less glycerol than the Wt WAT; when the WAT was stimulated with isoproteranol, a β-adrenergic receptor agonist known to increases lipolysis, the Tg WAT still released significantly less glycerol than the Wt counterpart (Fig. 9B). These data indicate that overexpression of the hVDR transgene suppresses lipolysis in adipose tissues. To confirm that ATGL and HSL are regulated by vitamin D, we treated differentiated NIH3T3-L1 adipocytes with ethanol or 1,25(OH) 2 D 3 , and found that ATGL and HSL transcripts were significantly suppressed by 1,25(OH) 2 D 3 in these cells (Fig. 9C). Together these observations strongly suggest that vitamin D regulates lipolysis by directly targeting ATGL and HSL. Glucose intolerance in aP2-hVDR Tg mice. Finally we examined the ability of the Tg mice to handle glucose. The non-fasting blood glucose level was higher in the Tg mice compared to the Wt mice (Fig. 10A). When these mice were subject to an intraperitoneal glucose tolerance test, the Tg mice were not able to clear glucose as effectively as the Wt mice, indicating glucose intolerance (Fig. 10B). Discussion In our previous studies of the VDR knockout mouse model we demonstrated a role of the VDR in the regulation of energy metabolism (17). These null mutant mice developed a lean phenotype that is resistant to high fat diet-induced obesity. These studies, however, have limitations because the expression of VDR was ablated from the entire animal, which cannot exclude the possibility that VDR inactivation in a number of tissues contributes to the metabolic phenotype. By overexpressing the hVDR specifically only in the adipose tissue we are able to correlate the role of the VDR in the adipose tissue to metabolic phenotypes. Furthermore, in the global knockout model, the VDR is missing from the adipocyte lineage throughout development; therefore, results might be confounded by developmental impairment. The aP2 gene is expressed late in adipocyte differentiation (18,24). Therefore, overexpressing the hVDR using the aP2 promoter enables us to evaluate the role of the VDR in mature adipocytes without interfering adipocyte differentiation. In this study we showed that overexpression of the hVDR in the adipose tissue in transgenic mice leads to obesity with increased body weight and fat mass. Because there was no difference in energy intake between the Wt and Tg mice, the increase in body weight and fat mass seen in the Tg mice was likely due to a decrease in energy expenditure. This notion was confirmed in calorimetric studies, which showed reduced energy expenditure and oxygen consumption in the Tg mice. Unexpectedly, the Tg mice had significantly less locomotive activity than the Wt mice during the dark cycle, when the animals are most active. This reduction in movement may partially explain the increased fat mass in the Tg animals. It is not yet clear why the Tg mice move less than their Wt counterparts. It is possible that the adipose tissue somehow signals the brain to decrease movement, but there is no evidence to support such a speculation at this time. Consistent with the reduced energy metabolism in Tg mice, hVDR overexpression in the adipocytes led to reduced fatty acid βoxidation and lipolysis, accompanied by suppression of key genes involved in these processes. HK, an enzyme in regulation of glycolysis, and CPT1 and CPT2, regulators of long chain fatty acid transport into the mitochondrial matrix for oxidation, were downregulated in both WAT and BAT in the Tg mice. Importantly, the rate-limiting enzymes involved in lipolysis, ATGL and HSL, were also suppressed in the WAT of the Tg mice. The suppression of β-oxidation and lipolysis by the VDR provides a critical explanation for the increased adipose mass seen in the Tg mice. These observations suggest a direct role of VDR in the regulation of β-oxidation and lipolysis in the adipose tissue. Further investigations are needed to understand the molecular mechanism whereby the vitamin D/VDR signaling regulates these regulatory proteins. Given that the VDR functions as a ligand activated transcription factor, it is not surprising that some of these metabolic regulators are directly regulated by the VDR in the adipose tissue. In this regard, the in vitro data that we obtained from the NIH3T3-L1 adipocytes support direct regulation of ATGL and HSL by the vitamin D hormone. The critical role of these two enzymes in lipolysis warrants further investigations to dissect the molecular mechanism underlying these important regulations. In addition, we also examined the effect of 1,25(OH) 2 D 3 on the expression of CPT1, CPT2, HK and PK in 3T3-L1 adipocytes as well as in an immortalized brown fat cell line (25), but observed no direct regulation of these genes by vitamin D in both adipocyte cell lines (data not shown). These observations suggest that vitamin D influences fatty acid transport and glycolysis by indirectly regulating these genes. Another factor that may contribute to the increased adipose mass is the suppression of UCP expression in the Tg mice. Previous studies have demonstrated increased expression of UCP1 and UCP3 in the WAT and BAT of VDR knockout mice, which contribute to the increased energy metabolism and a lean phenotype in these mutant mice (16,17). Our previous data from primary BAT culture and adipocyte cells indicated that UCP1 and UCP3 are under direct regulation by the vitamin D/VDR signaling (17). Consistently, in this study we showed that hVDR overexpression in the adipose tissue led to the suppression of UCP1, UCP2 and UCP3. Although the roles of the UCP proteins in the regulation of metabolism and body fat mass are controversial (26)(27)(28)(29), these data are consistent with the metabolic phenotypes seen in the Tg mice. The aP2-hVDR Tg mice showed glucose intolerance, which may be partly attributed to the decreased expression of adiponectin. Low levels of adiponectin are correlated with insulin resistance (30,31). Low adiponectin expression in the Tg mice is probably due to the increase in fat mass. Epidemiological studies have reported a relationship between obesity and vitamin D status in humans. Generally, adiposity or body mass index (BMI) is inversely correlated to the levels of serum 1,25(OH) 2 D 3 (32) and 25hydroxyvitamin D 3 (33)(34)(35) in healthy adults; however, whether low vitamin D levels play any roles in the development of obesity is unknown. One theory to explain this observation is that in obesity more vitamin D, which is highly fat-soluble, is trapped in the body fat, leading to low serum vitamin D status (34). It is speculated that the trapped vitamin D would increase the local vitamin D concentration in the fat, and based on the finding from the present study, activation of the VDR in adipocytes negatively affects energy expenditure, further promoting obesity. In summary, here we have provided further evidence that the VDR is able to regulate global metabolism by exerting its effects on the adipose tissue. As the energy sensing tissue of the body, the WAT is sensitive to changes in metabolic rate. Small changes in energy expenditure over time can have gross impacts on fat mass. This is probably the main reason for the Tg mice to develop obesity. Previous studies have shown that ablation of the VDR results in a lean phenotype (16,17), and this present work demonstrates that overexpression of the VDR specifically in the adipose tissue has an opposite effect. Together, these studies provide very strong evidence for a previously unrecognized role of the adipocyte VDR in energy

metabolism. The underlying mechanism for VDR regulation of energy metabolism involves the regulation of UCPs and enzymes involved in β-oxidation and lipolysis. The exact molecular basis of VDR regulation in these pathways, however, remains to be established. HER2 and EGFR amplification and expression in urothelial carcinoma occurs in distinct biological and molecular contexts We analyzed a cohort of 599 cases of urothelial carcinoma for EGFR, ERBB2, and ERBB3 gene expression and genomic alterations. The cohort consisted of a reference set (n = 292) comprising all stages and grades and one set (n = 307) of advanced tumors. All cases were previously classified into urothelial carcinoma molecular subtypes. Genomic amplifications were established by array-CGH or in-situ hybridization, and gene expression both at mRNA and protein levels. Clinical HER2 status was independently evaluated using standard clinical procedures. EGFR amplifications were observed in 14% and ERBB2 amplifications in 23% of the reference cohort. EGFR gains were enriched in the Basal/SCC-like and ERBB2 gains in the Genomically Unstable subtypes. The expression data suggests that the Genomically Unstable show high ERBB2/ERBB3 but low EGFR expression and that Basal/SCC-like tumors show high EGFR but low ERBB2/ERBB3 expression. Whereas the frequency of ERBB2 genomic amplification were similar for cases of the Genomically Unstable subtype in the two cohorts, the Urothelial-like subtype acquires ERBB2 amplifications and expression during progression. Even though a good correlation between gene amplification and ERBB2 gene expression was observed in the Urothelial-like and Genomically Unstable subtypes less than half of the Basal/SCC-like cases with ERBB2 amplification showed concomitant ERBB2 mRNA and protein expression. We conclude that clinical trials using ERBB2 (HER2) or EGFR as targets have not fully appreciated the molecular heterogeneity in which activated ERBB2 and EGFR systems operate. Proper tumor classification is likely to be critical for arriving at thorough conclusions regarding new HER2 and EGFR based treatment regimes. INTRODUCTION Bladder cancer is the fifth most common malignancy in the Western world. It is associated with a high rate of mortality in patients with advanced disease. In spite of this, few significant improvements in survival has been achieved during the last three decades [1]. While a proportion of patients respond to conventional chemotherapy these responses are rarely long-lasting. Even though molecular targeted therapy has been successful for other cancers types, few advances have been made in bladder cancer. In analogy with other tumor types, members of the ERBB receptor family have been considered as potential targets also for urothelial carcinomas e.g., ERBB2 (HER2) [2] and EGFR [3]. In urothelial carcinomas, overexpression of ERBB2 protein has been reported to vary considerably between studies [4] as well as between geographically distinct cohorts [5]. Several studies have also reported a low concordance between ERBB2 protein level and gene amplification in urothelial carcinomas [5,6]. Hansel et al. detected ERBB2 overexpression (IHC) in 36% but genomic amplification in only 10% of tumors [7]. Similarly low levels of ERBB2 genomic amplifications have also been reported by other investigators (e.g., [8]). Hence no consensus regarding the status of HER2 alterations in urothelial cancer has Research Paper been arrived at. Furthermore, even though several trials aiming for HER2 as a target in bladder cancer have been initiated, no general conclusions have been reached. A factor contributing to the disparate results is most likely the underlying heterogeneity of bladder cancer and the lack of adequate molecular descriptions of bladder cancer molecular subtypes; the context in which the ERBB2 and EGFR targets operate. We have described three major molecular subtypes of bladder cancer, Urothelial-like (Uro) (previously termed Urobasal [9]), Genomically Unstable (GU) and Basal/SCC-like [9], that accommodates fundamental differences in their molecular biology [10][11][12][13][14]. By combining global mRNA gene expression and extensive immunohistochemistry (IHC) analyses we have shown that the Uro, GU, and SCC-like subtypes dominate also the muscle invasive tumors, albeit in more progressed and infiltrated versions [15]. Here we investigate EGFR, ERBB2 (HER2), and ERBB3 (HER3) genomic alterations and expression in relation urothelial carcinoma molecular subtypes. To be able to study changes during tumor progression we analyze two cohorts, one (n = 292) dominated by non-muscle invasive tumors and a second (n = 307) dominated by muscle invasive tumors. We show that ERBB2 amplifications and expression, as well as clinically "HER2 positive" cases, may be of two fundamentally different molecular subtypes, of Uro or the GU subtypes, and that EGFR expression is associated with the SCC-like subtype. Co-occurring focal amplifications of EGFR, and ERBB2 or ERBB3 were not observed, while co-occurring gains of EGFR and ERBB2, and EGFR and ERBB3 were seen in eight and one cases, respectively. Five cases showed gains of all three genes. Examining the very high expressing cases (with Tumor Cell Score ≥ 2.5, indicating high protein expression throughout the tumor), EGFR+ERBB2 expression was only observed in one GU tumor, while two UroA tumors expressed both EGFR and ERBB3 at high levels. None of the tumors had strong simultaneous expression of all three proteins. Very high simultaneous expression of both ERBB2 and ERBB3 was seen in 11 tumors, where 6 were GU and 5 were UroA. EGFR, ERBB2 and ERBB3 in advanced urothelial carcinoma cohort The 307 samples in the advanced cohort were all from patients that underwent radical cystectomy. The tumors in this cohort have been classified into molecular subtypes through the combination of global mRNA expression and extensive immunohistochemistry [15]. The combined analyses produced robust subtype assignments into Urothelial-like (Uro), Urothelial-like B (UroB), Genomically Unstable (GU), Basal/SCClike, Mesenchymal Infiltrated (Mes-Inf), and Small Cell/ Neuroendocrine (Sc/NE) ( Figure 4). In addition, a small group (n = 6) of highly infiltrated, and thus unclassified, tumors were detected (Inf in Figure 4). The expression patterns defining the Basal/SCC-like subtype is illustrated in Figure 4A, with high KRT5 and KRT14, and low GATA3 and FOXA1 expression [9]. The majority of the non-Basal/SCC-like tumors belong to the Urotheliallike or Genomically Unstable category of tumors, which can be distinguished from each other by the expression patterns of FGFR3, CCND1, CDKN2A (p16), and RB1 ( Figure 4B). The combination of mRNA and IHC data also enabled us to resolve the remaining tumor cell phenotypes where examination of mRNA expression patterns alone is complicated by infiltrating immune cells or high stromal cell content [15]. The final subtype distribution was 109 Uro (35.5%), 24 UroB (7.8%), 66 GU (21.5%), 62 Basal/ SCC-like (20.2%), 16 Mes-Inf (5.2%), and 24 Small Cell/Neuroendocrine (Sc/NE) (7.8%) and 6 remaining Infiltrated (Inf) (2%). Increased EGFR mRNA and protein levels were largely confined to the Basal/SCC-like subtype (p = 10 -12 and p = 10 -13 , respectively) ( Figure 4C). ERBB2 mRNA expression was significantly higher within the Uro and GU subtypes (p = 10 -9 and p = 10 -13 , respectively), an enrichment also seen at the protein level (p = 10 -7 and p = 10 -10 ). The ERBB3 mRNA expression was highest in the Uro and GU tumors (p = 10 -16 and p = 10 -9 ), and was strongly correlated with the mRNA levels of ERBB2 (r = 0.69, p = 10 -16 ). The Uro and GU tumors also had the strongest ERBB3 protein expression. The EGFR, ERBB2, and ERBB3 protein expression showed a strong correlation with the mRNA levels (r = 0.65, 0.75, and 0.69, respectively, all p < 10 -16 ) ( Figure 4C). The ratio of ERBB2+ERRB3 to EGFR mRNA expression summarizes the overall expression patterns and identifies Uro and GU tumors as ERBB2, ERBB3 high and EGFR low, and the Basal/SCC-like tumors as EGFR high and ERBB2, ERBB3 low ( Figure 4C). No tumors had very high simultaneous EGFR and ERBB2 protein expression (IHC Tumor Cell Score ≥ 2.5), while one Uro tumor had high EGFR and ERBB3 protein expression. Similarly, no tumor G2, blue; G3, red. White indicates missing information. Panel 2: Gene expression of EGFR, ERBB2, and ERBB3 for 292 samples ordered by subtype. Red, high expression; green, low expression. Panel 3: The ratio of EGFR/ ERBB2 and ERBB3 mRNA expression emphasizes the subtype specific pattern of expression. Red, high ERBB2 and ERBB3 expression, and low EGFR expression; blue, low ERBB2 and ERBB3 expression, and high EGFR expression. Panel 4: EGFR, ERBB2, and ERBB3 immunohistochemistry scores. The score represents a tumor cell score (TCS), where the staining intensity is multiplied by the fraction of the tumor cells that show staining. Dark brown, high expression; white, no expression; gray, missing data. (B) Copy number alterations as determined by array-CGH. Red, focal amplification; orange, gain; blue, loss; gray, missing data. (C) Clinical HER2-status evaluated using clinical guidelines. Panel 1: ERBB2 (HER2) IHC intensity scores (0-3) in > 10% of tumor cells. White, score = 0; light brown, score = 1, brown, score = 2; dark brown, score = 3; gray, missing data. Panel 2: HER2 amplifications as determined by in-situ hybridization (ISH) (ERBB2/CEN17 ISH ratio ≥ 2 or ≥ 4 ERBB2 copies). Red, ERBB2 amplification; white, no amplification; gray, missing data. Panel 3: HER2-positive samples (IHC2+ with HER2 amplification). Blue, HER2 positive; white, HER2 negative; gray, missing data. All panels show the 292 samples. had simultaneous high expression of all three proteins. There were 22 tumors with very strong ERBB2 and ERBB3 protein expression (9 Uro, 12 GU, and 1 Sc/Ne). The advanced UC tumors were evaluated for clinical HER2 status ( Figure 5A). In total, 88 tumors (29%) were deemed HER2 positive. The proportion of HER2 positive tumors was 38% (41/109) in Uro, 21% (5/24) in UroB, 47% (31/66) in GU, 10% (6/62) Basal/SCC-like, 6% (1/16) in Mes-Inf, 17% (4/24) in Sc/NE, and none in the remaining six infiltrated tumors. Notably, HER2 positive Uro and GU patients showed higher ERBB3 mRNA and protein expression compared to the other subtypes, indicating a different context of ERBB2 amplification and overexpression ( Figure 5B). The Uro tumors may be divided into one group that contain the majority of the < T2 and < G3 tumors found in the dataset (indicated by green in the lower subtype bar in Figure 5), and in a more progressed UroC group (indicated by blue in the lower subtype bar in Figure 5) defined in Sjödahl et al. 2017 [15] with almost exclusively ≥ T2 G3 tumors. The latter group represents a progressed form of Uro distinct from UroB. The advanced Uro tumors exhibited a 34% (14/41) HER2 positive frequency while the UroC group showed a frequency of 54% (19/35), in stark contrast to the 11% seen in non-muscle invasive UroA tumors in the reference cohort ( Figure 1). In contrast, GU tumors were HER2 positive in 45% the reference cohort and in 47% in the advanced cohort. DISCUSSION To clarify the molecular context in which ERBB2 and EGFR gene amplifications and overexpression occur we performed an in-depth study of EGFR, ERBB2, and ERBB3 alterations in relation to existing urothelial carcinoma molecular subtypes. In essence, urothelial carcinomas could be divided into two distinct categories based on EGFR, ERBB2, and ERBB3 expression and genomic alterations. One category showed high ERBB2/ ERBB3 expression with concomitant high frequencies of clinically HER2 positive patients. This category coincided very well with the molecular subtypes Uro and GU (Summarized in Figure 6). The Uro and GU subtypes combined correspond to the "Luminal" subtypes of urothelial carcinoma as described by Choi et al. [16], Damrauer et al. [17], and the TCGA Clusters I and II [18]. The increased ERBB2 amplification and expression events seen in the Uro subtype seems to be a part of the Uro progression process from non-muscle invasive to muscle invasive growth as the frequency of HER2 positive cases increased from 11% in the non-muscle invasive UroA to 54% in the more advanced and muscle invasive UroC. In contrast ERBB2 amplification and expression seems to be a founding feature of the GU phenotype [10,19], even though no separate and distinct "HER2 subtype" of UC analogous the breast cancer "HER2 enriched" have emerged. Uro tumors differ from GU tumors by exhibiting high frequencies of FGFR3 and PIK3CA mutations [10] and in muscle invasive urothelial carcinoma FGFR3 mutations are heavily skewed towards the Uro subtype [19]. Furthermore, bi-allelic loss of the CDKN2A locus and overexpression of CCND1 are frequent events in the Uro subtype, indicating a dependency on mitogenic MAPK/PI3-K signaling for cell proliferation [19]. In contrast, genomic events associated with the GU subtype, i.e., RB1 loss,

TP53 mutation, and E2F3 gene amplification, allow for a more uncontrolled proliferation [19]. In addition, GU tumors, but not UroA tumors, strongly express FOXM1, a direct downstream target of ERBB2 (HER2) [20], and a FOXM1 related gene signature [13]. Furthermore, in breast cancer FOXM1 and HER2 expression is tightly correlated [21]. Consequently, it may be expected that "clinical HER2 positivity" may have different implications in Uro and GU due to context differences. Additionally, the Urothelial-like subtype is to some extent analogous to the ESR1 expressing breast cancer "Luminal A" subtype [13] that responds less well to HER2 targeted therapy [22,23]. We also noted that less than half of the ERRB2 amplifications found in tumors Red, high FOXA1 and GATA3 expression, and low KRT5 and KRT14 expression; blue, low FOXA1 and GATA3 expression, and high KRT5 and KRT14 expression. Panel 5: Immunohistochemistry scores (TCS) of KRT5, KRT14, GATA3 and FOXA1. Dark brown, high expression; white, no expression; gray, missing data. (B) Panel 1: mRNA heat map of key genes FGFR3, CCND1, CDKN2A, and RB1 that distinguish Uro from the GU subtype. Panel 2: Ratio between the mRNA levels of FGFR3, CCND1, and RB1, and CDKN2A (p16). Red, high FGFR3, CCND1, and RB1 expression, and low CDKN2A expression, blue, low FGFR3, CCND1, and RB1 expression, and high CDKN2A expression. Panel 3: Immunohistochemistry scores (TCS) of FGFR3, CCND1, CDKN2A (p16), and RB1. Dark brown, high expression; white, no expression; gray, missing data. (C) Panel 1: mRNA expression of EGFR, ERBB2, and ERBB3. Panel 2: Ratio between the mRNA expression of ERBB2 + ERBB3, and EGFR. Red, high ERBB2 and ERBB3 expression, and low EGFR expression; blue, low ERBB2 and ERBB3 expression, and high EGFR expression. Panel 3: IHC scores (TCS) for EGFR, ERBB2, and ERBB3. Dark brown, high expression; white, no expression; gray, missing data. Panel 4: ERBB2 amplifications. Orange, ERBB2/CEN17 ISH ratio ≥ 2 or ≥ 4 ERBB2 copies; red, ERBB2/CEN17 ISH ratios ≥ 3 or ≥ 6 ERBB2 copies. Panel 5: ERBB2 S310F/Y mutations indicated in black, ERBB3 V104M/L mutation indicated in black and M91I mutation in gray. of the Basal/SCC-like subtype led to increased mRNA and protein levels, indicating that genomic amplification of 17q12 in Basal/SCC-like, as detected by HER2 ISH analyses, does not have the same consequence as in Uro and GU tumors. The strong association between ERBB2 and ERBB3 overexpression in Uro and GU HER-positive cases, but not in the Basal/SCC-like HER2-positive cases, should also be considered, as ERBB3 has been shown to be essential for ERBB2 driven tumor formation and maintenance [24][25][26]. Hence, recruiting urothelial carcinoma patients for HER2 directed therapy almost certainly have to take the molecular context i.e., molecular subtypes, into consideration. Even though clinically positive HER2 tumors showed heterogeneity in respect to molecular context, the EGFR positive samples represents a subset of welldefined urothelial carcinoma tumors. EGFR expressing cases coincided well with tumors defined as Basal/SCClike using the consensus definition i.e., KRT5/KRT14 high and FOXA1/GATA3 low [9]. In line with this, urothelial tumors of the Basal/SCC-like subtype respond well to erlotinib in experimental systems [3]. The Basal/ SCC-like subtype show extensive molecular similarities with the breast cancer basal-like subtype as well as the SCC subtype of lung carcinomas [13,27]. Consequently, lessons learned from EGFR targeted treatment in these tumor types may be translated into treatment of Basal/ SCC-like urothelial carcinomas. The fact that the EGFR/ ERBB2 and ERBB3 expression ratio readily identified two distinct classes of urothelial carcinomas strongly associated with the independently determined molecular subtypes shows that the status of the EGFR, ERBB2, and ERBB3 receptors may represent fundamentally different tumor cell phenotypes ( Figure 6). In conclusion, we argue that future clinical trials using targeted therapies against ERBB2 (HER2) may have to take the molecular context i.e., molecular subtypes, into consideration. It may be that clinical trials up to now have not fully appreciated the molecular background, and thus the molecular heterogeneity, in which activated ERBB2 and EGFR signaling systems are operating in. We believe that a proper tumor classification is critical for arriving at thorough conclusions during the evaluation process of new treatment regimes. Tumor cohort and data availability The present investigation makes use of two cohorts, one reference cohort containing all tumor stages and one cohort of advanced tumors. The reference cohort consists of 292 cases established from previously published data [10,12] (GSE32894 and GSE32549) and includes all tumor stages (123 Ta, like, n = 32). Among the 292 samples, 249 had copy number array data derived from one or more platforms [12,14,28,29]. Tumors showing narrow amplifications peaks on the lower resolution array platforms were hybridized to a custom made high density zoom-array with an average genome-wide probe spacing of 17 500 bp and between 7 000 -12 000 bp in select target regions [30]. All copy number profiles and copy number calls for EGFR, ERBB2, and ERBB3 were manually evaluated. For 217 samples both copy number data and IHC data were available. The advanced cohort consists of 307 formalin-fixed paraffin-embedded TUR-B samples from patients that underwent radical cystectomy between 2006 and 2011. Of these, 243 were muscle invasive in the pathological TUR-B evaluation. Generation of gene expression and immunohistochemistry data is described in detail by Sjödahl et al. 2017 [15]. Briefly, tissue-microarray cores were positioned to yield high tumor cell content and to be morphologically and histologically representative of the tumor. RNA extraction was performed on macrodissected tissue areas located close to the TMA core sites, using HighPure isolation kits (Roche). RNA was amplified and labeled using SensationPlus kits (Affymetrix) and hybridized to the Gene ST 1.0 platform (Affymetrix). Raw-and normalized data is deposited in Gene Expression Omnibus under GSE83586. The tumors were classified using a combination of global mRNA gene expression analysis and IHC analysis into Urothelial-like, Urothelial-like B, Genomically Unstable, Basal/SCC-like, Mesenchymal Infiltrated, and Small cell/Neuroendocrine subtypes as detailed in Sjödahl et al. 2017 [15]. Six heavily infiltrated cases could not be assigned to any of the subtypes. These were categorized as "infiltrated", and referred to as "Inf" in the text. Array CGH data was not available for this cohort. Informed consent was obtained from all patients and the study was approved by the Local Ethical Committee of Lund University in accordance with the Helsinki declaration. Immunohistochemistry Tissue microarrays (two 1.0 mm cores per tumor, 4 µm sections) were analyzed with antibodies against EGFR (3C6, Ventana), ERBB2 (4B5, Ventana), ERBB3 (Dako, DAK-H3-IC). TMA cores were evaluated as blinded digitalized image files. Each TMA slide contained 120 randomly selected tumor cores to obtain an efficient distribution between low (blank) and high expressing cases. Antibody dilutions were tuned to obtain a staining ranging from blank to intensive staining. IHC stainings for EGFR, ERBB2, and ERBB3 were quantitatively scored from 0 to 3. When hetrogeneous staining was observed within a single core, the fraction of tumor cells with each score was recorded and multiplied with the intensity for a tumor cell score (TCS). The mean tumor cell score of core pairs from the same sample was calculated. includes GU Ta and T1 tumors; Uro MI, includes Uro, UroB, and UroC cases from the series of advanced tumors; GU MI, includes GU cases from the series of advanced tumors; Basal/SCCL MI, includes Basal/SCC-like cases from the series of advanced tumors. Non muscle invasive cases of Basal/SCC-like were too few to be included. Panel 1: Summary of EGFR, ERBB2, and ERBB3 mRNA expression. Red, high expression; green, low expression. Panel 2: Summary of EGFR, ERBB2, and ERBB3 protein expression. Dark brown, high protein expression; white, low protein expression. Panel 3: Summary of HER2 positive cases. Dark blue, high frequency of HER2 positive cases; white, low frequency of HER2 positive cases. Immunohistochemistry analyses of KRT5, KRT14, GATA3, FOXA1, FGFR3, CCND1, CDKN2A (p16), and RB1 was as reported by Sjödahl et al. 2017 [15]. Clinical HER2 status Clinical HER2 status was assessed using dual in-situ hybridization (ISH) and IHC according to Swedish guidelines for HER2 testing in breast cancer [31]. The clinical scoring was performed on a separate section, independently evaluated by an experienced bladder and breast cancer pathologist (GC) using Ventana Pathway HER2 4B5 and Ventana Dual ISH assay HER2 DNP and CHR17 DIG probes on the BenchMark Ultra platform. The IHC score (0, 1+, 2+, 3+) was determined by the highest staining observed in >10% of the tumor cells. An ERBB2/ CEN17 ISH ratio ≥ 2, or ≥ 4 ERBB2 copies when CEN17 was not determinable, was used for amplification calls. In samples with two TMA cores the highest ratio was used for calling amplifications. HER2 amplified tumors with IHC 2+ scores in > 10% of cells were considered HER2positive. The inter-observer agreement between the IHC core intensity scores by GS and the clinical re-evaluation by GC was excellent (kappa 0.807 with equal weights). Multicenter study on antibiotic susceptibility/resistance trends in the western region of Cameroon In the global frame aiming at assessing bacterial susceptibility for safer and cost-effective healthcare, the present survey was conducted in three hospitals: Bafoussam Regional Hospital (BRH), Bangwa Protestant Hospital (BPH) and Bangangté District Hospital (BDH). Sampling was performed by fingerprinting on culture media and swabbing of hospital devices or surfaces. Wards of interest included: Pediatrics, Medicine, Operating Theater, Intensive Care Unit, Maternity, and, in the BDH, Laboratory in addition. Culture, isolation, identification and susceptibility tests were conducted according to standard guidelines and assigned contamination rates. Seventeen antibacterial agents were chosen and included representatives of major families of antibiotics used in Cameroon. Analysis of 238 specimens revealed 90%, 86% and 92% contamination rates in the BRH, BPH, and BDH, respectively. On healthcare provider’s hands, the respective rates were 63%, 100% and 91%. Bacillus and Staphylococcus were predominant bacteria types in all settings (BPH: 92%; BDH: 86%; BRH: 81%). Susceptibility profiles indicated high resistance rates and clonal distribution in all settings; and most reduced susceptibility with common drugs. Further investigations and previous works alleged drug use and basic hygiene as crucial in addressing resistance issues for safer care. This would be achieved with State support to public and private institutions. © 2017 International Formulae Group. All rights reserved. INTRODUCTION Low-income and technologically advanced countries throughout the world currently face the double challenge of communicable and non communicable diseases that were previously overlooked in many areas. This became obvious thanks to advances in medical technologies and international cooperation. Though a global concern, the burdens associated are higher in developing countries due to several factors in connection with geography, demography, socio-economy, and related (Jasovsky et al., 2016;Laxminarayan and Chaudhury, 2016). Beyond the classic high mortality and morbidity rates recorded with communicable diseases in developing countries, the global burden in these areas actually encompasses the financial aspects related to prolong hospitalization and/or caretaking of Hospital Acquired Infections (HAIs) caused by resistant microbial strains (Bhutta et al., 2014;Jasovsky et al., 2016). In fact, infectious diseases (IDs) impede human development in lowincome communities and impact basic frames like education and health in general. In the preantibiotic era, IDs used to be serious threats because of lack or insufficient knowledge on microbial life and variants of host-microbe interactions. Inspired by the findings of Louis Pasteur and Robert Koch in the 1860s and Fleming in the 1929s, better understanding was acquired to cope with bacteria. Improving knowledge on microbial life and host-microbe interactions became a permanent challenge for IDs investigators in connection with the stochastic feature of outcomes emerging as combinatory events between inherent and acquired attributes from both partners (genome flexibility, other forms of life and host defense, for instance). This interest in bacterial-host interactions has consistently shifted over time, extending from bacteria as "professional pathogens", "antibiotic resistant professional pathogen", "resistant opportunistic IDs etiologies" to the famous current "resistant environmental microbial flora" recognized as potent threats in healthcare facilities (Barber, 1961;Spinosa et al., 2000;Tagoe et al., 2011;Martínez and Baquero, 2014;Baquero et al., 2015;Nurain et al., 2015;Simo Louokdom et al., 2016;Noukela Noumi et al., 2016). Otherwise, it becomes clearer in the light of advances in mastering host-parasite interactions that persistent relationship and fitness are tributary to several constraints which interact to provide conditions for sustainable coexistence. In this, and in order to express the inherent ability enabling perpetuation, all organisms respond to ranges of environmental signals that might vary both in types and magnitudes (Baquero et al., 2015;Teillant, 2015;Laxminarayan and Chaudhury, 2016). Disrupted

harmony generally results from unusual signals that an organism receives from the surrounding multi-factorial systems. This type of interactions have been important in microbial evolution, in connection with the genome litheness that regularly adapt to changes and more specifically, to bacteria in which the clearer picture depicts a putative common "tool sets" from which all species might select the genetic traits they require for survival. In this, antibiotic resistance (AbR) is one of the most relevant features that enable materialization of adaptive genetic changes for better fitness in microbial life. AbR was first noticed by the end of the Second World War in the 1840s with resistance to Penicillin. Resistance to inhibitor-resistant Penicillin (Methicillin) which later emerged in the 1961s (Barber, 1961) further drew awareness on the reduced efficacy of previous potent antibacterial agents around the time. It was however, only by the 1980s that AbR, recognized as threat for global health, excited greater interests on bacteria and AbR. Nowadays, it is proven that this natural phenomenon is typically amplified by human activities through misuse of antimicrobials in human health, animal husbandry and crops production (Smith et al., 2014;Islam, et al., 2014;Teillant, 2015;Ramanan and Ranjit, 2016). The impact of chemicals like heavy metals from industries through untreated or partially treated wastes released in the environment could also be alleged (Matyar et al., 2008). Mastering IDs has long been associated with control of known pathogenic types. Often omitted in AbR investigations, it is currently admitted that endogenous and environmental microbial flora might play significant roles in selecting, maintaining and spreading resistance traits (Spinosa et al., 2000;Tagoe et al., 2011;Oluyege et al., 2015;Simo Louokdom et al., 2016) undermining thereby, all infection prevention and control strategies in healthcare systems. Next to previous works and within the global framework that aims at investigating novel approaches in tracking, tracing and controlling AbR in order to mitigate resistance effect in healthcare facilities, the present work was conducted in three hospitals in which bacterial flora from healthcare workers and hospital environmental items were studied. This investigation conducted on common opportunistic and environmental bacteria will serve as groundwork to initiate and implement a stewardship program with holistic database. MATERIAL AND METHODS Ethical consideration, study population and collection sites Prior to field works, the go-ahead was obtained from the Université des Montagnes Ethical Committee under reference number 2015/074/UdM/PR/CAB/CIE. Authorizations were also obtained from the Regional Hospital of Bafoussam (high-ranked State Institution), the Protestant Hospital of Bangwa (Church property) and the District Hospital of Bangangté (low-ranked State Institution). Typically rural (Bangwa), semi-urban (Bangangé) and urban (Bafoussam), the general host populations share common traditional family and community values like assistance and physical presence by the side of a sick member. In these areas, traditional medicine practices and automedication with conventional drugs are so enrooted that most of the times only serious cases are taken to the hospital. Clean water resources are scarce and most people in these areas rely on wells and streams that also commonly serve as rubbish dump for local wastes. This scarcity extends to human resources for health and contrasts with fast growing populations accompanied by large networks of drugs from doubtful origins in all settings (unpublished). Specimens collected from inanimate items (surfaces and medical devices) and healthcare workers who consented to participate were processed at the Université des Montagnes Teaching Hospital's Infectious Diseases Laboratory. The number of collection sites per ward depended on the number of potential bacterial reservoirs that could be identified. Generally for this purpose, one item was selected out of three. Sampling, culture, isolation, identification and susceptibility testing Sampling After all necessary ethical and administrative requirements were fulfilled; samples were collected as follows: first, healthcare workers fingers were printed on appropriate agar in Petri dishes and secondly, humidified sterile swabs were used to rub approximately 2.5x4 cm 2 surface areas on inanimate surfaces and hospital devices. All collected materials, preserved in brain-heart infusion agar and Petri dishes were kept in refrigerated containers (4-8 °C) and immediately conveyed to the Laboratory for analysis. Culture, isolation and identification Culture and isolation were conducted on MacConkey agar (Liofichem ® ) for Enterobacteriaceae, Columbia agar (Liofichem ® ) with 5% sheep blood supplement and chocolate agar for fastidious bacteria like Streptococci; and mannitol-salt agar (Liofichem ® ) for Staphylococci. For non fastidious bacteria, incubation was performed aerobically for 18-24 h at 37 °C. Fastidious bacteria were searched for after incubation for 24-48 h under 5% CO 2 . When incubation was complete, the morphology of bacterial colonies was used for presumptive identification. Then, all suspected colonies were characterized on the basis of their biochemical properties, according to the standard guidelines (REMIC, 2007). A culture was regarded as positive (for high bacterial density on the item) when a total count of at least 8 CFU/cm 2 was obtained (modified Vandini et al., 2014). For identification and susceptibility tests, reference bacterial strains used in quality control were S. aureus: ATCC 29213, S. aureus: QC 1625, E. faecalis: ATCC 29212 and E. coli: ATCC 25922. Susceptibility testing This was carried out on 24 h bacterial pure culture. Prior to the test, bacterial isolates were streaked on fresh agar and incubated at 37 °C overnight. From the resulting bacterial population, a suspension to the density of a McFarland 0.5 turbidity standard in 0.9% saline was prepared and adjusted to the final opacity recommended for susceptibility tests by agar diffusion technique on Mueller Hinton or chocolate agar according to the "Comité de l'Antibiogramme de la Societé Française de Microbiologie, CA-SFM, EUCAST, 2014". RESULTS Overall, 109 items were targeted in the three hospitals. Hospital wards and inanimate materials of interest were summarized and displayed as shown in Table 1. Overall picture showed that types and numbers of targets varied from one hospital to the other for similar wards and that, within the same hospital, inter-ward variations were observed. In this, the largest numbers of potential reservoirs were recorded in the District Hospital of Bangangté. When contamination rates were assessed based on culture results and the number of these potential reservoirs, data were grouped once again per hospital and summarized as displayed in Table 2. It disclosed that a total of 238 samples were collected. In Bangwa all items' in the Operating Theater and the Maternity were contaminated above the assigned threshold, like all healthcare workers hands. Contamination due to Gram-negative rods ranged from 11-20% on inanimate hospital items and 9 -14% on healthcare workers'. With regards to Gram type from healthcare environments, the total of 428 Grampositive bacteria overwhelmed the isolations. Further details indicated that 92%, 86%, and 81% of bacterial isolates from Bangwa, Bangangté, and Bafoussam respectively, belonged to one of the genera Staphylococcus and Bacillus. In this, Bacillus spp. (≈ 62%) overwhelmed Staphylococcus spp. Otherwise and from specimens (collectively from hands and hospital inanimate items) 8%, 14% and 19% were other than Gram-positive bacteria in the respective healthcare institutions. In the BPH in addition, Clostridium spp. was isolated from the Operation Theater. Susceptibility tests were therefore performed on major Gram categories of isolates according to their shapes and origins; then displayed in Table 3 and Table 4 as data summary from hospital devices and healthcare workers hands, respectively. Most reduced susceptibility rates were obtained with beta-lactams headed by drugs from the penicillin sub-group (Penicillin and Oxacillin) third generation cephalosporins (Ceftriaxone, Ceftazidime) and Aztreonam regardless of healthcare facility. This reduced susceptibility was also recorded with Vancomycin (the Glycopeptide sub-group representative) and Nalidixic acid (first generation quinolone). In general, however, reduced susceptibility rates were relatively higher in Bacillus spp., while higher effectiveness was observed with Gentamicin and Ciprofloxacin; observed in both Bacillus spp., Staphylococcus spp. To a lesser extent, this effectiveness was recorded with Nitrofurantoin. The above tabled data further indicated varied levels of intermediate phenotypes. A similar frame used to address caregiver's hands' resulted in the total of 95 isolates (Staphylococcus spp. 52% and Bacillus spp. 48%) recovered and used for susceptibility testing (Table 4). Overall, reduced susceptibility was observed in both major bacterial types in all settings. The susceptibility rates in Bacillus spp., and Staphylococcus spp., were similar for Oxacillin and Ceftriaxon (<10%). Also, similar rates were recorded in Bacillus and Staphylococcus isolates recovered in Bafoussam for Penicillin; and Bangangté for Vancomycin and Cefoxitin. In the Staphylococcus sub-group especially, susceptibility rates lower than 50% was recorded for 87% of the antibiotics in Bangangté, 82% in Bangwa, and 75% in Bafoussam. DISCUSSION Data analysis from the present multicenter survey disclosed high isolation rates of Bacillus spp., and Staphylococcus spp. (>80%, collectively in all settings), in connection with the low rates of the Gramnegative bacteria which are most commonly alleged in HAIs and not stressed in the present discussion to minimize emphasis dilution. In a former study (Nagajothi et al., 2015) the contamination rate estimate, 97.5%, was dominated by Gram-negative rods and coagulase-negative cocci. Reduced susceptibility was also observed in the major bacterial Gram types which overwhelmed the isolations in that survey. These contamination rates were almost similar from one hospital to the other, and, within the same hospital amongst the investigated wards. Moreover, bacteria populations appeared similar in their proportions regardless of the location. The low isolation of Gram-negative rods could be associated with the conventional hygienic cleaning that might have been effective at a certain degree or, could be in connection with the large Bacillus populations. According to several authors in fact (Vandini et al., 2014;Martínez and Baquero et al., 2014;Amin et al., 2015), Bacillus can produce biocides which naturally and adversely affect the growth of Gram-negative bacteria. Further, and with regards to cultural values, hospital external environments and patient wards regularly and indiscriminately host visitors that likely reintroduce new germs soon after cleaning. This assertion is consistent and substantiated by data from the Pediatrics of all three hospitals where the contamination rates were found above 80%. Accordingly, the least contaminated wards were recorded at the Regional Hospital of Bafoussam (RHB). This public system-empowered healthcare facility is one of the high-ranked hospitals in Cameroon, before the BPH (a church property) and the BDH (the other public system hospital). Regarded as privileged amongst the three, it could also be observed that it is in the RHB's Operating Theater that the lowest contamination rate was recorded throughout; justifiable by the restricted in-andout movements of people amongst other restrictive measures. Admitting that the expected phenotype will only be delayed if the bacterial inoculum (number of individual bacteria per unit space or volume) is below a certain threshold, it can be understood how proper conventional cleaning that is more effective in some health facilities (especially resource-empowered ones) would reduce germ spread and mitigate HAIs. Viceversa, reduced load would facilitate effective cleaning with conventional techniques. However, because conventional cleaning is relatively costly, it cannot be sustainably afforded by many communities, especially, those in the context of resource limitation like the ones concerned in the present works. One of the key points that deserved emphasis appeared, therefore, to be awareness and education on microbe-host interactions in connection with microbial loads for healthcare personnel, for beneficiaries of healthcare services and for their relatives as well as the role of hygiene. Bacillus and Staphylococcus overwhelmed the isolations. Often omitted in general when addressing R/S in bacterial populations, the mastering of such parameters like population density, extreme flexibility (common in all bacteria communities) with regards to increasing numbers of immunedepressed human hosts (favored by advanced age, famine, malnutrition and other natural threats) make of these organisms, potent etiologies of human disorders and/or reservoirs for genetic elements dissemination (Spinosa et al., 2000;Yassin and Ahmad, 2012;Oluyege et al., 2015). This is consistent with the currently admitted concept of "bacteria tool set" that helps microbiologist communities to understand, at least partially, the unpredictable dynamics of bacterial genomes enabling the strain-adapting ability which is necessary for survival and explains the stochastic ranges of infections they can cause in their hosts. Moreover and with this dynamics, the Koch's Postulate used to identify IDs etiologies would prevail over the definition referring to specific genetic traits in "true or professional pathogens" when it comes to identify ID etiologies for several reasons two of which are: 1 st , the genetic map may undergo changes that elicit new putative phenotypes which might be harmful to the host and 2 nd , as not all known genetic elements

are fully described, advent of Bacillus ssp., as causative agents of HAIs is likely, theoretically, beyond their acknowledged implication in human disorders such as anthrax and other food-borne infections, for their predominant populations in all settings (overall ≈62%). This allegation would be supported by the increasingly recognized role of strains from the related bacteria genus, Clostridium, in healthcare facilities (Dubberke et al., 2009;Foster et al., 2011) and the increasing numbers of susceptible hosts as stated above. Resistance to several antibiotics was clearly highlighted in the present survey, independently of the locations. Acknowledged as facilitated by misuse of antimicrobial agents as discussed in previous works (Kouamouo et al., 2013;Fotsing Kwetche et al., 2015), positive selection and diversification promote fitness that results in emerging phenotypes with broaden host spectra, likely colonizers of new niches. This signal-dependent (type-and magnitude) natural phenomenon is known to occur in time and space scale throughout history and in all living systems (Planta, 2007;Martínez and Baquero, 2014;Laxminarayan and Chaudhury, 2016). The terms "Staphylococcus spp." used in the present works actually encompassed coagulase-positive and coagulase-negative strains recognized as major agents of IDs (Anonynous, 2015a). One of the famous members, S. aureus, is common etiology of infections in almost all bodily systems. With a few exceptions (etiologies of anthrax and foodborne infections, for instance), members of the genus Bacillus have long been regarded as common hosts of the environment therefore, ignored in human disorders. In the light of advances, it becomes clearer that they might be relevant as causative agents or as engine for bacterial persistence within a niche for their role as providers of resistance genetic traits to all bacterial communities (Spinosa et al., 2000;Tagoe et al., 2011;Anonynous, 2015a;Noukela Noumi et al., 2016). This assertion is consistent with the known role of members of the related Clustridium in case of immune-compromission (Dionisio et al., 2002;Forster et al., 2013;Singh et al., 2013;Brouwer et al., 2013;Shen et al., 2013;Barbosa et al., 2014;Martínez and Baquero, 2014;Anonynous, 2015a). The R/S patterns that appeared invariably high in all settings were also consistent with the shared socio-cultural values and, undoubtedly, in connection with the purchasing power of the local populations. This special view needs to be addressed in order to redirect and redesign disease prevention and the healthcare delivery concomitant practices that mitigate HAIs. Necessary measures would include amongst others, the provision of human and financial resources approaching the required healthcare provider/healthcare needs ratio. Separate previous studies reported the alarming trends of resistance in the West Region of Cameroon (Kouamouo et al., 2013;Simo Louokdom et al., 216). The most affected antibiotics belonged to the large family of beta-lactams, and first generation of quinolones while, Ciprofloxacine, Gentamicin and, to a lesser extent, Nitrofurantoin could be used with the highest likelihood of therapeutic success. These authors blamed poor sanitation and misuse of antibiotics in humans; and alleged the impact from animal husbandry. The consistent blame on alleged poor sanitation was partially substantiated with data from the Ndé Division where only 25% of the daily needs in water are provided by the company in charge, and the in-charge population/medical doctor ratio estimated at 8000, according to the Bangangté Municipal council's report in 2015 (Anonymous, 2015b). With this resource scarcity it can be understood why healthcare in general and IDs prevention and management could be negatively affected, as most needy that cannot afford medical attention would refer to auto-medication. Still previous investigations reported that different resistance mechanisms also interplayed in local bacterial population and that bacterial strains could disseminate from one hospital area to the other (Noukela Noumi et al., 2016), as observed in each of the three hospitals targeted. R/S profile for Staphylococcus spp. and Bacillus spp. were similar in many cases. Whether strains actually carried the same genetic elements is likely, but yet to be fully addressed. Conclusion With a few exceptions, all settings were contaminated with high rates of Staphylococcus spp., and Bacillus spp. Bacillus (≈62%) overwhelmed the isolation while the susceptibility rates were similar in both bacteria groups and indiscriminately high, especially with antibiotics that are commonly used. Gentamicin, Cirpofloxacin and Nitrofurantoin proved to be most effective. Strains dissemination from one place to the other also appeared frequent in all settings. Consistent with previous works, these common bacterial types could sustainably be used in a cost-effective active R/S stewardship program. But the State support to public and private healthcare facilities is crucial for the HAIs mitigation to be achieved, especially in high-risk-patient's environments. CO M PE T IN G IN T E R E S TS The authors declare that they have no competing interests. AUTHORS' CONTRIBUTIONS RFTN, MT, PRFK and DPNN did the field and laboratory investigation. JK, JSL, SGD, SHT, BPTK and KK performed the data analysis and revising manuscript. Preclinical evaluation of new α-radionuclide therapy targeting LAT1: 2-[211At]astato-α-methyl-L-phenylalanine in tumor-bearing model CURRENT STATUS: POSTED PURPOSE: Targeted α-radionuclide therapy has growing attention as a promising therapy for refractory cancers. However, the application is limited to certain types of cancer. Since L-type amino acids transporter 1 (LAT1) is highly expressed in various human cancers, we prepared LAT1-selective α-emitting amino acid analog, 2-[211At]astato-α-methyl-L-phenylalanine ([211At]-2-AAMP), and evaluated its potential as a therapeutic agent. METHODS: [211At]-2-AAMP was prepared from the stannyl precursor. Stability of [211At]-2-AAMP was evaluated by both in vitro and in vivo. In vitro studies using LAT1 expressing human ovary cancer cell line, SKOV3, were performed for evaluating cellular uptake and cytotoxicity of [211At]-2-AAMP. Biodistribution and therapeutic studies in SKOV3 bearing mice were performed after intravenous injection of [211At]-2-AAMP. RESULTS: [211At]-2-AAMP was stable in murine plasma in vitro and excreted into urine as intact. Cellular uptake of [211At]-2-AAMP was inhibited by treatment with LAT1-selective inhibitor. After 24 hours of incubation, [211At]-2-AAMP suppressed clonogenic growth at 10 kBq/ml, and induced cell death and DNA double strand break at 25 kBq/ml. When injected to mice, [211At]-2-AAMP exhibited the peak accumulation in the tumor at 30 min postinjection, and the radioactivity levels in the tumor retained up to 60 min. The majority of the radioactivity was rapidly eliminated from the body into the urine as an intact form immediately after injection. [211At]-2-AAMP significantly improved the survival of mice (P<0.05) without serious side effects. CONCLUSION: [211At]-2-AAMP showed α-radiation-dependent cellular growth inhibition after taking up via LAT1. Furthermore, [211At]-2-AAMP provided a beneficial effect on survival in vivo. These findings suggest that [211At]-2-AAMP might be useful for treatment of LAT1-positive significantly improved survival rates of mice in vivo. These results suggest that [ 211 At]-2-AAMP is possibly beneficial for treatment of LAT1-positive cancer. However, significant delay or reduction of tumor size was not observed by treatment with [ 211 At]-2-AAMP in vivo partially because small number of mice. Since the body weight loss after 2 MBq of [ 211 At]-2-AAMP injection was not large and the present administration dose did not reach the maximum tolerated dose, suggesting that higher doses would be safely administered to Introduction Targeted radionuclides therapy with α-emitters has drawn global attention. Linear energy transfer (LET) of α-particle is approximately 80 keV/μm, and thus α-radiation is considered to be high-LET radiation [1,2]. According to the lots of studies with external beam irradiation, high-LET radiation can induce irreparable DNA double-strand breaks and effectively lead cell death compared to low-LET measured with a well-type g-counter (ARC-7001, Hitachi-Aloka Medical, Tokyo, Japan). For examining extracellular release, SKOV3 cells were incubated in HBSS containing [ 211 At]-2-AAMP for 10 min. After 2 times washes with HBSS, cells were incubated in HBSS at 37 o C for 10, 20, 30, and 60 min. Then, supernatant was collected, and the radioactivity was measured with a well-type g-counter. Colony assay Cells (1×10 6 cells/dish) were pre-incubated in the growth medium for 24 h, and treated with 0, 10, 25, and 50 kBq/ml of [ 211 At]-2-AAMP for 24 h. After [ 211 At]-2-AAMP treatment, cells were washed with PBS, suspended in growth medium, and seeded at 200 cells/well in 6-well plate. Then, cells were incubated at 37 o C for 14 days. After incubation, cells were washed with PBS twice and stained with crystalviolet solution (6% glutaraldehyde, 0.5% crystalviolet). After washing cells with tap-water, the number of colonies (> 50 cells) were counted. Competitive inhibition of [ 211 At]-2-AAMP with various inhibitors showed that the uptake of [ 211 At]-2-AAMP was significantly inhibited by treatment with AMT, a selective inhibitor of LAT1, as well as substrates of LAT1, such as branched-chain amino acids and aromatic amino acids (Fig. 3d). As shown in Fig. 3e, [ 211 At]-2-AAMP was gradually released to extracellular space, and the released fraction reached more than 90% within 60 min after exchanging the buffer. Table 1 Therapeutic effects in tumor-bearing mice In our preliminary examination with small number of mice, weight loss more than 20% was not However, there was no significant difference of tumor volume between 2 groups ( Fig. 5a and b). Body weight was transiently reduced after treatment with [ 211 At]-2-AAMP ( Fig. 5c and d). The peak of weight loss was at 3 days after injection, and the maximum reduction was 14.5%. Kaplan-Meyer survival analysis showed that the survival of mice was significantly improved with [ 211 At]-2-AAMP treatment (Fig. 6, P<0.05). Discussion In the present study, we newly synthesized [ 211 At]-2-AAMP and evaluated its potential as an α- and its therapeutic effect in vivo [17,18]. However, L-phenylalanine is a substrate for not only LAT1, but also LAT2 [19]. In this regard, this is the first report of a LAT1-specific 211 At-labeled amino acid derivative. In general, astatinated aryl compounds can be prepared by the procedure similar to be used for radioiodinated and radiobrominated compounds. in pancreas compared to the other organs [21]. No significant uptake of [ 123 I]-3-iodo or [ 18 F]-3-fluoroα-methyltyrosine, other LAT1 specific tracers, was not observed in the human pancreas [22,23]. Therefore, the high distribution in the pancreas will not be expected in patients. Dehalogenation of Case Series of Herpes Zoster in Children ; Ultraviolet as a Trigger ? Background: The condition of immunosuppression is a known condition that triggers reactivation of varicella zoster virus, to its manifestation as herpes zoster. Excessive exposure to ultraviolet is shown to decrease the immune system, even the effects can be systemic. Children often experience excessive UV exposure during play. Is excessive UV exposure a precursor to HZ in children? Case Report: We observed five cases of HZ in children aged 3-14 years; three of them are HZ ophthalmic, one HZ thoracic and one HZ brachialis. All had never received vaccine immunization and no history of intrauterine infection. Only one of them has suffered from varicella. There was no pathologic immunosuppression (HIV infection, organ transplant, immunosuppressant therapy, systemic disease) in these five children, but all children experienced intense UV exposure 3-5 days before the cutaneous lesions appeared. After treatment according to the HZ protocol, all recover without complications. Discussion: Various experimental models prove that UV exposure can suppress the immune system in some conditions: after vaccination, infection and provocation of allergic contact dermatitis in skin exposed directly or indirectly. This condition is in accordance with the state of immunosuppression that can trigger the reactivation of VVZ to HZ. The incidence of HZ is also reported to increase in the summer. All the patients we observed had intense UV exposure a few days before the emergence of HZ cutaneous lesions. Although only one patient clearly showed a history of varicella, subclinical VVZ infection may explain the reactivation of VVZ in all four cases. Based on the various evidence mentioned above, we conclude that intense UV exposure is very likely the main precipitating factor of HZ reactivation in children. Introduction Herpes zoster (HZ) is caused by reactivation of varicella zoster virus (VVZ), it can occur at any age but is rare in children and adolescents, although recent studies have shown an increasing incidence in children [1][2][3]. In children usually without prodromal symptoms, skin lesions are lighter, shorter duration and less postherpetic neuralgia. The incidence of HZ in children is reported to be about 42-238,5 per 100,000 people per year [3]. In children aged less than 10 years of 0.25-1.15 per 1000 and 0.43-1.60 per 1000 HZ in adolescents. Impaired immunity such as a history of primary intrauterine infection, immunocompromised, and infected first-year varicella of childhood is a risk factor for HZ in children [2][3][4][5][6]. The

virus replicates in the sensory dorsal nerve ganglion and when a person's cellular immunity decreases, the activated VZV travels through the sensory nerve to the skin [7]. Ultraviolet (UV) exposure, especially UVB waves, can suppress the immune system in several ways. UVB inhibits antigen presentation, inducing the release of immunosuppressive cytokines and causing leukocyte apoptosis. However, UVB does not cause general immunosuppression but inhibits immune reactions through specific antigens. Application of contact allergens on UV exposed skin does not cause sensitization but induces specific antigen tolerance because such individuals cannot be sensitized to the same allergen in the future. This specific immunosuppression is mediated by a specific antigen suppressor/ regulatory T cell. UVB DNA damage is a major molecular trigger for immunosuppression. Presentation of antigens by Langerhans cells that have been damaged by UV in lymph nodes appears to be one of the essential requirements for the formation of regulatory T cells [7][8][9]. Case Report There were reported five cases of HZ in children. All the cases we observed had never received vaccine immunization and no history of intrauterine infection. Only one of them has suffered from varicella. No immunosuppressed condition (HIV infection, organ transplant, immunosuppressant therapy, or systemic disease) was observed in five children. Case 1 Boy, aged 3 years, according to his mother appeared a rash of multiple clusters on arms, fingers III-IV and part of the palm of the right hand suddenly since 2 days ago without fever and pain. On physical examination found vesicles clustered in the region brachialis, ulnar and medianus. History of UV exposure while swimming in the pool four days before the cutaneous lesions appear and has suffered from varicella at 2.4 years of age. Case 2 Girl, aged 7 years, 3-day history that occurs in reddened skin on the left eyelid, forehead and nose, burns without fever, no headache and no sore eyes. On physical examination vesicular eruptions were found in groups in the left ophthalmic region. History of UV exposure while playing on the beach three days before the cutaneous lesions appear. Case 3 Boy, aged 7 years, 5-day history unnoticed appeared rash filled liquids clustered, then broked and dried on the forehead, eyelids and nose; pain without fever, burning and headaches. On physical examination found erosion, crusting in the right ophthalmic region. History of UV exposure when swimming on the beach four days before the cutaneous lesions appear. Case 4 A 12-year-old boy, said 3 days ago experience pain fluid filled lession clustered in the left chest distribute to the left arm. On physical examination found vesicles clustered in the thoracic region (dermatomes T4-T5). History of UV exposure when follow fun-bike five days before cutaneous lesions appear. Case 5 A 14-year-old boy, a 5-day history of right side right headache and 4 day appeared a rash of multiple clusters liquids on the eyelid and right side of the forehead is mild and swollen. On examination found vesicular and crustal eruptions in the ophtalmic region. History of UV exposure when working on the field five days before skin lesions appear. Discussion Diagnosis of HZ is primarily clinical but can also be done with Tzank smear examination and to distinguish with herpes simplex can be done direct fluorescent monoclonal antibody test [2][3][4][5]10]. The case of HZ brakhialis that we observed was supported by Tzank smear examination and found multinucleated giant cells. The other four cases were diagnosed only clinically. All children treated with acyclovir 20 mg/ KgBB, four times daily for five days and symptomatic therapy for itching and pain [1][2][3]. The reactivation mechanism of VZV is not fully understood. It is believed that reactivation is the result of reduced cellular immunity specific to VZV, and periodic exposure to people with varicella or herpes zoster will enhance cell-specific immunity against VZV characterized by periodic subclinical reactivation. Many studies have shown no pattern of herpes zoster season, while other studies suggest a higher incidence in summer in the middle of the year, as a consequence of increased exposure to ultraviolet light [5,6]. The results of Wu et al. study found an increase in HZ incidence very strongly associated with UV. Recent research has reported an increase in the incidence of HZ in summer in countries with several different seasons of the year, this pattern is associated with UV peak summer exposure and possibly a HZ trigger [5][6][7]10,11]. All the children we reported experienced intense UV exposure 3-5 days before the cutaneous lesions appeared, so we concluded that UV exposure was the initiator of VZV reactivation. Although it is not argued that UV exposure can suppress immunity, it is clear that the way it works is very complex with the possibility of various pathways involved in local UV immunosuppression; an antigen is applied to the irradiated area, apparently local suppression due to contact hypersensitivity. This corresponds to the three cases of ophthalmic HZ we observed. Whereas under systemic UV immunosuppression conditions, the antigen is applied to an unexposed area but no alteration in function or antigen presenting cells amount in the lymph node adjacent to the contact site with the antigen. Thus, there may be other types of cells or the release of mediators from the exposed area. Various mediators are produced in areas exposed from keratinocytes to other cells. The first possibility is a platelet activation factor that can bind to receptors in monocytes, mast cells and the release of keratinocytes and the release of prostaglandins. This is followed by the release of various cytokines, including interleukin (IL)-4 and IL-10, both of which are immunosuppressive. Other molecules are also found locally, such as histamine, progtaglandin, TNF-α, IL-1β, neuropeptides and neurohormones, which can have systemic effects [8,9,11,12]. The allegations above may explain the effects of immunosuppression in one patient observed which suffered from HZ thoracic, but UV exposure did not occur directly in the area of the rash due to a closed rash. Complementary systems are also involved. UV-activated C3 is important for skin infiltration by monocytes/ macrophages (CD11b + cells), which contribute to immunosuppression [9,11]. All of the above components have significant effects on the migration and function of some immune cells, some of which are involved in antigen presentation. Some Langerhans cells, the main antigen presenting cells in the epidermis, migrate into the lymph nodes by passing the exposed area, or if the UV dose is high, then it may experience apoptosis in situ. Dermis dendritic cells also migrate to the drainning of the lymph nodes. Specific monocyte / macrophage populations capable of producing IL-10, migrate into the skin and then to lymph node drainage when stimulated with antigen. The end result is thought to be an abnormal change of antigen presentation in the epidermis, dermis and lymph node drain. In the latter case, there is a decrease in the production of IL-12 and IL-23, a key cytokine that normally promotes the activation of immune cells including T cells and which is capable of reducing DNA damage due to UV exposure. Simultaneously, the T helper (Th1) cytokine level was reduced, and a group of regulatory T cells (Treg) were stimulated, specific to antigens found at exposure to UV. These cells have CD4 +, CD25 +, Foxp3 + and CTLA4 + phenotypes. They are cytotoxic for antigen presenting cells, producing IL-10 at activation and may suppress activation, cytokine production and the proliferation of other immunostimulatory T cells. There is also little evidence of the involvement of natural killer cells (NK), which secrete IL-4 during activation. These cells represent a unique collection of lymphocytes that express NK cell markers plus T cell receptors [9,12]. Although the number of infections in humans affected by UV radiation exposure is quite limited, the most interesting in the field of UV infection and radiation is the question of whether UV exposure can adversely affect the immune response to vaccination. Four animal models have indicated that UV radiation exposure has the ability to alter the efficacy of vaccination in such a way that the response/ memory response produced by the vaccine is significantly reduced. There is little research on human vaccination and little data is available today regarding infection protection [8]. Only one of the cases we observed had varicella at 2.4 years and on the C5-T1 dermatome. A study concluded that children younger than 2 years with varicella infections have the highest and most rapid risk of HZ. Other studies suggest that immune status at the time of primary infection is important in determining the presence of HZ in childhood. Low lymphocyte levels, NK cells and cytokines are found in infants, along with viral-specific immunoglobulins. All of these can cause an inability to keep VZV latency so that zoster arises early in children [3]. This serial case observation proves that UV exposure as one of the precipitating factors for HZ reactivation in immunocompetent children determines the education that should be given to the public regarding their exposure to UV to prevent HZ in children. Further observation in larger populations is still needed. Giardia duodenalis infection in dogs affected by primary chronic enteropathy Background: Canine primary chronic enteropathy (CE) includes a heterogeneous group of diseases characterized by chronic gastrointestinal signs. Aim: This study evaluated the occurrence of Giardia duodenalis infection in primary CE-affected dogs. Methods: Forty-seven CE-affected dogs of different age and sex were enrolled in the study. For each dog, frequency of defecation, fecal consistency, and eventual fecal abnormalities were evaluated. A clinical scoring index of CE severity (clinical chronic enteropathy activity index) was also assessed, and the type of enteropathy was retrospectively classified. For parasitological analysis, fresh fecal samples collected from each dog were examined by fresh and Lugol stained smears, flotation test, and a rapid immunoassay. Giardia duodenalis genotypes were identified by molecular analysis. Differences of clinical parameters between G. duodenalis positive and G. duodenalis negative dogs were statistically evaluated. Results: Among the CE canine patients, 16 out of 47 (34%) dogs were found positive for G. duodenalis and assemblages C and D were identified. No statistical differences emerged according to the types of CE between G. duodenalis-positive and G. duodenalis-negative dog groups. The clinical index of CE severity was indicative of significant less severe clinical forms in G. duodenalis-positive dogs (p = 0.037). Conclusion: Results here obtained shows how G. duodenalis may be present in primary CE-affected dogs and further investigations are needed to clarify the real significance of mild clinical presentation in G. duodenalis-positive dogs affected by CE. Introduction Giardia duodenalis is a worldwide intestinal protozoan parasite that infects a wide range of hosts, including humans, domestic, and wild mammals (Monis et al., 2009). The localization site of G. duodenalis is the small intestine, mainly duodenum and jejunum, and in the infected hosts, it may be responsible for gastrointestinal signs (Hawrelak, 2003;Tangtrongsup and Scorza, 2010). Based on genetic analysis, G. duodenalis is considered as a species complex, which includes at least eight distinct genetic groups or assemblages, from A to H (Ryan and Cacciò, 2013). Zoonotic assemblages A and B and canine-specific assemblages C and D have been reported in dogs (Ballweber et al., 2010;Ryan and Cacciò, 2013;Sommer et al., 2018). The prevalence of G. duodenalis infection in dogs may vary depending on the population examined and the diagnostic method used (Ballweber et al., 2010;Epe et al., 2010). Younger animals and some dog communities, as stray and kennels dogs, have a higher risk of G. duodenalis infection than other dog populations (Huber et al., 2005;Tysnes et al., 2014), while dogs kept as pets are less likely to be positive (Bouzid et al., 2014). The prevalence of G. duodenalis in pet dogs in Italy is about 4%-29% (Riggio et al. 2013;Pipia et al., 2014;Zanzani et al., 2014), while it is about 25% in symptomatic dogs from Europe (Epe et al., 2010). Giardia duodenalis-infected dogs may show a spectrum of clinical signs ranging from subclinical forms to acute and intermittent forms or chronic diarrhea (Epe et al., 2010;Tangtrongsup and Scorza, 2010). Factors associated with the onset of the disease include the age and the clinical, nutritional, and immune status of the infected dog (Roxstrom-Lindquist et al., 2006). Although some studies suggested a possible relation between G. duodenalis assemblage and the severity of clinical disease, other studies suggested the opposite (Tysnes et al., 2014). The pathogenic mechanisms proposed for G. duodenalis infections include production of

toxins, disruption of normal intestinal microbiota, inhibition of normal enterocyte enzymatic function, blunting of microvilli, intestinal motility disorders, intestinal epithelial cell apoptosis, and intestinal inflammation (Tangtrongsup and Scorza, 2010). To explain variations in signs both among dogs and over time within the same dog, it has been suggested that microbiota modifications may influence the pathogenicity of G. duodenalis (Tysnes et al., 2014). Associations with bacterial pathogens have been also hypothesized (Tysnes et al., 2014) although results from a recent study suggest that G. duodenalis may also protect against gastrointestinal disease induced by a co-infecting bacterial enteropathogen (Manko et al., 2017). In dogs, chronic enteropathy (CE) includes a heterogeneous group of diseases characterized by chronic gastrointestinal signs lasting from longer than 3 weeks, as diarrhea, vomiting, hyporexia, abdominal pain, weight loss, with the exclusion of extra-intestinal or intestinal diseases of other etiology (Dandrieux, 2016). At present, chronic enteropathies are classified by treatment response in food-responsive enteropathy (FRE), antibiotic-responsive enteropathy (ARE), immunosuppressant-responsive enteropathy (IRE), and non-responsive enteropathy (NRE). Moreover, to avoid misclassification, an early treatment with fenbendazole for Giardia is performed even if G. duodenalis is not detected by routinely laboratory tests (Hall and Day, 2017). However, because the complexity of interactions between dietary components, microbiota, intestinal epithelial integrity, and immune system is supposed to play the main role in maintaining chronic inflammation, the role of G. duodenalis in dogs with chronic enteropathy needs to be clarified. The aim of this prospective study was to evaluate the occurrence of G. duodenalis infection in dogs affected by primary CE. Animals Forty-seven dogs presented at the Veterinary Teaching Hospital (VTH) "M. Modenato" of the University of Pisa from January 2016 to March 2017, were enrolled in the study. At presentation to our VTH, parasitological examination for intestinal parasites was performed by using the same methods reported below. All dogs were treated then with fenbendazole (50 mg/kg once daily, for five consecutive days) before to be enrolled in the study even if there were negative for parasites. All 47 dogs had a history of recurrent gastrointestinal signs and were diagnosed with a primary CE before the enrollment in the study, since extra-intestinal or intestinal diseases of other etiology had been excluded by previous diagnostic and therapeutic work-up. More specifically, all dogs underwent a complete blood cell count, a complete biochemical profile, including serum trypsin-like immunoreactivity and an abdominal ultrasound. At the time of inclusion in the study, when the diagnosis of CE was made, a fresh fecal sample was newly obtained for the parasitological and molecular analysis. Written informed consent was obtained from all the owners of dogs enrolled in this study. Clinical analysis For each dog enrolled in the study, the frequency of defecation (normal ≤ 3 times/day) and fecal abnormalities, including the presence of blood and/ or mucus, were evaluated. In addition, the use of a fecal consistency scoring system (Purina Prolan Veterinary Diets. Fecal Scoring Chart. https:// www.proplanveterinarydiets.ca/wp-content/ uploads/2018/05/180107_PPPVD-Fecal-Scoring-Chart-UPDATE-EN-FINAL.pdf) allowed us to assign to each dog fecal sample a different score on a scale from 1 to 7. A clinical index of CE severity was also assessed based on the clinical chronic enteropathy activity index (CCECAI) (Allenspach et al., 2007;2016). Moreover, the type of enteropathy was retrospectively classified as FRE, ARE, or IRE. More specifically, FRE diagnosis was based on a positive response to a change of diet using an elimination diet. ARE was diagnosed when a positive clinical response to antibiotic treatment (tylosin 10 mg/kg/bid) was observed, intestinal signs reoccurred after the discontinuation of the treatment, and no other underlying etiology was identified. Finally, IRE was diagnosed in dogs with persistent or recurrent gastrointestinal signs who had a histopathological evidence of intestinal inflammation and all the possible causes of this condition were excluded (Hall and Day, 2017). Based on serum albumin (Alb) concentrations, CE was further classified as protein-losing enteropathy (PLE; Alb < 2.0 g/dl) or non-protein-losing enteropathy. Parasitological and molecular analysis For parasitological analysis, individual fresh fecal samples collected from each dog were examined within 24 hours by fresh and Lugol stained fecal smears and by flotation test with a low-density solution (33% ZnSO4 solution, specific gravity 1.18). A commercial rapid immunoassay for the search of G. duodenalis and Cryptosporidium spp. fecal antigens (RIDA QUICK ® Cryptosporidium/Giardia Combi, R-Biopharm, Darmstadt, Germany) was also used. Positivity for G. duodenalis at least to one of these methods was assumed to indicate the positivity of examined samples. For molecular analysis, G. duodenalis-positive samples were processed by a commercial kit (QIAamp DNA Stool Mini Kit, QIAGEN, Valencia, CA) for DNA extraction. A nested PCR protocol was applied to amplify a fragment of the small subunit ribosomal RNA gene (Read et al., 2002). Amplification products were run on 2% ethidium bromide agarose gel and visualized under ultraviolet light. Positive amplicons were purified using mi-PCR Purification Kit, Metabion International AG. Amplification products were sent to an external laboratory for sequencing (Bio-Fab Research, Rome, Italy); sequence multiple alignment was carried out by ClustalW to identify G. duodenalis assemblages. After performing all clinical and parasitological evaluations, all dogs that in the present study scored positive for endoparasites were treated with previously recommended antiparasitic protocols. Statistical analysis For all parameters considered in this study, i.e. age and sex, CCECAI, fecal score index (FSI), frequency http://www.openveterinaryjournal.com S. Perrucci et al. Open Veterinary Journal, (2020), Vol. 10(1): 74-79 of defecation (FD), and type of enteropathy, including PLE, ARE, FRE, and IRE, possible statistical differences between G. duodenalis positive and G. duodenalis negative dogs were evaluated. Data analysis was performed using the statistical software GraphPad Prism 7 and data were analyzed by the Chi square test and the Fischer's exact test. The significance level was set at p < 0.05. Ethical approval In this study, only fecal samples were collected from dogs with primary chronic entheropathy, thus a formal ethics approval was not applicable. More specifically, all 16 dogs were positive at the immunoassay, while 12 and 14 dog fecal samples scored positive for G. duodenalis at the fresh fecal smear and at the remaining tests performed in the study, respectively. The mean age of the G. duodenalis positive group was 5.3 years, with a median of 3 years (6 months-14 years), and only three subjects were younger than one year. In the G. duodenalis negative group, the mean age was 4.8 years with a median of 4 years (5 months-13 years). Due to the small amount of fecal material available, PCR was performed only on 9/16 dog samples found positive for G. duodenalis at parasitological examination, while only for 6/9 positive dogs it was possible to identify also G. duodenalis assemblage. G. duodenalis assemblage D was identified in five of them and assemblage C in only one. From the evaluation of possible differences between G. duodenalis positive and G. duodenalis negative dogs according to the FSI, a higher frequency of fecal consistency alteration was observed in the G. duodenalis negative dog group, although no significant differences emerged at statistical analysis (Table 1). Indeed, fecal consistency was greatly reduced (FSI 6 or 7) in the 77.4% (24/31) of G. duodenalis negative dogs and in the 44% (7/16) of the G. duodenalis positive dogs. However, significant differences (p = 0.037) emerged at statistical analysis between G. duodenalis positive and G. duodenalis negative dogs according to the CCECAI index, since most of Giardia-negative dogs had a moderate/severe CCECAI score (20/31, 64.5%), while among Giardia-positive dogs the same index was mainly low/negligible (11/16, 68.8%) ( among G. duodenalis negative dogs (9/31, 29%) in respect to that observed among G. duodenalis positive dogs (2/16, 12.5%). Discussion To date, this is the first study evaluating G. duodenalis in a selected population of dogs with a diagnosis of primary CE. Our findings are taken from a population which underwent on a fenbenzadole treatment prior the inclusion. The prevalence of G. duodenalis in primary CE found in the present study (34%) is much higher than data reported in recent studies in symptomatic dogs in Europe (Epe et al., 2010;Volkmann et al., 2017). Furthermore, it is higher than prevalence of G. duodenalis found in different dog populations in Italy (Pipia et al., 2014;Zanzani et al., 2014;Paoletti et al., 2015;De Liberato et al., 2018), including privately owned (about 4%) and kennel (about 5%) dogs of the same area (Riggio et al., 2013;Sauda et al., 2018). In our opinion, this high prevalence which we found it is unlikely to be associated to a primary G. duodenalis infection. Although the reported efficacy of fenbenzadole treatment is very high (Barr et al., 1994;0 et al., 1998), a fenbendazole resistance may occur in dogs. Another possible explanation is that the risk of G. duodenalis infection seems to increase with high frequency of anthelmintic treatment. This may be due to the change of the intestinal niche caused by the anthelmintic therapy on major parasites (Bugg et al., 1999). However, it is not possible to exclude that the positive dogs may be newly infected due to incomplete animal and/or environmental disinfection (Raza et al., 2018). Finally, in our opinion, this result may be related to the dog population here considered because the intestinal alterations caused by chronic inflammation may predispose CE-affected dogs to the acquisition of G. duodenalis infections. However, it is unlikely that G. duodenalis may have a significant role in the gastrointestinal signs in our dogs. Indeed, from the evaluation and scoring of clinical parameters, the presence of G. duodenalis was not associated with more severe clinical forms. On the contrary, the CCECAI score was indicative of significantly more severe clinical forms in G. duodenalis negative group. In addition, FSI and the presence of mucus or blood in stools were not statistically associated with G. duodenalis infection. No significant differences emerged also between G. duodenalis-positive and G. duodenalis-negative dogs according to the type of enteropathies (e.g., ARE, FRE, IRE, and PLE). Giardia duodenalis infections can be asymptomatic, subclinical, and clinically evident forms may occur especially when other factors are also present, such as concurrent entero-pathogenic bacteria or parasites, food intolerance or decreased host defense ability and stress conditions (Ballweber et al., 2010;Tysnes et al., 2014). Interestingly, all clinical parameters used in this study to score the severity of signs were indicative of less severe clinical forms in the group of G. duodenalisinfected dogs, especially the CCECAI score that gave statistically significant results. To the best of our knowledge, no previous studies specifically investigated the prevalence of G. duodenalis in canine primary CE; its possible associations with the type of enteropathy and the severity of clinical presentation. In fact, only the stool consistence and the presence of diarrhea were previously evaluated in dogs infected by G. duodenalis (Epe et al., 2010;Upjohn et al., 2010;Pipia et al., 2014;Volkmann et al., 2017). However, it is also possible that some dogs considered negative in this study were instead positive, since only a single fecal sample per dog was examined. In fact, due to the inconstant shedding of the cysts into the feces, for the detection of G. duodenalis it is advisable to examine three samples collected in three non-consecutive days, while the examination of a single fecal sample has a lower sensitivity (Tangtrongsup and Scorza, 2010). In this study, only the canine assemblages C and D were identified at genotyping of some dog fecal samples found positive for G. duodenalis at molecular analysis. These data seem to confirm the higher frequency of these canine-specific assemblages in privately owned dogs in Italy (Pipia et al., 2014;Paoletti et al., 2015;Simonato et al., 2017;Sauda et al., 2018). Nevertheless, it is not possible to exclude the presence of the zoonotic assemblages A and B in the remaining positive dog samples for which molecular analysis and genotyping were not possible. Indeed, the assemblages A and B have been also identified in owned dogs from Italy (Riggio et al., 2013;Zanzani et al., 2014;Simonato et al., 2017) and a high possibility of transmission of these zoonotic assemblages between dogs and humans has been supposed (Marangi et al., 2010;Feng and Xiao, 2011). We acknowledge that this study has several limitations. First, the number of dogs examined in this study was low and follow-up for clinical conditions was not performed. Secondly, at inclusion in the study only a single fecal sample per dog to assess positivity or negativity for G. duodenalis was possible to analyze. In

conclusion, G. duodenalis may be present in primary CE-affected dogs and further investigations are needed to clarify the real significance of less severe clinical presentation observed in G. duodenalis-positive and CE-affected dogs and a possible higher diffusion of G. duodenalis in CE dog patients. Conflict of interest The authors declare that there is no conflict of interest. Authors contribution Stefania Perrucci and Veronica Marchetti conceived and designed the study. All authors contributed to perform research, interpreting the results, writing and critically revising the manuscript. All authors approved the final version of the manuscript. All-trans retinoic acid regulates the expression of the extracellular matrix protein fibulin-1 in the guinea pig sclera and human scleral fibroblasts. Purpose Fibulin-1 (FBLN1) mRNA is expressed in human sclera and is an important adhesion modulatory protein that can affect cell–matrix interactions and tissue remodeling. Scleral remodeling is influenced by all-trans retinoic acid (RA). Our purpose was to confirm the presence of fibulin-1 protein in guinea pig sclera and investigate the effect of RA on the expression of fibulin-1 in guinea pig sclera in vivo and in cultured human scleral fibroblasts (HSFs). Methods Confocal fluorescence microscopy was used to study fibulin-1 and aggrecan expression and localization in sclera from control guinea pigs and in animals given RA by daily gavage from 4 to 8 days of age. The effects of RA (from 10−9 to 10−5 M) on fibulin-1 expression in HSFs were observed by immunohistochemistry and assayed by real-time PCR and western blot analysis. Results Fibulin-1 protein expression was detected by confocal fluorescence microscopy in guinea pig sclera and in cultured HSFs. Upregulation of fibulin-1 in scleral tissue was observed after feeding with RA. In vitro, the level of Fbln1 mRNA was increased after treatment of HSFs with RA (at concentrations of 10−8 to 10−6 M; p<0.001), with a maximum effect at 10−7 M. Fibulin-1 protein levels were significantly increased after treatment of HSFs with 10−7 M of RA for 24 or 48 h (p<0.05). Conclusions Fibulin-1 protein was expressed in guinea pig sclera and cultured HSFs. Expression was regulated by RA, a molecule known to be involved in the regulation of eye growth. Further studies on the role of fibulin-1 in the regulation of eye growth, including during the development of myopia, are therefore warranted. regulate ECM metabolism in many systems [8]. It is an important molecular signal in the control of eye size, perhaps acting as a signal for the direction of ocular growth [9,10]. Although the mechanism by which RA regulates ocular size is not clear, it may affect scleral ECM during scleral remodeling [10]. RA can regulate aggrecan expression and skeletal growth and articular cartilage remodeling in skeletal systems [11]. We sought here to determine whether RA affects the expression of aggrecan and fibulin-1 in guinea pig sclera. It has been shown that six RA receptor subtypes are present in cultured human scleral fibroblasts (HSFs) [12], suggesting that RA may interact with these cells. However, whether RA treatment affects the expression of fibulin-1 in HSFs has not been studied. The purpose of this study was first to examine the localization of fibulin-1 protein in guinea pig sclera and the presence of fibulin-1 protein in cultured HSFs. Second, we examined whether the expression of fibulin-1 in guinea pig sclera (in vivo) was regulated by RA given by daily gavage [13] and in cultured HSFs treated with RA (in vitro). They were reared in litters with their mother until 9 days of age when scleral tissue was extracted. During rearing, animals were housed in plastic boxes (65×45×30 cm) lined with a bed of paper pellets and with stainless steel wire lids. Lighting was on a 12 h:12 h light-dark cycle and was provided by overhead incandescent lamps evenly diffused through a perspex barrier located 15 cm above the cages. The average light intensity in the cage was 300 lux. To treat the sclera with RA exogenously, during rearing, young guinea pigs were fed 0.5 ml of RA (24 mg/kg, n=8 eyes) or carrier alone (peanut oil [PO] n=4 eyes) from 4 to 8 days of age every day, 1 h after the beginning of the light cycle. Half of each litter was given either RA or PO to ensure a matched-pair design. Feeding was by gavage while the guinea pigs were lightly anesthetized with isoflurane (1.5% in O2). At 8 days of age, 24 h after the day 7 gavage and before the day 8 gavage, eyes were cyclopleged with 1% cyclopentolate. One hour later, refractive error was measured using streak retinoscopy followed by measurement of axial ocular dimensions in animals using A-scan ultrasonography. These measures were the same as previously described [9] and were used to confirm that the RA treatment had affected ocular elongation First, the animals were lightly anesthetized with 1.5% isoflurane in oxygen. Then, the anesthetized animals were restrained to allow alignment of the transducer probe to the center of the pupil along the optic axis. Approximately 10 measures were taken from each eye sequentially, with the probe realigned several times. On day 9, animals were euthanized by intraperitoneal overdose of barbiturate, eyes were removed on ice, and tissue extracted for immunohistochemistry. All procedures were in compliance with the New South Wales Animal Research Act and approved by the Animal Ethics Committee of the University of Newcastle, Newcastle, Australia. Immunohistochemistry on guinea pig sclera: To observe the expression and localization of fibulin-1 in scleral tissue and find the relationship between fibulin-1 and aggrecan, immunohistochemical staining was observed in scleras from animals fed PO or RA. On day 9 (24 h after the last gavage), eyes were removed and eye cups (without the vitreous) were fixed for 30 min at 20 °C in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Fixed samples were washed three times in PBS (0.05 M phosphate buffer, 195 mM NaCl, and 3 mM NaN3, pH 7.4), cryoprotected in PBS plus 30% sucrose, soaked in embedding medium (optimal cutting temperature compound) for 10 min, and freeze-mounted onto sectioning blocks. Vertical sections (10-12 μm thick) were cut from the posterior pole of the eye and thaw mounted onto gelatincoated glass slides. Sections from RA-and PO-treated eyes were placed together in pairs on each slide to ensure equal exposure to reagents. Sections were air dried, ringed with rubber cement to form a well for antibody solutions, and stored at −20 °C until use. Human scleral fibroblast isolation and identification: This part of the study was approved by the Ethics Committee of Sun Yat-sen University, Guangzhou, China, and complied with the tenets of the Declaration of Helsinki for Research Involving Human Tissue. Normal human eyes (n=8) from donors ranging from 10 to 20 years of age were obtained from the Eye Bank of Zhongshan Ophthalmic Center (Sun Yat-sen University) and were used to make cultures of HSFs. Eyes were washed immediately in Hank's balanced salt solution (HBSS; Gibco, Grand Island, NY) with penicillin (200 μg/ml penicillin-streptomycin) (Invitrogen, Carlsbad, CA) and gentamicin sulfate (400 μg/ml; Invitrogen). The retinas and choroids were removed from the sclera. The posterior sclera was trimmed into 1×1-mm 2 pieces, placed in DMEM/F12 (Gibco) with 1× antibiotic/antimycotic (penicillinstreptomycin; Invitrogen), and 10% fetal bovine serum (Gibco), and then incubated at 37 °C in a humidified incubator containing 5% CO2. The growth medium was changed every 3 days. When a heavy primary monolayer was achieved, cells were trypsinized for 2 min at room temperature in 0.25% trypsin/ethylene diamine tetraacetic acid (EDTA) solution in PBS (Gibco). Cells were subcultured at a split ratio of 1:3 in a 25-mm 2 plastic bottle. The third passage of fibroblasts was used for this experiment. The fibroblasts were grown on coverslips in six-well plates (Corning Ltd., Tokyo, Japan) to 70%-80% confluence. The cells were washed with PBS three times, fixed with acetone for 15 min, air dried, and kept frozen at −20 °C until use. The purity of fibroblast cultures was confirmed by staining for vimentin and stain resistance for cytokeratin, desmin, and S-100, using the indirect immunofluorescence procedure, as previously described [14]. All-trans retinoic acid (RA) treatment of cell cultures: RA was dissolved in dimethylsulfoxide (DMSO) to a concentration of 10 −2 M, and this solution was diluted into DMEM/F12 to different concentrations and stored at 4 °C for immediate use or frozen in aliquots at −20 °C. Medium containing 0.1% DMSO, which is the highest final concentration used in these experiments, served as the control. The numbers of cells in the cultures incubated with RA were estimated using the cell counting Kit-8 (Dojin Laboratories, Kumamoto, Japan) containing 2-(2-methoxy-4nitrophenyl) −3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2Htetrazolium (WST-8) and 1-methoxyphenazine methosulfate (1-methoxy-PMS). Briefly, HSFs were seeded into 96-well plates coated with rat tail tendon type I collagen (Sigma, St. Louis, MO) at an initial density of 10 5 cells/well and preincubated with DMEM/F12 without fetal bovine serum (FBS) for 24 h. Various concentrations of RA (from 10 −9 M to 10 −4 M) or control medium were added. The cultures were incubated for a further 24 or 48 h. The solution of WST-8 and 1-methoxy-PMS was then added to each well, and the plates were incubated for 1 h, after which the absorbance at 450 nm was measured to determine the amount of formazan dye generated by dehydrogenases in cells, which is directly proportional to the number of living cells. To extract total RNA or total protein for real-time PCR and western blot analysis, HSFs were plated in standard plates coated with rat tail tendon type I collagen with DMEM/F12 in 10% FBS at the same cell density for 24 h. The medium was changed to DMEM/F12 without FBS for another 24 h. Finally, these cells were grown in medium containing RA (from 10 −9 M to 10 −5 M) or in medium containing 0.1% DMSO as an untreated control. After 24 h of culture, total RNA was extracted. In a second experiment, cells were grown in control medium or in medium with RA at concentrations of either 10 −7 M or 10 −6 M. Total RNA was extracted after 12 h, 24 h, or 48 h, and total protein was extracted after 24 h or 48 h. Quantitative real-time PCR analysis and western blot analysis were undertaken on extracted total RNA and homogenized total protein, respectively. After 48 h of culture, the morphology of fibroblasts was also observed with light microscopy in control medium and in medium with 10 −7 M RA. Indirect immunofluorescence: Fibulin-1 was visualized using standard immunohistochemistry in guinea pig sclera from animals fed RA or PO, and in both control and RA-treated HSFs. Briefly, fixed sections or cells were washed three times with PBS, covered with 10% normal donkey serum diluted in PBS, and incubated for 20 min at 37 °C. The slides were incubated at 4 °C overnight with primary antibodies (antifibulin-1 diluted to 1:50 in PBS, anti-aggrecan diluted to 1:50 in PBS; Santa Cruz Biotechnology, Santa Cruz, CA). Cells and sections were incubated in PBS without primary antibodies as a negative control. The antibody-treated and negative control sample slides were washed with PBS and exposed to DyLight 488-conjugated anti-rabbit immunoglobulin G (IgG) antibodies or Dylight 594conjugated anti-goat IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) at 1:1,000 in PBS at 37 °C for 60 min. The slides were washed in PBS three times, and cell nuclei were stained with Hoechst 33342 or propidium iodide (Sigma). Immunofluorescence images were taken using a confocal microscope (LSM 510 META; Carl Zeiss, Jena, Germany). Quantitative real-time PCR analysis: Real-time PCR analysis was used to show the expression of fibulin-1 at the transcription level. Total RNA was extracted from cultured HSFs using Trizol Reagent (Invitrogen). Complemenary DNAs (cDNAs) were synthesized with 4 μg of total RNA, 0.4 μl random primer, 0.5 μl dNTPs, and 200 U MMLV reverse transcriptase, 5× RT buffer (4 μl) at 37 °C for 1 h, followed by 95 °C for 3 min, using a TaqMan Reverse-Transcription kit from Promega (Promega, Madison, WI). An aliquot of the resulting single-stranded cDNA was used in the PCR experiments. Based on the sequences reported in the GenBank database (Table 1), primers were designed for Fbln1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) with Primer Express Software (Applied Biosystems, Foster City, CA). Quantitative real-time PCR was performed with an ABI7500 Fast Real-Time PCR System (Applied Biosystems , Foster

City, CA). A typical reaction was performed in 50 μl, consisting of 5 μl of cDNA and 10 μl of 5× SYBR Green I PCR buffer, containing the specific primer pairs (final 10 pmol each). The PCR temperature cycle was 93 °C for 3 min, followed by 40 cycles at 93 °C for 30 s, 55 °C for 45 s, and 72 °C for 45 s, with SYBR Green fluorescence recorded at the end of each elongation segment. The PCR reaction was followed by a melting curve analysis: 93 °C for 10 s, 55 °C for 10 s, linear increase to 93 °C at 0.1 °C/s, with continuous SYBR Green fluorescence recording. The primer pairs used in this study for Fbln1 and GAPDH were validated as follows: they gave a single PCR product, as verified by melting curve analysis, agarose gel electrophoresis, and DNA sequencing; and the distribution of the PCR sigmoids was linear (r was 0.99 to 1) over 5 log units of template concentration with an efficiency of 1.85-1.98. The critical cycle of each sigmoid PCR curve was calculated by the ABI 7500 Fast Real-Time PCR System as the PCR cycle corresponding to the maximum of the second derivative. Total cDNA copy number from each cell culture sample was analyzed by the 7500 Fast Real-Time PCR Systems for Fbln1 and GAPDH. All copy numbers from each sample were compared to the GAPDH cDNA copy number from corresponding samples. Western blot analysis: Western blot analysis was used to show the expression of fibulin-1 at the protein level. Briefly, cells were washed with PBS and lysed in ice-cold lysis buffer (Shanghai Xinghan Sci & Tech Co. Ltd, Shanghai, China). After cell debris was removed by centrifugation at 15,000g at 4 °C for 30 min, the protein concentration was detected by BAC kits (Shenneng Bocai Biotechnology Co. Ltd, Shanghai, China). Protein (40 μg) was loaded in each lane of 10% sodium dodecyl sulfate-polyacrylamide gels, transferred onto polyvinylidene difluoride membranes for electrophoresis, and blocked in Tris Buffered Saline with Tween (TBST) (5% fatfree dry milk, 0.1% Tween-20, 150 mM NaCl, and 50 mM Tris at pH 7.5) for 1 h. The membranes were exposed to 1:500 anti-fibulin-1 antibodies and incubated overnight at 4 °C. The same blots were then stripped and reanalyzed using anti-GAPDH antibodies (Beijing Biosynthesis Biotechnology Co. Ltd., Beijing, China) as an internal control. The membranes were then incubated with a secondary horseradish peroxidase (HRP)-labeled antibody (Beijing Biosynthesis Biotechnology Co. Ltd.) for 1 h. Protein bands were visualized with the use of a chemiluminescence Phototope (R)-HRP Western Blot detection system (Cell Signaling Technology, Inc., Danvers, MA) and exposed to a negative film, developed, and fixed. The film was scanned and analyzed with Quantity One Analyzer Software (Bio-Rad Laboratories, Santa Cruz, CA). The relative level of protein expression was expressed as the density ratio of the protein compared to GAPDH in the same sample. Independent experiments were performed and repeated three times. Effects of gavage with retinoic acid on ocular size and refraction: Ocular size was increased by treatment with RA with a difference in axial length (axial distance from the cornea to the retina) of 97.5 µm (RA, 7.743±0.005 mm; PO, 7.628±0.009 mm, p<0.01), with most of the elongation due to an increase in the vitreous chamber depth (difference of 101.8 µm). For animals of this age, a 50-µm increase in ocular length is equal to approximately a −2D refractive shift [15]. We found here that RA treatment caused a myopic shift in refractive error of −4.8D (RA, 1.73±0.8D; PO, +6.5±1.6D, p<0.01). Expression of fibulin-1 and aggrecan in guinea pig sclera: Fibulin-1 and aggrecan immunohistochemical staining were observed in scleras from RA-or PO-treated guinea pigs. In all eyes, fibulin-1 was localized in scleral extracellular matrices in parallel with the distribution of aggrecan. For the PO-fed guinea pigs, aggrecan was most strongly labeled in posterior sclera and was irregularly arranged between collagen fibrils and collagenous lamellae, with scleral fibroblasts located between them. In PO eyes, fibulin-1 immunoreactivity was low and was detectable in and around fibroblasts ( Figure 1E-H). In RA-fed guinea pigs, the expression of fibulin-1 in the sclera was substantially increased, whereas the expression of aggrecan was depressed ( Figure 1I-L). where density values were compared to GAPDH density. The asterisk indicates a significant difference relative to the control (p<0.05). Figure 3A). Concentrations of RA that reduced cell numbers were less effective in the upregulation of Fbln1 mRNA. To find the effective time that RA took to upregulate Fbln1 mRNA expression in HSFs, total RNA prepared from cells treated with 10 −7 M RA for different times (12 h, 24 h, 48 h) was analyzed and compared with total RNA from control cultures ( Figure 3B). RA at 10 −7 M induced the most marked expression of Fbln1 mRNA in HSFs, and the effect was time dependent. There were no significant changes in Fbln1 mRNA in HSFs after incubation with RA for 12 h, but Fbln1 mRNA levels were significantly increased after treatment of HSFs with 10 −7 M RA for 24 h and 48 h (p<0.001 in both cases), with the latter showing a dramatic increase of 9.4 times the control ( Figure 3B). Retinoic acid-induced changes in fibulin-1 protein levels in cultured human scleral fibroblasts: Protein prepared from the cells treated with RA concentrations of 10 −7 or 10 −6 M for either 24 h or 48 h were analyzed and compared with controls (medium with 0.1% DMSO). The relative protein levels of fibulin-1 in HSFs incubated with RA are represented in Figure 4. RA upregulated the fibulin-1 protein level in a timedependent manner. The fibulin-1 protein expression was significantly increased after the cells were treated with RA at 10 −7 M at both 24 h and 48 h (p<0.05). Morphological changes in human scleral fibroblasts and fibulin-1 immunohistochemical staining after retinoic acid treatment: HSFs seeded onto collagen I and cultured in 10 −7 M RA or medium with 0.1% DMSO (control) for 48 h showed morphological changes. Under the inverted phase contrast microscope, control HSFs displayed many spreading protrusions on collagen-coated standard plates ( Figure 5A). After RA treatment HSFs became smaller with little evidence of any protrusions ( Figure 5B). With fibulin-1 immunohistchemistry, fibulin-1 was found to be weakly expressed in the cytoplasm of control cells with a uniform distribution ( Figure 6D-F). In cells cultured in a medium containing 10 −7 M RA for 48 h, fibulin-1 expression was significantly increased ( Figure 6G-I). DISCUSSION Fibulin-1, which is an important ECM molecule, is associated with tissue remodeling in many systems [1,2]. Recent work has shown that fibulin-1 is expressed in human sclera [7]. Here, we provide evidence that fibulin-1 protein is expressed in vivo in guinea pig sclera and that gavage with RA leads to increased expression of fibulin-1 and decreased expression of aggrecan. We further show that fibulin-1 is expressed in HSFs grown in culture, where expression is also regulated by RA. These results suggest that RA regulates the expression of fibulin-1 similarly in vivo and in vitro and that the molecular pathways involved may therefore be further studied using cultured fibroblasts. RA can affect scleral ECM during scleral remodeling [10]. RA appears to regulate the expression of aggrecan in skeletal systems and affect cartilage matrix homeostasis [11]. However, the effect of RA on aggrecan and fibulin-1 in the sclera has not been previously reported. In this study we fed RA to normal guinea pigs, and the eyes rapidly elongated and became myopic. Using immunohistochemical staining, we found that fibulin-1 and aggrecan were distributed in a similar fashion in normal guinea pig sclera, but fibulin-1 staining was weak, whereas aggrecan staining was stronger. In the sclera from RA-fed animals, the expression of fibulin-1 was increased, while there was a loss of scleral aggrecan staining. Similar results have been reported on cultured cartilage tissue where exposure to RA led to the extensive loss of aggrecan from the tissue [16]. The reciprocal relationship between fibulin and aggrecan observed in our study may be due to the ability of fibulin-1 to enhance the cleavage of aggrecan by a disintegrin-like and metalloprotease with thromobospondin type-1 [6,17]. This suggests that the pathway leading to increased eye growth may involve upregulation of fibulin-1 by RA and degradation of aggrecan. The expression of fibulin-1 was induced by RA in a dosedependent manner in cultured HSFs, as shown by both realtime PCR and western blot analysis. A 10 −7 M concentration of RA maximally increased the expression of fibulin-1 protein in HSFs. This concentration is similar to the concentration of RA shown to inhibit proteoglycan synthesis in cultured chick sclera [10], and we have shown here that this concentration does not reduce the number of cells that grow under cultured conditions. Since RA appears to be an important molecular signal in the control of eye size [9,10,[18][19][20], it is tempting to relate the capacity of RA to inhibit aggrecan synthesis in cultured sclera to its effect on eye growth. Although the functions of fibulin-1 in the sclera remain unclear, our finding of fibulin-1 expression induced by RA, which is a molecule known to be involved in the regulation of eye growth, suggests that it is worth looking at the role of fibulin-1 in scleral remodeling. Interactions between fibulin-1 and aggrecan may be important, and the modulation of aggrecan levels and distribution could play a key role in changing the scleral creep rate [5], which is associated with axial elongation of the eye and the development of myopia [21]. Further research on the function of fibulin-1 and its interaction with aggrecan in sclera during the development of myopia is warranted. Modulation of Microglial Activation by Adenosine A2a Receptor in Animal Models of Perinatal Brain Injury Neuroinflammation has a key role in the pathogenesis of perinatal brain injury. Caffeine, a nonspecific antagonist of adenosine receptors (ARs), is widely used to treat apnea of prematurity and has been linked to a decrease in the incidence of cerebral palsy in premature infants. The mechanisms explaining its neuroprotective effect have not yet been elucidated. The objective of this study was to characterize the expression of adenosine and ARs in two neonatal rat models of neuroinflammation and to determine the effect of A2aR blockade on microglial activation assessed through inflammatory cytokine gene expression. We have used two rat models of microglial activation: the gestational low protein diet (LPD) model, associated with chronic brain injury, and postnatal ibotenate intracerebral injections, responsible for acute excitotoxicity injury. Adenosine blood levels have been measured by Tandem Mass Spectrometry. The expression of ARs in vivo was assessed using qPCR and immunohistochemistry. In vivo models have been replicated in vitro on primary microglial cell cultures exposed to A2aR agonist CGS-21680 or antagonist SCH-58261. The effects of these treatments have been assessed on the M1/M2 cytokine expressions measured by RT-qPCR. LPD during pregnancy was associated with higher adenosine levels in pups at postnatal day 1 and 4. A2aR mRNA expression was significantly increased in both cortex and magnetically sorted microglial cells from LPD animals compared to controls. CD73 expression, responsible for extracellular production of brain adenosine, was significantly increased in LPD cortex and sorted microglia cells. Moreover, CD73 protein level was increased in ibotenate treated animals. In vitro experiments confirmed that LPD or control microglial cells exposed to ibotenate display an increased expression, at both protein and molecular levels, of A2aR and M1 markers (IL-1β, IL-6, iNOS, TNFα). This pro-inflammatory profile was significantly reduced by SCH-58261, which reduces M1 markers in both LPD and ibotenate-exposed cells, with no effect on control cells. In the same experimental conditions, a partial increased of M1 cytokines was observed in response to A2aR agonist CGS-21680. These results support the involvement of adenosine and particularly of its receptor A2aR in the regulation of microglia in two different animal models of neuroinflammation. Neuroinflammation has a key role in the pathogenesis of perinatal brain injury. Caffeine, a nonspecific antagonist of adenosine receptors (ARs), is widely used to treat apnea of prematurity and has been linked to a decrease in the incidence of cerebral palsy in premature infants. The mechanisms explaining its neuroprotective effect have not yet been elucidated. The objective of this study was to characterize the expression of adenosine and ARs in two neonatal rat models of neuroinflammation

and to determine the effect of A2aR blockade on microglial activation assessed through inflammatory cytokine gene expression. We have used two rat models of microglial activation: the gestational low protein diet (LPD) model, associated with chronic brain injury, and postnatal ibotenate intracerebral injections, responsible for acute excitotoxicity injury. Adenosine blood levels have been measured by Tandem Mass Spectrometry. The expression of ARs in vivo was assessed using qPCR and immunohistochemistry. In vivo models have been replicated in vitro on primary microglial cell cultures exposed to A2aR agonist CGS-21680 or antagonist SCH-58261. The effects of these treatments have been assessed on the M1/M2 cytokine expressions measured by RT-qPCR. LPD during pregnancy was associated with higher adenosine levels in pups at postnatal day 1 and 4. A2aR mRNA expression was significantly increased in both cortex and magnetically sorted microglial cells from LPD animals compared to controls. CD73 expression, responsible for extracellular production of brain adenosine, was significantly increased in LPD cortex and sorted microglia cells. Moreover, CD73 protein level was increased in ibotenate treated animals. In vitro experiments confirmed that LPD or control microglial cells exposed to ibotenate display an increased expression, at both protein and molecular levels, of A2aR and M1 markers (IL-1β, IL-6, iNOS, TNFα). This pro-inflammatory profile was significantly reduced by SCH-58261, which reduces M1 markers in both LPD and ibotenate-exposed cells, with no effect on control cells. In the same experimental conditions, a partial increased of M1 cytokines was observed in response to A2aR agonist CGS-21680. These results support the involvement of adenosine and particularly of its receptor A2aR in the regulation of microglia in two different animal models of neuroinflammation. INTRODUCTION Brain injury is one of the most important complication related to preterm birth (1). From 25 to 50% preterm infants display neurodevelopmental disabilities (2), with dramatic consequences in terms of cost and impact on quality of life. The most implicated adenosine receptor in neuroinflammation is A2aR (14). Its expression in microglia is usually low but increases following brain insults. In microglial cells, activation of A2aRs has facilitating effects on the release of cytokines (15) and on the change into amoeboid morphology (16). Conversely, A2aR antagonists suppress microglia activation, as described using in vitro (17,18) and in vivo (18) studies. To our knowledge, there are no data regarding the adenosine pathway and neuroinflammation in preterm infants, but nevertheless, the involvement of adenosine signaling in prematurity is suggested by the clinical use of caffeine. Indeed, caffeine, a non-specific antagonist of ARs widely used to treat apnea of prematurity, not only improves survival and reduces the duration of respiratory support, but also reduces the incidence of cerebral palsy and cognitive delay (19). Recently, a retrospective study demonstrated the existence of high blood levels of adenosine in premature infants (20), with the highest adenosine concentrations associated with the lowest birthweight. These data suggest a possible link between caffeine action, adenosine plasma levels and an imbalance between the proand anti-inflammatory profiles in very preterm infants usually delivered following a perinatal inflammatory event. Whether a similar link exists for adverse neurological outcomes in preterm infants is not known and there is still little evidence relating to effects of caffeine on brain development, especially at the cellular and molecular levels (21). Therefore, this study was aimed to characterize the synthesis and expression of adenosine and its receptors in two experimental animal models of neonatal neuroinflammation. The effect of A2aR blockade was also studied in vitro using a specific antagonist on microglial activation assessed through pro-and anti-inflammatory cytokine gene expressions. Animals and Models All experiments were carried out according to INSERM ethical rules and approved by the institutional review board (Robert Debré ethics committee, Paris, France, approval number Big Project 01542.01). Sprague-Dawley rats (Janvier SAS, Le Genest-St-Isle, France) were housed in temperature-controlled rooms (24 • C), with 12 h light cycling and free access to chow and water ad libitum. Low Protein Diet (LPD) Model After mating, dams were randomly allocated to either isocaloric low-protein diet (LPD) (9% casein; as previously described (22,23), SAFE-diets Augy, France) or control diet (CTL) (23% casein) during the gestational period. The control and LP diets are balanced for energy intake assuming equivalent consumption rates. At birth, dams were returned to standard diet. Sex, birthweight and postnatal growth rates were determined. Experiments research plan using this model is summarized in Supplemental Figure 1. Ibotenate (IBO) Model Ten µg IBO diluted in Phosphate Buffered Saline (PBS) was injected intracerebrally (i.c.) at postnatal day 5 (P5) to rat pups of both sexes as previously described (24). Experiments research plan using this model is summarized in Supplemental Figure 2. The rat pups were killed and dissected at different postnatal days (P1, P4, and P5). Blood was collected by exsanguination on filter paper. Brains were collected, immediately snap frozen and stored at −80 • C or immediately dissociated for microglia cells isolation. Microglia Cell Isolation and Primary Culture Brains were collected from control and LPD animals at P1 and P4 removing the cerebellum and the olfactory bulbs. The tissues dissociation was performed using the Neural Tissue Dissociation Kit and the gentleMACS Octo-Dissociator with Heaters accordingly to the manufacturer's instruction (Miltenyi Biotec, Germany). CD11b positive cells were isolated from the resulting homogenates using an anti-CD11b (microglia marker) MicroBeads (Miltenyi Biotec, Germany) and multiMACSCell-24 separator (Miltenyi Biotec, Germany). After elution the sorted microglia cells were stored at −80 for RNA extraction. In a second set of experiments, microglia cells were magnetically sorted from control and LPD animals at P4 and after elution pellet was isolated by centrifugation (300 g -10 min). Following re-suspension in Dulbecco's modified Eagle's minimum essential medium/Nutrient mixture F-12 (DMEM/F-12, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (P/S), cells were maintained in DMEM/F-12 supplemented with 10% FBS and 1% P/S at a concentration of 5 × 10 5 cells/ml in 12-well culture plates. The purity of isolated microglia cells was verified by Iba1 immunostaining (dilution 1/1,000). A medium change was performed after 24 h and cells were treated as follows after 48 h. Microglial cells were treated with SCH-58261 (A2aR antagonist) at 50 nM (17,25) or CGS-21680 (A2aR agonist) at 10 µM (26,27) or DMSO. For the IBO model, cells sorted at P4 were treated with SCH, CGS or DMSO 20 min before adding ibotenate 300 µM (28). After 6 h, cells were harvested and RNA extracted for gene expression analysis. For cytokine levels, supernatant (conditioned media) was collected after a longer exposure time (12 h) and stored at −80 • C until analysis. RNA Extraction, Retro-Transcription and Real-Time PCR Total RNA was extracted from cortex using Qiazol reagent and RNeasy mini kit (Qiagen, France) and from microglia cells using the NucleoSpin RNA Plus extraction kit (Macherey-Nagel, France) according to the manufacturer's instructions. RNA yield and purity were determined by spectrophotometry absorption at 260 and 280 nm by means of a NanodropTM apparatus (Thermofisher Scientific, MA, USA). Five hundred ng of mRNA from cortex and 150 ng from microglia were used to perform reverse transcription (iScript TM cDNA synthesis kit, Biorad, France), respectively. qPCR measurements were performed in duplicate using SYBR Green Super-mix (Bio-Rad). The reaction conditions were as follows: 98 • C for 10 min (Polymerase activation), followed by 45 cycles at 95 • C for 5 min, 60 • C for 10 min and 72 • C for 10 min. The specificity of used primers was assessed with a melting curve analysis and the results were quantified using the relative standard curve methods. The relative mRNA expression for each target gene was calculated after normalization respect to the Rpl13 references gene. The primers sequences are available in Supplemental Table 1. Multiplex Cytokine Assay Cytokines were measured using the Bio-Plex rat cytokine multiplex kit (Bio-Rad). Calibration curves from recombinant cytokine standards were prepared with serial dilutions in the same media as the culture supernatant (DMEM/F-12 supplemented with 10% FBS and 1% P/S). Standards and samples were analyzed in duplicate and blank values were subtracted from all readings. All assays were carried out directly in a 96-well filtration plate (Bio-Rad) at room temperature and protected from light. Briefly, wells were pre-wetted with culture supernatant, then beads together with either standard, sample, or blank were added in a final volume of 50 µl, and incubated together at room temperature for 30 min with continuous shaking. Beads were washed three times with 100 µl Bio-Plex wash buffer. A cocktail of biotinylated antibodies (25 µl/well) was added to the beads for a further 30-min incubation with continuous shaking. Beads were washed three times, then streptavidin-phycoerythrin was added for 10 min. Beads were again washed three times and resuspended in 125 µl assay buffer. The fluorescence intensity of the beads was measured using the Bio-Plex array reader. Bio-plex manager software with five-parametric-curve fitting was used for data analysis. Immunofluorescence Assay and Quantification For histological analysis after ibotenate i.c., injections, animals were anesthetized with pentobarbital and transcardially perfused with 4% paraformaldehyde in PBS. Brains were collected, postfixed in 4% paraformaldehyde overnight, cryoprotected, cut coronally in 10 µm-thick slices, and stained according to standard protocols. After three washings of the slices with PBS, the non-specific binding was blocked by incubating the tissue sections with PBS-Triton 0.5%-gelatin 0.2% for 45 min at room temperature. Incubation with primary antibodies (rabbit anti-CD73 1/1,000; goat anti-Iba1 1/1,000) was performed overnight at 4 • C in PBS-Triton 0.5%-gelatin 0.2%. After rinsing three times in PBS for 5 min each, sections were exposed (1 h, room temperature) to secondary species-specific antibodies (all at 1/1,000 dilution in PBS-Triton 0.5%-gelatin 0.2%) conjugated to Alexa Fluor R 488 or to Cy3. Nuclei were then labeled with the fluorescent DAPI dye (1/10,000 in PBS). Stained sections were mounted on microscope slides with Fluoromount-G (SouthernBiotech). Primary microglia cells cultured in micro-slide 8-well chamber (Ibidi, Germany) and treated as reported above were fixed in 4% paraformaldehyde for 30 min at room temperature. Each well was incubated with a blocking solution (PBS with 1% BSA) for 1 h at room temperature and incubated overnight at 4 • C with goat anti-Iba1 (1/500) and anti-A2aAR (1/250). The following day, after rinsing three times in PBS for 5 min, cells were incubated with secondary antibodies coupled to the green and red fluorescence markers (1/500 dilution) for 1 h at room temperature. Nuclei were visualized by staining the cells with DAPI dye (1/10,000). Cells were analyzed using a fluorescent microscope (Nikon Eclipse Ti-E) and images captured with a 20X objective (4 wells/group and 5 images/well). Fields used for quantitation were randomly selected throughout the dish and focused using phase contrast optics. Images from different emission specters were acquired separately using the same parameters and superimposed in the aftermath. For the analysis Image J software (Research Service Branch, National Institutes of Health, Bethesda, MD; http://rsb.info.nih.gov/ij/) was used. Images were converted to binary images using an automatic threshold function. The cells in each image were then defined by outlining a mask image (ROI). For LPD experiments, as the cells showed a significant difference in cell size due to microglial activation, the fluorescence power was calculated as the mean of all the pixel intensities of each individual cell. For the ibotenate experiment, the fluorescence power was calculated as an integrated density (i.e., the product of the mean of all the pixel intensities of each individual cell and the ROI area). Cell size was calculated using Iba1-positive cells as the product of number of pixels in ROI and the conversion factor 0,103. Finally, the sums of the values for each condition were normalized to control values for the statistical analysis. Statistical Analysis The graphs and the statistical analysis were performed with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). Appropriate statistical analyses were carried out with a two-sided alpha level of 0.05 or 95% confidence interval. For continuous data, descriptive analyses were carried out employing means and ± SEM, Parametric or non-parametric tests were used to compare quantitative variables (Student's t-test for independent samples for comparisons between two groups and one-way ANOVA with Newman Keuls post-test for comparisons of more than two groups). A p-value < 0.05 was considered statistically significant. The LPD Model Influences Adenosine Blood Levels and the Cerebral Expression of Enzymes and Receptors Implicated in Adenosine Metabolism Prenatal LPD-induced malnutrition resulted in the absence of postnatal mortality but

fetal growth restriction with body weight of the pups significantly lower from P0 to at least P4 compared to the controls (Supplemental Figure 4). LPD pups at P1 and P4 showed significantly higher adenosine blood levels compared to the control pups (0.822 ± 0.088 µM in Controls vs. 1.850 ± 0.355 µM in LPD at P1, 0.598 ± 0.051 µM in Controls vs. 1.464 ± 0.215 µM LPD at P4, see Figure 1). In the pre-frontal cortex of LPD animals, the expression of adenosine deaminase (ADA), adenosine kinase (ADK), ectonucleoside triphosphate diphosphohydrolade-1 (Entpd1) and cluster of differentiation 73 (CD73), enzymes involved in the regulation of extracellular and intracellular adenosine levels, were significantly increased at P1, compared to controls (Figure 2A). The relative gene expression of intracellular enzymes ADA and ADK were comparable between LPD and controls at P4, whereas the two ectonucleotidases Entpd1 and CD73 are persistently increased at this age in the LPD group, indicating an increased extracellular production of adenosine. A2aR and A2bR, the two main ARs with pro-inflammatory functions, were significantly up-regulated at P1 and P4 in the cortex of LPD pups (Figure 2B), suggesting a pro-inflammatory status of the LPD brains. In contrast, no significant differences were observed at a transcriptional level for A1 and A3 receptors, except for an increased expression of A3R at P4 in the LPD group ( Figure 2B). A2aR Expression Is Increased in Microglial Cells Sorted From the Rat Pups Subjected to Antenatal LPD While most of the receptors display a similar pattern of expression, A2aR expression was found significantly increased in microglia cells sorted from LPD brains compared to controls, both at P1 and P4 (Figure 3). Finally, no statistically significant difference was observed in the expression of genes encoding for the enzymes involved in the adenosine extracellular metabolism, except for a transient and mild increase in the expression of CD73 at P1 in LPD animals. The mean density of A2aR immunoreactivity in microglial cells after 48 h in culture was significantly higher in LPD compared to control group (Figures 4A,B). Moreover, Iba1 immunoreactivity was found increased and cell size reduced in LPD microglial cells, compared to control cells (Figures 4C,D). A2aR Antagonist Exposure Changes Microglial Reactivity in vitro After 2 day, significant increases in gene expression of both M1 (IL-1β, IL-6, iNOS, TNFα) and M2 markers (IL-10, IL-4ra) were detected in microglial cells sorted from LPD rat pups, compared to control cells ( Figure 5). SCH-58261, an A2aR antagonist, induced a significant reduction in the expression of M1 markers while no effect on M2 markers was detected in LPD microglial cells. Conversely, the A2aR agonist CGS-21680 was able to increase the mRNA levels of iNOS, TNFα and IL-4ra in LPD microglial cells. No substantial effect of either SCH-58261 or CGS-21680 was observed in control cells. A2aR Plays a Role in Excitotoxic-Induced Microglial Activation In the IBO model assessed 24 h after ibotenate injection, 1/3 of Iba-positive cells surrounded the white matter lesion expressed CD73, an enzyme responsible for the extracellular production of adenosine in the brain (Supplemental Figure 5). Frontiers in Neurology | www.frontiersin.org FIGURE 3 | Comparisons of gene expression of enzymes involved in the production of adenosine and its receptors in microglial cells sorted from animals exposed to LPD (black bars) and from control animals (white bars). Data are shown as relative expression of control values normalized to 1 (*p < 0.05, using unpaired Student's t-test). magnetically sorted microglial cells from P4 control pups were exposed to 300 µM IBO for different lengths of time from 2 to 12 h. The exposure to IBO was unable to induce an increase in A1 and A2a receptor expression levels after 2-h exposure (Figure 6). In contrast, a significant and progressive increase in the gene expression of both receptors was observed after 6-and 12-h IBO exposure. CD73 was also found to be slightly but significantly increased, while no difference was observed regarding A2b and A3 receptors gene expression. These results were confirmed at protein level using immunocytochemistry (Figure 7). IBO exposure for 6 h was associated with a significant increase in A2aR immunoreactivity in microglial cells, while cell size and the Iba1 immunoreactivity were found similar with and without IBO. A2aR Modulation Is Involved in the Regulation of M1-M2 Microglia Phenotype in Excitotoxic-Induced Inflammation In our in vitro model of excitotoxic activation of primary microglial cells, highly significant increase in gene expressions of IL1β, IL6, iNOS, TNFα, and IL10 were was observed in response to 300 µM IBO exposure for 6 h (Figure 8). This effect was significantly reduced by SCH-58261 pre-treatment, 20 min before IBO challenge. Interestingly, SCH-IBO treated microglia display a higher expression level of the M2 cytokine IL-4ra. Pre-treatment with CGS-21680 induced down-regulation of TNFα and IL4ra and up-regulation of IL6 gene expression in IBO-treated cells, but had no effect on gene expressions of IL1β, iNOS and IL10. In another set of experiments, conditioned culture media were collected after 12-h IBO exposure with or without SCH-58261 pre-treatment, and cytokine concentrations were assessed (Figure 9). While IBO induced higher IL1β and TNFα concentrations in the culture media, the A2aR antagonist SCH-58261 exposure was associated with a reduced cytokine production in response to excitotoxic challenge. DISCUSSION This study strongly suggests that adenosine and the regulation of its receptor A2aR play a role in neonatal brain inflammation and microglial activation in rat. Abundance of literature has demonstrated both in clinical and preclinical studies that neuroinflammation is a relevant component in the pathogenesis of prematurity-related brain injury (3,4,9,(29)(30)(31)(32). Interestingly, adenosine exerts a role in proteins in microglial cells treated with or without SCH-58261 (A2aR antagonist) and CGS-21680 (A2aR agonist) sorted from control or LPD animals (*p < 0.05, **p < 0.01, ***p < 0.001, using one-way ANOVA followed by Newman-Keuls post-hoc analysis). this context and is able to orchestrate the inflammatory response (33)(34)(35). However, its role in the neonatal brain has not yet been elucidated, even if caffeine, a non-specific adenosine antagonist, has shown an important neuroprotective role in premature infants. The adenosine system appears to be involved in regulating inflammation in both acute (IBO) and chronic (LPD) neuroinflammation in vivo. The two animal models used in the present study display alterations that occur at a developmental stage of the rat brain that corresponds to the human brain at 28-32 gestation weeks (GW) (36). This window is recognized as a period of high vulnerability for the developing brain to either excitotoxic or inflammatory insults (37). Interestingly, these effects are exerted through the modulation of microglia reactivity, that, as previously reported (23,38), characterizes the two animal models. In conditions of inflammation, oxidative stress, excitotoxicity or cellular necrosis, the purinergic system is the first to be involved (39). Indeed, under pathological conditions, extracellular ATP is produced by both neurons and glial cells (40) and is rapidly converted to ADP and AMP by Entpd1 and by CD73, which convert AMP to adenosine (41). Despite its very short half-life (few seconds), increased adenosine brain levels contributes to induction and modulation of neuro-inflammation (42). To assess the adenosine implication in the LPD model, an adenosine assay on whole blood was performed. Since deliveries can span a 12-h and to avoid the stressful peak related to delivery, the blood samples were collected at P1, when LPD pups showed higher level of adenosine blood levels compared to controls. As described for human neonates (43), delivery is responsible for a physiological increase in adenosine blood levels in the newborn. Interestingly, adenosine blood levels remained higher in LPD at P4 in our study, when the effects of delivery have disappeared. Remarkably, the increasing adenosine blood levels in LPD rats are similar to those found in premature infants (20) and suggest a pro-inflammatory condition that characterizes these babies from birth until the first month of life. Similarly to neonates, LPD pups also showed a chronic inflammatory condition, as a result of maternal malnutrition, with a detrimental effect on neurodevelopment (23). In the brain, CD73, also known as Nt5e, is considered as the principal enzyme involved in the production of extracellular adenosine (44). The mRNA expression of the two ectonucleotidases Entpd1 and CD73, responsible for the final step of ATP catabolism into adenosine, is significantly increased in LPD both at P1 and P4. These results are in agreement with the study conducted by Chen et al. who reported that the activity of the ectonucleotidases is stimulated by inflammatory conditions (45). Adenosine elicits its physiological responses by binding to and activating one or more of the four transmembrane ARs. Solid evidence demonstrated that the four receptors are all expressed in the brain (33) and our results confirm the expression of all ARs in the rat pup cortex. Interestingly, LPD animals displayed an increased expression of A2aR and A2bR, known to exert a pro-inflammatory action (14,46), in the pre-frontal cortex both at P1 and P4. In the LPD model, which induces fetal growth restriction, the main alteration consists in disturbance of oligodendrocytes progenitor cells (OPC) maturation conducing to a deficit of myelination, that occurred in combination with a proinflammatory state evidenced by transcriptomic analysis performed in sorted microglia (23). Our in vitro results revealed that microglial cells sorted from rat pups subjected to LPD have abnormal reactivity with increased Iba1 staining and smaller size, when compared to control cells. Iba1 is constitutively expressed by microglia and is involved in the actin-crosslinking associated with membrane ruffling of microglial cells, an event essential for the morphological changes from quiescent ramified microglia to activated amoeboid microglia (47). Furthermore, the reduction in cell size has been shown to be strongly correlated to microglial activation (48,49). Regarding the potential role of adenosine in the modulation on microglia activity, we reported an increase in A2aR transcripts and protein levels in LPD-exposed microglia cells. A2aR has an important role in the control of inflammatory events (14) by regulating microglial reactivity, changing microglial FIGURE 8 | A2aR modulation regulates the phenotype of microglial sorted cells exposed to 300 µM ibotenate (IBO) for 6 h. Quantification of gene relative expressions of M1 (A) and M2 (B) proteins in microglial cells treated with or without SCH-58261 (A2aR antagonist) and CGS-21680 (A2aR agonist) sorted from control or LPD animals (*p < 0.05, **p < 0.01, ***p < 0.001, using one-way ANOVA followed by Newman-Keuls post-hoc analysis). morphology (16), increasing the release of cytokines and prostaglandin E2 (15) and nitric oxide synthase activity. In addition to A2aR, CD73 has been shown to have an important role in modulation of microglia ramification and activation (51). These data are consistent with our findings and support a possible role of adenosine in the regulation of the inflammatory response following brain injury. The results of in vitro studies conducted using an A2aR pharmacological approach clearly evidenced the involvement of A2aR in the regulation of inflammatory response in LPD model. These results are well supported by previous studies, which demonstrated that A2aR antagonists suppress microglia activation and IL-1β secretion in murine microglial cells exposed to an inflammatory stimulus induced by LPS (17,18). A2aR gene disruption in mice showed a lower severity of inflammatory response and subsequent damage in different models of brain injury including ischemia/hypoxia and traumatic brain injury (50)(51)(52). On the other hand, treatment with CGS-21680, an A2aR selective agonist, promotes the increase in M1 markers in LPD microglial cells suggesting that chronic inflammation causes microglial cells to be more susceptible to the pro-inflammatory effect of adenosine via A2a receptors. The Ibotenate model closely mimics the pathological features observed in periventricular white matter (38). In this model of acute brain injury and inflammation, both in vivo and in vitro findings support the involvement of A2aR. However, some features of microglial activation in vitro appear to be different from those observed in the LPD model, as microglial cells treated by ibotenate showed no difference in cell size and Iba1 expression. In conclusion this study gives evidence of the involvement of adenosine and in particular of its receptor A2aR in the regulation of microglia in two models of perinatal brain injury associated with neuro-inflammation. The present study focused only on the relation between adenosine formation and A2aR inactivation; it remains to be explored whether the other receptors might play a role in the regulation of microglia in the same models of perinatal brain injury. Indeed, all the 4 receptors subtypes

are non-specifically inactivated by caffeine. Further studies may provide a functional role for caffeine and specific antagonists of the remaining ARs in the limitation of perinatal brain injury associated with neuro-inflammation. In summary, the present study suggests that A2aR, up-regulated as consequence of inflammation, can influence the microglia phenotype, building up a potential for A2aR antagonist as a therapeutic strategy for neonatal brain damage. AUTHOR CONTRIBUTIONS MC, JM, LR, and OB designed the study. MC, MZ, and JP performed experiments. MCa performed adenosine measurements. MC, JM, and OB wrote the paper. All authors revised and approved final version of the manuscript. FUNDING This study was supported by Inserm, France and by University of Genoa, Italy. ACKNOWLEDGMENTS We thank Audrey Toulotte-Aebi for the editing of the manuscript draft. Malignant peripheral nerve sheath tumor associated with neurofibromatosis type 1, with metastasis to the heart: a case report A rare case is presented of a 61-year-old man with a malignant peripheral nerve sheath tumor associated with neurofibromatosis type 1, with metastasis to the heart. The primary tumor originated in the right thigh in 1982. Since then, the patient has had repeated local recurrences in spite of repeated surgical treatment and adjuvant chemotherapy. He has developed previous metastases of the lung and heart. The patient died of cardiac involvement. Background Malignant peripheral nerve sheath tumor (MPNST) is an aggressive and uncommon neoplasm that develops within a peripheral nerve; most cases of which are associated with neurofibromatosis type 1 (NF1). Metastasis of MPNST usually occurs in the lung [1], whereas cardiac metastasis of MPNST is quite rare [2][3][4]. In this report, we describe a case of MPNST that metastasized to the heart, with a review of the literature. Case Presentation A 61-year-old man presented with clinical stigmata of NF1. Genetic analysis had not been performed. He had no family history and had developed multiple neurofibromas and café-au-lait spots on his body ( Figure 1). He underwent surgical treatment for a growing tumor (details are unknown) on the posterior surface of his right thigh at another hospital in April 1982, and was referred to our institution because of regrowth of the tumor in October 1990. Wide resection was performed in November 1990 ( Figure 2). The histological diagnosis was MPNST developed on a background of NF1, and he received adjuvant chemotherapy. Since 1990, the patient has had repeated local recurrences (7 times, until May 2003) and has received surgical treatment for each recurrent tumor. In December 2003, the patient presented with a skin tumor of the head, and the lesion was resected. The histological diagnosis was MPNST that arose in a neurofibroma of the head skin. Two years later, the right leg was amputated at the hip joint because of local recurrence of the leg lesion, with aggressive growth and rupture. Computed tomography (CT) performed in August 2005 showed multiple lung lesions, suggestive of tumor metastases. Some of the lung lesions were resected surgically in January 2006 and diagnosed histologically as MPNST metastases. After the operation, he developed new lung metastases. In May 2007, he was admitted to our institution with sudden dyspnea and palpitations. Echocardiography revealed an enlarged heart and cardiac effusion, which did not suggest cardiac tamponade. Magnetic resonance imaging showed the presence of a large tumor (9 cm) in the inferior-posterior wall of the heart ( Figure 3). CT revealed an infarction of the inferior lobe of the left lung, which was derived from tumor emboli. As a result of the worsening condition, irradiation was applied to the cardiac mass (2 × 25 Gy). The patient's condition still worsened, with increasing cardiac effusion and arrhythmia. Regardless of drainage of the effusion, he died of acute heart failure soon after. At autopsy, the cardiac mass was confirmed histologically as MPNST ( Figure 4). The cardiac lesion was thought to be a metastatic tumor, because neurofibroma was not seen as possible precursor within the lesion. Serial sections suggested that the tumor involved the conduction system. MPNST affected both lungs. Histological images of cardiac and pulmonary tumor were similar to those of the tumor of the right leg that was amputated in 2005 (Figure 2). Discussion Patients with NF are at greatest risk for developing sarcomas, including MPNST. The incidence of MPNST arising in NF is 4.6%, which is much higher than the 0.001% in general population [5]. The most common metastatic site of MPNST is the lung [1,4]. Cardiac involvement from metastatic MPNST is extremely rare, whether with or without a background of NF [2][3][4]. Our patient is one of the extremely rare cases of cardiac metastasis of MPNST associated with NF1. Most of the cardiac metastases are preceded by other metastatic lesions, such as in the lung. Although the possibility of cardiac involvement becomes higher as the tumor progresses, details of the histopathological features specific to cardiac metastasis remain to be investigated. Recent studies have in part revealed the genomic imbalance in sporadic and NF1-associated MPNST [6,7]. The biology The prognosis of patients with MPNST is generally poor. Aggressive surgery significantly improved diseasefree survival [5,8]. Adjuvant chemotherapy and radiotherapy has not been proven to prolong patient survival, but it is effective as a palliative option [4,8,9]. If clinical symptoms of cardiac dysfunction occur during the progression of MPNST, it might be that the heart is involved. In such cases of MPNST, it is necessary to exclude cardiac involvement, even if it is rare, by occasional echocardiography. Early diagnosis can allow timely surgical intervention, if the patient is operable, which may improve results, as in the case described here. Clinical features vary according to the site of cardiac involvement [10], such as pericardium, epicardium, myocardium or endocardium. The present case showed a large mass in the myocardial region, which was accompanied with increasing pericardial effusion and arrhythmia. Serial histological sections revealed that the metastatic tumor markedly affected the common bundle of His, in addition to the ordinary myocardium. Based on these findings, we surmised that complete atrioventricular block was attributable to cardiac metastasis of MPNST, which is causative of circulatory failure [10]. Related features of cardiac involvement of MPNST are uncertain. More cases should be reported to elucidate the clinical entity associated with cardiac involvement of MPNST and to formulate an appropriate treatment strategy. Most NF1 patients carry a constitutional mutation of the NF1 tumor suppressor gene [11]. Biallelic inactivation of NF1 and mutations of numerous additional tumor suppressor genes within the p19 ARF -MDM2-TP53 and p16 INK4A -Rb signaling cascades have been identified in MPNSTs [12,13]. These abnormalities of suppressor genes, except for NF1, are not present in neurofibromas. It is therefore thought that the development of neurofibromas and their subsequent progression to become MPNSTs involves a sequential series of tumor suppressor mutations. Deletions and other mutations that result in loss of function of the TP53 tumor suppressor gene are some of the more common abnormalities found in MPNSTs. Biallelic inactivation of the TP53 locus is found rarely in MPNST, which has led to the suggestion that hemizygous TP53 mutations may suffice for neurofibromas to progress and become MPNSTs. A recent study has demonstrated that two MPNST cell lines derived from sporadically occurring MPNSTs have functional and intact NF1 genes [14]. Paradoxically, however, microarray studies that have compared the transcriptomes of sporadic and NF1-associated MPNSTs have not found a molecular signature that distinguishes these neoplasms [14,15]. For understanding of these complex neoplasms and the development of the effective new therapy, further investigation will be needed into the clinical features and the basic science. Conclusion Although cardiac involvement of MPNST is rare, precise examination including occasional echocardiography is necessary when clinical signs of tumor development in the heart are suspected. Consent Written informed consent was obtained from the patient's family for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. Problems associated with the right of access in the context of the rights of the child in Latvia The child is a special right-holder. In the legal system, the status of the child is characterised by substantive rights that are typical only for children and by different means of exercising and safeguarding rights, i.e. the rights of a child are exercised by custodians acting on behalf of the child. It should be noted that the state, which uses the levers of public power, is also involved in the protection and safeguarding of the rights of the child in the most direct way. The right of access should be mentioned as a special right of the child. This comprises the right of a child to direct contacts with the child's parents and siblings. Communication forms an integral part of daily lives of people as social beings. Communication with parents is essential for a child undergoing the process of personality development. The right of access is an absolute right, which may be restricted only in cases specified in laws, provided that access is harmful to a child. Although the right of access is provided for by law, the existing legal framework is still deficient, which is confirmed by frequent disputes arising over the exercise of this right. Objectives and methodology of the thesis The objective of the thesis is to research the legal framework dealing with the right of access in the context of the rights of the child by finding a way to make the exercise of the right of access solely in the interests of the child more effective. The research has employed descriptive and analytical, inductive and deductive methods. Laws, the opinions of legal scholars and case-law have been analysed using these methods; conclusions and suggestions have been formulated. It has become evident nowadays that the rights of children need to be guaranteed. However, the best and most effective ways of securing children's rights, including the right of access, still remain an open issue. The case-law shows a trend that, in certain cases, the right of access may be used by parents to settle their mutual relationships. This is evidence that the right of access does not always serve its main goal, which is to ensure the best interests of the child. The right of access is referred to in Article 181 of the Civil Law [1], whereby a child may maintain direct contact and personal relations with any of the parents, in both narrow and broad sense, and siblings, grandparents and other persons with whom the child has lived in a common household and developed a close relationship. In assessing the role of access for the development of a child, the concept of parents should be viewed more broadly, namely that a child may maintain direct contact with both biological parents and other relatives with whom the child has developed a close relationship (or an emotional bond). This is an essential aspect because it must always be established with whom of the relatives a child would prefer Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 to maintain contact when deciding on the procedure for exercising the right of access. The notion "close relationship" is open and must be given substance in each particular case, but it must be interpreted broadly in order to safeguard the best interests of a child. In practice, there is a great deal of discussion as to whether a child should meet and communicate with a new spouse of the divorced parent and spouse's relatives. It should be concluded that this matter is not dealt with by law. It is not prohibited. The author believes that contact with members of the new family of the parent is permissible provided it is accepted by the child and does not hinder the effective communication between the child and the parent, i.e. their travels together attending cultural events and other activities should be facilitated. The law clearly implies two things. First, it is a child's right rather than a duty; second, it is a right of a child rather than of child's parents. Parents must maintain direct contact with their child, thus ensuring the right of access. It is the child who decides on how this right will be exercised. It should be stressed that a child does not possess legal capacity and, accordingly, exercises his/her rights through a guardian who, as a rule, is a

parent with day-to-day care of the child. The fact that a parent and not the child brings an action regarding the right of access against the other parent creates a false impression that the dispute involves only the child's parents and not the child and a parent. The author is of the opinion that, in certain cases, this misunderstanding makes it more difficult to settle the dispute because the child's right of access may be used by parents to settle problems existing between them. Therefore, such disputes should be solved by means of mediation, with the participation of a family court that would defend the interests of a child. According to the existing legislation, a court must suggest mediation but, as regards family courts, a family court's opinion is necessary, but it is not required that their official attends court hearings. Practice shows that legislation does not always achieve its aim. Divorce, which quite often implies conflicts between parents, causes an adverse impact on the impartiality of hearings regarding the right of access. Mediation would help understand the causes of a dispute, while a family court's official would deliver an outside view and defend the child's interests effectively. Consequently, it would be necessary to amend the existing legislation and define mediation as a mandatory phase of the pre-trial dispute settlement procedure requiring the participation of a family court's official in the proceedings. This conclusion is consistent with the opinion expressed by legal scholars that a child has turned from a person to be cared for and protected into a right-holder [2]; therefore, the involvement of a child and any adult right-holder in the exercise of their rights is equally fully-fledged. It is clear that, depending on age and social maturity, children are not able to carry out procedural actions independently, and guardians of a child are persons who perform specific procedural actions required for the protection of the child's rights on behalf and in lieu of the child. This duty is a sort of "representation by legal nature" for guardians of the child [3]. It is established in the case-law that, although a child exercises the right of access independently, the child may need the other parent's assistance, with no obstacles put to exercising this right [4]. Guardians in their capacity as child's representatives must act for the child with the utmost care. It should therefore be concluded that the essence of the institution of representation consolidates the parents' duties laid down in family law. However, the dual nature of the right of access, namely the right of a child to contact and a parent's duty to ensure such contact, should always be borne in mind. There can be no right of access, unless both these components are present simultaneously. A child must be willing to see his/her parent. For obvious reasons, legislation cannot deal with a situation when a child does not want to exercise his/her right of access. In this case, the child is considered to have waived this right voluntarily. It should be emphasised that parent's refusal to see the child cannot be treated equally. For parents, access is a duty rather than a right. Parents may not waive their duty to maintain contact with their child. The legislator, however, has not introduced any responsibility for failure to fulfil this duty. It is questionable whether Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 any legal responsibility can be provided in the case of a duty requiring extremely personal involvement. Punishment for breach of legal duties is what makes legislation different from other types of social regulation (for example, moral standards). The author considers that legal responsibility of a parent should be introduced for refusal to exercise the right of access. The solution of this matter can be twofold. Apart from the above, competent authorities already have legal tools available for penalising parents, namely, in situations when communication is ordered by a court ruling, a parent who fails to comply with the court ruling and maintain contact with the child will be prosecuted. Article 168 of the Criminal Law lays down criminal liability for failure to comply with a court ruling regarding the right of access to the child, where a criminal offence has the following objective indicators: evasion of compliance with a court ruling, failure to comply with or delaying the execution of a court ruling in bad faith [5]. The objective aspect covers all possible forms of failure to comply with a ruling. Such criminal offences are punishable by temporary imprisonment, or community service, or a monetary fine. Punishment is adequate and serious enough. The author has analysed the Latvian case-law and established that no person has ever been prosecuted according to this article of the Criminal Law. This raises the question of whether the right of access is exercised effectively and parents do not evade their duty or law is not applied in situations when parents fail to fulfil their duty of access and, accordingly, a court ruling. Available information suggests that there are problems with the execution of court rulings. For example, 1,163 civil cases dealing with the right of access were heard in Latvia in 2018 [6]. Based on an analysis of statistical data in conjunction with publicly available information, it can be assumed that the respective provisions of the Criminal Law are not applied effectively. Considering that criminal proceedings are normally initiated on the basis of a statement lodged by a party concerned, it should be concluded that parties involved in the protection of the rights of children do not fulfil their duties properly. This mainly refers to parents providing day-today care of their children, whose duty would be to facilitate compliance with court rulings, including criminal prosecution of parents breaching the duty of access. Meanwhile, a family court should control parents and interfere in situations when they do not respond. From this viewpoint, cooperation between parents ensuring day-to-day care of their children and family courts should be improved. The right of access is absolute. It means that a right-holder may not be deprived of this right; however, the right of access may be restricted or suspended temporarily under certain circumstances. This is due to the absolute nature of the right of access. Temporary suspension of access should not be regarded as a sort of punishment of parents. Its legal aim is to restrict any contacts between a child and child's parents that have become harmful to the child for a certain period of time during which the parents may eliminate violations and improve their behaviour. It should be emphasised that, in fact, the restriction of the right of access is a process, which should be divided into the following three phases: assessment of circumstances that have led to the suspension of access, assistance to parents in the elimination of such circumstances and, finally, evaluation of the achieved result, namely whether the parents have rectified the deficiencies. Circumstances are assessed by a family court and a psychologist invited by them, whose opinions are relied on by a court. The legislator has not defined the duration of such restriction or suspension. An analysis of legislation demonstrates that the right of access may be restricted or suspended on the grounds that access affects a child's interests. Circumstances should be assessed only from the standpoint of a child's best interests. Disputes over the right of access are normally referred to courts, and that is why the goal of a court is to identify relevant circumstances, i.e. whether access is detrimental to a child's interests and should therefore be restricted or suspended, and determine the duration of such restriction or suspension. Circumstances leading to the suspension or restriction of the right of access must substantially affect a child's interests and Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 be supported by relevant evidence. Impairment of a child's interests can have both a legal and emotional or ethical nature. A child's interests can be adversely affected by communication with a parent who is striving to set the child against the other parent or encouraging the child, by either words or conduct, to act contrary to generally accepted behavioural rules (for example, not to go to school, use foul language) or commit administrative offences (for example, to beg, drink alcohol or use drugs); there is no doubt that parent violence, both emotional and physical, is harmful to the child's interests. We should agree with the opinion expressed in the case-law that violence against a child is a circumstance giving grounds for the restriction or suspension of access. A parent's subjective interpretation or suspicions cannot be accepted as sufficient because it cannot be excluded that the parent is driven by personal negative relations with the other parent. In order to be in a position to rule on the restriction of the right of access, a court needs strong evidence that a parent has acted contrary to a child's interests. In addition to general evidence that can be normally accepted for civil proceedings, such as witness statements, pleadings of parties or material evidence, there exist two significant kinds of means of evidence for cases concerning the right of access: a child's statements and a psychologist's opinion. When necessary, reports on psychological examination can also be accepted as evidence. It is set forth in Article 13(3) of the Law of the Republic of Latvia on the Protection of Children's Rights that a child may express his/her opinion about any matters affecting the child, and the child's opinion must be given proper attention, considering the age and maturity of the child [7]. Any ruling that is contrary to the opinion of a child must be explained by a court. When deciding on the restriction or suspension of the right of access, a court must determine the duration of such restriction or suspension. In accordance with Article 1495 of the Civil Law [8], duration can be specified as a certain period in months or years. Nevertheless, it should be acknowledged that duration defined as a certain period is not relevant for the right of access because this right can be restored as soon as circumstances that gave rise to the restriction or suspension of the right concerned no longer apply rather than upon expiry of the set duration. A psychologist, family court and the child concerned can be involved to establish these circumstances. Accordingly, the restoration of the right of access depends on whether special conditions are met. For instance, the Supreme Court of the Republic of Latvia has held that it can be difficult to enforce a judgment by which the right of access is suspended for three years provided that it can be restored in situations where, based on a psychologist's opinion, access will not be detrimental to the child's safety and health [9]. Therefore, the restoration of the right of access should be linked with a certain condition, which is the time when circumstances that gave rise to the restriction or suspension of access cease to exist. It should, however, be noted that the wording "when circumstances that gave rise to the restriction of access cease to exist", which does not refer to any specific timeframe, is too general and may unjustifiably affect the rights of both a child and a parent. It is indeed an open question who would be entitled to seek revision of the restriction imposed. How often may it be sought? While accepting the aforementioned conclusion of the court that duration based on a pre-requisite that the restoration of access will no longer be harmful to a child can hardly be defined because, to some extent, this wording is contradictory, the author still considers that special regulation is necessary. This means that legislation could define frequency of the revision of circumstances that gave rise to the restriction of access. Similar provisions have been introduced for the restriction of the capacity to act. The second paragraph of Article 364 1 of the Civil Law lays down that a court judgment regarding the restriction of the capacity to act may be reviewed at any time but not less than once every seven years from the date that the judgment has become final [1]. Certainly, a seven-year period is too long for cases dealing with the right of access because a

child would become Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 adult over this time. In this case, a period of two years would be reasonable. Therefore, a court would check whether circumstances that gave rise to the restriction of access have ceased to exist on its own initiative once every two years. The next step is assistance to parents provided by the state. Children are interested in restoring contact with their parents, and it is the state's job to help parents improve their behaviour and correct errors they have made in communication with children. The state should formulate a clear programme of how customised assistance could be provided to parents whose right of access has been restricted or suspended by a court. It is a state's duty because the state has assumed responsibility for safeguarding the interests of children. The state is not active enough in this respect. Such a programme should incorporate mandatory state-sponsored psychological assistance, social workers' support, involvement of employment officials, namely the assistance of specialists who would be able to both identify deficiencies and provide practical advice for the rectification of erroneous behaviour, which, as a rule, takes the form of alcohol abuse, fits of rage and hatred, unemployment or lack of understanding about a child's needs. At present, parents are those who solve these matters. If parents are motivated, state's assistance can be provided, but it is only possible in situations when parents are really motivated. Meanwhile, parents who lack motivation, who do not understand why their right of access has been suspended and do not want to change their behaviour remain in the grey area. State support cannot reach these people. Cooperation and feedback among courts and municipalities are necessary for improving the existing situation. Court rulings restricting the right of access should be included in a unified register and communicated to competent municipalities, who would organise assistance for parents whose right of access has been restricted or suspended. Systemic support activities would be organised owing to these efforts. The involvement of a parent in support activities and the progress achieved would also help evaluate whether grounds for the suspension of access have ceased to exist. It is also important that decisions on the restoration of the right of access would be much more objective and require less time. Parent information would be dynamic, covering a longer period of time. Certainly, the question remains how parents who lack motivation could be forced to take part in support activities because it can be assumed that a parent lacking self-motivation will not be willing to participate in mandatory activities voluntarily. Legislation should provide for a relevant enforcement mechanism. Interference with private rights of a child's parents is permissible for the sake of protection of the child's interests. Similar provisions have been adopted with respect to child support defaulters, whose car and boat driving licences can be suspended, except for situations when the parent is disabled or unfit for work over a long period of time, or suspension may considerably harm the child's interests (for example, a parent needs this special licence for the performance of his/her professional duties). This is provided by Article 7 of the Maintenance Guarantee Fund Law [10]. Analogous provisions could be introduced with regard to parents refusing to participate in support programmes with the aim of restoring effective communication. A child's rights to maintenance and contact are equally important; accordingly, there is no reason to consider that the adoption of a law with respect to a certain right (the right of access) in addition to that dealing with the other right would lead to excessive restriction of parents' rights. The right of access is not identical to, but should be linked with, the right of custody. The right of access is a component of the right of custody. Custody has a broader scope than access. It should be noted, however, that the right of access can be viewed as an independent legal institution, which has a theoretical link with custody. Deprivation of the right of custody per se does not affect the right of access. A parent having no right of custody will still have the right of access. Practice has shown, however, that Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 actions are quite often brought for the deprivation of the right of custody and the restriction of the right of access simultaneously. This is because one of parents is confident that any contact with the other parent with whom the child does not usually reside may produce an adverse effect on the child. This concern is well-founded in certain cases but this cannot be a priori sufficient to give rise to the presumption that the parent who does not ensure dayto-day custody is not able to maintain effective communication. Deprivation of the right of custody means that a parent with whom the child does not usually reside will have no right to decide on the child's daily needs and will be released from childcare responsibilities. This, however, should not affect communication between the child and the parent and their leisure time together. In some cases, a parent may be deprived of the right of custody for objective reasons, for example, because the parentis living far away from the child, and not because the parent is unable to exercise custody or has some negative traits. Therefore, there is no reason to believe that a parent whose right of custody has been withdrawn is ipso facto unable to maintain contact or has bad traits that would prevent effective communication and affect the child's interests. The right of custody may be withdrawn in the following two cases: based on a request of either the other parent or an institution that is responsible for the protection of a child's rights, i.e., a family court. Where the deprivation of the right of custody is sought by the other parent, the case is normally handled in a single set of proceedings and the action brought concerns access frequency and procedure. Actions brought for the withdrawal of the right of custody by family courts normally deal with both the restriction of the right of access and placing a child in a care institution because family courts are involved in serious cases, where the lowest possible point has been reached with respect to the safeguarding of a child's interests, namely in situations where the parents of the child are unable to ensure the child's rights or even abuse them, and it is therefore understandable why the suspension of the right of access is sought apart from the withdrawal of the right of custody. This situation is rather ambiguous. The author believes that the restriction of access should be used as a last resort. As a result of being placed in an extra familial care institution subject to restricted access, the child feels estranged from his/her parents, while the parents are less motivated to restore their right of custody. Placing a child in an extra-familial care institution may lead to a situation when the right of access cannot be exercised physically. The existing legislation does not promote the exercise of the right of access in such situations because no law requires that a child be placed in an extra-familial care institution that is nearer to the parents' home. The state must promote meetings between children and their parents. Researcher A. Haružka has also noted that the European Court of Human Rights has recognised in a judgment delivered on 26 February 2002 in Case 46544/99 Kutzner v. Germany that "the fact that a child could be placed in a more beneficial environment for his or her upbringing will not on its own justify a compulsory measure of removal from the care of the biological parents; there must exist other circumstances pointing to the "necessity" for such an interference with the parents' right to enjoy a family life with their child." [11]. It is clearly indicated that a child must be brought closer to his/her parents. Wherever the right of access is exercised, it is important to define how a child's right of access can be secured in the most effective way, and it is essential to understand the legal scope of the right of access. What falls within that scope, what specific actions are provided and what their frequency is. Practice demonstrates that access frequency and procedure may be different: at weekends, during school holidays or daily. This mainly depends on the physical possibilities of parents, such as distance, financial situation, etc. It cannot be clearly concluded that daily visits would be more useful than spending time with the other parent at weekends. It should be stressed, however, that the absence of even minimal access should be Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 viewed negatively. The author believes that guidelines that would define the basic principles of access should be formulated. For example, the legislation dealing with minimum child support has evolved analogously, from general rules whereby parents merely had to pay child maintenance depending on their possibilities and financial situation through to the specific minimum child support payment to be made regardless of the relevant parent's possibilities and financial situation. Minimum access and minimum time to be spent with a child could be defined in a similar manner. It should be stressed that access is a two-way right, where a child has a right, while parents have both a duty and right to maintain personal relationships and direct contact with their child. Both prerequisites must be met simultaneously for the effective exercise of the right: both parents and the child must be willing to establish and maintain direct contact. According to scientific literature, removed choices become diktat and rights are abused when interpreted too literally and forced upon people. The interpretation affecting children should be selected appropriately because children have insufficient control over their lives, which makes them susceptible to being controlled by over-interpretation [12]. Hence, it is important to establish whether a child really wishes to maintain contacts with a parent with whom the child does not live permanently and take into account the frequency and time of access desired by the child and expected activities during the time they spend together. Legislation dealing with the right of access has been left open by the legislator. Imperatively, legislation provides for the duty and specifies persons covered by this right, such as parents, siblings, grandparents and other persons with whom the child has lived in a common household for a long time provided it is in the best interests of the child. This could be defined as a quantitative component, but the qualitative component of access, which is the frequency and quality of contacts, is just as important. Practice shows that there are hardly any disputes over the right of access within families. This is evidence that communication with grandparents, siblings and other family members is less important, namely: access to these persons is not considered individually, it is assumed when deciding on contact with a divorced parent that communication with that parent's relatives is guaranteed by default. This, however, should not be regarded as self-evident, considering the interests of a child. The right of access with respect to persons with whom the child has lived in a common household is just as important as communication with a parent; therefore, family courts should carry out both awareness-raising and control activities. Without protecting the right of access to the fullest extent possible, the right of access is just formalism; accordingly, the safeguarding of children's rights is formal. As regards the right of access, A. Vorobjovs, professor says that any interaction between people can be called access. Means of communication include facial expressions, gestures, actions, words, results of actions [13]. It is concluded that the right of access presumes active cooperation of persons involved over a certain period of time, which includes personal presence, exchange of thoughts, talks and expression of emotions. It should be emphasised that in today's world remote access can also be implemented in the electronic environment, by using modern technologies. This, however, is surrogate access, which is not a substitute for direct physical contact. Remote communication cannot be regarded as due access. This conclusion is also indirectly confirmed by legislation, namely Article 2(10) of the Brussels Ibis Regulation

provides for a special right to take a child to a place other than his or her habitual residence in order to ensure direct contact [14]. Parents seeking to agree on the exercise of the right of access or bring an action must specify the time and place of access. It follows from both the essence of the right of access and Article 244 2 of the Civil Procedure Law [15]. The frequency and duration of access should be defined on an individual basis. It should be borne Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 in mind, however, that a child's opinion must be taken into consideration when deciding not only on the establishment of access but also on the procedure of access, with child's interests being paramount in this respect. The author shares the opinion of Associate Professor John Tobin that sometimes children lack the ability to exercise their rights but a child's rights are his/her interests rather than the ability to exercise them [16]. The frequency of communication, which should be regular, also in the form of contact, embodies the substantive nature of the right of access. Regularity means that access must be regular and take place in certain periods of time. This is associated with the nature of access, which is to ensure dynamic information about a child and allow participation of a divorced parent throughout all stages of child's development, while the child would enjoy parent's attention based on the needs that are relevant for a child of a certain age. It also means the right of a child to see persons with whom the child has lived in a common household because any of these persons (siblings, grandparents) are able to produce a different, but no less significant, impact on the development of the child. It should be concluded that the right of access ensures an educational effect, facilitating various qualities of a child, such as empathy, warm-heartedness and understanding of the family as a value. Results The right of access is a complicated legal institution. The right of access cannot be viewed only from the standpoint of law as it comprises psychological and emotional aspects because of which enforcement is not possible. Both children and their parents have inalienable rights to build and maintain direct contact with each other. Practice shows that it is mostly parents who do not want or are unable to ensure contact with children with whom they do not usually reside. It is also acknowledged that it is a parent's duty to ensure communication when the parent is not willing to do so. Meanwhile, a child is free to choose whether contact with a parent will be maintained, it is a child's right. This is because the rights of children have priority over those of parents as special attitude is guaranteed for children, considering that they are special right-holders. As a result of the research, the author has arrived at the following: 1) the right of access should be exercised solely in the interests of the child; 2) the right of access is an absolute right, which should be used regardless of whether parents have custody of the child; 3) the right of access is not part of the right of custody; 4) there is a lack of legal instruments that would prevent the use of the right of access contrary to its goal; 5) the state has no tool available that would ensure systemic parent support aimed at removing obstacles that led to the restriction or suspension of access. Conclusions The right of access exercised by parents neglecting the best interests of their child, when the child is used merely as a tool to settle their mutual relationships, is a major concern. As regards existing disputes, it is more likely that a decision on the exercise of the right of access might not fully meet the best interests of the child. The improvement of the existing legal framework must be focused more on laws that would eliminate the sources of disputes. It should be borne in mind that the right of access is an independent legal institution governing legal relationships between parents and children in the building and maintenance Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 of their direct contact. The state has a duty to ensure the respect of this right in the event of non-compliance for objective or subjective reasons. A child as a special right-holder has priority status in these legal relationships. Measures associated with the right of access can be qualified as ensuring access, the restriction or temporary suspension of access, the removal of obstacles that led to the restriction or suspension of access, the restoration of access. Officials involved in these stages should respect the interests of children, and their decisions should be in line with these interests. It should be borne in mind that, in fact, parents do not have their own interests in legal relationships as their interests arise from and are subordinate to the interests of children. Therefore, a policy pursued by the state in this respect should support parents, meanwhile safeguarding the interests of children in order to respect their rights. Proposals The legal framework dealing with access will definitely evolve. The first issue would be that parents should be encouraged to fulfil their duty to ensure access. Both motivational and enforcement mechanisms should be used, which would allow for the selection of the most appropriate tool depending on specific circumstances based on a parent's behaviour and the extent of the parent's motivation and involvement. In some situations, a parent who is striving to improve behaviour and restore contact with a child but is unable to do so due to external circumstances needs only motivational support. If a parent maliciously evades his/her responsibilities, enforcement or punishment is necessary. The state has various social support tools available, such as extra benefits, development of leisure facilities, etc. It is essential to introduce a requirement that, if necessary, a child be placed in an extra-familial care institution that is nearer to the parents' home. This would facilitate communication between parents and their children, and it would be easier to implement this requirement. The second issue to be considered by the legislator would be the introduction of adverse consequences for parents avoiding the fulfilment of their duty to provide access. Considering that administrative penalties would not be directly applicable in this case, other solutions must be found. A child as a right-holder requiring special protection has a right to state's interference with private relationships, namely parent-child relationships; therefore, proper legislation is permissible. But, before introducing any punishment mechanisms, the state should formulate and implement a support policy, whereby the assistance of psychologists, social workers and employment officials would be provided for parents. Parents may be held responsible for their non-compliance with support activities only in the presence of due support. The interruption of the right of access should be as short as possible. An emotional bond can be lost, and, as a result, the longer the interruption period, the harder the restoration of contact. Therefore, activities of a preventive nature should rather be facilitated. As regards the divorce of parents with minor children, a family court should be involved during the divorce process. The role of a family court would be to promote mediation, also if parents have voluntarily agreed on their divorce and all related matters, such as child maintenance, right of access and custody. Divorcing without conflict is not a guarantee that there will be no dispute over the right of access. Mediation is a means of identifying the cause of a dispute and preventing any controversies at their source. The legal framework governing access has both the potential and need for development. It should be admitted that the existing legislation is rather declaratory. The state has defined that the right of access is an inalienable right of children, while the subsequent implementation of this right is left freely to the discretion of parties concerned. The above means that it Int. Conf. SOCIETY. HEALTH. WELFARE. 2018 is parents who must ensure contact. Given that the protection of children's rights is a state priority, greater involvement of the state would be necessary for the solution of these matters. The author believes that the involvement of the state might take the form of legislation that would support parents and help them to overcome problems that led to the restriction or suspension of access or, as regards parents who lack motivation and do not want to cooperate, introduce adverse legal consequences. Within the frame of this process, a family court should be an authority representing a child and safeguarding the child's interests apart from a natural guardian. Only complex solutions involving parents, state and municipal officials and, certainly, children will bring the best result for the assurance of the right of access. Activity of M3814, an Oral DNA-PK Inhibitor, In Combination with Topoisomerase II Inhibitors in Ovarian Cancer Models DNA-dependent protein kinase (DNA-PK) has been shown to play a crucial role in repair of DNA double-strand breaks, facilitating nonhomologous end-joining. DNA-PK inhibitors have the potential to block DNA repair and therefore enhance DNA-damaging agents. M3814 is a DNA-PK inhibitor that has shown preclinical activity in combination with DNA-damaging agents, including radiotherapy and topoisomerase II inhibitors. Here we evaluated the activity of M3814 in combination with multiple topoisomerase II inhibitors, doxorubicin, etoposide, and pegylated liposomal doxorubicin (PLD) in vivo, utilizing ovarian cancer xenografts. Using cell lines representative of P53 wild-type ovarian cancer (A2780), and P53 mutant ovarian cancer (SKOV3), cells were implanted in the flank of athymic nude female mice. Mice were treated with vehicle, M3814 alone, topoisomerase II inhibitor alone, and M3814 in combination with topoisomerase II inhibitor, and change in tumor volume over time was documented. The addition of M3814 was well tolerated. We demonstrated that M3814 shows limited efficacy as a single agent in ovarian cancer models. The combination of M3814 with PLD showed enhanced activity over PLD as a single agent. Further study of this combination is warranted. outcome, but a number of trials suggest that combining chemotherapy agents in the platinum-resistant setting leads to increased toxicity with little increase in efficacy 19 . DNA-dependent protein kinase (DNA-PK) has been shown to play a crucial role in repair of DNA double-strand breaks, facilitating nonhomologous end-joining (NHEJ) [20][21][22][23] . This includes repair of double-strand breaks resulting from oxidative stress, oncogene induced transcription, or therapeutic treatment of cancer with chemotherapy or radiation 24 . The active DNA-PK complex is composed of a catalytic serine/threonine protein kinase (DNA-PKcs) and two heterodimeric subunits (KU80 and KU70) that bind to the double-strand break to direct the catalytic subunit to the site requiring repair 24 . The discovery of this necessary role in DNA damage repair highlighted the potential of DNA-PK inhibitors to block DNA repair and therefore enhance DNA-damaging agents. M3814 is a highly potent and selective inhibitor of DNA-PK 25 . M3814 has shown preclinical activity in combination with DNA-damaging agents, including radiotherapy across multiple cancer models 26 . M3814 has also shown activity in combination with the topoisomerase II inhibitor, etoposide, across multiple cell lines derived from lung cancers as well as increased efficacy when added to etoposide and cisplatin in a small cell lung cancer xenograft model, vs etoposide and cisplatin alone 26,27 . In addition, elevated DNA-PKcs expression has been associated with poor cancer specific survival in ovarian cancer patients 28 . We aimed to determine if DNA-PK inhibitors could enhance the efficacy of type II topoisomerase inhibitors in models of ovarian cancer. We hypothesized that the combination of M3814 and Top2 inhibitors will provide ovarian cancer patients with new treatment options that are both efficacious and tolerable. Results Ovarian cancer cell lines show etoposide sensitivity in vitro. Because multiple passaging of cell lines has been associated with loss of etoposide sensitivity 29 , we first sought to confirm sensitivity of ovarian cancer cell lines to etoposide in vitro. The effect of etoposide on cell proliferation was determined by MTT assay after 72 hours of drug treatment. A2780 and SKOV3 cell lines demonstrated sensitivity to etoposide inhibition with an IC50 of 112 nM and 1.9 uM, respectively. The OVCAR3 cell line was resistant to etoposide inhibition with an IC50 of 29.1 uM (Fig.

1). The IC50 of each cell line was statistically significant from each other (P < 0.0001). DNA-PK inhibitor M3814 enhances activity of select topoisomerase II inhibitors in vivo. The DNA-PK inhibitor M3814 has shown activity in combination with etoposide in lung cancer models 27 . To test the efficacy of M3814 in combination with topoisomerase II inhibitors, we performed in vivo xenograft studies using cell lines representative of P53 wild-type ovarian cancer (A2780), and P53 mutant ovarian cancer (SKOV3). Cells were implanted in the flank of athymic nude female mice, and drug treatment was started once tumor volume reached approximately 100 mm 3 . Mice were treated with vehicle, M3814 alone, topoisomerase II inhibitor alone, and M3814 in combination with topoisomerase II inhibitor. M3814 treatment alone showed no difference in tumor volume compared to vehicle, in both A2780 and SKOV3 xenograft models (Fig. 2). As shown in Fig. 3, A2780 cells demonstrated decreased tumor growth in response to treatment with etoposide, doxorubicin (Adriamycin), and pegylated liposomal doxorubicin (PLD, Doxil) compared to vehicle. Of the single agents, cells were most sensitive to PLD, with a mean tumor volume of 1227 mm 3 at day 31 compared to a mean tumor volume of 2208 mm 3 for vehicle alone. Although A2780 cells displayed sensitivity to etoposide in vitro, tumor growth in vivo did not show inhibition to a similar extent. However, combination of M3814 with etoposide trended toward improved growth inhibition with a mean tumor volume of 1542 mm 3 at day 31 compared to a mean tumor volume of 1784.1 mm 3 for etoposide alone, although the difference was not statistically significant (P = 0.8088) (Fig. 3A,B). Similarly, combination of M3814 with PLD also trended toward reduced tumor growth, although not statistically significant, with a mean tumor volume of 1109 mm 3 at day 31 compared to 1227 mm 3 for PLD alone (P = 0.9732) (Fig. 3G,H). A2780 showed limited sensitivity to doxorubicin alone in www.nature.com/scientificreports www.nature.com/scientificreports/ vivo, with little improvement seen with combination therapy (Fig. 3D,E). Body weights remained stable throughout the experiment (Fig. 3C,F,I). Xenograft experiments with the SKOV3 cell line exhibited slightly more resistance to etoposide and doxorubicin compared to the A2780 line, which recapitulated our in vitro results. As a result, combination of M3814 with either etoposide or doxorubicin had little effect on SKOV3 tumor growth compared to etoposide or doxorubicin alone (P > 0.9999, P = 0.9934, respectively) ( Fig. 4A-E). In contrast, SKOV3 cells were sensitive to PLD, with a mean tumor volume of 593 mm 3 at day 54 compared to a mean tumor volume of 1257 mm 3 at day 44 for vehicle. Combination of M3814 with PLD led to a further reduction in tumor growth, with a mean tumor volume of 345 mm 3 at day 54, although not statistically significant from M3814 alone (P = 0.2143) (Fig. 4G,H). Body weights remained stable throughout the experiment (Fig. 4C,F,I). Discussion Treatment options for platinum-resistant ovarian cancer patients remain limited and, although PLD has activity, single agent response rates are low. Viable combination therapy options are necessary to improve the efficacy of available treatment options. DNA-PK inhibitors have been shown activity with DNA-damaging agents, highlighting their potential to improve the efficacy of these agents while remaining tolerable for patients. We studied the effects of M3814 in combination with topoisomerase II inhibitors. M3814 showed no efficacy as a single agent in ovarian cancer models. This is consistent with the functional mechanism of DNA-PK; inhibiting this protein in the absence of DNA damage should have no effect on the cell. It is only in the presence of DNA damage that DNA-PK inhibition prevents DNA damage repair, exacerbating cell death. The importance of combining DNA-PK inhibition with therapies that effectively induce DNA damage was emphasized by our in vitro and in vivo results. In cell lines that demonstrated sensitivity to topoisomerase II inhibitors in vitro, combination therapy of M3814 and topoisomerase II inhibitor generally led to reduced tumor growth compared to topoisomerase II inhibitor alone. However, cell lines that were resistant to topoisomerase II inhibitors in vitro showed no benefit from combination therapy in vivo. The combination of M3814 with PLD produced a more pronounced affect when compared to vehicle in the SKOV3 cell line, which is p53 null, than in the A2780 cell line which is p53 wild-type (p < 0.0001 vs p = 0.003). Prior in vitro work of M3814 in combination with irradiation has suggested that p53 mutation may serve as a possible marker for response to M3814 30 . When moving forward with clinical trials for combination therapy, it will be crucial to use DNA-PK inhibitors in combination with DNA-damaging agents that have single agent efficacy, before attempting combination therapy. If the DNA-damaging agent alone does not have an impact on cell growth, then combination therapy with DNA-PK inhibitors will have little success. In addition to using the right DNA-damaging agents, our results highlight the importance of determining the optimal dosing schedule in therapy to obtain the best result. While the A2780 cell line was sensitive to etoposide inhibition in vitro, etoposide had little effect on tumor growth in vivo. Because the efficacy of the single DNA-damaging agent alone is critical in effecting any improvement in combination therapy, it will be essential to determine dosing schedules that maximize the efficacy of the single agent. Although there are important considerations in DNA-PK combination therapy, our results demonstrate that M3814 enhances the efficacy of PLD, leading to reduction in tumor growth. While combination therapy seems necessary to improve upon the available treatment options, recent studies have shown that many combination therapies lead to increased toxicity. Bevacizumab, a humanized monoclonal antibody targeting vascular endothelial growth factor, was recently approved by the FDA for use in combination with chemotherapies (PLD, paclitaxel, and topotecan) in platinum-resistant advanced ovarian cancer, based on a phase III study that demonstrated an improved progression-free survival 31 . However, bevacizumab combination therapy does carry additional risk of toxicity, including risk of hypertension, thromboembolic complications, and bowel perforation, thus precluding its use in certain patient populations 32 . We predict that the DNA-PK inhibitor M3418 may improve the efficacy of PLD in ovarian cancer patients; further clinical trials are warranted. www.nature.com/scientificreports www.nature.com/scientificreports/ experiments were approved by the Institutional Animal Care and Use Committee (protocol # 04-03-009). For each cell line, two million cells with Matrigel were injected subcutaneously into a single flank of athymic nude female mice. Treatment was started when tumors measured approximately 100 mm 3 . Each treatment group (vehicle, M3814 alone, chemotherapy alone, or combination) consisted of 7 mice. M3814 was administered at 50 mg/ kg by oral gavage once daily, five days per week. M3814 vehicle control was 0.5% Methocel ™ , 0.25% Tween20, 300 mM Na-Citrate Buffer, pH 2.5. Etoposide was administered at 8 mg/kg by intraperitoneal injection once daily, three days per week. Etoposide vehicle was saline. Doxorubicin was administered at 2.5 mg/kg by intraperitoneal injection once weekly. Doxorubicin vehicle was water. PLD was administered at 6 mg/kg by intravenous tail injection once weekly. PLD vehicle was 0.5% dextrose. Tumor growth and animal weight were monitored twice per week. Tumor volume at end of treatment was compared for treatment with vehicle vs chemo and vehicle vs combination therapy for each experiment using an unpaired T-test (Prism) (Figs. 3 and 4). Critical Success Factors for Safety Training in the Construction Industry : Construction is a hazardous industry. The project-based nature and fragmentation in the industry lead to change and uncertainty requiring special expertise. To handle those, construction firms must develop strategies and action plans along with the experience gained from lessons learned. Among the risks, safety risks are of critical importance leading to accidents. Hence, firms need to strengthen their safety programs, review their strategies for safety management, and develop effective safety training sessions to protect their workers. This study focuses on the success factors promoting safety performance. In this respect, a questionnaire was designed and administered to the Engineering News-Record (ENR) 2020 Top 400 Contractors. The questionnaire data was utilized in conducting a factor analysis to group and name the factors considering the total variance. The analysis of the factors resulted in six-factor groups; namely, project and firm-related factors, demographic factors, practical factors, motivational factors, organizational factors, and human-related factors. Project and firm-related factors were found to be the most essential factor group in terms of promoting the effectiveness of safety training. The results of this study are expected to guide industry practitioners in terms of reviewing and revising their safety training programs. Introduction The construction industry is one of the riskiest industries due to the high number of work-related hazards and injuries [1][2][3][4][5]. The U.S. Bureau of Labor Statistics (BLS) [6] reported that the total number of fatal work injuries in the U.S. was 5250, 1008 of which were in construction. Moreover, the total number of non-fatal work injuries was recorded as 2,834,500, 199,100 of which were in construction [6]. The work-related injuries and fatalities mostly stem from the fact that workers fail to comply with the rules in safety programs [7]. Heinrich's (1931) [8] study on safety indicated that a sequence of factors on worker mistakes combined with dangerous or unsafe behavior is the main cause of accidents. Bird and Germain (1990) [9] further mentioned that work accidents are preventable provided the underlying factors of accidents are well determined. Alarcón et al. (2016) [10] reported that the occurrence of work accidents is not random and due to several controllable factors. Durdyev et al. (2017) [11] mentioned that contractors might have taken early action towards promoting construction safety, if the factors affecting construction safety performance were studied. Similarly, Øien et al. (2011) [12] stated that studying the factors leading to accidents beforehand shall be effective in preventing major accidents. As a result, these studies raised a growing interest in construction safety (Guo et al., 2016) [13]. The study of factors leading to a high rate of injuries and fatalities in the construction industry revealed that inadequate training is one of the most important factors causing accidents [14]. To this end, Occupational Health and Safety Administration (OSHA) has various programs Materials and Methods The majority of work-related accidents on construction sites stem from unsafe actions and conditions [24,25]. Both researchers and industry practitioners are seeking new methods against the high number of accidents in the construction industry [3]. There are various methods mentioned in the previous works to enhance the performance of safety practices. For example, Lai et al. (2011) [26] focused on human resource (HR) practices (recruitment, incentives and rewards, safety training, communication, and feedback, etc.). In their study, they pointed out the importance of establishing safety policies based on effective HR practices to improve safety management on construction sites. Hinze et al. (2013) [27] focused on key indicators to assess construction safety and proposed a simple formula for "percent of worker observations that were safe," and "the number of positive reinforcements provided per 200,000 h". A considerable portion of the studies developed models or present strategies for increased safety performance such as the safety culture interaction (SCI) model [28], plan-prevent-protect strategy [29], fatigue assessment scale for construction workers (FASCW) [30], and social-ecological model of safety performance improvement (SEM-SPI) [31]. Being the main focus of this study, safety training holds an important place among the factors affecting safety performance. Several studies have already implied the critical role of the training in reducing the number of unsafe acts and behavior. For example, Sollis-Carcadio and Franco-Poot (2014) [40] found that workers who had limited safety training and developed a poor safety culture are more vulnerable to being involved in accidents. Moreover, workers who had proper safety training are more likely to identify hazards and develop a safety risk perception [41,42]. Tam and Fung (2012) [43] determined that compulsory training raises construction workers' interest in safety. Training designed following the needs of workers having different levels of knowledge yields a higher efficiency in terms of safety [3,43]. Hence, workers who received safety training were critical to the development of safety programs and management [14]. Ho and Dzeng (2010) [44] proved that a suitable safety training has the potential to promote safe behavior. Kaskutas et al. (2013) [45] concluded that foremen who had completed safety training developed more

effective safety communication practices. A significant portion of previous studies shows that incorporating technology and specifically virtual reality (VR) in a safety training session is an effective means of enhancing the efficiency and quality of the training [46][47][48][49][50][51][52]. VR is effective in attracting trainees' attention and strengthening learning in construction safety training [46]. Previous studies also focused on the impact of safety training on immigrant construction workers [53][54][55]. Williams et al. (2010) [53] revealed that immigrant workers are willing to improve their knowledge of safety. The study also found that the number of hazardous behaviors among immigrant workers had reduced after an effective safety training session. Furthermore, Han et al. (2008) [54] Learning constitutes an important part of an effective safety training [3]. However, workers might fail to practice what they learned from the training when they returned to work [5,56]. Therefore, several studies highlighted the critical role of transfer of training, concluding that the involvement of stakeholders positively affects the training transfer [57,58]. Given this general background, this study focuses on determining the critical success factors (CSF) of an effective safety training program. In this context, both qualitative and quantitative methods were adopted. As part of the qualitative methods, a content analysis and an in-depth literature review were conducted. Moreover, expert interviews were conducted to create a consensus for the identification of the safety training variables. Then, a survey study with factor analysis was employed. The survey helped assess and quantify the variables identified. Moreover, statistical data was analyzed based on the responses collected. The survey also revealed information regarding the position, background, and experience of the respondents. The factor analysis was conducted to group the variables in a coherent factor and name as a categorical group. The scarcity of resources investigating the variables making a safety training program effective and the gap in the literature are the main motivations for the study. The study assesses these critical factors with a careful consideration of the recent research in this domain, revealing the need for such research in the construction industry. Table 1 presents the safety training success variables adopted for this study based on an in-depth analysis of the previous studies. Variable Explanation References Age Age is listed as an essential parameter in terms of affecting the success of safety training for the fact that there may be a significant difference in the safety training perceptions of younger and older workers. [14,36,42,44,53,55,[59][60][61] Gender Gender is indicated as an important parameter in the success of safety training since differences might be observed in safety training success depending on workers' genders. [5,14,44,59,61] Country of Origin Country of origin is another essential factor affecting the success of training sessions since people from different origins might develop different perceptions of safety. [14,53,55,59] Educational Background The level of education might be a strong indicator of safe behaviors. Hence, educational background is listed as an important parameter for safety training success. [14,36,44,53,58,61] Language There is strong evidence that language might become a barrier in safety training sessions. Thus, language is considered as an important component of safety training success. [3,14,26,55,62,63] Work Experience Work experience is a directly related parameter with safety training since experienced workers are more likely to promote safety training success. [14,36,53,55,59,62] Perception of Safety Training Perception of safety training is critical in terms of putting the training into practice. A strong/positive perception of safety training leads to a successful implementation of what was learned in a safety training session. [3,5,43,46] Hands-on Training Hands-on training is an effective practice in terms of promoting safety training since it is linked with real cases on construction sites leading to a better training session considering worker needs. [3,5,43,46,58,64] Training Frequency Training frequency is important for the success of safety training sessions. More frequent training results in better reinforced learning of safety. [5,43,44,61,65,66] Methods and Materials Methods and materials used in safety training sessions are of utmost importance in terms of promoting safety. They help elucidate essential safety information for workers through learning, strategies, and equipment. [3,5,14,43,63,64,66,67] Training Satisfaction Training satisfaction is directly related to safety training as the level of satisfaction, comprehension, and emotional engagement in training contributes to the success of safety training sessions. [20,44,61] Safety Awareness and Motivation Safety awareness and motivation are crucial for the success of safety training. This is simply described as the level of knowledge concerning unsafe acts or behaviors and the need for motivation for safe practices. [3,38,46,58,64,65,68] Number of Unsafe Acts and Accidents Developing a prior knowledge of unsafe acts and accidents constitutes an important part of safety training success. Hence, the number of unsafe acts and incidents affects the effectiveness of safety training sessions. [44,46,61] Effectiveness of Training The success of safety training depends on the effectiveness of training. Effectiveness represents the developed understating of the severity and consequences of safety-related accidents after the training. [14,36,44,46,53,62] Coordination and Collaboration Coordination and collaboration in safety training sessions lead to information sharing and observational environment among trainees. This eventually contributes to the success of safety training. [3,44,46,50,64] Feedback Feedback is an essential part of safety training sessions. Providing feedback after a training session, especially regarding the trainee performance, leads to a higher success in the training. [3,51,58,64] Management Support Effective training is only possible with strong management support. This contributes to higher performance in the training sessions. [3,58,64] Use of Personal Protective Equipment (PPE) One of the most important subjects in safety is to promote the use of PPEs through safety training sessions. Providing workers with a core training section of the PPE use is strongly related to safety training success. [38,43,46,48,49,55,58,63,66,69] Firm Size Firm size is associated with the scope and content of safety training considering the budget and resources spared for the training sessions. This in turn affects the success of training sessions. [3,5,48,59,60,66] Project Type Safety training sessions differ depending on project type. Hence, it is important to design safety training content by the different requirements of project types [42,44,48,55] Project Duration Project duration determines the ability to design training sessions in varying durations and content. This directly affects the success of safety training. [44,52] Leadership Leadership is effective in terms of promoting safer practices. Exemplary leadership has a positive impact on the success of safety training sessions. [13,15,33,37,38] Project size Project size refers to the scale of projects in regards to budget and workload. Project size shapes the content and duration of safety training, which in turn affects the success of training sessions. [26,31,42,70] Incentives for Safety Incentives for safety have the potential to promote safety training sessions and safe behaviors through positive reinforcement. [26,37,38,64,65,70] Training Language Training language is directly related to the success of safety training sessions since some trainees fail to practice what they learned in training sessions when the training language is not their mother tongue. [3,55,58,66,71] Research Methodology This study explores the CSFs in safety training sessions in the construction industry. To determine these factors and cluster them in a coherent scheme, an online survey was designed and administered to the construction firms listed in the Engineering News-Record (ENR)'s 2019 Top Contractors List. According to the data on the ENR website, the firms listed in the Top 400 Contractors in 2018 generated a revenue of US$405 billion in 2018, where this was US$373.98 billion in 2017 [72]. This indicates that these firms on the list generate most of the construction value and shows their extensive presence in the industry. The survey consists of two sections, namely the general information about the respondents and the success factors in safety training. The survey designed is presented in Appendix A. The first sections aimed at collecting data regarding the characteristics of the respondents such as respondent age, total turnover, number of employees, educational background, and years of operation in the construction industry. The second part consists of questions about the success factors in safety training sessions. In the second part, a 5-point Likert Scale was adopted to rate the success factors, where 1 represents very low and 5 represents very high. A total of 400 surveys were sent out and 93 responses were collected at a response rate of 23%. The survey data was analyzed using the Factor Analysis tool of SPSS 23. The factor analysis was utilized to identify which underlying factors are measured by a considerably larger number of observed variables. The analysis of the data collected through the survey is provided in the following sections. General Information about the Respondents The respondent characteristics were sought to better interpret the results. Table 2 presents the respondents' characteristics such as the age of the firm, annual turnover, and the number of employees. According to Table 2, the average firm age is 45, whereas the maximum and minimum ages are 69 and 27, respectively. The annual turnover of the responding firms was reported as US$181.81 million on average, where the maximum turnover is US$806 million, and the minimum is US$1 million. Finally, the total number of employees was reported as 876 on average, where the maximum number is 3700 and the minimum is 58. The first part of the survey also collected information regarding the respondent's gender, origin, and educational level. Figure 1 presents data regarding the gender of the respondents. According to Figure 1, 82% of the respondents are male, whereas 18% of the respondents are female. Figure 2 presents the origin of the respondents. According to Figure 2, the majority of the respondents' origin is United States (81.7%) and the remaining origins are from other countries. Figure 3 exhibits the level of education of the respondents, where 63% of the respondents hold an MSc. degree, and 29.0% BSc., 6.5% Ph.D., and 1.5% of the respondents are high school graduates, respectively. Figure 4 presents the respondents' current role in the construction industry. According to this, 39.8% of the respondents are project managers, 23.7% are engineers, 12.9% are general managers, and 23.6% have other roles such as foreman, safety directors, safety supervisors, and architects. The analysis of both figures revealed that the majority of the respondents hold an MSc degree and work at the project manager position. Figure 5 shows the years of operation of the responding firms in the construction industry. According to this, 37.6% of the firms have been operating in the construction industry for more than 20 years, where 30.1% have been operating in the industry between 15-20 years, 22.6% have been operating in the industry between 10-15 years, 8.6% have been operating in the industry between 5-10 years, and 1.1% of the firms reported the years of operation less than 5 years. The respondents were further categorized according to their years of experience in the construction industry as shown in Figure 6. The majority of the respondents (33.3%) have been working in the construction industry for 15-20 years. Twenty-nine percent of the respondents reported an experience interval of 10-15 years. A considerable percentage of the respondents (17.2%) reported that they have more than 20 years of experience in the construction industry. The remaining respondents reported 5-10 years of experience (10.8%) and less than 5 years of experience (9.7%) in the construction industry. The analysis of both figures indicated that the majority of the firms are quite experienced in the construction industry and a major portion of the respondents have much experience in the industry. In the survey, the respondents were asked whether there were previously involved in a safety training session. All respondents indicated they have previously participated in a safety training session. Moreover, the number of safety training attended by the respondents was assessed. According to the data collected, 52.7% of the respondents have previously participated in safety training sessions one to three times, 36.6% of the respondents reported three to five times, and 10.8% indicated that they had been involved in a safety training session more than five times. The survey also collected the satisfaction level of the respondents regarding the previous safety training that they had. Figure 7 presents the satisfaction level reported by the respondents. According to this, a major portion (90.3%) of the respondents implied that their satisfaction was high with the previous training they had. Of the respondents, 4.3% indicated that they had very high satisfaction with the previous training, whereas 5.4% reported that they had a medium satisfaction level with

the training they had received. The respondents responded to the main causes of unsafe behaviors or acts such as low perception of safety, inattention, and improper use of personal protective equipment (PPE). The respondents emphasized that effective safety training helped them avoid unsafe behaviors and increased their ability to detect unsafe conditions. Moreover, the respondents were asked to report whether they had been involved in a work-related accident. The analysis of the results indicated that 35.5% of the respondents had already been involved in a work-related accident, whereas 64.5% reported they had not been involved in a workrelated accident before. The respondents who reported that they had been involved in a work-related accident, also stated that safety training helped them detect mistakes and avoid unsafe behavior to a great extent. The survey further assessed the increase in safety awareness after safety training. Figure 8 presents that the safety awareness increased by 51-70% after having at least one safety training session for the majority of the respondents (57%). Of the respondents, 35.5% indicated that the safety awareness increased by more than 71% after participating in a safety training session; 7.5% reported an increase between 31-50% in safety awareness as a result of a completed safety training session. Another question in the survey aimed to reveal whether the respondents had had any language problems in the safety training sessions. The analysis of the data showed that 11.8% of the respondents had problems in understanding the language of the safety training, where 88.2% reported that they did not have such a problem. Moreover, the respondents were asked to indicate whether they needed motivation to eliminate unsafe behavior. All respondents reported the need for motivation to eliminate unsafe behavior. The survey further assessed the safety training methods utilized in the training sessions ( Figure 9). All respondents reported that more than one training method was utilized in the training sessions. Accordingly, computer-based training (94.6%) and on-the-job training (91.4%) are the most commonly used methods in the training sessions. Success Factors in Safety Training Sessions Safety training is challenging for several reasons such as content and audience issues. Considering the impact of safety training in increasing safety performance, more emphasis needs to be put on improving safety training effectiveness. This requires a critical analysis of the underlying factors for successful safety training sessions. Within this context, this study identifies a set of critical success factors (CSFs) for successful implementation of safety training. The CSFs of safety training sessions were identified through an in-depth analysis of the previous studies and interviews with the safety directors and managers. The interviews were conducted with safety directors and safety managers. The stratified sampling was used to identify the interviewees. A total of eight experts were selected for the interviews, where five of them are safety managers and three are safety directors. The inclusion criteria were to have experience with at least three safety training programmes in a project and a minimum of 5 years of experience in construction safety. Each interview was allotted 30 min and was conducted online. The interviews with the experts revealed the need for improvement regarding the training sessions so that trainees better learn and conceive the content and implement what they learned in the construction sites. In the initial step, a total of 41 variables were identified such as reward mechanisms for safety, training content, and training tools and techniques. After reviewing these 41 factors with industry experts and academics, some of the variables were merged into one variable, or some were removed to represent a coherent list of variables. For example, reward mechanisms and promotions were merged into one single variable as safety incentives. Moreover, communication was synthesized with coordination and collaboration variable. A similar approach was followed for the remaining variables. A final list composed of 25 safety training variables was produced after carefully revising the list (Table 3). Table 3 presents descriptive statistics for the CSFs based on the 93 responses collected. The results indicate that the use of PPE, leadership, safety awareness, and motivation are the most important success drivers in safety training sessions, whereas demographic variables such as age, country of origin, and gender are less important in terms of impacting the success of safety training sessions. A further investigation of the descriptive statistics indicated that the coefficient of variation (CV) for the variables lies between 0.05-0.30, representing a significant variability and dispersion in the data set. Results In the second phase, the factor analysis method was carried out to determine the underlying factors for safety training success. Factor analysis is a multivariate statistical method to reveal the correlated variables along with some independent factors, which represent a combination of the original variables [73][74][75]. The factor analysis helps group the variables into a meaningful and considerably smaller number of factors, which is also called factor extraction, by representing the relations among a set of interrelated variables. Factor analysis has two types, namely the exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). EFA is carried out by identifying a set of factors maximizing the amount of variance required. CFA rather considers the factors based on an a priori hypothesis, which is an indication of the evaluation of the factors being mostly driven by the theory. This study adopts the EFA to identify the factors for safety training success. In this respect, the suitability of the data was assessed first by conducting the Kaiser-Meyer-Olkin (KMO) and Barlett's tests for sphericity. The KMO test presents the sampling adequacy with values lying between 0-1, indicating the extent of prediction of one variable by other variables without error. Higher values of the test prove that the factor analysis is likely to produce more reliable and distinct factors. According to the scale defined by [70], KMO test values that are greater than 0.5 are acceptable, whereas the values between 0.5-0.7 are mediocre, and the values between 0.7-0.8 are considered to be good [76]. Bartlett's test checks whether the correlation matrix is an identity matrix [77]. The results of the KMO statistic were found to be 0.818 in this study. This value proves that the dataset is suitable for the factor analysis. The Barlett's test generated a Chi-square value of 2533.228 and the level of significance was found to be small (p < 0.0005) [78]. These values indicate that the factor analysis is an appropriate method for the dataset. The factor analysis was conducted using the SPSS 23 software. The analysis generated several files such as correlation matrix, total variance, component matrix, rotated component matrix, and the component plot in rotated space. The determinant value for the correlation matrix helps assess the singularity effect. It is considered that there is no singularity effect for the determinant values greater than 0.00001 [79]. The analysis of the data generated a determinant value of 5.22 × 10 −3 , which is greater than 0.00001, proving that there is no singularity effect. This also removes the need for eliminating any variables from the defined list of variables. The correlation matrix presents the total variance explained by a set of factors. However, the motivation behind the factor analysis is to come up with an optimal number of factors to explain a significant portion of the variance. The high intercorrelations detected between some variables also led to the dimension reduction of the matrix. Table 4 presents the total variance explained by components. The components are shown based on the initial Eigenvalues, variance, and cumulative variance. The table proves that the components having an Eigenvalue of greater than 1 explain the major portion of the variance, which accounts for 77.070% of the total variance. This amount is acceptable for the analysis based on the values provided in the literature [79]. Hence, a total of six components were investigated for the analysis. Table 5 shows the rotated components matrix, which was generated with the Varimax rotation method. The matrix is composed of six groups with 25 variables. The first common group is called "project and firm related factors", explaining 39,041% of the total variance. This factor corresponds to the variables impacting the trainee performance in the success of safety training sessions. The second common group is called "demographic factors," explaining 13,083% of the total variance. The third common group is called "practical factors" explaining 9952% of the total variance. The fourth common factor group is called "motivational factors", corresponding to 5504% of the total variance. The fifth-factor group is called "organizational factors" (5301%) focusing on the factors related to an organization in terms of affecting safety training success. Finally, the sixth-factor group is called "humanrelated factors" comprising of the variables for human-related parameters on safety training success with a variance of 4.190%. The values in bold in the rotated component matrix represent the factor loadings of the variables. Discussion of Findings In this study, a total of 25 variables were determined through an in-depth literature review and interviews with safety experts to reveal the success factors for safety training sessions. Then, the factor analysis was executed to find the common factor groups. The analysis resulted in six common factor groups namely, the project and firm-related factors, demographic factors, practical factors, motivational factors, organizational factors, and human-related factors. The discussion of each factor group is presented below. Project and Firm Related Factors This factor group constitutes 39,041% of the total variance based on the data presented in Table 4. Project type (mean: 3.59), project size (mean: 3.58), and project duration (mean: 3.58) were found to be the most significant variables in this factor group with factor loadings of 0.920, 0.921, and 0.923, respectively (Table 5). According to 75% of the survey participants, the project type, size, and duration are listed as the most important components in a successful safety training session. Teo and Feng (2009) [79] stated that project duration and size have an impact on safety culture and climate. Moreover, Jannadi and Assaf (1998) [80] explained that safety awareness differs depending on construction project size and safety practices, which is considerably lower for small-sized construction projects. Love et al. (2018) [81] further stated project type has a strong correlation with the number of injuries and accidents. This may be due to the fact that the innate nature of particular project types forces organizations to allocate more resources and attention to the training sessions. For instance, Kang et al. (2018) [82] implied that fall accidents are more common in residential projects than other project types. Firm Size (mean: 3.54) was determined as the fourth important variable among the other variables. Khosravi et al. (2014) [34] mentioned that firm size is directly associated with unsafe acts and accidents. Several other studies also emphasized that large-sized firms are less likely to experience unsafe acts and accidents [45,83]. This may be due to the fact that large-sized firms generally have more established safety management practices overall. With the evidence from previous studies and the analysis of the survey data, one may assert that project and firm-related factors have a strong effect on the success of a safety training session. Hence, both practitioners and organizations in the industry must be aware of this for their safety management practices. Demographic Factors The demographic factors strongly affect unsafe acts and accidents [34]. Moreover, Wilkins (2011) [14] stated that the demographic composition of workers is important in terms of construction safety training. According to Table 4, the demographic factors constitute 13,083% of the total variance. Among the variables of the demographic factors group, gender (mean: 3.37) was listed as the most important variable with a factor loading of 0.838. Loosemore and Malouf (2019) [61] underlined that gender differences are of great importance for the improvement of accident prevention strategies such as safety training. Gender might also act as a parameter to discriminate between the severity levels of work accidents [84]. Country of origin (mean: 3.35) was determined as the second most important variable among demographic factors with a factor loading of 0.815. Mahalingam and Levitt (2007) stated that the origin country of workers is an important factor in construction safety due to cultural differences. Some other studies implied that immigrant workers are more likely to be involved in accidents than local workers [85,86]. Byler (2013) [87] found that the number of fatal accidents is 66% higher for Latin American

workers who are foreign-born than for US-born ones. Hence, one may claim that country of origin is an essential factor needing careful consideration for safety training sessions. Age (mean: 3.32) is the third most important variable with a factor loading of 0.769. Khosravi et al. (2015) [34] stated that age is closely related to the number of unsafe acts and accidents. Khodabandeh et al., 2016 [84] indicated that the age of workers might be investigated to distinguish between different severity levels of work accidents. Kale and Baradan (2020) [39] emphasized that the frequency of work injuries decreases as the age of workers increases. According to Zhang and An (2012) [88], as workers grow older, safety observation and intelligibility decrease. Moreover, Lossemore and Malouf (2019) [61] underlined that older workers are less motivated by the safety training than other age groups. Hence, safety training content and approaches tailored for different age groups should be considered. The educational background (mean: 3.77) of workers was determined as the fourth important variable (factor loading: 0.703). Educational background is one of the essential factors for the formation of risk perception [89][90][91]. Some other studies found that the educational background of construction workers is directly related to safety awareness [82,85]. Given this background, educational background is considered as an important factor affecting the success of safety training. Thus, one may advocate that as educational level increases, safety awareness becomes higher. Finally, language (mean: 3.68; factor loading: 0.509) of workers was determined as the last important variable among the demographic factors. Keng and Razak (2014) [62] reported that one of the problems in safety applications on construction sites is the language barrier among workers. However, immigrant workers might be motivated to conduct bilingual safety training [62]. Hence, language is an important factor that needs to be carefully considered in safety training sessions to increase the success of safety training. Practical Implementation Factors The practical implementation factors constitute 9.952% of the total variance as shown in Table 4. Among this factor group, hands-on training (mean: 4.69; factor loading: 0.838) was determined as the first important variable. Burke et al. (2006) [92] reported that safety training becomes more effective with hands-on training. According to [58], instructors of safety training must try to develop not only traditional paper-based but also on-site assessment and engagement techniques. Cameron et al. (2011) [93] indicated that learning methods driven by visual cues and tools facilitate safety training. Since handson-training provides a practical dimension for safety, it has the potential to increase safety training success. Perception of training (mean: 4.60; factor loading: 0.794) was found to be the second important variable of the practical factors. According to 97% of the survey participants, the perception of training constitutes an important place in safety training implementation. Mushayi et al. (2018) [94] highlighted that employees' perceptions of health and safety training create a considerable impact on their health and safety behavior and compliance. Demirkesen and Arditi (2015) [3] also highlighted the importance of safety training perception in developing a safe behavior. Employees perceiving the benefits of training sessions have the potential to develop safe behavior and avoid themselves from being involved in hazardous situations. They may also pay better attention during the sessions. Therefore, developing a positive safety training perception is a critical variable in terms of conducting more effective safety training and must be considered by safety professionals. Methods and materials (mean: 4.90; factor loading: 0.565) were determined as the fourth important variable of the practical factors. Jeelani et al. (2017) [51] revealed that there is a great need to improve the methods and efficiency for safety training. Hussain et al. (2018) [58] reported that a minor increase in training transfer results in a major increase in workers' safety performance. Therefore, authorities should focus on additional methods to maximize the level of safety training [58]. Tam and Fung (2012) [43] stated that alternative training methods should be included in safety training sessions. They further mentioned that different methods (such as multimedia aids) should be used in safety training sessions to provide a more effective training and attract participants' attention. Jeelani et al. (2017) [51] recommend realistic and immersive training environments to improve the efficiency of safety training. Furthermore, safety procedures must be well written and understandable to all workers [43]. Hence, the methods and materials used in training sessions should be wisely selected and implemented to attract workers' interest in safety training. This is also linked with workers' perception of safety training. Work experience (mean: 4.41; factor loading: 0.511) of workers was determined as the fifth important variable on the rotated component matrix. Previous studies proved that the experience of workers affects work safety and safety perception. In Hare et al. (2013)'s study [71], it was found that there are significant differences among workers in terms of understanding safety pictures. They further indicated that workers with more than 5 years of experience were successful in defining safety pictures than those with less experience. According to [39], the possibility of work injuries decreases as the experience of workers increases. Aligned with these studies, it is of utmost importance to grow experience in safety and safe practices in parallel to the experience in construction. Therefore, one needs to carefully consider the impact of the experience of trainees on safety training success while tailoring safety training sessions. Organizational Factors The organizational factors constitute 5.504% of the total variance as shown in Table 4. Effectiveness of training (mean: 4.83; factor loading: 0.842) was determined as the first important variable explaining the motivational factors group. Haslam et al. (2005) [95] stated that poor safety knowledge is the main cause of at least 70% of the accidents on construction sites. According to [58], effective training increases workers' knowledge of safety. Moreover, it was further implied that effective training practices and training transfer elements should be adopted together to maximize risk perception and hazard recognition [41,42]. Demirkesen and Arditi (2015) [3] underlined that necessary attention must be given to the effectiveness of learning in training sessions to improve construction site safety. Tam and Fung (2012) [43] further mentioned that training content and methods must be carefully designed and reevaluated when necessary to increase the effectiveness of training sessions. As proven by the previous studies, increasing the effectiveness of training is one of the critical elements to provide a more successful safety training session. This variable should be considered carefully by practitioners and organizations. Feedback (mean: 4.83; factor loading: 0.715) was determined as the second important variable explaining the motivational factor group. Providing feedback is critical to improving the safety performance of workers [26,96]. Lai et al. (2011) [26] reported that feedback is an important reminder to workers about their unsafe acts. Demirkesen and Arditi (2015) [3] also mentioned that feedback is one of the most essential components of effective safety training. Therefore, safety trainers must ensure having effective feedback mechanisms in their safety training programs. Coordination and collaboration (mean: 4.74; factor loading: 0.689) was determined as the third important variable of the motivational factors. According to 99% of the survey participants, coordination and collaboration play a critical role in safety training implementation. Sun et al. (2017) [50] showed that construction firms will not be able to provide vocational (skill and safety) training without government involvement concluding that the government, workers, professional organizations, and firms need to cooperate for vocational training. A good coordination and a collaborative environment lead to effective safety training sessions, allowing trainees to better conceive the essence of a training session, where they feel empowered and motivated to learn about safety. Management support (mean: 4.71; factor loading: 0.662) was found to be the fourth important variable of the motivational factors group. Keng and Razak (2014) [62] reported that the problems on construction sites usually stem from the lack of safety awareness and financial resources for safety management. Copper (2006) [97] indicated that safety performance is better when management support is high. Several studies have already highlighted the critical role of management commitment to reducing the number of safety-related accidents [26,32,34,[98][99][100]. As evidenced by previous studies, management support is critical in terms of creating a workplace safety culture and promoting safety training practices. Therefore, firms aiming to increase their safety performance through training sessions must beware of the fact that they need to commit to safety management programs with full support. Safety awareness and motivation (mean: 4.91; factor loading: 0.559) was found to be the fifth important variable of the motivational factors. A major portion of the survey respondents indicated that safety awareness and motivation constitute an important part of safety training implementation. [41] found that motivation is one of the necessary transfer elements for hazard recognition and safety training. Demirkesen and Arditi (2015) [3] expressed that large construction firms in the U.S.A. are responsive to worker issues such as motivating workers and raising awareness for safety. Kim (2009) [91] showed that worker personality affects safety awareness of construction workers. Hence, safety awareness and motivation are essential elements in achieving a successful safety training session. Trainees would better learn in an environment, where awareness for safety is raised and motivation is promoted. The number of unsafe acts and accidents (mean: 4.62; factor loading: 0.518) was determined as the last important variable of the motivational factors. Kale and Baradan (2020) [33] showed that unsafe conditions and unsafe acts influence work injury severity. Hence, awareness of unsafe acts and accident numbers leads trainees to being alerted during safety training sessions. Moreover, trainees would better concentrate during training sessions, where the number of unsafe acts and accidents are presented in the training sessions. This raises the potential for taking preventive measures towards a specific type of accident or develop measures for different unsafe acts. Motivational Factors The organizational factors constitute 5.301% of the total variance as presented in Table 4. Incentives for safety (mean, 4.65; factor loading: 0.825) was found to be the first important variable of the organizational factors. Khosravi et al. (2015) [34] mentioned that incentives are directly associated with unsafe acts and accidents. Sectoral incentives such as rewards and punishments encourage construction firms to provide vocational (skill and safety) training [50]. Teo et al. (2005) [98] concluded that one of the most important factors affecting construction site safety is the incentives. Hence, the presence of incentives deserves special emphasis to raise motivation for safety training and to increase the effectiveness of training sessions. Firms seriously considering incentives to reward workers for their safe behavior are more likely to experience enhanced safety performance. Thus, organizations should take incentives into account and consider having an incentive structure linked with their safety training programs. Training language (mean: 4.61; factor loading: 0.780) was found to be the second important variable for the organizational factors group. A significant portion of studies in the literature proves that immigrant construction workers experience difficulties in understanding the language of training in safety training sessions. Demirkesen and Arditi (2015) [3] advocated that safety training participants' language problems in their training sessions might be solved using translators and visual aids. Moreover, Hussain et al. (2018) [58] proposed that immigrant workers can be encouraged to go through bilingual safety training provided their firms are able to offer such training opportunities. Training frequency (mean: 4.54; factor loading: 0.724) was determined as the third important variable of the practical factors. Pandit et al. (2019) [60] emphasized that a positive safety climate affects construction workers' perception of risk, leading to increased awareness for risks and hazards. Furthermore, as the frequency of training increases, construction workers' attitude towards safer practices changes positively [55]. Tam and Fung (2012) [43] further implied that regular training is essential for promoting safety and positively affecting workers' safety behavior. Therefore, regular/frequent safety training arrangements must be considered carefully. Training satisfaction (mean: 4.86; factor loading: 0.601) was found to be the third important variable of the organizational factors group. The statistics obtained from this research showed that 95% of the participants in the study, who are mostly managers, indicated that they are satisfied with the safety training they had before. Nevertheless, Wilkins (2011) [14] evaluated the satisfaction of construction workers with safety training and stated that a high proportion of workers are actually not satisfied with the training provided to them. Moreover, Lossemore and Malouf (2019) [61] mentioned that there are

minimum positive changes in workers' behavioral intentions before and after safety training, indicating that workers' emotional engagement with safety is high at the beginning of the safety training. Hence, it is essential that trainees are satisfied with the training provided to them so that they commit to the training efforts. Measuring the training satisfaction during or after a safety training session and acting for change and improvement are critical to tailor effective training sessions. Because training satisfaction is a variable affecting training success, it is of greater importance to show efforts in measuring training satisfaction. Human Related and Behavioral Factors The final factor group is the human-related and behavioral factors, which constitute 4.190% of the total variance. This factor consists of the use of personnel protective equipment (PPE) (mean: 4.95; factor loading: 0.731) and leadership (mean: 4.92; factor loading: 0.543) variables. According to all survey participants, using PPE and leadership have high importance in safety training implementation. Previous studies also show the importance of these two factors in construction safety. Yap and Lee (2019) [33] stated that the most critical variable for safety awareness is the use of PPE. Wu et al. (2017) [101] revealed that leadership positively affects construction safety management. The use of PPE in training sessions is a way to demonstrate to trainees how life-saving equipment might be correctly used. This would reinforce the safety knowledge of the trainees leading to a visual representation of what needs to be done on-site. By utilizing PPE in training sessions, one might increase the session's success. Leadership is another essential variable in terms of affecting the success of safety training. Leaders are people who have the potential to affect a broad community with essential safety information by championing safety and setting an example. Good leadership results in good practice and enhanced safety performance. Consequently, training success would be higher in the existence of good leaders and leadership skills. Conclusions The construction industry suffers from a high number of work-related injuries and accidents. Even though various efforts have been put in place to eliminate work-related accidents, there still needs to be serious measures to completely protect workers from being involved in hazardous situations. One of the most commonly utilized efforts in this respect is safety training provided within an organization. Knowing this great potential for preventing work-related injuries and fatalities, organizations must develop ways to improve their safety programs and specifically, their safety training sessions. However, there are still difficulties in designing effective safety training sessions and enhancing safety performance through these sessions. Especially, different learning styles of construction workers render safety trainings more challenging and risk their safety knowledge gained through those trainings. One other concern is that the workers are not still motivated and fostered towards employing safeguards such as use of PPE, following safety standards, and stopping the line in unsafe situations. The language barrier also presents a critical risk for foreign workers, leading to poor safety learning. Therefore, this study explores the CFs for effective safety training sessions for the construction industry. In this respect, a total of 25 variables were identified for the success of safety training based on an in-depth literature review and interviews with safety experts. Then, a survey was designed and administered to the ENR Top 400 Contractors List. The survey participants were selected from this group since the top contractors are the ones that provide regular safety training to their employees. The survey resulted in a response rate of 23%. This study targets the responses provided by the top contractors for the fact that their successful practices become examples to those aiming to promote their safety programs. Moreover, the study puts safety training to the forefront of safety programs since different learning styles leading to safe and unsafe acts are best observed during training sessions. Based on the analyses of the survey responses, it was found that most of the responses were collected from construction professionals having over 20 years of experience in the construction industry. The respondents also reported that they are highly satisfied with the previous safety training they had and implied that safety training helped them increase their safety awareness. The respondents further mentioned that they were trained with different safety training methods such as computer-based training, on-the-job training, and equipment-use simulation. For the safety training success variables, the factor analysis was executed to group the factors and name them accordingly. The analyses of the factors resulted in six-factor groups, namely the project and firm-related factors, demographic factors, practical factors, motivational factors, organizational factors, and human-related factors. Among the factor groups, project and firm-related factors were found to be the most influential factors for safety training success. This factor group consists of variables such as project type, project size, project duration, and firm size. The other factor groups were also found to be significant, but their relative importance is lower compared to the project and firm related factors group. The motivation behind this study is to assess the factors leading to a higher success in safety training sessions for the fact that effective training provides trainees with better safety performance, eventually reducing work-related injuries and fatalities. Since this is a serious concern in the construction industry, the investigation of the parameters affecting safety performance plays a critical role in rendering a better safety performance. To this end, the results of this study are expected to guide construction industry practitioners in terms of reviewing their safety programs and revise their safety training assessment accordingly. This also contributes to the reputation of an organization and increases their work reliability. However, this study has also some limitations as the results generated are based on the perceptions and experiences of the 400 Top Contractors in the U.S. The results could differ with another data set collected in a different region. Moreover, the survey was conducted with a relatively small sample reflecting the perceptions of a small group. On the other hand, the results are generalizable since the 400 Top Contractors are industry leaders with their well-developed safety programs and safety management experience. Hence, their insights into safety training might provide a roadmap for a broader community. As future work, the results of this study might lead researchers to investigate safety training practices in other countries/regions and the results might be compared. Similar studies focusing on different firm sizes and project types might be useful. Moreover, the performance of safety training sessions from both the trainer's and trainee's perspectives might be assessed. The success factors provided in this study might be used to assess whether firms considering these parameters in designing their safety training perform better or not. Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions. Conflicts of Interest: The authors declare no conflict of interest. Appendix A Survey of critical success factors in safety training 1 Optimized Extraction of Polysaccharides from Grateloupia livida (Harv.) Yamada and Biological Activities Polysaccharides from Grateloupia livida (Harv.) Yamada (GL) were extracted by a heating circumfluence method. Single-factor experiments were performed for the three parameters: extraction time (X1), extraction temperature (X2) and the ratio of water to raw material (X3) and their test range. From preliminary experimental results, one type of the response surface methodology, the Box-Behnken design was applied for the optimizing polysaccharide extraction conditions. The experimental data obtained were fitted to a second-order polynomial equation. The optimal conditions were extraction time 5 h, extraction temperature 100 °C and ratio of water to raw material 70 mL/g. Under these conditions, the experimental yield was 39.22% ± 0.09%, which well matched the predicted value (39.25%), with 0.9774 coefficient of determination (R2). GL polysaccharides had scavenging activities for DPPH and hydroxyl radicals in vitro. The scavenging rates for both radicals peaked at 20 mg/mL GL concentration. However, the positive standard, VC (ascorbic acid), possessed stronger antioxidant activities than GL polysaccharides. Furthermore, the anticancer activity of GL polysaccharides on HepG2 cell proliferation increased dose- and time-dependently, but the positive standard, 5-fluorouracil (5-fu, showed more significant anticancer activity in this study. Overall, GL polysaccharides may have potential applications in the medical and food industries. arrange and interpret results for than other methods [20]. It is widely used in optimizing extraction process variables, such as polysaccharides, protein, anthocyanins and phenolic compounds [21]. According to our preliminary research, we separated the ethanol extract, petroleum ether, ethyl acetate, n-butyl alcohol and aqueous fractions from GL and discovered antioxidant, antibacterial and antischistosomal bioactivities [8]. In this experiment, we studied the extraction of aqueous fractions, or polysaccharides. The extraction conditions of polysaccharides were optimized to increase the extraction yield. Furthermore, we evaluated the scavenging and anticancer activities of GL polysaccharides with the aim of exploitation for medicines and foods. Single-Factor Experimental Analysis of G. livida (Harv.) Yamada (GL) Polysaccharide Extraction The extraction time, extraction temperature and ratio of water to raw material were studied and the results of single-factor experiments are shown in Figure 1. The three factors were effective for the yield of GL polysaccharides. The most suitable conditions were chosen for use in the BBD design. Effect of Extraction Time on Extraction Yield of GL Polysaccharides The effect of extraction time on the yield of GL polysaccharides was investigated with extraction time 3 to 7 h ( Figure 1A), with other extraction parameters fixed, such as an extraction temperature of 90 °C and ratio of water to raw material of 70 mL/g. The extraction yield of GL polysaccharides significantly increased from 9.02% to 22.42% with extraction time varying from 3 h to 5 h; it began to plateau with proceeding extraction. The polysaccharides need much time to swell so that they can be released and diffused into water [22]. However, extended extraction time may lead to the degradation of the polysaccharides [23]. Therefore, an extraction time ranging from 4 h to 6 h is the best for polysaccharide extraction. Effect of Extraction Temperature on Extraction Yield of GL Polysaccharides The extraction yield of GL polysaccharides affected by extraction temperature is in Figure 1B. The extraction process was performed with temperatures from 60 °C to 100 °C, with the other extraction variables such as ratio of water to raw material and extraction time fixed at 70 mL/g and 5 h, respectively. The extraction yield of GL polysaccharides increased with increasing extraction temperature from 60 °C to 80 °C, then peaked at 26.73% ± 2.25% at 100 °C. With increasing extraction temperature, the polysaccharide diffusion coefficient increases accordingly, for enhanced solubility of polysaccharides in the extracting solvent [24]. We selected temperatures from 80 °C to 100 °C as the appropriate extraction temperature. Effect of Ratio of Water to Raw Material on Extraction Yield of GL Polysaccharides Different ratios of water to raw material have an effect on extraction yield ( Figure 1C). In this study, we used ratios of water to raw material from 50 mL/g to 90 mL/g, with other parameters (extraction time and extraction temperature) fixed at 0 level (5 h and 90 °C). Extraction yield increased with ratio of water to material from 50 mL/g to 70 mL/g and decreased before it peaked at 27.19% ± 2.41% at 90 mL/g. Under the appropriate condition of ratio of water to raw material, the polysaccharides can swell thoroughly, and more polysaccharide molecules could dissolve in water to improve extraction yield [22]. If ratio of water to raw material is too small, polysaccharides cannot be completely extracted, and if too high, the cost will increase [25]. Thus, we chose ratios of water to material from 60 mL/g to 80 mL/g as the optimal conditions. According to the single-factor study, the following conditions could be used for the BBD (Box-Behnken design) experiments: extraction time of 4-6 h, extraction temperature of 80-100 °C and ratio of water to raw material of 60-80 mL/g. Table 1 shows the BBD with three factors and three levels examined to optimize the mutual effect of three independent variables (extraction time, extraction temperature and ratio of water to material) on extraction yield of GL polysaccharides. The design matrix and experimental and predicted extraction yield of polysaccharides are in Table 2. The yield of GL polysaccharides ranged from 20.3% to 39.48% and the maximum yield was with ratio of water to

raw material 70 mL/g at 100 °C and 4 h extraction time. On multiple regression analysis, the quadratic polynomial equation for the independent variables and response variable can be expressed as follows: A summary of the ANOVA parameters for the fitted quadratic polynomial model for extraction yield of GL polysaccharides is in Table 3. The F-test was used to check the statistical significance of the regression equation and p values to check the significance of each coefficient, which in turn may indicate the pattern of the interactions between the variables [26]. The model f-value for the model was 24.01 and the p-value was 0.0014, so the model was significant. The probability that the large f-value could occur due to noise was 0.14%. The smaller the p-values, the larger the significance of the corresponding coefficient [27]. The linear coefficients (X2 and X3), interaction terms (X1X2, X1X3 and X2X3) and quadratic term (X 2 3 ) were significant (p < 0.05) ( Table 3). The adequacy of the model could be confirmed by the determination coefficient (R 2 ) [28]. The determination coefficient (R 2 = 0.9774) suggested that the model was valid, implying that 97.74% of the variation could be explained by the fitted model. The adjusted determination coefficient was used to evaluate the correlation between the experimental values and predicted values, and the outcome (R 2 Adj = 0.9367) suggested that the correlation was significant. The f-value (6.12) for "the lack of fit" indicated that the "lack of fit" was not significant relative to the pure error (p > 0.05). As a general rule, the coefficient of variation (CV) should not be >10% [29]. The CV for yield of GL polysaccharides was 4.25%, which indicated a good reliability of the experimental values. Adequate precision measures the signal to noise ratio. Its desired value is ≥4 [30]. The adequate precision value for our model of 15.554 indicated an adequate signal and that the model could be used to navigate the design space. Optimization of GL Polysaccharide Extraction The response surface methodology was used to investigate the interactions of the variables and determine the optimal level of each variable for the maximal response [31]. The 3D response surface and 2D contour plots provided graphical representations of the regression equation (Figure 2), providing a means of visualizing the mutual effect of independent variables at different levels on the extraction yield of polysaccharides and the interactions between the independent variables. The 3D response surface plot demonstrated the mutual effect of two variables, with the other variable maintained at the respective zero level. Figure 2A depicts the interaction effect of extraction time (X1) and extraction temperature (X2) on the yield of GL polysaccharides with the ratio of water to raw material (X3) fixed at 70 mL/g. The yield of GL polysaccharides increased slightly with the increasing extraction time and temperature. Furthermore, the contour plot ( Figure 2B) showed significant interactions between two variables (p < 0.05, Table 3). The extraction time slightly increased yield of GL polysaccharides, whereas yield increased rapidly with ratio of water to raw material from 60 mL/g to 68 mL/g ( Figure 2C). However, with a further increase of the ratio of water to raw material, the yield decreased. Maximal extraction yield of GL polysaccharides could be achieved with extraction time and ratio of water to raw material 5 h and 70 mL/g, respectively. The contour plot ( Figure 2D) and the results from Table 3 (p < 0.05) show significant reciprocal interactions between the two variables. The yield of GL polysaccharides increased with increasing ratio of water to raw material and after yield peaked, it decreased slowly ( Figure 2E). However, yield increased little with increased extraction temperature when extraction time was set at 5 h. The relationships between the two variables were significant (p < 0.05, Table 3, Figure 2F). In conclusion, the highest yield of GL polysaccharides could be obtained with the extraction temperature and ratio of water to raw material of about 100 °C and 70 mL/g, respectively. Verification of Predictive Model According to the Figure 2 and Equation (6), the optimal extraction conditions for GL polysaccharides were as follows: extraction time (X1) 4.63 h, extraction temperature (X2) 100 °C and ratio of water to raw material (X3) 69.35 mL/g. Under these optimal extraction conditions, the theoretical maximal extraction yield of GL polysaccharides was predicted to be 39.25% by the mathematical model. Considering operability in the experiment, the optimal conditions could be adjusted to the following conditions: extraction time (X1) 5 h, extraction temperature (X2) 100 °C and ratio of water to raw material (X3) 70 mL/g. To ensure the validation of the model equation, we performed three independent experiments under the modified polysaccharide extraction conditions. The mean polysaccharide extraction was 39.22% ± 0.09%, which agrees well with the predicted value for the good correlation between experimental and predicted values, and the response model is adequate for optimization. DPPH Radical Scavenging Activity The free radical scavenging activity of GL polysaccharides were measured by DPPH assay. DPPH, a stable N-centered free radical, has been often used to analyze the ability of free-radical scavenging properties or hydrogen donation of compounds and medicine materials [32]. As shown in Figure 3, DPPH scavenging by GL polysaccharides increased rapidly with increasing concentrations from 4 mg/mL to 8 mg/mL, and then grew steadily to peak at 80.71% with 20 mg/mL concentration. Therefore, GL polysaccharides, especially at high concentration, had a noticeable scavenging activity on DPPH radicals. Nevertheless, compared to the scavenging rate of the positive standard (VC; ascorbic acid), the antioxidant activity of GL polysaccharides was not as good as VC. The possible mechanism may be their electron donation power to free radicals, thereby terminating the radical chain reaction [33]. Hydroxyl Radical Scavenging Activity The hydroxyl radical is considered the most reactive and poisonous free radical and induces severe damage to adjacent biomolecules as an active initiator for lipids peroxidation [34]. The reaction system containing Fe 2+ -H2O2-salicylic acid in the aqueous phase was used to generate OH [35] to measure the scavenging activity of GL polysaccharides for hydroxyl radicals (Figure 4). With increasing concentration of GL polysaccharides, the hydroxyl radical scavenging activity increased, and at 20 mg/mL concentration, its scavenging rate peaked. Furthermore, the antioxidant activity of VC was greater than with GL polysaccharides. Previous studies have suggested that the antioxidant mechanism may be due to the supply of hydrogen by polysaccharides, which combine with radicals and form a stable radical to terminate the radical chain reaction [36]. GL polysaccharides may scavenge hydroxyl radicals by providing hydrogen. However, the exact mechanism relating to the hydroxyl radical scavenging activity by polysaccharides needs further study. Anticancer Activity We evaluated the cell inhibition rates of GL polysaccharides on HepG2 cells after incubation at three intervals (24 h, 48 h, and 72 h) via MTT assay. As shown in Figure 5, the growth of HepG2 cells was subject to different inhibition rates at different times with different concentrations of GL polysaccharides, representing the time-effect relationship. With increasing incubation time, the inhibition rate of HepG2 cells increased. The results were significant except for cells incubated between 48 h and 72 h at 0.06 mg/mL. Otherwise, the GL polysaccharides inhibited the proliferation of HepG2 cells dose-dependently. However, with 24 h incubation, cell growth was not inhibited significantly at a low concentration of GL polysaccharides, so the polysaccharides had no useful effects until they were at the suitable concentration. As a positive standard, 5-fluorouracil (5-fu) had a greater effect on HepG2 cells than GL polysaccharides did. However, incubation of GL polysaccharides and 5-fu at low concentrations (0.06 mg/mL and 0.6 mg/mL) were not significant (p ˃ 0.05). Oxidative stress is among the main causes of cancer-related death. If compounds can enhance the level of anti-oxidation and reduce reactive oxygen species content in cancer cells, they may inhibit cell growth [37]. Therefore, the anticancer activity of GL polysaccharides might be ascribed to its effect on scavenging free radicals. Plant and Cell Materials Grateloupia livida (Harv.) Yamada (GL) was obtained from Nan Ao Island, Shantou, Guangdong Province, China, and identified by the Nan Ao Marine Biological Research Institute of Shantou University, China. The human liver carcinoma cell line HepG2 was obtained from the Medical College of Shantou University, China. Preparation of GL GL was washed, smashed, and dried at 60 °C to prepare the experimental sample. To remove lipid materials and other micromolecule matter thoroughly, GL powders were prepared by operating heating circumfluence with 95% ethanol, at 90 °C twice and 3 h each time. The purified GL powders were then filtrated and dried for use. Polysaccharide Extraction from GL The prepared GL powders (0.5 g) were extracted by water in a designed extraction time (3-7 h), extraction temperature (60-100 °C), and ratio of water to raw material (50-90 mL/g) and the extract liquid was concentrated in a rotary evaporator. The solution was precipitated by 80% ethanol and incubated for 12 h at 4 °C in the refrigerator before it was centrifuged at 3500 rpm for 15 min [38]. Then the precipitant was dissolved in water and frozen at −80 °C. Crude GL polysaccharides were obtained by lyophilization. The content of GL polysaccharides was measured by the common phenol-sulfuric acid method, with D-glucose used as the standard to construct a standard curve. The percentage polysaccharide yield (%) was calculated as follows [39]: where X (mg/mL) is the concentration of GL polysaccharide solution, V (mL) is the volume of GL polysaccharide solution and W (g) is the dried powder of GL weight. Single Factor Experimental Design Single factor experiments were performed to investigate the effect of extraction time, extraction temperature, and ratio of water to raw material on GL polysaccharide extraction yield. During the optimization of experimental factors, one factor was changed and the other factors were kept constant in each experiment. All experiments were repeated three times. BBD and Statistical Analysis From the preliminary single factor experiment design, BBD was performed for further optimization of GL polysaccharide extraction. Three parameters, including extraction time, extraction temperature, and ratio of water to raw material, were designated X1, X2, and X3, respectively ( Table 1). Each of the independent variables had coded levels of −1 (low), 0 (central), +1 (high).The quadratic polynomial equation reflecting the relationship between three independent variables and the response variable (extraction yield of GL polysaccharides) is as follows [40]: where Y is the response variable; β0, βi, βii, and βij are regression coefficients of variables for intercept, linearity, quadratic, and interaction terms, respectively; and Xi and Xj are the coded independent variables (i ≠ j). In the whole design, 15 experimental runs were performed in random order, with 12 factorial runs and three center runs. The factorial point is the 3D vertex consisting of independent variables (X1, X2 and X3). Three experiments were performed at the center point (0) to evaluate the pure error [41]. The design of the experiments is in Table 2. All trials were performed in triplicate. Statistical analysis of data was performed to establish the mathematical model by the BBD method, evaluating the effects of each independent variable on the response. DPPH Radical Scavenging Activity The DPPH radical scavenging activity assay was performed as described [42], with a few modifications. Briefly, 250 μL of 0.1 mM freshly prepared DPPH solution (in 95% ethanol) was added to 50 μL GL polysaccharide solution at 1 mg/mL, 4 mg/mL, 8 mg/mL, 16 mg/mL and 20 mg/mL, respectively. The mixture was shaken vigorously and incubated at 25 °C for 30 min in the dark, before measuring the absorbance at 517 nm. VC (ascorbic acid) was used as the positive standard. The GL polysaccharide ability to scavenge DPPH was calculated by the following equation: where Ai is the absorbance obtained from GL polysaccharides mixed with DPPH solution, Aj is the absorbance of GL polysaccharides with 95% ethanol and A0 is the absorbance of DPPH solution with 95% ethanol. Data are expressed as mean ± SD of three independent experiments. Hydroxyl Radical Scavenging Assay The hydroxyl radical scavenging ability of GL polysaccharides was analyzed by the hydroxyl radical system generated by the Fenton reaction [43] with minor changes. In brief, GL polysaccharides were dissolved in distilled water at 1 mg/mL, 4 mg/mL, 8 mg/mL, 16 mg/mL and 20 mg/mL. The

reaction mixture contained 50 μL FeSO4 (6 mM), 50 μL H2O2 (6 mM) and 100 μL GL polysaccharides of varying concentrations. After shaking and incubation at room temperature for 10 min, 50 μL salicylic acid was added to the mixture, then shaken and incubated at room temperature for another 30 min, and the absorbance of the mixture was measured at 510 nm. VC was considered the positive standard. The hydroxyl radical scavenging rate was calculated by the following equation: where Ai is the absorbance obtained from the reaction mixture (FeSO4 and H2O2) with GL polysaccharides and salicylic acid, Aj is the absorbance of the reaction mixture with GL polysaccharides and A0 is the absorbance of reaction mixture with salicylic acid. Data are expressed as mean ± SD of three independent experiments. Cell Cytotoxicity Assay in Vitro 3.8.1. Cell Culture DMEM medium was supplemented with 10% fetal bovine serum, penicillin (100 units/mL) and streptomycin (100 units/mL). HepG2 Cells were cultured in DMEM medium and incubated at 37 °C in a humidified 5% CO2 incubator. MTT Assay The cytotoxic effect of GL polysaccharides on HepG2 cells was evaluated by the standard MTT assay. HepG2 cells were seeded at 5 × 10 4 per well in 96-well microplates and incubated for 24 h to allow cell attachment. Then different concentrations of GL polysaccharides dissolved (0.06 mg/mL, 0.6 mg/mL, 6 mg/mL) in DMEM were added to each well in the plates. 5-fluorouracil (5-fu) was considered the positive standard. The cells cultured with DMEM medium were the control. The plates were incubated in a humidified 5% CO2 incubator at 37 °C for 24 h, 48 h, and 72 h, respectively. Then, 20 μL (5 mg/mL) MTT solution was added to the medium and further incubated for 4 h in the dark. The medium in each well was removed and 150 μL DMSO was added to dissolve the purple formazan crystals. The absorbance at 490 nm (A490) was read on microplate reader and the cell cytotoxicity rate was calculated as follows: Cell cytotoxicity rate (%) = where Ai is the absorbance from the treatment group and Aj is the absorbance of the control group. Data are expressed as mean ± SD from three independent experiments. Conclusions In summary, the polysaccharides from G. livida were extracted by the heating circumfluence method and the extraction process was successfully optimized by the BBD method. From the single-factor experiments, the optimization of the extraction conditions, including extraction time, extraction temperature and ratio of water to material, were determined as extraction time 5 h, extraction temperature 100 °C and ratio of water to raw material 70 mL/g. Under these conditions, the experimental yield of polysaccharides was 39.22% ± 0.09%, close to the predicted yield of 39.25%. Hence, the model established by BBD was adequate for GL polysaccharide extraction. The antioxidant and anticancer activities of GL polysaccharides were evaluated. GL polysaccharides possessed certain inhibitory effects on DPPH radical and hydroxyl radicals. For anticancer activity, GL polysaccharides can suppress HepG2 cell proliferation somewhat, but are not as good as 5-fu. The results obtained in this experiment should be useful for the further exploitation of the seaweed. Ulinastatin Suppresses Burn-Induced Lipid Peroxidation and Reduces Fluid Requirements in a Swine Model Objective. Lipid peroxidation plays a critical role in burn-induced plasma leakage, and ulinastatin has been reported to reduce lipid peroxidation in various models. This study aims to examine whether ulinastatin reduces fluid requirements through inhibition of lipid peroxidation in a swine burn model. Methods. Forty miniature swine were subjected to 40% TBSA burns and were randomly allocated to the following four groups: immediate lactated Ringer's resuscitation (ILR), immediate LR containing ulinastatin (ILR/ULI), delayed LR resuscitation (DLR), and delayed LR containing ulinastatin (DLR/ULI). Hemodynamic variables, net fluid accumulation, and plasma thiobarbituric acid reactive substances (TBARS) concentrations were measured. Heart, liver, lung, skeletal muscle, and ileum were harvested at 48 hours after burn for evaluation of TBARS concentrations, activities of antioxidant enzymes, and tissue water content. Results. Ulinastatin significantly reduced pulmonary vascular permeability index (PVPI) and extravascular lung water index (ELWI), net fluid accumulation, and water content of heart, lung, and ileum in both immediate or delayed resuscitation groups. Furthermore, ulinastatin infusion significantly reduced plasma and tissue concentrations of TBARS in both immediate or delayed resuscitation groups. Conclusions. These results indicate that ulinastatin can reduce fluid requirements through inhibition of lipid peroxidation. Introduction Burn injury is one of the most severe forms of injuries that evokes both strong physical and emotional responses, producing considerable morbidity and mortality. Data supplied by the National Center for Injury Prevention and Control indicate that in 2010 there were 412,256 (133.53/100,000) nonfatal fire/burn injuries and 3,194 (1.03/100,000) fire/burn deaths in the United States [1,2]. The cost of fire/burn injuries, including both medical costs and cost of lost productivity, is very high, costing a total of 7.546 billion dollars in 2000 [3]. Hypovolemic shock can develop rapidly after major burn injury, and current treatment for burn shock mainly focuses on maintaining a sufficient tissue perfusion with early, adequate fluid resuscitation. Currently, the most wildly used formula of fluid resuscitation in burn injury is the Parkland formula, which advocates providing 4 mL of Ringer's lactate/kg/%TBSA (total body surface area) burned/24 hours after burn, with one-half of the fluid expected to be infused over the first 8 hours and the remaining infused over the next 16 hours [4]. However, burn shock can develop and progress despite fluid resuscitation because much of the infused fluid leaks into the extravascular space, and sometimes extensive fluid resuscitation can exacerbate the interstitial edema, producing life-threatening complication such as abdominal compartment syndrome [5,6]. Furthermore, although an early, adequate fluid resuscitation is practically achievable under normal conditions, effective treatment is a challenging issue in mass casualty incidents caused by forces of nature or by accidental or intentional explosions and conflagrations, where the environmental conditions, the presence of mass casualties, and logistic constraints preclude the availability of intensive fluid resuscitation. Thus, the development of pharmacologic resuscitation strategies that could reduce the fluid requirements for burn injury would be beneficial both in civilian burns or in burn disasters. Previous studies have indicated that the burn-induced hypovolemic shock is mainly due to the increase in total body capillary permeability and the subsequent plasma leakage [4,7]. Previous studies also suggest that reactive oxygen species (ROS) contribute to the increased microvascular permeability, edema formation, and tissue damage after burn injury [8][9][10][11][12]. After major injury, the peripheral perfusion and oxygen are decreased; however, the restoration of oxygen delivery during aggressive fluid resuscitation will initiate a deleterious cascade of events that results in the burst of oxygen radicals and causes lipid peroxidation which influences numerous cellular functions [13]. Physiological functions of cell membranes change because lipid peroxidation modifies properties of the membrane such as membrane potential, fluidity, and permeability to different substance [14]. Increased plasma and tissue levels of lipid peroxidation products, such as malondialdehyde (MDA), have been well documented after thermal injury [13,[15][16][17][18][19][20]. In addition, the use of antioxidants has been found to be efficacious in reducing fluid requirements after burn injury [21][22][23][24][25][26][27]. Ulinastatin is a protease inhibitor obtained from human urine, and it has been reported to have free radicalscavenging properties in various models [28][29][30]. Our recent study also showed that ulinastatin attenuated vasopermeability both in vivo and in vitro [31]. The present study tested the hypothesis that ulinastatin would attenuate lipid peroxidation and tissue edema, thereby reducing fluid requirement after major burn injury in a swine model. Swine Burn Model and Treatment. Forty female, inbred Chinese Wuzhishan miniature swine (4-6 months, 20-25 kg, purchased from the Institute of Animal Sciences, Chinese Academy of Agricultural Sciences, Beijing, China) were used. They were acclimatized in the animal quarter of our laboratory for two weeks. All the animals were fasted for 16 h, and water was withheld for 4 h before the surgery. Under anesthesia with intramuscular injection of ketamine (Gu-Tian Pharmacy, Fu Jian Province, China), animals were then instrumented with a thermodilution catheter of PICCO-PLUS monitor (Pulsion Co., Germany) in the aorta for the measurement of hemodynamic variables, and a vascular catheter was positioned in the superior vena cava for drawing blood samples and fluid infusion. After surgery, all animals were monitored for 1 hour to assure hemodynamic stabilization, and the baseline data were then recorded. The animals were then infused with 5 mg/kg of propofol to assure adequate anesthesia and were subjected to a 40% TBSA fullthickness flam burn injury. A urinary catheter was inserted in the bladder and connected to a commercial urine collection bag. Animals were given buprenorphine (10 micrograms/kg i.v., Sigma, St. Louis, MO) immediately after burn injury and every 12 hours thereafter for sedation and pain control. The animals were kept in special slings for monitoring. All animal experiments were approved by the Committee of Scientific Research of First Hospital Affiliated to General Hospital of PLA, China, and were conducted in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals. The injured animals were randomly allocated to the following four groups: immediate resuscitation with lactated Ringer's (ILR), immediate resuscitation with lactated Ringer's containing ulinastatin (ILR-ULI), delayed resuscitation with lactated Ringer's (DLR); delayed resuscitation with lactated Ringer's containing ulinastatin (DLR-ULI), and with 10 pigs in each group. For ILR and ILR-ULI groups, each pig was infused with 4 mL/kg lactated Ringer's alone or lactated Ringer's containing ulinastatin (80,000 U/kg) over 30 minutes immediately after burn injury, followed by continuous infusion of lactated Ringer's, with the infusion rate adjusted each hour to maintain a urine output of 1-2 mg/kg/h. For DLR and DLR-ULI groups, each swine was initially infused with 4 mL/kg or lactated Ringer's containing ulinastatin (80,000 U/kg) over 30 minutes at 6 hours after burn, followed by continuous infusion of lactated Ringer's, with the infusion rate adjusted each hour to maintain a urine output of 1-2 mg/kg/h. 24 hours after burn injury, a second dose of 40,000 U/kg ulinastatin was given simultaneously with continuous lactated Ringer's in ILR-ULI and DLR-ULI groups. Net fluid accumulation was defined as cumulative infused fluid volume minus cumulative urine output and was measured hourly as previously described by Dubick et al. [24]. Blood samples were collected for determination of hematocrit and thiobarbituric acid reactive substances (TBARS) concentration. Hematocri was determined by the clinical laboratory in our hospital. TBARS were determined as an index of lipid peroxidation using a commercial kit (Nanjing Jiancheng Science and Technology Co., Ltd, Nanjing, China) according to manufacturer's instruction. At 48 hours after burn, pigs were euthanized, and heart, liver, lung, muscle and ileum were harvested for determining concentrations of TBARS, activities of antioxidant enzymes (catalase, copperzinc, and manganese superoxide dismutase) using commercial kits (all purchased from Nanjing Jiancheng Science and Technology Co., Ltd, Nanjing, China). Water content rate of the harvested tissues was measured in percentage of dry/wet ratio as previously described [31]. Statistical Analysis. SPSS 13.0 statistical software was used, and all results were expressed as mean ± SD. One-way ANOVA was used for comparison among all groups, followed by the Student-Newman-Keuls (SNK) test for comparison between two groups. Differences were considered to be statistically significant when < 0.05. Hemodynamic Variables. After thermal injury, no significant changes in MAP or CO in immediate resuscitation groups were observed (Table 1). MAP and CO were decreased in the delayed resuscitation groups before resuscitation is initiated. After resuscitation, the MAP and CO in both delayed resuscitation groups were gradually restored to near baseline levels; however, the CO was slightly higher in DLR-ULI group when compared to that of DLR group ( Table 1). The PVPI and ELWI were increased after thermal injury despite resuscitation; however, ulinastatin significantly attenuated the increase in PVPI and ELWI both in immediate or delayed resuscitation groups (Table 1). There were no significant changes in hematocrit in immediate resuscitation groups after thermal injury (Figure 1). In delayed resuscitation groups, hematocrit increased after burn injury due to loss of plasma volume, however, it was returned to near baseline levels after resuscitation, showing adequate restoration of plasma volume in both groups (Figure 1). Effect of Ulinastatin on Net Fluid Accumulation. After thermal injury, pigs were given either immediate or delayed resuscitation with adjusted infusion rate to maintain urine output between 1 and 2 mL/kg/h. Burn injury resulted in a decrease in urine

outputs in both delayed resuscitation groups before resuscitation; however, when resuscitation initiated, urine outputs were similar in all groups (Figure 2(a)). Resuscitation resulted in different extent of net fluid accumulation in all groups; however, ulinastatin significantly reduced net fluid accumulation both in immediate and delayed resuscitation groups (Figure 2(b)). Effect of Ulinastatin on TBARS Concentrations and Antioxidant Enzymes Activities. Burn insult and resuscitation resulted in an increase in plasma TBARS concentrations, which was more prominent in delayed resuscitated animals early after burn (Figure 3). Ulinastatin significantly prevented the increase in plasma TBARS both in immediate and delayed resuscitation groups (Figure 3). The TBARS concentrations in different organs, especially in lung and liver tissues, were lower in ulinastatin-treated animals, although some of them did not reach statistical significance (Figure 4). However, activities of antioxidant enzymes (superoxide dismutase and catalase) in heart, liver, lung, skeletal muscle, and ileum were similar in all groups (Table 2). Effect of Ulinastatin on Water Content in Different Organs. Ulinastatin significantly reduced the water content of heart, lung, and ileum in both immediate or delayed resuscitation groups ( Figure 5). However, there was no significant difference among all groups in the water content of liver and skeletal muscle tissues ( Figure 5). Discussion Hypovolemic shock is a key factor influencing the mortality rate during the early phase of major burn injury. Current Oxidative Medicine and Cellular Longevity 5 Figure 4: TBARS concentrations in heart, liver, lung, muscle, and ileum 48 hours after thermal injury in pigs. Data were expressed as mean ± SD. = 7-10 in each group. * ULI versus ILR; # DLR/ULI versus DLR; + ILR versus DLR; and $ ILR/ULI versus DLR/ULI at < 0.05. efforts to improve burn shock outcome mainly focus on early and adequate fluid resuscitation. Intensive fluid resuscitation, however, may exacerbate interstitial edema and cause serious adverse events, such as abdominal compartment syndrome. Furthermore, intensive fluid resuscitation from burn shock is usually difficult in austere environments (battlefield, forest conflagration, or earthquake) due to the environmental conditions, the presence of mass casualties, and logistic constraints. Thus, pharmacologic agents that could reduce the fluid requirements for burn injury by attenuating plasma leakage would benefit casualties both in civilian burns or in burn disasters. Emerging evidence suggests that lipid peroxidation after thermal injury plays a critical role in the increase in vascular permeability and the subsequent plasma leakage [4]. Ulinastatin is a protease inhibitor obtained from human urine, and it has been reported to reduce lipid peroxidation in various models [29,[32][33][34][35]. Our recent study also showed that ulinastatin attenuated vasopermeability both in vivo and in vitro [31]. However, it remains unknown whether ulinastatin treatment would reduce lipid peroxidation and fluid requirements in swine model of major burn injury. In this study we adopted a swine model of 40% TBSA burn injury to investigate the effects of ulinastatin on lipid peroxidation and fluid requirements. We have shown that in this swine burn model, ulinastatin treatment attenuates lipid peroxidation, tissue edema, and net fluid accumulation and thereby reducing fluid requirements. Ulinastatin is a relatively safe drug, and the dosages from 5,000 to 1,000,000 U/kg have been reported in different animal models [36][37][38]. In this study, we used a high dosage of ulinastatin (80,000 U/kg in the first 24 hours after burn and then another 40, 000 U/kg in the second 24 hours after burn) in order to obtain more significant protective effects. We first evaluated the effects of ulinastatin on fluid requirements in burnt pigs followed by immediate resuscitation or delayed resuscitation in an adjusted rate according to urine output. The hemodynamic response to burn injury and resuscitation was similar to previous reports [24,39,40]. There was no significant changes in MAP and CO in immediate resuscitation groups. Although there was a reduction in MAP and CO in delayed resuscitation groups after burn injury, they were gradually restored to near baseline levels when resuscitation initiated. Similarly, hematocrit increased after burn injury due to loss of plasma volume, however, it was returned to near baseline levels after resuscitation in delayed resuscitation groups. These results suggest that adequate restoration of plasma volume was achieved in both resuscitation strategies. An increase in PVPI and ELWI was observed after thermal injury despite resuscitation, however, ulinastatin significantly attenuated the burn-induced increase in PVPI and ELWI both in immediate or delayed resuscitation groups. This suggests that ulinastatin could attenuate burn-induced lung injury and edema which was supported by other studies [41,42]. PVPI and ELWI in DLR and DLR/ULI groups were significantly higher than those in ILR and ILR/ULI groups at 6 hours after injury when DLR and DLR/ULI groups were not given fluid resuscitation. It is possible that ELWI and PVPI are overestimated by PiCCO system because of hypovolemia which is one of the limitations of PiCCO system [43]. In this experiment, the urine output was maintained at 1-2 mL/kg/h by adjusting the infusion rate. Urine output was similar in animals treated with or without ulinastatin. However, ulinastatin significantly attenuated the net fluid accumulation in both immediate and delayed resuscitation groups. Furthermore, water content of heart, lung, and ileum was significantly reduced in ulinastatin-treated animals. These findings, together with previous studies by ours and others [31,42], indicate that ulinastatin is able to attenuate the burn-induced increase in vascular permeability and plasma volume loss and thereby reducing fluid requirements. Since free radical-induced lipid peroxidation is suggested to be implicated in burn-induced increase in vascular permeability and the subsequent plasma leakage [4], ulinastatin has been reported to reduce lipid peroxidation in various models, including burn models [29,[32][33][34][35]44]. Thus, we further investigated whether the protective effects of ulinastatin on burn-induced increase in vascular permeability and plasma volume loss were associated with reduced lipid peroxidation. We measured the plasma and tissue concentrations of TBARS as an index of lipid peroxidation. In consistent with previous study [44], we found that burn insult that resulted in an increase in lipid peroxidation and ulinastatin administration effectively attenuated the burn-induced lipid peroxidation. We further measured the antioxidant enzymes activities in heart, liver, lung, muscle, and ileum harvested at 48 hours after burn. However, in contrast to previous investigation by Shimazaki et al. [44], no significant difference in antioxidant enzymes activities was observed among all groups. Differences between Shimazaki's results and ours could be due to differences in animals, tissues, or time of tissue harvest, and further study is needed to confirm the effects of ulinastatin on antioxidant enzymes activities. Conclusions In summary, ulinastatin, a protease inhibitor, attenuates burn-induced increase in vascular permeability and net fluid accumulation and has a therapeutic role in reducing fluid requirements of thermal injuries. These protective effects of ulinastatin may be mediated in part through the inhibition of burn-induced lipid peroxidation. This study offers a potential small-volume fluid resuscitation strategy in combating major burn injury. Lenient rate control versus strict rate control for atrial fibrillation: a protocol for the Danish Atrial Fibrillation (DanAF) randomised clinical trial Introduction Atrial fibrillation is the most common heart arrhythmia with a prevalence of approximately 2% in the western world. Atrial fibrillation is associated with an increased risk of death and morbidity. In many patients, a rate control strategy is recommended. The optimal heart rate target is disputed despite the results of the the RAte Control Efficacy in permanent atrial fibrillation: a comparison between lenient vs strict rate control II (RACE II) trial. Our primary objective will be to investigate the effect of lenient rate control strategy (<110 beats per minute (bpm) at rest) compared with strict rate control strategy (<80 bpm at rest) on quality of life in patients with persistent or permanent atrial fibrillation. Methods and analysis We plan a two-group, superiority randomised clinical trial. 350 outpatients with persistent or permanent atrial fibrillation will be recruited from four hospitals, across three regions in Denmark. Participants will be randomised 1:1 to a lenient medical rate control strategy (<110 bpm at rest) or a strict medical rate control strategy (<80 bpm at rest). The recruitment phase is planned to be 2 years with 3 years of follow-up. Recruitment is expected to start in January 2021. The primary outcome will be quality of life using the Short Form-36 (SF-36) questionnaire (physical component score). Secondary outcomes will be days alive outside hospital, symptom control using the Atrial Fibrillation Effect on Quality of Life, quality of life using the SF-36 questionnaire (mental component score) and serious adverse events. The primary assessment time point for all outcomes will be 1 year after randomisation. Ethics and dissemination Ethics approval was obtained through the ethics committee in Region Zealand. The design and findings will be published in peer-reviewed journals as well as be made available on ClinicalTrials.gov. Trial registration number NCT04542785. INTRODUCTION Atrial fibrillation is the most common arrhythmia of the heart with a prevalence of approximately 2% in the western world. 1 2 Atrial fibrillation is associated with an increased risk of death and a number of morbidities. [3][4][5][6][7][8][9] The risks of both cerebral stroke and heart failure are increased nearly fivefold in patients with atrial fibrillation, and about 20% of all strokes may be due to atrial fibrillation. [3][4][5][6][7][8] Atrial fibrillation also has a significant impact on healthcare costs and accounts for approximately 1% of the National Health Service budget in the UK and approximately $26 dollars of annual expenses in the USA. 10 11 Two different overall intervention strategies may be used for atrial fibrillation: a rhythm control strategy or a rate control strategy. [12][13][14] We have previously shown in a systematic review with meta-analysis and trial sequential analysis that rhythm control strategies compared with rate control strategies seem to significantly increase the risk of serious adverse events in Strengths and limitations of this study ► First trial assessing lenient versus strict rate control in patients who upon inclusion are considered as having persistent atrial fibrillation. Hence, this trial is expected to provide data on patients who upon inclusion have a relatively short duration of atrial fibrillation. ► First superiority trial with quality of life as primary outcome in patients with both permanent atrial fibrillation and persistent atrial fibrillation on inclusion. ► Pragmatic trial with multiple sites ensuring high external validity. ► Treatment providers are not blinded in a trial that is otherwise expected to have low risk of bias regarding blinding of other domains. ► Trial will not have enough power to assess 'hard outcomes' such as mortality and serious adverse events. Open access patients with atrial fibrillation. 13 14 Based on current evidence as well as guidelines, it seems that most patients with atrial fibrillation should be treated with a rate control strategy unless there are specific reasons justifying a rhythm control strategy. 13 14 The resting heart rate target for rate control has recently changed from below 80 beats per minute (bpm) to below 100-110 bpm at rest depending on the guideline. 12 14 15 This change was a result of the the RAte Control Efficacy in permanent atrial fibrillation: a comparison between lenient vs strict rate control II (RACE II) trial, which randomised 614 participants to a lenient rate control strategy (<110 bpm at rest) versus a strict rate control strategy (<80 bpm at rest). 16 The participants were outpatients with permanent atrial fibrillation. The RACE II trial showed that the lenient rate control strategy was non-inferior compared with the strict rate control strategy on the risk of a composite outcome of mortality, stroke, cardiac arrest, arrhythmic events, systematic emboli or major bleeding. Furthermore, the HR of 0.84 (90% CI 0.58 to 1.21) suggested that the lenient rate control group might decrease the risk of the composite outcome. The RACE II trial also showed no difference of the two strategies on quality life, but this analysis has questionable validity. 17 A theoretical concern when using a lenient control strategy is that patients may develop heart failure if the heart rate is too fast. [18][19][20] The RACE II trial found that the lenient strategy was also non-inferior for heart failure patients but the majority of the participants had preserved EF at baseline. 21 We searched the Cochrane Central Register of Controlled Trials, MEDLINE and ClinicalTrials. gov on 26 September 2019. Our literature search identified only the RACE II trial assessing the effect of lenient rate control versus strict rate control in atrial fibrillation. We found no systematic reviews or

meta-analyses on the topic. Trial rationale Currently, lenient rate control is the guideline recommended initial rate control strategy. 14 However, this recommendation is primarily based on the RACE II trial, which had two major limitations. First, the validity of the RACE II trial results when assessing symptoms and quality of life were questionable mainly because of substantial problems with missing data. Regarding quality of life and symptom severity, only 437/614 (71%) participants had data available at maximum follow-up. 17 Furthermore, the authors did not use multiple imputation or other valid methods to handle the missing data. 22 Second, the RACE II trial only showed a lenient rate control strategy was non-inferior but could not answer if a lenient rate control strategy is superior to a strict rate control strategy. The RACE II trial was not adequately powered to confirm or reject minimal important differences between the two strategies. Conducting a superiority randomised clinical trial and afterwards performing a systematic review with meta-analysis will give us the possibility of confirming or rejecting that there is a difference in effect between the two strategies, at least on quality of life. Health-related quality of life as an outcome There are many definitions of health-related quality of life. 23 24 In general, quality of life questionnaires can be designed in two ways. 23 Generic questionnaires assess multiple domains applicable to a variety of health domains. 23 They more readily permit comparison across different disease and seem to have unquestionable patient relevance. 23 25 Generic quality of life scales are often criticised for being less sensitive to change than disease-specific quality of life scales, but when outcome results show no difference, it is most often unknown whether the lack of difference is caused by non-sensitive outcome scales or if the results demonstrate that there is no 'true' difference between the compared interventions when assessing 'generic' quality of life. 23 25 The opposite holds true for disease-specific questions, which in general are thought to be more responsive to change in the clinical condition than generic disease questionnaires but may be less patient relevant. The disease-specific questionnaires tend to focus more narrowly on the disease. Any increase in quality of life as a result of a treatment for a specific disease may be off set by unforeseen negative consequences of the treatment that the questionnaire by design will not capture. We will therefore supplement the general assessment using Short Form-36 (SF-36) with a disease-specific questionnaire. Currently, there seems to be no optimal questionnaire. 25 26 The Atrial Fibrillation Effect on Quality of Life (AFEQT) is a validated, disease-specific questionnaire, which aims to capture the objective and subjective burden of disease. 27 It contains 20 items that aim to assess four domains: symptoms, activities, treatment concern and treatment satisfaction. It also includes a summary score that summarises the first three domains. It assesses the burden of the atrial fibrillation symptoms. 27 28 When assessing quality of life, it is important to focus on a minimally important difference, which typically can be done using an anchor-based method or a distribution-based method, or a mix of the two. 29 30 To interpret the clinical significance of future trial results, we will carefully define minimal important differences for all primary and secondary outcomes (see 'Statistical plan and data analyses'). 31 Objectives Our primary objective will be to investigate the effect of a lenient rate control strategy (<110 bpm at rest) compared with a strict rate control strategy (<80 bpm at rest) on quality of life in patients with persistent or permanent atrial fibrillation. METHODS AND ANALYSIS Trial design The design of the Danish Atrial Fibrillation (DanAF) trial will be a randomised, two-group, superiority trial of lenient rate control versus strict rate control in patients with persistent or permanent atrial fibrillation at inclusion who accept rate control as the main strategy. Treatment providers responsible for the rate control treatment will not be blinded. Any other treatment providers (i.e. those Open access managing co-morbidities) will be attempted blinded as well as participants. Three hundred and fifty outpatients will be recruited from four university hospitals in Denmark: Holbaek University Hospital, Hvidovre University Hospital, Region Zealand University Hospital -Roskilde and Odense University Hospital. The present protocol follows the recommendation in the Standard Protocol Items: Recommendations for Interventional Trials guideline including all items from the WHO Trial Registration Data Set (online supplemental files 1 and 2). Trial conduct This trial will be conducted according to good clinical research practice and the latest Declaration of Helsinki. 32 33 Randomisation Participants will be randomised 1:1 to a lenient or a strict medical rate control strategy. The trial will use centralised randomisation at OPEN. Prior to the trial, a computer will generate randomisation sequences with varying block sizes between 6 and 10 that are unknown to the investigators. An internet-based randomisation system will be set up conducting randomisation stratified according to site, type of atrial fibrillation at inclusion (persistent vs permanent) and left ventricular ejection fraction (LVEF) (ejection fraction (EF) ≥40% and EF <40%). The randomising investigator will get access to the internet site through a personal password. The randomising investigator will not be an outcome assessor. Blinding The investigator prescribing the rate control medication (treatment provider) will not be blinded, as the treatment requires knowledge of the group the participant is randomised to. All other treatment providers, outcome assessors, data managers, statisticians and participants will be sought blinded (the participants will neither be informed of their rate control target nor their allocated intervention group). Blinded data will be sent to OPEN for blinded data management. Statistical analyses will be performed with the two intervention groups coded as 'A' and 'B' by two independent blinded statisticians. Two blinded conclusions will be drawn by the steering group: one assuming 'A' is the experimental group and 'B' is the control group-and one assuming the opposite. Based on these two blinded conclusions, two abstracts will be written (will be published as a supplement to the main publication). When the blinding is broken, the 'correct' abstract will be chosen, and the conclusions in this abstract will not be revised. As all medical procedures are available to any treatment provider, we cannot foresee any reason for unblinding participants. If, however, any medical personnel deem it necessary to unblind a participant, the participant will be unblinded. Other factors such as echocardiographic assessments, stability of the disease and similar will be factored in when judging if a participant is dependent on a high ventricular rate. Such a decision will be made before randomisation by the treatment provider. 5. Participants who are haemodynamically unstable and therefore require immediate electrical cardioversion. Participant withdrawal Participants can withdraw his or her consent at any time point for any reason but will be invited to still participate in the follow-up assessments. Lenient rate control The heart rate will be assessed on a 12-lead resting ECG measured over 1 min after 5 min of rest. The treatment provider will target the highest tolerable resting heart rate <110 bpm. Treatment providers are encouraged not to attempt to lower the heart rate if already below 110 unless symptoms or other reasons necessitates this. If the heart rate is below 90, the treatment provider is encouraged to reduce rate limiting treatment. If the patient remains symptomatic due to atrial fibrillation after achieving this definition of heart rate control, Holter monitoring or exercise tests may be deemed necessary by the treatment provider. These evaluations may be followed by adjustment of rate control drugs, rhythm control (electrical cardioversion, arrhythmia surgery and rhythm control medications) Open access or atrioventricular node ablation. In case of the need for rhythm control or atrioventricular node ablation, the allocated heart rate target is no longer relevant in management. Strict rate control Strict rate control achieved by using rate control medication (see further) will be defined as a mean resting heart rate <80 bpm with a general recommendation of targeting 70 bpm on a 12-lead resting ECG measured over 1 min after 5 min of rest. Exercise test to determine activity heart rates or Holter monitoring will only be performed if the treatment provider believes this is indicated. These evaluations may also be followed by adjustment of rate control medications, electrical cardioversion, arrhythmia surgery or atrioventricular node ablation (treatment provider's choice). Rate control medications Treatment will be provided according to current guidelines, and as such, the algorithm for treatment will be differentiated based on the status of left ventricular ejection fraction. 14 For participants with reduced LVEF, betablockers (metoprolol and bisoprolol) will be the primary therapy. Secondary therapies may include digoxin or amiodarone. For participants with preserved LVEF, the primary therapy will be beta-blockers (metoprolol and bisoprolol) or non-dihydropyridine calcium-channel blockers (verapamil) with secondary therapy consisting of digoxin or amiodarone. We briefly summarise the pharmacological treatment in the DanAF trial (table 1). Concomitant medication Besides rate control, the treatment provider will be free to prescribe any other standard medical cointervention such as the need for anticoagulation (based on the CHA 2 DS 2 -VASc score and comorbidity, 14 hypertension management, heart failure management or lipid lowering drugs as long as the prescriptions adhere to guidelines. 14 This also includes recommendations regarding modifiable risk factors that may have adverse effects on atrial fibrillation management (excess alcohol, smoking and sleep apnoea). 14 35 A brief description of what is considered standard management of comorbidities and risk factors are given in online supplemental file 3. All other interventions are allowed if they are administered evenly in all intervention arms. Follow-up and outcome events All participants will attend a minimum of two follow-up visits within 2 months after randomisation. Further visits are possible with 2-week intervals until adequate titration of rate control therapy is as required or for other reasons such as participants having inadequate symptom control, management of comorbidities and so on. Treatment providers may plan a visit sooner or later if clinically indicated. To assess if the ECG guided heart rate target is representative of the heart rate under normal conditions, we will perform 24-hour Holter monitoring at the end of the titration phase and after 1 year of follow-up for documentation purposes. After the initial adequate titration of rate control, participants are to follow the normal referral system in the Danish healthcare system. A hotline will be established where treatment providers may call and ask for the participant's rate control target. If treatment providers themselves do not contact the trial treatment provider, participants are encouraged to contact the trial treatment provider. If possible, a treatment provider involved in the trial will be the managing treatment provider of the referral, if the referral is to a participating department. Primary outcome ► Quality of life using the SF-36 questionnaire (physical component score), continuous outcome. 36 Secondary outcomes ► Days alive outside hospital, count outcome. ► Symptoms due to atrial fibrillation using the AFEQT, continuous outcome. 27 ► Quality of life using the SF-36 questionnaire (mental component score), continuous outcome. 36 ► Serious adverse events, dichotomous outcome. We will define a serious adverse event as any untoward medical occurrence that resulted in death, was lifethreatening, required hospitalisation or prolongation of existing hospitalisation and resulted in persistent or significant disability or jeopardised the patient. 33 Exploratory outcomes ► All-cause mortality, dichotomous outcome. ► Composite of all-cause mortality, stroke, myocardial infarction and cardiac arrest, dichotomous outcome. ► Cardiac mortality, dichotomous outcome. ► Stroke, dichotomous outcome. ► Hospitalisation for worsening of heart failure, dichotomous outcome. ► Number of hospital admissions, count outcome. ► Six-minute walking distance, continuous outcome. ► Healthcare costs. ► Various biomarkers (N-terminal pro-brain natriuretic peptide (nt-proBNP), high-sensitivity C reactive protein (hsCRP), high-sensitivity troponin I (hsTnI), growth differentiation factor-15 (GDF-15), interleukin 6 (IL6), cystatin-C, YKL40, soluble urokinase plasminogen activator receptor (suPAR) and fibulin-1). All secondary, exploratory and echocardiographic outcomes will only be hypothesis generating. Adverse events Participants will be asked during visits to the clinic if they had experienced any undesirable medical events. Suspected unexpected serious adverse reactions (SUSAR) will be reported to the ethics committee within 7 days of investigators being aware of the event. Once a year, a report of all serious adverse events and serious adverse reaction will be submitted to the ethics committee. Assessment time point The primary assessment time point for all outcomes will be 1 year after randomisation. Procedures for

screening Potential participants according to inclusion and exclusion criteria at Holbaek University Hospital, Hvidovre University Hospital, Region Zealand University Hospital -Roskilde and Odense University Hospital will receive an invitation to participate in the trial on a routine visit in the clinic or hospitalisation for atrial fibrillation. Possible participants will be identified by trial staff employed at the site. Procedures for informed consent Participants will receive printed material containing details of each study visit, the design and rational of the trial, participant rights (such as the right to withdraw), possible adverse reactions of medication and more. The printed material will be given either immediately after being identified as a possible candidate or during a private, information session where verbal information is given and the participants can ask any questions they may have. The information session will take place in an undisturbed environment. The information will be given by the project coordinator on site or medical personnel with equivalent prerequisites for conveying the project. Potential participants will be informed that they can bring a third party if they wish so. The participants will be given up to 3 weeks to consider participation depending on when they choose to schedule the information session. There will be a minimum of 48 hours from the information session to the obtaining of informed consent. Data collection Data will be attempted to be collected from all participants regardless of protocol adherence. Study plan and data will be as shown in table 2. Echocardiography will be performed according to current international guidelines. 38 A detailed plan for the echocardiographic examination and recordings has been developed. The echocardiograms will be sent to a core echocardiographic reading centre at Holbaek Hospital to be assessed by one of two assessors that will be blinded. Biobank We will collect blood samples for a research biobank and measure: Nt-proBNP, hsCRP, hsTnI, GDF-15, IL6, Cystatin-C, YKL40, suPAR and fibulin-1. In addition to the above blood samples, we will collect the following three types of blood samples: 5 mL serum, 5 mL plasma and 5 mL citrat plasma to be stored for future research. Participants will be given separate information on this blood collection as well as be required to give a separate informed consent (online supplemental file 4). Data management All data will be sent encrypted to OPEN for management. All data on paper will be securely stored, and a copy will be sent to a computerised database. The computerised database will be continuously checked for missing values and errors at 1-month intervals. Before a trial site begins recruitment, an internal monitoring of the following procedures will be checked: validation of inclusion and exclusion criteria, informed consent procedure, randomisation procedure and data entry into REDcap. Statistical plan and data analyses Sample size: quality of life using the SF-36 questionnaire (physical component score) Using a minimal important difference of 3 points on the physical component score, an SD of 10, power of 80% and a significance level of 5% and a total of 350 participants will be needed. 17 39 40 Based on this sample size, we have estimated the power of all remaining outcomes (see online supplemental file 5). Recruitment plans We will involve key medical personnel at the different departments as well as hold sessions at the different departments informing of the trial. Statistical analyses A detailed statistical analysis plan will be published around 1 month after the trial has been launched. In short, our primary conclusions will be based on the results of our single primary outcome. Hence, we will consider a p value of 0.05 as our threshold for statistical significance. 31 The results of secondary outcomes, exploratory outcomes, subgroup analyses and possible per protocol analyses will be hypothesis generating only. We will assess whether the thresholds for statistical and clinical significance are crossed according to the five-step procedure proposed by Jakobsen et al. 31 The analyses of the outcomes will be based on the 'intention to treat' principle, that is, all randomised participants will be included in the analysis regardless of how much treatment they have received. In case of more than 5% not receiving the allocated heart rate target, we will secondarily analyse all outcomes according to the actual heart rate achieved (per protocol analysis) defined as the average heart rate on ECG after 5 min of rest. Participants who receive a rhythm control strategy (assessed by the treating physician) at our primary assessment time point will be excluded from this analysis. If outcomes are not present due to retraction of informed consent or dropout, the pattern of the missing data will be investigated. Missing data will be handled according to the recommendations proposed by Jakobsen et al. 22 In short, we will conduct a worst-best and best-worst case scenario, testing the potential impact of missing data. 22 If the pattern of missing data allows it, we will also conduct multiple imputations. 22 Analysis methods Continuous outcomes will be presented as means and SD with 95% CIs. Count outcomes will be presented as medians and IQRs. We will analyse continuous outcomes using mixed effects linear regression with 'site' as a random intercept using an exchangeable covariance matrix and type of atrial fibrillation at inclusion (persistent vs permanent) and LVEF (EF ≥40% and EF <40%) as a fixed effect. 41 We will analyse count data using the van Elteren's test stratifying for 'site'. 42 Dichotomous outcomes will be presented as proportions of participants in each group with the event, as well as risk ratios with 95% CIs. Dichotomous outcomes will be analysed using mixed effects generalised linear models using a log link function with 'site' as a random intercept using an exchangeable covariance matrix, and type of atrial fibrillation will be included as a fixed effect. 42 All outcomes will be analysed according to final value. Subgroup analyses All subgroup analyses will be regarded as hypothesis generating only, and we will not base any conclusions on these. We will in the planned statistical analysis plan (see 'Statistical analysis') in detail describe each planned subgroup analysis. In short, we will in each publication compare: ► Patients with heart failure compared with patients without heart failure (including subtypes). ► Men compared with women. ► Different durations of atrial fibrillation at randomisation. -Less than 1 year. -1-2 years. -More than 2 years. ► Patients with age above compared with below 75 years. ► Patients according to the European Heart Rhythm Association symptoms score. Data monitoring A data safety monitoring committee (DSMC) independent from the sponsor and the investigators will be created. The DSMC will be free of conflicts of interest. The DSMC will be responsible for conducting an interim analysis after 50% of participants have been included and monitor if the trial still holds scientific merit. The DSMC will decide when/if a new interim analysis should be performed. The DSMC will make recommendations to the steering committee whether the trial should stop or continue (further details in online supplemental file 6). Auditing The trial can be audited by the regional ethics committee, which is independent from the investigators and sponsor. Patient and public involvement Patients were invited to a workshop after the initial draft was accepted by all participating departments. They were asked to give inputs to the chosen outcomes, the written material, the relevance of the objective of the trial and any other aspects they found relevant. Patients are anticipated to work as ambassadors after the trial results are available. We will therefore perform a second workshop to involve patients in the best strategy for dissemination. Ethics and dissemination The management of patients is in accordance with standard care, and as such, patients are at no greater risk compared with receiving standard care outside the trial. It is therefore ethical for patients to be part of the trial. The potential benefit for future patients is that we may uncover a superior heart target to be the goal of future management of patients with atrial fibrillation. The trial protocol has been approved by the regional ethics committee, which is a branch of the Danish ethics committee, the regulatory body approving research in Denmark. As such, the committees are independent from the trial. The committee reviewed the full protocol, the written material for the participants, the consent form and the administered questionnaires before giving approval. The ethics committee has the option of conducting an audit of the trial if it wishes to do so. The committee must be provided with a notification of any serious adverse events including Suspected unexpected serious adverse reactions within a week as well as a yearly report of serious adverse events. Any changes to the approved protocol will be submitted and approved before continuing the trial. Site investigators or personnel with equivalent skills will obtain informed consent from possible participants (online supplemental file 7). Additional consent will be obtained in order to store blood samples for future research. Before enrolment of participants, screening will be done by personnel employed at the study site using the local electronic journal system. Any information collected on potential and enrolled participants will be entered directly into REDcap, using a secure connection. The project and its data have been registered at the Region Zealand, who is the data controller. Study investigators will have access to the full data set. OPEN, who is in charge of storing the data, will also have access to the full data set. Ethics review will also have access to data on request. Participants, who suffer harm during the trial, are insured by the the Danish Patient Compensation Association. Trial results will be sought published in a peer-reviewed journal. We will also communicate results directly to relevant patient advocacy groups, relevant medical associations and attempted presented at relevant congresses. Aggregate data analysis will be published in a clinical trial register no later than 3 years after trial results have been collected. Data sharing will be made available on request after approval from ethics committee. Authorship will be granted according to the recommendations from the International Committee of Medical Journal Editors. 43 DISCUSSION Our trial has several strengths. It is a pragmatic trial assessing the benefits and harms of a lenient versus a strict rate control strategy on quality of life in patients with persistent or permanent atrial fibrillation. The number of inclusion and exclusion criteria is low, and hence, the external validity will be high. Participants will be recruited from more than one site, which will further increase the external validity. We have performed a sample size estimation based on previous evidence with realistic intervention effects, we will adjust the thresholds for statistical significance if the sample size is not reached, and we have chosen only one outcome we will base conclusion on. The remaining outcomes will be considered hypothesis generating only thereby taking into account problems with multiplicity. Furthermore, we have taken measures to reduce the risks of bias from the Open access allocation sequence generation, allocation concealment, blinding of outcome assessors and participants, selective outcome reporting, for-profit bias and missing outcome data. Hence, our trial will be conducted with a low risk of random errors ('play of chance') and with as low risk of systematic errors ('bias') as the trial design allows (see further). 31 44 In Denmark, a complete follow-up of all participants for death and hospitalisations is secured by an unique number given to all born in Denmark, Central Person Register. Our trial also has limitations. The treatment providers responsible for the rate control intervention will not be blinded, which may bias our results. We will use 12-lead ECG to guide rate control therapy. Holter monitoring and measurement of the heart rate during exercise will only be used at the discretion of the investigator if deemed necessary. As such, there may be fluctuations in the heart rate we do not detect. Another limitation is that we do not have sufficient power to assess 'hard outcomes' such as mortality and serious adverse events. This will be explored in a future meta-analysis with individual patient data from the RACE II trial and other trials. The consequence may ultimately be that a superiority trial in terms of 'hard outcomes' is needed. Our results will only be generalisable to a population where rate control is considered appropriate as the main strategy going forward. The results of the EAST

trial is expected to delay the initiation of rate control for many patients, and hence, our results will need to be interpreted in light of this. Yet another limitation is that participants presumably will receive different medications and procedures in the compared groups. If we show a difference (or lack of a difference) between the groups, it will be difficult to interpret what part of the treatment algorithm for reaching a certain rate target caused this difference. We expect the results of this trial will play a part of future recommendations for rate control treatment in patients with both persistent and permanent atrial fibrillation. Protocol version and amendments This abbreviated version of the full protocol is based on V.2.0 of the protocol (January 2020). Any changes to the original protocol will be submitted to the regional ethics committee. After approval, changes will be conveyed to all investigators, participants and trial registries. The findings will be published in a peer-reviewed journal as well as be made available on ClinicalTrials. gov. committee at Holbaek Hospital. We would also like to thank Lise Pedersen Patient consent for publication Not required. Provenance and peer review Not commissioned; externally peer reviewed. Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise. Description of trial design including type of trial (eg, parallel group, crossover, factorial, single group), allocation ratio, and framework (eg, superiority, equivalence, noninferiority, exploratory) 7 Methods: Participants, interventions, and outcomes Study setting 9 Description of study settings (eg, community clinic, academic hospital) and list of countries where data will be collected. Reference to where list of study sites can be obtained 7 Eligibility criteria 10 Inclusion and exclusion criteria for participants. If applicable, eligibility criteria for study centres and individuals who will perform the interventions (eg, surgeons, psychotherapists) 9-10 Interventions 11a Interventions for each group with sufficient detail to allow replication, including how and when they will be administered 10-12 11b Criteria for discontinuing or modifying allocated interventions for a given trial participant (eg, drug dose change in response to harms, participant request, or improving/worsening disease) 10 11c Strategies to improve adherence to intervention protocols, and any procedures for monitoring adherence (eg, drug tablet return, laboratory tests) 11d Relevant concomitant care and interventions that are permitted or prohibited during the trial 10-12 Outcomes 12 Primary, secondary, and other outcomes, including the specific measurement variable (eg, systolic blood pressure), analysis metric (eg, change from baseline, final value, time to event), method of aggregation (eg, median, proportion), and time point for each outcome. Explanation of the clinical relevance of chosen efficacy and harm outcomes is strongly recommended 13-15 Participant timeline 13 Time schedule of enrolment, interventions (including any run-ins and washouts), assessments, and visits for participants. A schematic diagram is highly recommended (see Figure) 16-18 Methods: Assignment of interventions (for controlled trials) Allocation: Sequence generation 16a Method of generating the allocation sequence (eg, computer-generated random numbers), and list of any factors for stratification. To reduce predictability of a random sequence, details of any planned restriction (eg, blocking) should be provided in a separate document that is unavailable to those who enrol participants or assign interventions 8 Allocation concealment mechanism 16b Mechanism of implementing the allocation sequence (eg, central telephone; sequentially numbered, opaque, sealed envelopes), describing any steps to conceal the sequence until interventions are assigned 8 Implementation 16c Who will generate the allocation sequence, who will enrol participants, and who will assign participants to interventions 8 Blinding (masking) 17a Who will be blinded after assignment to interventions (eg, trial participants, care providers, outcome assessors, data analysts), and how 8-9 17b If blinded, circumstances under which unblinding is permissible, and procedure for revealing a participant's allocated intervention during the trial 9 Data collection methods 18a Plans for assessment and collection of outcome, baseline, and other trial data, including any related processes to promote data quality (eg, duplicate measurements, training of assessors) and a description of study instruments (eg, questionnaires, laboratory tests) along with their reliability and validity, if known. Reference to where data collection forms can be found, if not in the protocol 18-19 18b Plans to promote participant retention and complete follow-up, including list of any outcome data to be collected for participants who discontinue or deviate from intervention protocols How personal information about potential and enrolled participants will be collected, shared, and maintained in order to protect confidentiality before, during, and after the trial 22-23 Declaration of interests 28 Financial and other competing interests for principal investigators for the overall trial and each study site 26 Access to data 29 Statement of who will have access to the final trial dataset, and disclosure of contractual agreements that limit such access for investigators 22-23 Ancillary and posttrial care 30 Provisions, if any, for ancillary and post-trial care, and for compensation to those who suffer harm from trial participation Management of heart failure and hypertension Management of heart failure will follow the recommendations of the European Society of Cardiology. Briefly, the table below summarizes the recommendations for medical therapy. Ultimately, any management is at the discretion of the treatment providers and participants. Sleep apnea Participants will be systematically screen for signs of sleep apnea. If signs and symptoms of sleep apnea are discovered, participants will be referred to treatment if appropriate. Obesity Weight loss will be encouraged if BMI > 25. General advice will be provided and involvement of participants in local municipal programs will be discussed. Smoking Participants will be asked about their smoking habits as part of the initial work-up. Participants will be informed of the detrimental effects of smoking on health. Current smokers will be encouraged to quit and will be informed of available support programs through the municipals. Alcohol Participants will be asked about their alcohol habits as part of the initial work-up. Participants will be informed of current evidence regarding alcohol in atrial fibrillation and will be encouraged to abstain from alcohol or alternatively reduce their alcohol intake. Special emphasis will be put on participants who drink above 10 standard drinks/week. 1 2 BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) BMJ Open doi: 10.1136/bmjopen-2020-044744 :e044744. Supplementary file 4 -biobank We will further collect blood samples for a research biobank and measure: Nt-proBNP, hsCRP, hsTni, GDF-15, IL6, Cystatin-C, YKL40, suPAR and Fibulin-1. Due to the manner of which these analysis have to be analysed and the variations in the measurement depending on blood sample kit is used, blood samples will be collected at the first visit, after 6 months, and at follow-up after 1 year and analysed together. Follow up after two and three years will be analysed together. These analyses will require 10 mL of blood per collection. The blood samples are expected to be analysed either at a laboratory in Sweden or a laboratory in Denmark, but may end up being analysed in another EU country. The storage of data will abide by the Danish General Data Protection Regulation and the Danish Data Protection Act in Denmark. Any spare blood that is collected will be stored in a biobank in Denmark for future unspecified research purposes. The storage of data will still abide by the Danish General Data Protection Regulation and the Danish Data Protection Act in Denmark. In addition to the above blood samples, we will collect three different types of blood samples: 7 ml. serum, 7 ml plasma and 7 ml citrat plasma to be stored for future research. This will total approximately 31 mL of blood. The blood samples are expected to be analysed either at a laboratory in Sweden or a laboratory in Denmark, but may end up being analysed in another EU country. Participants will be given separate information on this blood collection as well as be required to give a separate informed consent. The storage of data will abide by the Danish General Data Protection Regulation and the Danish Data Protection Act in Denmark. Supplementary file 5 -Power estimations of secondary outcomes The below power calculations are based on a sample size of 350 participants as specified in the main document. Days alive outside hospital Using a minimal important difference of 3 days, a standard deviation of 9, a risk of type I error of 5%, and accounting for the fact that the data is expected not to be normal distributed, we will be able to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) of 82.1%. 1 The Atrial Fibrillation Effect on Quality-of-Life (AFEQT) In previous trials the observed difference between groups was normally distributed with a standard deviation of 21. 2 3 Using a minimal important difference of 7, we will be able to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) of 87.5%. The Type I error probability associated with this test of this null hypothesis is 5%. Quality of life using the SF-36 questionnaire (mental component score) In previous trials the observed difference between groups was normally distributed with a standard deviation 10. 4-6 Using a minimal important difference of 4, we will be able to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) of 96%. The Type I error probability associated with this test of this null hypothesis is 5%. Serious adverse events We anticipate a failure rate among control of 20%. If we anticipate a relative risk reduction of 60%, we will be able to reject the null hypothesis with probability (power) of 90.2%. The Type I error probability associated with this test of this null hypothesis is 5%. All-cause mortality Prior data indicate that the mortality rate among controls is about 5%. 7 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 5.7%. The Type I error probability associated with this test of this null hypothesis is 5%. Composite of all-cause mortality, stroke, myocardial infarction and cardiac arrest Prior data indicate that this outcome occurs in controls in about 8%. 7 8 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 5.9%. The Type I error probability associated with this test of this null hypothesis is 5%. Cardiac mortality Prior data indicate that the failure rate among controls is 3.9%. 7 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 5.4%. The Type I error probability associated with this test of this null hypothesis is 5%. Stroke Prior data indicate that cardiac mortality among controls is 3.9%. 7 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 5.4%. The Type I error probability associated with this test of this null hypothesis is 5%. Hospitalisation for worsening of heart failure Prior data indicate that heart failure among controls is 27.4%. 7 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 9.0%. The Type I error probability associated with this test of this null hypothesis is 5%. Number of hospital admissions Prior data indicate that number of participant who are hospitalised is 27.4%. 7 If we anticipate a relative risk reduction of 10%, we will be able to reject the null hypothesis with probability (power) of 9%. The Type I error probability associated with this test of this null hypothesis is 5%. Six-minute

walking distance In previous trials the observed difference between groups was normally distributed with a standard deviation 75. 9-11 Using a minimal important difference of 40, we will be able to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) of 99.9%. The Type I error probability associated with this test of this null hypothesis is 5%. Physical activity using trial accelerometer Prior data indicates that the standard deviation among groups was 65 minutes pr. Day when measuring sedentary behaviour. Assuming a difference in groups of 20 minutes/day, we will be able to reject the null hypothesis with a probability of 81.9%. The type 1 error probability associated with this test of this null hypothesis is 5%. 12 13 Supplementary file 6. Short description of the independent Data Safety and Monitoring Committee (DSMC) Introduction This Charter defines the primary responsibilities for the independent Data safety and monitoring Committee (DSMC) of the randomised clinical trial DanAF. This includes the relationships with other aspects of the trial. Primary responsibility of the DSMC The DSMC will ensure the safety of trial participants. This will be achieved by the following tasks:  Performing planned analyses of outcomes related to the safety of participants from the two rate control strategies during the trial.  Continuously monitoring if the trial still holds scientific merit Members of the DSMC The exact composition of the DSMC will be specified later but is expected to consist of two clinicians and one person with adequate statistical knowledge to conduct the interim analysis. One member will be chosen as the committee chair. Recommendations are recommended to be anonymous. However, in case of members not coming to an agreement, members will vote. The points of discussion will be part of the discussion of the DSMC report to the Steering Committee (SC). The members of the DSMC will be free of conflicts of interest. Assessment if members are free of conflict of interest will be decided by the SC. Meetings This is the initial DSMC charter. The final charter will be determined and signed as the last part of the first meeting of the DSMC (see below). Meeting The first meeting will be a finalization of the DSMC role during the trial. The following will be agreed on and finalized.  How DSMC can request additional (unblinded) data  How meetings will be held (virtually, physical meeting, phone)  How many meetings are necessary.  Decision on whether a test run is necessary.  Finally, the charter will be finalised and signed. meeting The second meeting will take place as part of an interim analysis after 50% of the participants (n=175) have been recruited. The DSMC will be allowed to conduct additional interim analyses independently of the SC. The following meeting may take place virtually, in person or by phone. Communication Different formats will be used in order to secure proper communication is established. The formats include open and closed reports as well as open and closed sessions. Closed Sessions These sessions will involve only DSMC members. Discussions will be based on a closed report that will be based on blinded data provided by the data manager. A single member will be in charge of preparing the report but may receive input from the other two members before finalizing the closed report. If the DSMC deems it necessary, they may ask for unblinding of the data from the steering committee. Data for review will be the composite outcome all-cause mortality, stroke, myocardial infarction and cardiac arrest mortality (and its individual components), serious adverse events including any serious adverse reactions. Recommendations to the steering committee (open report) The DSMC will report its recommendations to the SC based on safety considerations. If the DSMC recommends anything other than continuing the trial, there will be held a virtual meeting between the DSMC and the SC. The DSMC will here present the reasoning behind its recommendations. The SC ultimately makes the decisions regarding all aspects of the trial. Data The DSMC will be provided with data on the following variables 1. Randomisation code (this will not reveal the allocated heart rate target) 2. The composite outcome of all-cause mortality, stroke, myocardial infarction and cardiac arrest and the individual components: a. All-cause mortality b. Stroke c. Myocardial infarction d. Cardiac arrest 3. Serious adverse events including subcategories of individual events 4. Numbers of participants lost to follow up The DSMC will not be provided with data on site or any identifier the data is considered anonymized. Analyses The DSMC is recommended to use Lan-DeMets sequential monitoring boundaries. Meta data The DSMC will be provided with a detailed codebook that explains all the coding in the data set. If during the research project significant information regarding your health, you will be informed. If you would like not to be informed of any new information regarding your health that comes to our attention during the trial, we ask that you mark here: __________ (mark with an x) Do you wish to be informed of the results of the trial and possible consequences for you?: Yes _____ (mark with an x) No _____ (mark with an x) Statement from the person providing information to the participant: I declare that the participant has received written and verbal information about the trial. To my knowledge there has been given enough information to make a decision to participate in the trial. Printed name of the person, who has given the information: Date: _______________ Signature: ____________________________________________ Regional ethics commitee project identification: Securing that the GDPR is complied with (by interaction with the Regional data controller) Site investigators Joshua Buron Feinberg (Holbaek University Hospital), Axel Brandes (Odense University Hospital), Ulrik Dixen (Hvidovre University Hospital) and Ole Dyg Pedersen (Region of Zealand University Hospital -Roskilde) Responsible for the proper conduct at respective sites. In charge of reporting Serious adverse events (SAE) including Suspected unexpected serious adverse reactions (SUSAR) to PI in a timely manner as well as reporting serious adverse events for annual review by the regional ethics committee. Steering committee (SC) All authors of the protocol will be invited to be part of the steering committee. Agreement of final protocolReviewing progress of study and if necessary agreeing changes to the protocol. In charge of reviewing proper conduct of the trial according to GCP, Helsinki-declaration and ethics review demands. Providing advice to lead investigators and personnel. Data manager Maintenance of trial IT system and data entry (OPEN). Data verification (OPEN in collaboration with PI) Providing data to the DSMC Providing data to the blinded statistician Outcome adjudication committee Responsible for adjudicating serious adverse events. Data safety monitoring committee Responsible for the safety of trial participants and the continuous scientific merit for the trial. Will report findings to the SC. Blinded statistician Prepare analysis for the steering committee to review Regional data controller (independent from trial) Data controller for the study hence must keep record of the type of data kept, data processor agreements and any other requirements needed to comply with GDPR Regional ethics committee (independent from trial) Approve the trial by review of protocol, written participant material, informed consent forms, etc. Monitor trial through reports of SAE and SUSAR reported to them by the daily management team as well as the yearly report submitted by the PI. Comparison of Saccharina japonica–Undaria pinnatifida Mixture and Minoxidil on Hair Growth Promoting Effect in Mice Background Algae have traditionally been used for promotion of hair growth. Use of hair regrowth drugs, such as minoxidil, is limited due to side effects. The aim of this study was to examine a mixture of Saccharina japonica and Undaria pinnatifida (L-U mixture) on hair growth and to compare the promoting effect of hair growth by a 3% minoxidil and a L-U mixture. Methods To evaluate the hair growth-promoting activity, saline, 50% ethanol, 3% minoxidil, and the L-U mixture were applied 2 times a day for a total of 14 days on the dorsal skin of C57BL/6 mice after depilation. Analysis was determined by using a high-resolution hair analysis system, real-time polymerase chain reaction, and H&E staining. Results On day 14, the hair growth effect of the L-U mixture was the same as that of the 3% minoxidil treatment. The L-U mixture significantly (P<0.05) stimulated hair growth-promoting genes, as vascular endothelial growth factor (VEGF) and insulin-like growth factor -1. Increase of VEGF was observed in the L-U mixture group compared with minoxidil and the negative control. In contrast, the L-U mixture suppressed the expression of transforming growth factor-β1, which is the hair loss-related gene. In histological examination in the L-U mixture and minoxidil groups, the induction of an anagen stage of hair follicles was faster than that of control groups. Conclusions This study provides evidence that the L-U mixture can promote hair growth in mice, similar to the effect from minoxidil, and suggests that there is potential application for hair loss treatments. INTRODUCTION The use of Food and Drug Adminstration (FDA) approved hair regrowth drugs such as finasteride (Propecia, Merck Sharp & Dohme Corp., Rahway, NJ, USA) and minoxidil (Rogain, Phar-macia and Upjohn, Kalamazoo, MI, USA), is limited and transient due to side effects such as irregular heartbeat or weight gain [1][2][3]. Hair growth is a complex interaction mechanism between epithelial cells and dermal cells of hair follicles. The hair follicle is remodeled during cyclical periods of growth (anagen), regression (catagen), and rest (talogen) [4,5]. The dermal papilla cell, epithelial cell, and enclosed hair follicle each release growth factors or inhibition factors during a cycle of hair growth state or degeneration state. Basic fibroblast growth factor (b-FGF), insulin-like growth factor-1 (IGF-1), keratinocyte growth factor, and vascular endothelia growth factor (VEGF) are well-studied hair growth factors. The epithelial growth factor (EGF) and transforming growth factor-β (TGF-β) are inhibition factors [6,7]. Companies are developing hair growth materials and hair growth drugs using natural resources. Algae have traditionally been used for promotion of hair growth. Studies have focused on understanding the relationship between hair growth and algae for potential therapeutic applications. For example, anagen progression of the hair shaft was induced after topical application of Ishige sinicola extract on the back of mice, [8]. Grateloupia elliptica, the red algae, has been reported to prevent hair loss in vitro [9]. The brown algae, Ecklonia cava contains dioxinodehydroeckol, which promotes hair growth through the stimulation of dermal papilla cells and outer root sheath cells [10]. Laminaria japonica and Undaria pinnatifida, also brown algae, have significant anti-inflammatory activities without serious toxic effects at moderate doses. These anti-inflammatory activities are related to treatment of hair loss, because scalp inflammation is one of the causes of hair loss [11]. Single taxon studies about the effects of algal treatment on hair growth have been published, but the hair growth effect of a single algal taxon does not yet surpass the therapeutic effect of minoxidil. We investigated the hair-growth effect on C57BL/6 mice using a combined mixture of the brown algae L. japonica and U. pinnatifida. If L-U mixture increases hair regrowth more than existing hair growth solution or single-alga extract, this study can have a positive impact on future hair-growth solution development. Materials L-U mixture (the extract of L. japonica and U. pinnatifida) was supplied by JW Bio Co. (Daegu, Korea), and 3% minoxidil was purchased from Nanopharm Corp. (Seoul, Korea) Saline and 50% EtOH are used for respectively negative control and solvent control. Experiment animals Five-week-old male C57BL/6 mice with black hair were used in this experiment. In the animal laboratory, temperatures were maintained at 23°C ± 3°C, relative humidity was maintained at 50% ± 10%, and day and night lengths were each maintained for a period of 12 hours. Mice were reared in isolation in mice breeding cages and allowed free access to food. The mice were divided into four groups (normal saline treated group, 50% ethanol treated group, 3% minoxidil treated group, L-U mixture treated group), and each group consisted of four mice (a total of 16 mice). Hair growth solution was applied topically to the shaved skin area; sample solutions (150 μL) were applied daily for a total of 14 days. All experimental protocols were conducted in accordance with guidelines for care and use of laboratory animals approved by the Institutional Animal Care and Use Committee of

Korea (2012-0216-CU-AEC-01-Y). Hair analysis using a high-resolution hair analysis system A high resolution hair analysis system (Professional Scientific Instrument Co., Ltd., Yongin, Korea) was used for the analysis of hair. Hair analysis was performed on days 1, 4, 7, 10, and 14 of the experiment and measured the density of hair, hair follicle lengths, and hair thicknesses of the treated areas of the six mice in each of the four groups. Hair densities were measured by counting the number of hairs per 0.4 mm 2 , and hair follicle lengths and hair thicknesses were measured from 20 hairs collected from the treated area of each mouse using pincers. Measured values were averaged by treatment group for each day (i.e., 1, 4, 7, 10, and 14 days) for analysis. Histological analysis On day 14 of the experiment, following anesthetization, the skin in the treated areas where samples were applied were surgically removed using a no. 15 scalpel blade, fixed immediately using 10% formalin, dehydrated using alcohol and xylene, embedded in paraffin, and sectioned into width of 5 μm, followed by removal of paraffin using alcohol and xylene. Slices were stained with hematoxylin and eosin (H&E), followed by observation using a light microscope. Sixteen tissue samples were used for staining (taken from 4 mice in the 4 groups, for a total of 16 mice). Real-time polymerase chain reaction analysis Real-time polymerase chain reaction (RT-PCR) was used to analyze the expression of mRNA for IGF-1, transforming growth factor TGF-β1, and glyceraldehyde-3-phosphate dehydrogenase (internal control). Skin was taken from each mouse in four groups (a total of 16 mice) for RT-PCR and histological samples for the same time. Real-time PCR was performed three times per skin sample. Total RNA was extracted from skin samples with TRIzol reagent (Takara Bio Inc., Kusatsu, Japan) and RT-PCR was performed using an oligo(dT) RT premix kit (iN-tRON Biotechnology, Daejeon, Korea). Real-time PCR was performed with a MJ Mini Thermal Cycler real time system (Bio-Rad, Hercules, CA, USA) using SYBR Premix Ex Taq (Takara Bio Inc.). The thermal cycle reaction was performed at 95°C for 30 seconds, followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds. The dissociation stage was initiated at 95°C for 15 seconds, followed by 1 cycle of 60°C for 30 seconds and 95°C for 15 seconds. Gas chromatography and mass spectrometry analysis Extract concentrate (8.4 g) from the L-U mixture was obtained by vacuum evaporation method and dissolved in 200 mL H2O. And then 200 mL methylene chloride (CH2Cl2) was added to this aqueous solution, which was then left to separate the CH2Cl2 layer and water layer (separation performed 3 times). The yield (0.2182 g) was collected by vacuum evaporation. The partite CH2Cl2 solvent was extracted from the L-U mixture, and gas chromatography with mass spectrometry (GC-MS) was used for analysis. The gas chromatography system (GC; Agilent Technologies, PA, USA) analysis was done using a DB-5 capillary column ( J&W Science Inc., Folsom, CA, USA), 30 m length × 0.25 mm id × 0.25 μm film. Helium was used as the carrier gas at a flow rate of 1 mL/min. The injection and detector temperatures were 280°C, and the oven temperature was programmed initial 40°C for 2 minutes, raised to 250°C at a rate of 10°C/min, and held at that temperature for 10 minutes. MS ( JEOL GC MateII mass spectrometer, JEOL Ltd., Tokyo, Japan) analysis was done under electron impact ionization at 70 eV of electron energy with a range from 10 to 500 at a rate of 1 scan/sec. Ion source temperature was 250°C. The chemical structure of each constituent was identified by searching and comparing mass data with the National Institute of Standards and Technology research library. Statistical analysis All statistical analyses were performed using the SPSS software package for Windows ver. 18.0 (SPSS Inc., Chicago, IL, USA). The Kruskal-Wallis test was used as the nonparametric method to compare each score by group. Multiple comparison was performed by the Dunn procedure. All data were presented by boxplot with median, first quartile, third quartile, upper fence, and lower fence. Statistical significance level was set at P < 0.05 and all data analysis was performed by a medical statistician. Effect of L-U mixture on hair regrowth The saline, 50% ethanol, L-U mixture, and 3% minoxidil were administered to C57BL/6 mice for two weeks (14 days) after depilation. Pink-colored skin was observed on day 4 of the experiment in all groups. But on day 7 of the experiment, in the L-U mixture group and 3% minoxidil group, a light-black skin color began to appear. On day 10, hair density was greater compared to the normal saline and 50% ethanol groups (Fig. 1B). On day 14, the depilated areas were almost covered with fullygrown hair in the L-U mixture group (Fig. 1A). Hair growth scores of the L-U mixture and 3% minoxidil groups were significantly higher compared with the normal saline group (L-U mixture, P = 0.008; 3% minoxidil, P = 0.034) (Fig. 1C). Results of hair analysis performed using a highresolution hair analysis system On day 14 of the experiment, hair width and length of follicle were measured in all groups using a hair analysis system. Increase in hair width was observed in the L-U mixture and 3% minoxidil groups, compared to the normal saline and 50% ethanol groups (Fig. 2A). The width of follicles on day 14 in the L-U mixture group was significantly increased, compared to the normal saline group (P = 0.020) (Fig. 2B). The length of follicles on day 14 was increased in the L-U mixture and 3% minoxidil groups, compared to the normal saline and 50% ethanol groups. The length of the follicles of the L-U mixture group were similar to those of the 3% minoxidil group, although not significantly (P = 0.53) (Fig. 2C). Histological characteristics Staining with H&E showed that the roots of hair follicles of the control groups (negative control: saline, solvent control: 50% ethanol) remained mostly small and round, while the roots of hair follicles treated with the L-U mixture and 3% minoxidil showed more growth. The numbers of basal cells of the hair root of the L-U mixture and 3% minoxidil treated groups increased compared to the control groups; in addition, hair follicles treated with L-U mixture and 3% minoxidil displayed large hair bulbs and well-differentiated hair shafts. In contrast, only a few well-formed hair follicles were found in the groups treated with saline and 50% ethanol (Fig. 3). Molecular biological analysis Groups treated with the L-U mixture and 3% minoxidil showed significantly increased gene expression of IGF-1, compared to the control groups (P < 0.05). The 3% minoxidil group showed C57BL/6 mice were treated with normal saline, 50% ethanol, the extract of Laminaria japonica and Undaria pinnatifida (L-U mixture), and 3% minoxidil for a period of 2 weeks. Photometric observation after 1, 4, 7, 10, and 14 days (A) and comparison of hair density using a high resolution hair analysis system ( × 300) after topical application of normal saline, 50% ethanol, L-U mixture and 3% minoxidil at 10 and 14 days (B). (C) Hair growth was evaluated by the scoring index: 0 (0%), 1 (1%-19%), 2 (20%-39%), 3 (40%-59%), 4 (60%-79%), and 5 (80%-100%). The Kruskal-Wallis test was used to compare the value of hair growth score. The result was statistically significant (P = 0.007). The multiple comparison was performed by Dunn procedure. a) P-value, significantly different than the normal saline group; b) P-value, significantly different than the 50% EtOH group. a greater increase of gene expression of IGF-1 than that of the L-U mixture treated group (Fig. 4A). Decreased gene expression of TGF-β1 was observed in the group treated with L-U mixture, compared to the normal saline group (Fig. 4B). L-U mixture and 3% minoxidil treated groups showed increased gene expression of VEGF compared to the normal saline group (Fig. 4C). GC-MS analysis The CH2Cl2 fraction had 18 components. Nonacosan-10-ol and β-sitosterol were major components; (-)-loliolide was the third component and has been identified in marine algae (red, brown, and green algae) and also in U. pinnatifida (Table 1). DISCUSSION Treatment of hair loss can be divided into two groups: surgical treatments, such as hair implants, and medical treatments, such as the FDA approved drugs minoxidil and finasteride [3]. Although surgical treatment guarantees optimal improvement (A) Picture of collected hair follicles ( × 300). (B) Analysis of hair follicles, measured at the greatest width. The Kruskal-Wallis test was used to compare the value of hair growth score. The result was statistically significant (P = 0.040). The multiple comparison was performed by the Dunn procedure. (C) Analysis of hair follicles, measured at the greatest length. L-U mixture, the extract of Laminaria japonica and Undaria pinnatifida. a) P-value, significantly different than the normal saline group. than medical treatment, surgical treatments might have some obstacles, such as the time or the cost of the treatment. On the other hand, medical treatments can save not only the time used in hair loss treatment, but could also provide convenience to the patients, thus enabling most patients to treat their hair loss with hair growth solution [12]. Studies have reported hair growth promotion effects of natural materials, including plants [13][14][15]. In addition, the use of algae in medical applications, is increasing. Marine organisms are known to be rich sources of biologically active substances [16]. Study of Gloiopeltis furcata has demonstrated a therapeutic effect on hepatoma cancer; treatment with U. pinnatifida extract reduced edema and immune reaction [17][18][19]. The study of fucoidan from brown algae has shown it is as effective as dexamethasone in treatment of atopic dermatitis symptoms in mice [20]. In a hair loss prevention study, I. sinicola extract induced significant inhibition of the activity of 5α-reductase [8]. G. elliptica extract has potential in the treatment of alopecia via the proliferation of dermal papilla, 5α-reductase inhibition, increase of prostaglandin E2 production, decrease of lipopolysaccharidestimulated pro-inflammatory cytokines, and inhibitory activity against Pityrosporum ovale ( = Malassezia furfur) [9]. In this study, the L-U mixture was analyzed by GC-MS (Table 1). The CH2Cl2 fraction had 18 components and major components were nonacosan-10-ol and β-sitosterol. A third component, (-)-loliolide, was determined in marine algae including red, brown and green algae and also including U. pinnatifida [21]. (-)-Loliolide has been reported to protect cells against H2O2induced cell damage or apoptosis [22]. We performed histological and molecular biological analyses (A) Insulin like growth factor type 1 (IGF-1) mRNA expression. (B) Transforming growth factor (TGF-β1) mRNA expression. (C) Vascular endothelial growth factor (VEGF) mRNA expression. Glyceraldehyde-3-phosphate dehydrogenase was used as a comparative control. The Kruskal-Wallis test was used to compare the value of hair growth score. The result was statistically significant (A, P = 0.007; B, P = 0.008; C, P = 0.009). The multiple comparison was performed by Dunn procedure. a) P-value, significantly different than the normal saline group; b) P-value, significantly different than the 50% EtOH group. for detection of hair regrowth effects. TGF-β regulates the growth, apoptosis, and specialization of cells. TGF-β includes TGF-β1, TGF-β2, and TGF-β3, all of which suppress keratinocyte proliferation and take part in extracellular matrix protein degradation by fibroblasts. A recent study reported that TGF-β1 plays a significant role accelerating hair loss. TGF-β1 hinders growth of anagen phase hair, causing hair to enter catagen phase, thereby inducing alopecia [7]. Synthesis of TGF-β2 was stimulated by dihydrotestosterone in dermal papilla cells [23]. In our study, the expression of TGF-β1 for the topical treatment of the L-U mixture was less than 3% minoxidil, which means the L-U mixture has potential for hair loss prevention. Growth factors such as IGF-1, EGF, and FGF promote hair growth and prevent involution of hair cells, playing an important role in hair follicle cycling and development, as well as specifying cell proliferation [24]. Our results showed increased expression of IGF-1 in L-U mixture treated mice, and increased expression of VEGF in 3% minoxidil treated mice. The RT-PCR data showed the same pattern of increase or decrease with both L-U mixture and minoxidil. Mice treated with the L-U mixture showed an increase of hair regrowth, follicle size, and number. Results from H&E staining provided evidence to support promotion of hair growth by treatment with the L-U mixture and 3% minoxidil. The L-U mixture had a hair growth

effect similar to that of minoxidil. Based on these results, we suggest that the L-U mixture may promote hair growth and prevent involution of hair cells. Discovery of a rich gene pool of bat SARS-related coronaviruses provides new insights into the origin of SARS coronavirus A large number of SARS-related coronaviruses (SARSr-CoV) have been detected in horseshoe bats since 2005 in different areas of China. However, these bat SARSr-CoVs show sequence differences from SARS coronavirus (SARS-CoV) in different genes (S, ORF8, ORF3, etc) and are considered unlikely to represent the direct progenitor of SARS-CoV. Herein, we report the findings of our 5-year surveillance of SARSr-CoVs in a cave inhabited by multiple species of horseshoe bats in Yunnan Province, China. The full-length genomes of 11 newly discovered SARSr-CoV strains, together with our previous findings, reveals that the SARSr-CoVs circulating in this single location are highly diverse in the S gene, ORF3 and ORF8. Importantly, strains with high genetic similarity to SARS-CoV in the hypervariable N-terminal domain (NTD) and receptor-binding domain (RBD) of the S1 gene, the ORF3 and ORF8 region, respectively, were all discovered in this cave. In addition, we report the first discovery of bat SARSr-CoVs highly similar to human SARS-CoV in ORF3b and in the split ORF8a and 8b. Moreover, SARSr-CoV strains from this cave were more closely related to SARS-CoV in the non-structural protein genes ORF1a and 1b compared with those detected elsewhere. Recombination analysis shows evidence of frequent recombination events within the S gene and around the ORF8 between these SARSr-CoVs. We hypothesize that the direct progenitor of SARS-CoV may have originated after sequential recombination events between the precursors of these SARSr-CoVs. Cell entry studies demonstrated that three newly identified SARSr-CoVs with different S protein sequences are all able to use human ACE2 as the receptor, further exhibiting the close relationship between strains in this cave and SARS-CoV. This work provides new insights into the origin and evolution of SARS-CoV and highlights the necessity of preparedness for future emergence of SARS-like diseases. Introduction Severe Acute Respiratory Syndrome (SARS) is a severe emerging viral disease with high fatality characterized by fever, headache and severe respiratory symptoms including cough, dyspnea and pneumonia [1]. Due to its high transmissibility among humans, after its first emergence in southern China in late 2002, it rapidly led to a global pandemic in 2003 and was marked as one of the most significant public health threats in the 21 st century [2,3]. The causative agent, SARS coronavirus (SARS-CoV), has been previously assigned to group 2b CoV and is now a member of the lineage B of genus Betacoronavirus in the family Coronaviridae [4]. It shares similar genome organization with other coronaviruses, but exhibits a unique genomic structure which includes a number of specific accessory genes, including ORF3a, 3b, ORF6, ORF7a, 7b, ORF8a, 8b and 9b [5,6]. Masked palm civets (Paguma larvata) were initially hypothesized to be the animal origin of SARS-CoV [7,8]. However, since a large number of genetically diverse SARS-related coronaviruses (SARSr-CoV) have been detected in multiple species of horseshoe bats (genus Rhinolophus) from different areas of China and Europe in the aftermath of SARS, it is prevailingly considered that SARS-CoV originated in horseshoe bats with civets acting as the intermediate amplifying and transmitting host [9][10][11][12][13][14][15][16]. Recently we have reported four novel SARSr-CoVs from Chinese horseshoe bats that shared much higher genomic sequence similarity to the epidemic strains, particularly in their S gene, of which two strains (termed WIV1 and WIV16) have been successfully cultured in vitro [17,18]. These newly identified SARSr-CoVs have been demonstrated to use the same cellular receptor (angiotensin converting enzyme-2 [ACE-2]) as SARS-CoV does and replicate efficiently in primary human airway cells [17][18][19]. Despite the cumulative evidence for the emergence of SARS-CoV from bats, all bat SARSr-CoVs described so far are clearly distinct from SARS-CoV in the S gene and/or one or more accessory genes such as ORF3 and ORF8, suggesting they are likely not the direct ancestor of SARS-CoV. Thus a critical gap remains in our understanding of how and where SARS-CoV originated from bat reservoirs. Previously, we reported a number of bat SARSr-CoVs with diverse S protein sequences from a single cave in Yunnan Province, including the four strains mentioned above most closely related to SARS-CoV [17,18]. Here we report the latest results of our 5-year longitudinal surveillance of bat SARSr-CoVs in this single location and systematic evolutionary analysis using full-length genome sequences of 15 SARSr-CoV strains (11 novel ones and 4 from previous studies). Efficiency of human ACE2 usage and the functions of accessory genes ORF8 and 8a were also evaluated for some of the newly identified strains. Continued circulation of diverse SARSr-CoVs in bats from a single location We have carried out a five-year longitudinal surveillance (April 2011 to October 2015) on SARSr-CoVs in bats from a single habitat in proximity to Kunming city, Yunnan province, China, which was mainly inhabited by horseshoe bats. A total of 602 alimentary specimens (anal swabs or feces) were collected and tested for the presence of CoVs by a Pan-CoV RT-PCR targeting the 440-nt RdRp fragment that is conserved among all known α-and β-CoVs [20]. In total, 84 samples tested positive for CoVs. Sequencing of the PCR amplicons revealed the presence of SARSr-CoVs in the majority (64/84) of the CoV-positive samples ( Table 1). Host species identification by amplification of either Cytb or ND1 gene suggested that most (57/64) of the SARSr-CoV positive samples were from Rhinolophus sinicus, while the remaining 7 samples were from Rhinolophus ferrumequinum, Rhinolophus affinis and from Aselliscus stoliczkanus which belongs to the family Hipposideridae. Based on the preliminary analysis of the partial RdRp sequences, all of the 64 bat SARSr-CoV sequences showed high similarity among themselves and with other reported bat SARSr-CoVs and SARS-CoVs from humans and civets. To understand the genetic diversity of these bat SARSr-CoVs, the most variable region of the SARSr-CoV S gene, corresponding to the receptor-binding domain (RBD) of SARS-CoV, were amplified and sequenced. Due to low viral load in some samples, RBD sequences were successfully amplified only from 49 samples. These RBD sequences displayed high genetic diversity and could be divided into two large clades, both of which included multiple genotypes. Clade 1 strains shared an identical size and higher amino acid (aa) sequence identity with SARS-CoV RBD, while clade 2 had a shorter size than SARS-CoV S due to two deletions (5 and 12-13 aa, respectively) (S1 Fig). Co-infections by two strains of different clades were detected in two samples, Rs3262 and Rs4087 (S1 Fig). [9,10,12,14,21,22]. The genome sequence similarity among the 15 SARSr-CoVs and SARS-CoV SZ3 strain was examined by Simplot analysis (Fig 1). The 15 SARSr-CoVs are In contrast, a considerable genetic diversity is shown in the S gene (corresponding to SZ3 genome position 21477 to 25244) and ORF8 (corresponding to SZ3 genome position 27764 to 28132) (Fig 1). The 11 novel SARSr-CoVs identified from this single location generally shared similar genome organization with SARS-CoV and other bat SARSr-CoVs. In our previous study, we identified an additional ORF termed ORFx present between ORF6 and ORF7 in strain WIV1 and WIV16 [18,23]. In this study, ORFx was also found in the genomes of Rs7327 and Rs4874. Compared with that of WIV1 and WIV16, the length of ORFx in Rs7327 and Rs4874 was extended to 510 nt due to a deletion of 2 nt in a poly-T sequence that resulted in a shift of reading frame (Fig 2 and S2 Fig). Co-circulation of different bat SARSr-CoVs with S, ORF8 and ORF3 sequences similar to those in SARS-CoV at a single location The primary difference between SARS-CoV and most bat SARSr-CoVs is located in S gene. The S protein is functionally divided into two subunits, denoted S1 and S2, which is responsible for receptor binding and cellular membrane fusion, respectively. S1 consists of two domains, the N-terminal domain (NTD) and C-terminal domain (CTD) which is also known as the RBD in SARS-CoV [24]. SARS-CoV and bat SARSr-CoVs share high sequence identity in the S2 region in contrast to the S1 region. Among the 15 SARSr-CoVs identified from bats in the surveyed cave, six strains with deletions in their RBD regions (Rs4081, Rs4237, Rs4247, Rs4255, Rf4092 and As6526) showed 78.2% to 80.2% aa sequence identity to SARS-CoV in the S protein, while the other nine strains without deletions were much more closely related to SARS-CoV, with 90.0% (Rs4084) to 97.2% (Rs4874) aa sequence identity. These nine SARSr-CoVs can be further divided into four genotypes according to their S1 sequences (Fig 2): RsSHC014/Rs4084 showed more genetic differences from SARS-CoV in both NTD and RBD regions; The RBD sequences of SARSr-CoV Rs7327, Rs9401 and previously reported WIV1/ Rs3367 closely resembled that of SARS-CoV. However, they were distinct from SARS-CoV but similar to RsSHC014 in NTD. In contrast, we found a novel SARSr-CoV, termed Rs4231, which shared highly similar NTD, but not RBD sequence with SARS-CoV (Figs 2 and 3). Its S protein showed 94.6% to 95% aa sequence identity to those of human and civet SARS-CoVs (S1 Table). Strains with both NTD and RBD highly homologous to those of SARS-CoV were also present in this cave. In addition to WIV16 which we described previously [18], Rs4874 was also found to have the S protein closest to SARS-CoV S (> 97% aa sequence identity) of all the bat SARSr-CoVs reported to date (Figs 2 and 3). In addition to the SARSr-CoVs subjected to full-length genome sequencing, we also obtained the RBD and NTD sequences from other samples collected in this cave. The sequences with high identity to SARS-CoV RBD were amplified from 10 more R. sinicus samples. SARSr-CoVs with this genotype of RBD were detected in different seasons throughout the five years. Strains containing the NTD similar to SARS-CoV were only found in 2013 (S2 Table). ORF8 is another highly variable gene among different SARS-CoV and SARSr-CoV strains [25,26]. We aligned the ORF8 nt sequences of the representative SARSr-CoVs discovered in this surveillance with those of other SARSr-CoVs and SARS-CoVs (Fig 4). Though WIV16, WIV1, Rs4231 and RsSHC014 were genetically closer to SARS-CoV in S gene, they contained a single 366-nt ORF8 without the 29-nt deletion present in most human SARS-CoVs and showed only 47.1% to 51.0% nt sequence identity to human and civet SARS-CoVs. However, the ORF8 of strain Rf4092 from R. ferrumequinum exhibited high similarity to that of civet SARS-CoV. It possessed a single long ORF8 of the same length (369 nt) as that of civet SARS--CoV strain SZ3, with only 10 nt mutations and 3 aa mutations detected (Fig 4). Similar ORF8 sequences were also amplified from other 7 samples collected in the cave during 2011 to 2013, from both R. ferrumequinum and R. sinicus (S2 Table). The ORF8 of Rs4084 was highly similar to Rf4092's but was split into two overlapping ORFs, ORF8a and ORF8b, due to a short 5-nt deletion (Figs 2 and 4). The position of start codons and stop codons of the two ORFs were consistent with those in most human SARS-CoV strains. Excluding the 8-aa insertion, Rs4084 and SARS-CoV strain BJ01 displayed identical aa sequence of ORF8a, and only three different residues involved in the interaction with human ACE2 are numbered on top of the aligned sequences. SARS-CoV GZ02, BJ01 and Tor2 were isolated from patients in the early, middle and late phase, respectively, of the SARS outbreak in 2003. SARS-CoV SZ3 was identified from civets in 2003. SARSr-CoV Rs 672 and YN2013 were identified from R. sinicus collected in Guizhou and Yunnan Province, respectively. SARSr-CoV Rf1 and JL2012 were identified from R. ferrumequinum collected in Hubei and Jilin Province, respectively. WIV1, WIV16, RsSHC014, Rs4081, Rs4084, Rs4231, Rs4237, Rs4247, Rs7327 and Rs4874 were identified from R.sinicus, and Rf4092 from R. ferrumequinum in the cave surveyed in this study. Another key difference between SARS-CoV and bat SARSr-CoV genomes is the ORF3 coding region [10,17,21]. We analyzed the ORF3a sequences amplified from 42 samples and found that most of the SARSr-CoVs closely related to SARS-CoV in the S gene shared higher ORF3a sequence similarity (96.4% to 98.9% aa identity) with SARS-CoV (S3 Fig and S2 Table). The ORF3b of SARS CoV, sharing a large part of its coding sequence with the ORF3a, encodes a 154-aa protein [27], but it is truncated to different extents at

the C-terminal in previously described bat SARSr-CoVs including WIV1 and WIV16 (S4 Fig). In the current study, we identified a non-truncated ORF3b for the first time (Rs7327), which maintained the nuclear localization signal at its C-terminal. Moreover, it shared 98.1% aa sequence identity with SARS-CoV strain Tor2 with only three aa substitutions (S4 Fig). Thus, Rs7327 is the bat SARSr-CoV most similar to SARS-CoV in the ORF3 region known to date. Recombination analysis The full-length genome sequences of all 15 SARSr-CoVs from the surveyed cave were screened for evidence of potential recombination events. Both similarity plot and bootscan analyses revealed frequent recombination events among these SARSr-CoV strains. It was suggested that WIV16, the closest progenitor of human SARS-CoV known to date [18], was likely to be a recombinant strain from three SARSr-CoVs harbored by bats in the same cave, namely WIV1, Rs4231 and Rs4081, with strong P value (<10 −30 ). Breakpoints were identified at genome positions nt 18391, 22615 and 28160 ( Fig 5A). In the genomic region between nt 22615 and 28160, which contained the region encoding the RBD and the S2 subunit of the S protein, WIV16 was highly similar to WIV1, sharing 99% sequence identity. In contrast, in the region between nt 18391 and 22615, which covered a part of ORF1b and the region encoding the NTD of the S gene, WIV16 showed substantially closer relationship to Rs4231. Meanwhile, the ORF1ab sequences upstream from nt 18391 of WIV16 displayed the highest genetic similarity (99.8% nt sequence identity) to that of Rs4081. Evidence of recombination event was also detected in the genome of the novel SARSr-CoV Rs4084, which had a unique genome organization with split ORF8a and 8b. The previously reported strain RsSHC014 and the newly identified strain Rf4092 were suggested to be the major and minor parent of Rs4084, respectively (P value < 10 −80 ). The breakpoint was located at nt 26796 (S5 Fig). In the region downstream of the breakpoint including ORF8, Rs4084 showed closet genetic relationship with Rf4092, sharing 98.9% nt sequence identity, while it shared the highest nt sequence identity (99.4%) with RsSHC014 in the majority of its genome upstream from the breakpoint. When civet SARS-CoV SZ3 was used as the query sequence in similarity plot and bootscan analysis, evidence for recombination events was also detected ( Fig 5B). In the region between the two breakpoints at the genome positions nt 21161 and nt 27766, including the S gene, closer genetic relationship between SZ3 and WIV16 was observed. However, from position nt 27766 towards the 3' end of its genome, a notably close genetic relationship was observed between SZ3 and Rf4092 instead. Throughout the non-structural gene, moreover, SZ3 shared a similarly high sequence identity with WIV16 and Rf4092. It indicates that civet SARS-CoV was likely to be the descendent from a recombinant of the precursors of WIV16 and Rf4092, or that the SARSr-CoVs found in this cave, like WIV16 or Rf4092, may have been the descendants of the SARS-CoV lineage. Phylogenetic analysis Phylogenetic trees were constructed using the nt sequences of nonstructural protein gene ORF1a and ORF1b. Unlike the high genetic diversity in the S gene, nearly all SARSr-CoVs from the bat cave we surveyed were closely clustered, and showed closer phylogenetic relationship to SARS-CoV than the majority of currently known bat SARSr-CoVs discovered from other locations, except YNLF_31C and 34C which were recently reported in greater horseshoe bats from another location in Yunnan [22] (Fig 6). The phylogeny of SARSr-CoVs in ORF1a and ORF1b appeared to be associated with their geographical distribution rather than with host species. Regardless of different host bat species, SARS-CoV and SARSr-CoVs detected in bats from southwestern China (Yunnan, Guizhou and Guangxi province) formed one clade, in which SARSr-CoV strains showing closer relationship to SARS-CoV were all from Yunnan. SARSr-CoVs detected in southeastern, central and northern provinces, such as Hong Kong, Hubei and Shaanxi, formed the other clade which was phylogenetically distant to human and civet SARS-CoVs (Fig 6 and S6 Fig). Rescue of bat SARSr-CoVs and virus infectivity experiments In the current study, we successfully cultured an additional novel SARSr-CoV Rs4874 from a single fecal sample using an optimized protocol and Vero E6 cells [17]. Its S protein shared 99.9% aa sequence identity with that of previously isolated WIV16 and it was identical to WIV16 in RBD. Using the reverse genetics technique we previously developed for WIV1 [23], we constructed a group of infectious bacterial artificial chromosome (BAC) clones with the backbone of WIV1 and variants of S genes from 8 different bat SARSr-CoVs. Only the infectious clones for Rs4231 and Rs7327 led to cytopathic effects in Vero E6 cells after transfection (S7 Fig). The other six strains with deletions in the RBD region, Rf4075, Rs4081, Rs4085, Rs4235, As6526 and Rp3 (S1 Fig) failed to be rescued, as no cytopathic effects was observed and viral replication cannot be detected by immunofluorescence assay in Vero E6 cells (S7 Fig). In contrast, when Vero E6 cells were respectively infected with the two successfully rescued chimeric SARSr-CoVs, WIV1-Rs4231S and WIV1-Rs7327S, and the newly isolated Rs4874, efficient virus replication was detected in all infections (Fig 7). To assess whether the three novel SARSr-CoVs can use human ACE2 as a cellular entry receptor, we conducted virus infectivity studies using HeLa cells with or without the expression of human ACE2. All viruses replicated efficiently in the human ACE2-expressing cells. The results were further confirmed by quantification of viral RNA using real-time RT-PCR (Fig 8). Activation of activating transcription factor 6 (ATF6) by the ORF8 proteins of different bat SARSr-CoVs The induction of the ATF6-dependent transcription by the ORF8s of SARS-CoV and bat SARSr-CoVs were investigated using a luciferase reporter, 5×ATF6-GL3. In HeLa cells transiently transfected with the expression plasmids of the ORF8s of bat SARSr-CoV Rf1, Rf4092 and WIV1, the relative luciferase activities of the 5×ATF6-GL3 reporter was enhanced by 5.56 to 9.26 folds compared with cells transfected with the pCAGGS empty vector, while it was (Fig 9A). The results suggests that various ORF8 proteins of bat SARSr-CoVs can activate ATF6, and those of some strains have a stronger effect than the SARS-CoV ORF8. Induction of apoptosis by the ORF8a of the newly identified bat SARSr-CoV We conducted transient transfection to examine whether the ORF8a of SARSr-CoV Rs4084 triggered apoptosis. As shown in Fig 9B, 11.76% and 9.40% of the 293T cells transfected with the SARSr-CoV Rs4084-ORF8a and SARS-CoV Tor2-ORF8a expression plasmid underwent apoptosis, respectively. In contrast, transfection with the empty vector resulted in apoptosis in only 2.79% of the cells. The results indicate that Rs4084 ORF8a has an apoptosis induction activity similar to that of SARS-CoV [28]. Discussion Genetically diverse SARSr-CoVs have been detected in various horseshoe bat species across a wide geographic range in China in the past decade [9][10][11][12]14,29]. However, most bat SARSr-CoVs show considerable genetic distance to SARS-CoV, particularly in the highly variable S1, ORF8 and ORF3 regions [10,25]. Recently, several novel SARSr-CoVs have been described to be more closely related to SARS-CoV, either in the S gene or in ORF8. The S proteins of RsSHC014, Rs3367, WIV1 and WIV16, which were reported in our previous studies, shared 90% to 97% aa sequence identities to those of human/civet SARS-CoVs [17,18]. Another strain from Rhinolophus affinis in Yunnan termed LYRa11 showed 90% aa sequence identity to SARS-CoV in the S gene [13]. In addition, two studies have described 4 novel SARSr-CoVs (YNLF_31C/34C and GX2013/YN2013) which possessed a full-length ORF8 with substantially higher similarity to that of SARS-CoV [22,30]. These findings provide strong genetic evidence for the bat origin of SARS-CoV with regard to the S gene or ORF8. However, all of these SARSr-CoVs were distinct from SARS-CoV in at least one other gene, suggesting that none of them was the immediate progenitor of SARS-CoV. Moreover, these SARSr-CoVs were discovered in bat populations from physically distinct locations. The site of origin of the true progenitor of SARS-CoV and the evolutionary origin of SARS-CoV have until now remained elusive. In the current study, we have identified a bat habitat potentially important for SARSr-CoV evolution where a series of recombination events have likely occurred among different SARSr-CoV strains, which provides new insights into the origin of SARS-CoV. SARS first emerged in Guangdong province in late 2002 [7]. However, SARSr-CoVs discovered in bats from neighboring areas of Guangdong to date have shown phylogenetic disparity from SARS-CoV especially in the S gene [9,10,14], suggesting SARS-CoV may have originated from another region. Our analysis of the phylogeny of SARS-CoVs and all known bat SARSr-CoVs using the nt sequence of their non-structural ORF1a and ORF1b genes, which constitute the majority of the genome, shows that SARSr-CoV evolution is strongly correlated with their geographical origin, but not host species. It is noteworthy that SARSr-CoVs detected in Yunnan are more closely related to SARS-CoV than strains from other regions in China. This finding implies that Yunnan, or southwestern China, is more likely to be the geographical source of SARS-CoV than other regions in China, but data from more extensive surveillance are yet needed to support this inference. In our longitudinal surveillance of SARSr-CoVs in a single cave in Yunnan where we discovered Rs3367, RsSHC014, WIV1 and WIV16, the CoV prevalence in fecal samples varied among different sampling time. Generally, a higher prevalence was observed in autumn (September and October) than in spring and early summer (April and May). This may be due to the establishment of a susceptible subpopulation of newborn bats which had not developed their own immunity after the parturition period [31]. Another factor may be the changes in the composition of bat species in the cave at different sampling dates. For example, in September 2012 when the CoV prevalence reached 51.3%, the majority of samples were from R. sinicus, but in May 2015 when only 3 out of the 145 samples tested positive, Aselliscus stoliczkanus was the predominant bat species in the cave. We failed to amplify the RBD sequences from 15 of the 64 SARSr-CoV positive samples. Most of these samples had comparatively low viral concentration (< 10 7 copies/g) (S8 Fig), as revealed by our previous quantitative studies [32]. The unsuccessful amplification of RBD in some samples with high viral concentration was probably because of the more divergent sequences in this region of these SARSr-CoV genomes. In this cave, we have now obtained full-length genome sequences of additional 11 novel SARSr-CoVs from bats. Our findings suggest the co-circulation of different bat SARSr-CoVs highly similar to SARS-CoV in the most variable S1 (NTD and RBD), ORF8 and ORF3 regions, respectively, in this single location. In the ORF1a, ORF1b, E, M and N genes, the SARSr-CoVs circulating in this cave also shared > 98% aa sequence identities with human/ civet SARS-CoVs. Thus, all of the building blocks of the SARS-CoV genome were present in SARSr-CoVs from this single location in Yunnan during our sampling period. Furthermore, strains closely related to different representative bat SARSr-CoVs from other provinces (e.g. Rs672, HKU3 and Rf1) in the RBD region were also detected there. Therefore, this cave could be regarded as a rich gene pool of bat SARSr-CoVs, wherein concurrent circulation of a high diversity of SARSr-CoV strains has led to an unusually diverse assemblage of SARSr-CoVs. During our 5-year surveillance in this single cave, we first reported Rs3367 and WIV1 in 2013, with RBD sequence closely resembling that of SARS-CoV [17]. More recently, we discovered WIV16 which had an RBD almost identical to WIV1's but shared much higher similarity with SARS-CoV than WIV1 in the NTD region of S1, making it the closest SARSr-CoV to the epidemic strains identified to date [18]. In this study, we found a novel strain Rs4231 from the same location sharing almost identical NTD sequence with WIV16 but distinct from it in the RBD, with evidence of a recombination event. Our recombination analysis indicated that a recombination event may have taken place at the junction between the coding region of NTD and RBD in the Rs4231 and WIV1 genomes and resulted in WIV16. Recombination at this genomic position also happened among other SARSr-CoVs relatively distant to SARS-CoV found in this location (e.g. Rs4081 and Rs4247, S5 Fig). The frequent recombination at this hotspot in the S gene increased the genetic diversity of SARSr-CoVs harbored in these bat populations and might have

been responsible for the generation of the S gene of the direct progenitor strain of SARS-CoV. The genomes of SARS-CoVs from patients during the early epidemic phase and civet SARS--CoVs all contained a single full-length ORF8 [3,7]. We have found that a number of bat SARSr-CoVs from this cave possessed a complete ORF8 highly similar to that of early human/civet SARS-CoV (>97% nt sequence identity), represented by strain Rf4092 (S3C Fig). This provided further evidence for the source of human SARS-CoV ORF8 in bats [22,30]. In contrast, the ORF8 was split into overlapping ORF8a and ORF8b in most human SARS-CoV strains from later-phase patients due to the acquisition of a 29-nt deletion [8,26]. In this study, we have discovered for the first time a bat SARSr-CoV with ORF8a and ORF8b highly similar to the laterphase human SARS-CoVs, though the split of ORF8 in the bat SARSr-CoV and that in human SARS-CoV were two independent events. Our recombination analysis suggests that this strain, Rs4084, likely acquired its ORF8 from Rf4092 through recombination, followed by the development of the 5-nt deletion which led to the splitting. It suggests that ORF8 region in bat SARSr-CoV genomes is prone to deletions as in human SARS-CoV [3,25]. Finally, the recombination analysis suggests that an ancestral strain of SARS-CoV SZ3 would have been generated if the recombination around ORF8 had occurred between the lineages that led to WIV16 and Rf4092. Taken together, the evidence of recombination events among SARSr-CoVs harbored by bats in this single location suggests that the direct progenitor of SARS-CoV may have originated as a result of a series of recombination within the S gene and around ORF8. This could have been followed by the spillover from bats to civets and people either in the region, or during movement of infected animals through the wildlife trade. However, given the paucity of data on animal trade prior to the SARS outbreak, the likely high geographical sampling bias in bat surveillance for SARSr-CoVs in southern China, and the possibility that other caves harbor similar bat species assemblages and a rich diversity of SARSr-CoVs, a definite conclusion about the geographical origin of SARS-CoV cannot be drawn at this point. R. sinicus are regarded as the primary natural host of SARS-CoV, as all SARSr-CoVs highly homologous to SARS-CoV in the S gene were predominantly found in this species. However, it is noted that two SARSr-CoVs previously reported from R. ferrumequinum showed the closest phylogenetic position to SARS-CoV in the ORF1a/1b trees. These strains were discovered in another location in Yunnan 80 km from the cave surveyed in the current study [22]. This information also supports the speculation that SARS-CoV may have originated from this region. Nonetheless, since the correlation between the host species and the phylogeny of SARSr-CoV ORF1ab seems limited, more SARSr-CoV sequences need to be obtained from different Rhinolophus bat species in both locations in Yunnan, and from other locations in southern China. In particular, it will be important to assess whether R. ferrumequinum played a more important role in the evolution of SARS-CoV ORF1ab. The cave we studied is located approximately 60 km from the city of Kunming. Beside a number of rhinolophid and hipposiderid species from which SARSr-CoVs have been detected, other bats like myotis were also present there. The temperature in the cave is around 22-25˚C and the humidity around 85%-90%. The physical nature of the cave is not unique, but it does appear to host a particularly dense population of bats in the reproductive season. Similar caves co-inhabited by bat populations of different species are not rare in other areas in Yunnan. We propose that efforts to study the ecology, host species diversity, and viral strain populations of these caves may provide critical information on what drives SARSr-CoV evolution. Our previous studies demonstrated the capacity of both WIV1 and WIV16 to use ACE2 orthologs for cell entry and to efficiently replicate in human cells [17,18]. In this study, we confirmed the use of human ACE2 as receptor of two novel SARSr-CoVs by using chimeric viruses with the WIV1 backbone replaced with the S gene of the newly identified SARSr-CoVs. Rs7327's S protein varied from that of WIV1 and WIV16 at three aa residues in the receptor-binding motif, including one contact residue (aa 484) with human ACE2. This difference did not seem to affect its entry and replication efficiency in human ACE2-expressing cells. A previous study using the SARS-CoV infectious clone showed that the RsSHC014 S protein could efficiently utilize human ACE2 [33], despite being distinct from SARS-CoV and WIV1 in the RBD (S1 Fig). We examined the infectivity of Rs4231, which shared similar RBD sequence with RsSHC014 but had a distinct NTD sequence, and found the chimeric virus WIV1-Rs4231S also readily replicated in HeLa cells expressing human ACE2 molecule. The novel live SARSr-CoV we isolated in the current study (Rs4874) has an S gene almost identical to that of WIV16. As expected, it is also capable of utilizing human ACE2. These results indicate that diverse variants of SARSr-CoV S protein without deletions in their RBD are able to use human ACE2. In contrast, our previous study revealed that the S protein of a R. sinicus SARSr-CoV with deletions (Rp3) failed to use human, civet and bat ACE2 for cell entry [34]. In this study, in addition to Rs4231 and Rs7327, we also constructed infectious clones with the S gene of Rs4081, Rf4075, Rs4085, Rs4235 and As6526, which all contained the deletions in their RBD. These 7 strains, plus Rs4874 and the previously studied WIV1 and RsSHC014, could represent all types of S variants of SARSr-CoVs in this location (S3A Fig). However, none of the strains with deletions in the RBD could be rescued from Vero E6 cells. Therefore, the two distinct clades of SARSr-CoV S gene may represent the usage of different receptors in their bat hosts. The full-length ORF8 protein of SARS-CoV is a luminal endoplasmic reticulum (ER) membrane-associated protein that induces the activation of ATF6, an ER stress-regulated transcription factor that activates the transcription of ER chaperones involved in protein folding [35]. We amplified the ORF8 genes of Rf1, Rf4092 and WIV1, which represent three different genotypes of bat SARSr-CoV ORF8 (S3C Fig), and constructed the expression plasmids. All of the three ORF8 proteins transiently expressed in HeLa cells can stimulate the ATF6-dependent transcription. Among them, the WIV1 ORF8, which is highly divergent from the SARS-CoV ORF8, exhibited the strongest activation. The results indicate that the variants of bat SARSr-CoV ORF8 proteins may play a role in modulating ER stress by activating the ATF6 pathway. In addition, the ORF8a protein of SARS-CoV from the later phase has been demonstrated to induce apoptosis [28]. In this study, we have found that the ORF8a protein of the newly identified SARSr-CoV Rs4084, which contained an 8-aa insertion compared with the SARS-CoV ORF8a, significantly triggered apoptosis in 293T cells as well. Compared with the 154-aa ORF3b of SARS-CoV, the ORF3b proteins of all previously identified bat SARSr-CoVs were smaller in size due to the early translation termination. However, for the first time, we discovered an ORF3b without the C-terminal truncation in a bat SARSr-CoV, Rs7327, which differed from the ORF 3b of SARS-CoV GZ02 strain at only one aa residue. The SARS-CoV ORF3b antagonizes interferon function by modulating the activity of IFN regulatory factor 3 (IRF3) [27]. As previous studies suggested, the nuclear localization signal-containing C-terminal may not be required for the IFN antagonist activity of ORF3b [36]. Our previous studies also demonstrated that the ORF3b protein of a bat SARSr-CoV, termed Rm1, which was C-terminally truncated to 56 aa and shared 62% aa sequence identity with SARS-CoV, still displayed the IFN antagonist activity [37]. It is very interesting to investigate in further studies whether Rs7327's ORF3b and other versions of truncated ORF3b such as WIV1 and WIV16 also show IFN antagonism profiles. As a whole, our findings from a 5-year longitudinal study conclusively demonstrate that all building blocks of the pandemic SARS-CoV genome are present in bat SARSr-CoVs from a single location in Yunnan. The data show that frequent recombination events have happened among those SARSr-CoVs in the same cave. While we cannot rule out the possibility that similar gene pools of SARSr-CoVs exist elsewhere, we have provided sufficient evidence to conclude that SARS-CoV most likely originated from horseshoe bats via recombination events among existing SARSr-CoVs. In addition, we have also revealed that various SARSr-CoVs capable of using human ACE2 are still circulating among bats in this region. Thus, the risk of spillover into people and emergence of a disease similar to SARS is possible. This is particularly important given that the nearest village to the bat cave we surveyed is only 1.1 km away, which indicates a potential risk of exposure to bats for the local residents. Thus, we propose that monitoring of SARSr-CoV evolution at this and other sites should continue, as well as examination of human behavioral risk for infection and serological surveys of people, to determine if spillover is already occurring at these sites and to design intervention strategies to avoid future disease emergence. Ethics statement All sampling procedures were performed by veterinarians with approval from Animal Ethics Committee of the Wuhan Institute of Virology (WIVH05210201). The study was conducted in accordance with the Guide for the Care and Use of Wild Mammals in Research of the People's Republic of China. Sampling Bat samplings were conducted ten times from April 2011 to October 2015 at different seasons in their natural habitat at a single location (cave) in Kunming, Yunnan Province, China. All members of field teams wore appropriate personal protective equipment, including N95 masks, tear-resistant gloves, disposable outerwear, and safety glasses. Bats were trapped and fecal swab samples were collected as described previously [9]. Clean plastic sheets measuring 2.0 by 2.0 m were placed under known bat roosting sites at about 18:00 h each evening for collection of fecal samples. Fresh fecal pellets were collected from sheets early in the next morning. Each sample (approximately 1 gram of fecal pellet) was collected in 1ml of viral transport medium composed of Hank's balanced salt solution at pH7.4 containing BSA (1%), amphotericin (15 μg/ml), penicillin G (100 units/ml), and streptomycin (50 μg/ml), and were stored at -80˚C until processing. Bats trapped for this study were released back into their habitat. RNA extraction, PCR screening and sequencing Fecal swab or pellet samples were vortexed for 1 min, and 140 μl of supernatant was collected from each sample after centrifuge at 3000 rpm under 4˚C for 1min. Viral RNA was extracted with Viral RNA Mini Kit (Qiagen) following the manufacturer's instructions. RNA was eluted in 60 μl of buffer AVE (RNase-free water with 0.04% sodium azide, Qiagen), aliquoted, and stored at -80˚C. One-step hemi-nested RT-PCR (Invitrogen) was employed to detect the presence of coronavirus sequences as described previously using a set of primers that target a 440-nt fragment in the RNA-dependent RNA polymerase gene (RdRp) of all known alphaand betacoronaviruses [20]. For the first round PCR, the 25 μl reaction mix contained 12.5 μl PCR 2 × reaction mix buffer, 10 pmol of each primer, 2.5 mM MgSO 4 , 20 U RNase inhibitor, 1 μl SuperScript III/Platinum Taq Enzyme Mix and 5 μl RNA template. The amplification was performed as follows: 50˚C for 30 min, 94˚C for 2 min, followed by 40 cycles consisting of 94˚C for 15 sec, 52˚C for 30 sec, 68˚C for 40 sec, and a final extension of 68˚C for 5 min. For the second round PCR, the 25 μl reaction mix contained 2.5 μl PCR reaction buffer, 5 pmol of each primer, 50 mM MgCl 2 , 0.5mM dNTP, 0.1 μl Platinum Taq Enzyme (Invitrogen) and 1 μl product of the first round PCR. The amplification was performed as follows: 94˚C for 3 min followed by 35 cycles consisting of 94˚C for 30 sec, 52˚C for 30 sec, 72˚C for 40 sec, and a final extension of 72˚C for 7 min. The RBD region was amplified using the one-step nested RT-PCR method previously described [17]. PCR products were gel purified and sequenced with an ABI Prism 3730 DNA analyzer (Applied Biosystems, USA). PCR products with low concentration or generating heterogeneity in the sequencing chromatograms were cloned into pGEM-T Easy Vector (Promega) for sequencing. The

positive samples in this study were termed using the abbreviated name of bat species plus the sample ID number (e.g. Rs4081). To confirm the bat species of individual sample, PCR amplification of cytochrome b (Cytob) or NADH dehydrogenase subunit 1 (ND1) gene was performed using DNA extracted from the feces or swabs [38,39]. Sequencing of full-length genomes Full genomic sequences of 11 SARSr-CoVs were determined by One-step PCR (Invitrogen) amplification of overlapping genomic fragments with degenerate primers designed by multiple alignment of available SARS-CoV and bat SARSr-CoV sequences deposited in GenBank, and additional specific primers designed from the results of previous rounds of sequencing in this study. Primer sequences are available upon request. Sequences of the 5' and 3' genomic ends were obtained by 5' and 3' RACE (Roche), respectively. PCR products with expected size were gel-purified and subjected directly to sequencing. Each fragment was sequenced at least twice. The sequencing chromatogram of each product was thoroughly examined and sequence heterogeneity was not observed. For some fragments with low concentration of amplicons, the PCR products were cloned into pGEM-T Easy Vector (Promega) for sequencing. At least five independent clones were sequenced to obtain a consensus sequence. Co-presence of sequences of distinct SARSr-CoVs was not found in any of the amplicons. The sequences of overlapping genomic fragments were assembled to obtain the full-length genome sequences, with each overlapping sequence longer than 100 bp. Evolution analysis Full-length genome sequences of the 15 SARSr-CoVs detected from bats in the cave surveyed in this study were aligned with those of selected SARS-CoVs using MUSCLE [40]. The aligned sequences were scanned for recombination events by Recombination Detection Program (RDP) [41]. The potential recombination events suggested by strong P values (<10 −20 ) were further confirmed using similarity plot and bootscan analyses implemented in Simplot 3.5.1 [42]. Phylogenetic trees based on nucleotide sequences were constructed using the Maximum Likelihood algorithm under the LG model with bootstrap values determined by 1000 replicates in the PhyML (version 3.0) software package [43]. Virus isolation The Vero E6 cell line was kindly provided by Australian Animal Health Laboratory, CSIRO (Geelong, Australia). Vero E6 monolayer was maintained in DMEM medium supplemented with 10% fetal calf serum (FCS). Fecal samples (in 200 μl buffer) were gradient centrifuged at 3,000-12,000 g, and the supernatant was diluted 1:10 in DMEM before being added to Vero E6 cells. After incubation at 37˚C for 1 h, the inoculum was removed and replaced with fresh DMEM medium with 2% FCS. The cells were incubated at 37˚C and checked daily for cytopathic effect. All tissue culture media were supplemented with triple antibiotics penicillin/ streptomycin/amphotericin (Gibco) (penicillin 200 IU/ml, streptomycin 0.2 mg/ml, amphotericin 0.5 μg/ml). Three blind passages were carried out for each sample. After each passage, both the culture supernatant and cell pellet were examined for presence of SARSr-CoV by RT-PCR using specific primers targeting the RdRp or S gene. The viruses which caused obvious cytopathic effect and could be detected in three blind passages by RT-PCR were further confirmed by electron microscopy. Determination of virus infectivity by immunofluorescence assay The HeLa cell line was kindly provided by Australian Animal Health Laboratory, CSIRO (Geelong, Australia). HeLa cells expressing human ACE2 were constructed as described previously [17]. HeLa cells expressing human ACE2 and Vero E6 cells were cultured on coverslips in 24-well plates (Corning) incubated with the newly isolated or recombinant bat SARSr-CoVs at a multiplicity of infection (MOI) = 1.0 for 1h. The inoculum was removed and the cells were washed twice with PBS and supplemented with medium. Vero E6 cells without virus inoculation and HeLa cells without ACE2 were used as negative control. Twenty-four hours after infection, cells were rinsed with PBS and fixed with 4% formaldehyde in PBS (pH7.4) at 4˚C for 20 min. ACE2 expression was detected by using goat anti-human ACE2 immunoglobulin followed by FITC-labelled donkey anti-goat immunoglobulin (PTGLab). Virus replication was detected by using rabbit antibody against the nucleocapsid protein of bat SARSr-CoV Rp3 followed by Cy3-conjugated mouse anti-rabbit IgG. Nuclei were stained with DAPI. Staining patterns were observed under an FV1200 confocal microscope (Olympus). Determination of virus replication in Vero E6 cells by plaque assay Vero E6 cells were infected with WIV1, Rs4874, WIV1-Rs4231S, and WIV1-Rs7327S at an MOI of 1.0 and 0.01. After incubation for an hour, the cells were washed with DHanks for three times and supplied with DMEM containing 2% FCS. Samples were collected at 0, 10, 27, and 48 h post infection. The viral titers were determined by plaque assay. Determination of virus replication in HeLa cells expressing human ACE2 by quantitative RT-PCR HeLa cells expressing human ACE2 were inoculated with WIV1, Rs4874, WIV1-Rs4231S, and WIV1-Rs7327S at an MOI of 1.0, and were incubated for 1h at 37˚C. After the inoculum was removed, the cells were supplemented with medium containing 1% FBS. Supernatants were collected at 0, 12, 24 and 48h. Virus titers were determined using quantitative RT-PCR targeting the partial N gene with a standard curve which expresses the correlation between Ct value and virus titer (shown as TCID50/ml). The standard curve was made using RNA dilutions from the purified Rs4874 virus stock (with a titer of 2.15 × 10 6 TCID50/ml). For qPCR, RNA was extracted from 140 μl of each supernatant with Viral RNA Mini Kit (Qiagen) following manufacturer's instructions and eluted in 60 μl AVE buffer. The PCR was performed with the TaqMan AgPath-ID One-Step RT-PCR Kit (Applied Biosystems) in a 25 μl reaction mix containing 4 μl RNA, 1 × RT-PCR enzyme mix, 1 × RT-PCR buffer, 40 pmol forward primer (5'-GTGGTGGTGACGGCA AAATG-3'), 40 pmol reverse primer (5'-AAGTGAAGCTTCTGG GCCAG-3') and 12 pmol probe (5'-FAM-AAAGAGCTCAGCCCCAGATG-BHQ1-3'). The amplification was performed as follows: 50˚C for 10 min, 95˚C for 10 min followed by 50 cycles consisting of 95˚C for 15 sec and 60˚C for 20 sec. Plasmids The ORF8 genes of bat SARSr-CoV WIV1 and Rf4092 and the ORF8a gene of bat SARSr-CoV Rs4084 were amplified by PCR from the viral RNA extracted from the isolated virus or fecal samples. The ORF8 gene of SARS-CoV GZ02 and bat SARSr-CoV Rf1, and the ORF8a gene of SARS-CoV Tor2 were synthesized by Tsingke Biological Technology Co., Ltd (Wuhan, China). All genes were cloned into the pCAGGS vector constructed with a C-terminal HA tag. Expression of the proteins was confirmed by Western blotting using a mAb against the HA tag. Five tandem copies of the ATF6 consensus binding sites were synthesized and inserted into the pGL3-Basic vector to construct the luciferase reporter plasmid 5×ATF6-GL3, in which the luciferase gene is under the control of the c-fos minimal promoter and the ATF6 consensus binding sites. Luciferase reporter assay HeLa cells in 24-well plates were transfected using Lipofectamine 3000 reagent (Life Technologies) following the manufacturer's instruction. Cells per well were co-transfected with 600ng of the 5×ATF6-GL3 reporter plasmid, with 300ng of each expression plasmid of SARS-CoV and SARSr-CoV ORF8 or empty vector and 20ng of pRL-TK (Promega) which served as an internal control. The cells were incubated for 24h, and were treated with or without 2μg/ml tunicamycin for 16h. Cells were harvested and lysed. Luciferase activity was determined using a dual-luciferase assay system (Promega). The experiment was performed in triplicate wells. Quantification of apoptotic cells 293T cells in 12-well plates were transfected using Lipofectamine 3000 reagent (Life Technologies) following the manufacturer's instruction. Cells per well were transfected with 3μg of the expression plasmid of SARS-CoV Tor2 or SARSr-CoV Rs4084 ORF8a, or the empty vector. 24h post transfection, apoptotic cells were quantified by using the Annexin V-fluorescein isothiocyanate (FITC)/PI Apoptosis Detection Kit (Yeasen Biotech, Shanghai) in accordance with the manufacturer's instruction. Apoptosis was analyzed by flow cytometry. The experiment was performed in triplicate wells. Accession numbers The complete genome sequences of bat SARS-related coronavirus strains As6526, Rs4081, Rs4084, Rf4092, Rs4231, Rs4237, Rs4247, Rs4255, Rs4874, Rs7327 and Rs9401 have been deposited in the GenBank database with the accession numbers from KY417142 to KY417152, respectively. Table. Distribution of SARSr-CoVs highly similar to SARS-CoV in the variable S, ORF3 and ORF8 genes in the single cave. Therapeutic Activity of Partially Purified Fractions of Emblica officinalis (Syn. Phyllanthus emblica) Dried Fruits against Trypanosoma evansi Emblica officinalis (E. officinalis) dried fruits were evaluated for its antitrypanosomal activity and cytotoxic effects. Vero cell line maintained in DMEM (Dubecco’s Modified Eagle Medium) and incubated with Trypanosoma evansi for more than 12 h. MPE was added to the Vero cell culture medium at different concentrations (250~1,000 μg/mL) with trypanosomes concentration (1 × 10 trypanosomes/mL in each ELISA plate well) and incubated at appropriate conditions for 72 h. In-vitro cytotoxicity of MPE of E. officinalis was determined on Vero cells at concentrations ((1.56~100 μg/mL). Acute toxicity and in-vivo infectivity tests were done in mice. Obtained MPE of E. officinalis underwent process of purification via column chromatography, preparative chromatography and HPLC (higher performance liquid chromatography) with bioassay at different strata on Alsever’s medium. In-vivo assay for trypanocidal activity, MPE and PPFs (partially purified fractions) of E. officinalis with two sets of mice, each mouse was inoculated with 1 × 10/mL of trypanosomes and treated (48 h post inoculation) at concentrations (12.5, 25, 50, 100 and 200 mg/kg body weight) were administered at dose rate of 100 μL per mouse via intraperitoneal route (in treating parassitemic mice) to different groups of mice, 6 mice per concentration. HPLC of partially purified fractions of E. officinalis was carried out with mobile phase of acetonitrile: water (40:60) in gradient mode. In vitro, MPE induced immobilization and killing of the parasites in concentration-time dependent manner. Significant reduction of trypanosomes counts from concentration of 250 μg/mL and complete killing of trypanosomes at 5th hour of observation, which was statistically equivalent to 4th hour of Diminazine Aceturate (Berenil), standard reference drug used. HPLC of the partially purified fractions revealed two major prominent peaks at retention time of 1~4 min. In vivo, both MPE and PPFs of test material did prolong lives of mice by 6~9 days but could not cure them. At concentration of 2,000 kg/kg body weight of MPE in acute test, all mice survived. For in-vivo infectivity test, mice injected with immobilized trypanosomes developed parasitemia and died while, the other group survived. MPE, PPFs and Diminazine Aceturate were toxic to Vero cells at all concentrations exception of 1.56, 1.56~3.13 and 1.56~6.25 μg/mL, respectively. From this report, PPFs of E. officinalis dried fruits demonstrated potential pathway for a new development of trypanocide in near future if additional investigations are put in place. Introduction  Trypanosomosis, a disease caused by blood protozoan parasites of genus Trypanosoma, is on the increase in endemic regions, (e.g., Africa and Latin America), where millions of population and cattle are Plant Material Emblica officinalis dried fruits of the family euphorbiaceae were obtained from reputable Ayurvedic shop from hilly region of Palampur, Himachal Pradesh. Plant material was identified by Institute of Himalaya Biosource Technology, Palampur, Himachal Pradesh, India. Extraction Twenty grammes of E. officinalis dried fruits were pounded into powder with pestle and mortar and cold extracted twice with 200 mL of methanol (analytical grade) according to Stahl [24]. The filtrates were dried at 37 °C and stored at 4 °C until used. Applied Solvent Systems The following solvent systems were tested to develop the TLC plates to obtain a more suitable system for both extract and fractions according to the method of Stahl [24]:  Chloroform/hexane/acetic acid (50:50:1);  Chloroform/ethyl acetate/acetic acid (50:50:1);  Methanol and chloroform (20: 80). Test Organism Trypanosoma evansi was obtained from the Division of Parasitology, IVRI (Indian Veterinary Research In-vitro Tryponocidal Activity In-vitro trypanocidal activity was carried out on two media: (1) On modified method of Oliveira et al. [26]: In this method, a Vero cell line was grown in DMEM (Dulbecco's Modified Eagle Medium) (Sigma) in 96-well flat bottom micro culture plates (Nunc, Denmark). Each well received 100 μL of DMEM containing 5 × 10 5 cells/mL. The plates were incubated at 37 °C under 5% CO 2 for 48 h to complete development of monolayer. After the formation of confluent monolayer, the medium (DMEM) was discarded and replaced with a fresh DMEM. And the medium was supplemented with 20~40% FCS (fetal calf serum), Gibco USA and antibiotics (100 units penicillin, 100 µg streptomycin and 40 µg gentamycin). A high parasitemic blood from mouse was diluted

with DMEM to obtain a final parasite of 1 × 10 6 parasites/mL. The suspension (100 mL of medium with trypanosomes was added at rate of 1:1 to MPE of E. officinalis dried fruits at concentrations (250~1,000 µg/mL). The suspension (100 mL of medium with trypanosomes) was added at rate of 1:1 to test extract and the plates were incubated at 37 °C under 5% CO 2. The mixture was incubated for 9 h. The test was repeated at least thrice .and the plate was incubated under the same conditions mentioned above. The test was repeated at least thrice. Stock of test MPE was solubilized in 1% DMSO (dimethyl suphoxide); (2) On Alsever's medium: Trypanosomes were suspended in Alsever's solution with inactivated bovine serum at 58 °C for 1 h. Trypanosomes concentration was 1 × 10 6 parasites/mL. 180 µL of the medium was added to the test extract of E. officinalis dried fruits (20 µL) and incubated at 37 °C with 5% carbon dioxide for 5 h. On hourly basis, drops of the incubated mixture were observed under inverted microscope for antitrypanosomal activity [27]. The concentration of DMSO in the experiment had no deleterious effect by itself on host cells or parasites. 1% DMSO in distilled water was used as control [28]. In-vivo Infectivity Assessment In-vivo infectivity of MPE of E. officinalis dried fruits was carried out after successful completion of anti-trypanosomal activity. Contents of microculture plate wells that contained reduced and apparently killed trypanosomes with MPE of the test material were inoculated (0.1 mL per mouse) into two groups of mice (six per group) via intra-peritoneal, and observed for more than 60 days for parasitaemia [29,30]. In-vitro Cytotoxicity Test Cytotoxic effects of the MPE and pooled PPFs of E. officinalis dried fruits were determined according to the method described by Sidwell and Hoffman [31]. Vero cell line was grown in DMEM (Dulbecco's Modified Eagle Medium) (Sigma) Gibco, USA antibiotics (100 units penicillin, 100 µg streptomycin and 40 µg gentamycin) in 96-well flat bottom micro culture plates (Nunc, Denmark). Each well received 100 µL of DMEM containing 5 × 10 5 cells/mL. The plates were incubated at 37 °C under 5% CO 2 for 48 h. After the formation of confluent monolayer, the medium was discarded and replaced with a fresh one. A high parasitaemic blood from mouse was diluted with DMEM to obtain a final parasite of 1 × 10 6 parasites/mL. Confluent monolayers of Vero cells were treated with serial dilutions of MPE and pooled PPFs of the test material at concentrations (1.56~100 µg/mL) in triplicate and incubated under the same conditions described previously. After 24 h of incubation, the culture plates were observed for evidence of cytotoxic effects. The plates were incubated for 72 h and observed daily. It was repeated thrice. In each case, after the 72 h of incubation, the culture media of the After 24 h of incubation, the culture plates were observed for evidence of cytotoxic effects. Acute Toxicity Test Acute toxicity test of E. officinalis dried fruits was carried out according to the method of Madubunyi [32]. In this method, powdered test material was dissolved in either distilled water or vegetable oil pending on it solubility at a dose rate of 2,000 mg/kg body weight and administered to six mice according to the body weight. Mice were observed for at least two weeks for any sign of toxicity and mortality. Column Chromatography of Methanolic Plant Extract of E. officinalis Dried Fruits This was done according to the method of Stahl [24]. Obtained MPE of E. officinalis dried fruits were used for bioassay-guided purification. Residues (11.623 g of MPE) obtained from methanol extraction of 60 g E. officinalis dried fruits were used for column chromatography. Extract was loaded into already packed glass column with silica gel (60~120 mesh) for column chromatography. The extract was eluted with varied ratios of chloroform/methanol. In-vivo Trypanocidal Activity of Methanolic Plant Extract and Partially Purified Fractions of E. officinalis Dried Fruits This was carried out as per the method of Freiburghaus et al. [9]. Six mice in a group were inoculated with trypanosomes (1 × 10 4 /mL). Infected mice were treated with both MPE and pooled PPFs of E. officinalis dried fruits at concentrations (12.5~200 mg/kg body weight) intraperitoneally 48 h post on set of parasitemia. 1% of DMSO was added to the extract, PPFs and diluted with DMEM. A drop of blood was taken from the tail-end of the mice daily and parasites were counted as previously described. HPLC (Higher Performance Liquid Chromatography) Analysis of Representative Pooled Partially Purified Fractions from E. officinalis Dried Fruits HPLC analysis was done according to Sharma et al. [33]. HPLC (Waters) analysis by injecting 20 µL of dissolved representative fractions of E. officinalis in HPLC graded methanol via 18 columns. Gradient of methanol: water (40:60) to methanol (100%) for 30 min was used. At a zero minute, the ration of acetonitrile to water percentages was 10:90. But at 25 min, the ratio percentages were 64:36, respectively. Institute Committee on Welfare and Cruelty to Animals Indian Veterinary Research Institute Committee on Welfare and Cruelty to Animals received and approved application for the usage of mice in this research. Statistical Analysis Results of trypanocidal activity were expressed as "mean ± SEM". Statistical significance was determined by Sigma Stat (Jandel), USA. Results In this current research, results are presented in Tables 1-9 and Figs. 1-3 accordingly. Extraction Solvent, methanol, was used in extraction of E. officinalis dried powdered fruits. It appeared that methanol was suitable for its extraction as per the bioactive constituents present in the MPE of sample material observed on the TLC plates. Solvent System Out of four solvent systems tested in the analysis of TLC plates with applied aliquots of MPE and fractions of the test material, solvent systems, In-vivo Infectivity Test One group of mice inoculated with contents of ELISA plate wells with completely killed trypanosomes survived for more than 60 days as to the other group inoculated with contents of ELISA plate wells with reduced trypanosomes count that died of parasitemia. In-vitro Trypanocidal Activity of Methanolic Plant Extract of E. officinalis Dried Fruits Results of in-vitro trypanocidal activity of MPE of E. officinalis dried fruits at different concentrations (250~1,000 µg/mL) were as given in Fig. 1. In this result, strong trypanocidal activity was observed with complete killing of the trypanosomes at 5 h of incubation at 250 µg/mL, which was statistically the same as Diminazine Aceturate (50 µg/mL) standard drug at 4 h of incubation (Fig. 1). Trypanocidal activity was concentration-time dependent faction. Average mean trypanosomes counts from 37.67 ± 0.58 and below is significant between the treatment groups and negative control (P ≤ 0.05 to 0.01). In-vitro Cytotoxicity Test As shown in Tables 1 and 2, MPE and PPFs of E. officinalis dried fruits, and Diminazine Aceturate were cytotoxic to Vero cells in all concentrations except at 1.56, 1.56~3.13 and 1.56~6.25 µg/mL, respectively. Cytotoxic effects, such as distortion, sloughing, swelling and dead of the affected cells, were observed compared to normal cells. Acute Toxicity Test At concentration of 2,000 kg/kg body weight, MPE of E. officinalis was not toxic to mice in different groups. All mice survived in acute toxicity test as given in Table 3. Fig. 1 In-vitro trypanocidal activity of methanolic extract of E. officinalis on Vero cell line. Fractions I (1~7 and 11~40) did depict presence of small amount of bioactive components and were combined as a result of similarity of TLC profile. Fractions II (8~10) displayed three layers of broad bands on TLC plates and nothing remained at the origin of applications. In Fractions III (41~96), there was a little mobility of bioactive components from the origin of applications. But, Fractions IV (97~581) mostly did not move from origin of applications of aliquots as depicted on TLC plates, which gradually increased in intensity. Corresponding in-vitro trypanocidal activities were given in Tables 4-7. In-vitro Trypanocidal Activity of PPFs (Partially Purified Fractions) of E. officinalis Dried Fruits As shown in Table 8, distinct pooled fractions 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 Bioassay status: Significant reduction of parasites counts started from concentration of 250 µg/mL and complete killing of parasites at 1,000 µg/mL at 4th hour of incubation .An average mean parasites count of 37.67 ± 0.58 is statistically critical value. Average mean from 37.67 ± 0.58 and below is significant between the treatment groups and negative control (P ≤ 0.05 to 0.01). .00 ± 0.0 Bioassay status: Significant reduction of parasites counts started from concentration 250 µg/mL and complete killing of parasites at 500 µg/mL at 5th hour of observation. An average mean parasites count of 37.67 ± 0.58 is statistically critical value. Average meanparasites counts from 37.67 ± 0.58 and below is significant between the treatment groups and negative control (P ≤ 0.05 to 0.01). 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 40.00 ± 0.0 Bioassay status: Significant reduction of parasites counts from concentration of 250 µg/mL and complete killing of parasites at 750 µg/mL at 5th hour of observation. An average mean parasites count of 37.67 ± 0.58 is statistically critical value. Average mean parasites counts from 37.67 ± 0.58 and below is significant between the treatment groups and negative control (P ≤ 0.05 to 0.01). Control (negative control) 6.167 ± 0.31 13.50 ± 0.56 39.50 ± 0.43 At dose rate of 200 mg/kg body weight, the mice in this group survived for 6 days post on set of parasitemia. There was significant difference (P < 0.05) between treated groups with test material in comparison to negative control that survived for only three days. exhibited significant trypanocidal activity in all concentrations (250~1,000 µg/mL) with significant difference (P < 0.01). In-vivo Trypanocidal Activity of Methanolic Plant Extract of E. officinalis Dried Fruits In-vivo trypanocidal activity of MPE of E. officinalis dried fruits at different concentrations were given in Table 8. Mice in distinct groups treated with MPE of the test material at concentrations (12.5, 25, 50, 100 and 200 mg/mL) after the onset of parasitemia survived up to Days 4, 5 and 6, respectively, in comparison to Day 4 of untreated control with significant difference (P < 0.05). In-vivo Trypanocidal Activity of Partially Purified Fractions of E. officinalis Dried Fruits in Mice Partially purified fractions of E. officinalis fruits with high content of desired unidentified compounds with maximum in-vitro trypanocidal activity were used to treat parasitemic mice, which exhibited different levels of trypanocidal activity, were as given in Table 9 and Fig. 2. Mice treated with the obtained pooled fractions of bioactive constituents of interest survived for maximum 9 days post infection as to 4 days of untreated control with significant difference (P ≤ 0.05 to 0.01). HPLC (Higher Performance Liquid Chromatography) Analysis of Pooled Fractions of E. officinalis Dried Fruits HPLC analysis of representative pooled fractions of E. officinalis dried fruits that contained bioactive constituents of interest revealed two prominent peaks depicting more than one compounds at two detections (210 and 320 nm), as given in Fig. 3. Detection at more than one wavelength depicted glaringly impurity of the At doses of 12.5, 25, 50, 100 and 200 mg/kg body weight, the mice in each group survived for 5, 5, 6, 8 and 10 days post on set of parasitemia. There was degree of significant difference between treated groups with test material compared to negative control that survived for only three days (P ≤ 0.05 to 0.01). Extraction In this report, methanol used in the extraction of E. officinalis dried fruits is similar to previous work documented by Shaba et al. [12,30,34]. Solvent System Solvents, methanol and chloroform/ethyl acetate/acetic acid (50:50:1) used in analysis of TLC plates applied with MPE and fractions of E. officinalis dried fruits are comparable to that used in bioassay guided isolation of a diastereoisomer of kolavenol from Entada absyssinica [9]. In-vivo Infectivity Test In-vivo infectivity test of MPE of the test is in line with work done by Igweh et al. [29] and Shaba et al. [35] where groups of mice inoculated with contents of ELISA plate wells with apparent killed trypanosomes survived. In-vitro Cytotoxicity Test Cytotoxic effects of E. officinalis dried fruits are comparable to cytotoxic effects of Terminalia arjuna

bark extract with distortion and apoptosis of human hepatoma cell line (HEPG2) [36] and Terminalia belirica dried fruits with similar cytotoxic effects observed in this report [37]. Acute Toxicity Test Acute toxicity test of MPE of the test material is comparable to that of Nuclea latifolia, in which no toxic sign was observed in rats at concentrations of 100~300 mg/kg body weight. But, fatalities were observed at 400 and 800 mg/kg body weight [32]. Column Chromatography of Methanolic Extract of E. officials Dried Fruits Fractionation of MPE of E. officinalis dried fruits via column chromatography is similar to fractionation of Cannabis sativa with two fractions active against T. brucei rhodesiense [38,39]. In-vitro Trypanocidal Activity of PFs (Pooled Fractions) of E. officinalis Dried Fruits Results of varied trypanocidal activity of PFs of E. officinalis dried fruits is in line with that of fractionation of Cannabis sativa with two fractions active against T. brucei rhodesiense [39] and bioassay guided isolation of a diastereoisomer of kolavenol from Entada absyssinica active on Trypanosoma brucei rdesiense [9]. Higher Performance Liquid Chromatography Analysis of Pooled Fractions of E. officinalis Dried Fruits HPLC analysis of pooled PPFs of E. officinalis dried fruit is comparable to fractionation and purification of bioassay guided isolation of a diastereoisomer of kolavenol from Entada absyssinica [9]. In-vivo Trypanocidal Activity of PPFs (Partially Purified Fractions) of E. officinalis Dried Fruits in Mice In-vivo trypanocidal activity of PPFs of E. officinalis dried fruits is in line with fractionation of Cannabis sativa with two fractions active against T. brucei rhodesiense [39] and bioassay guided isolation of a diastereoisomer of kolavenol from Entada absyssinica active on Trypanosoma brucei rdesiense [9]. Mechanism of action of E. officinalis dried fruits may be due to gallic acid that has been isolated already from it and its corresponding trypanocidal activity has been documented [40]. Also, it could be due to intercalation of obtained extracts/fractions/isolated compounds of E. officinalis with the DNA of trypanosomes of which such actions have been reported [10]. Representative pooled fractions of PPFs of E. officinalis could not cure the mice but prolonged its lifespan to Day 9 post infection at maximum dose of 200 mg/kg body weight. This may be due to inability of the PPFs to sustain sufficient blood plasma to kill the trypanosomes and possibly, the ease of its being degraded in the body of the mice, which is common to such fractions/isolated compounds [10]. Conclusions In conclusion, E. officinalis dried fruits exhibited significant in-vitro antitrypanosomal activity. Even though strong results of in-vitro trypanocidal activity was not completely transformed during in-vivo testing due to differences in physiological status, attained level bioassay-guided purification decreased cytotoxic effects level and increased trypanocidal activity as shown during in-vivo testing with PPFs of the test material. In near future, this could lead to development of urgently needed new trypanocide against menacing trypanosomes in both animals and humans. Further purification of PPFs of E. officinalis dried fruits is needed to determine its maximum trypanocidal status. A Novel Cytoplasmic Male Sterility in Brassica napus (inap CMS) with Carpelloid Stamens via Protoplast Fusion with Chinese Woad A novel cytoplasmic male sterility (CMS) in Brassica napus (inap CMS) was selected from the somatic hybrid with Isatis indigotica (Chinese woad) by recurrent backcrossing. The male sterility was caused by the conversion of tetradynamous stamens into carpelloid structures with stigmatoid tissues at their tips and ovule-like tissues in the margins, and the two shorter stamens into filaments without anthers. The feminized development of the stamens resulted in the complete lack of pollen grains, which was stable in different years and environments. The pistils of inap CMS displayed normal morphology and good seed-set after pollinated by B. napus. Histological sections showed that the developmental alteration of the stamens initiated at the stage of stamen primordium differentiation. AFLP analysis of the nuclear genomic composition with 23 pairs of selective primers detected no woad DNA bands in inap CMS. Twenty out of 25 mitochondrial genes originated from I. indigotica, except for cox2-2 which was the recombinant between cox2 from woad and cox2-2 from rapeseed. The novel cox2-2 was transcribed in flower buds of inap CMS weakly and comparatively with the fertile B. napus addition line Me harboring one particular woad chromosome. The restorers of other autoplasmic and alloplasmic CMS systems in rapeseed failed to restore the fertility of inap CMS and the screening of B. napus wide resources found no fertility restoration variety, showing its distinct origin and the related mechanism of sterility. The reasons for the mitochondrial rearrangements and the breeding of the restorer for the novel CMS system were discussed. INTRODUCTION Cytoplasmic male sterility (CMS) is a maternally inherited trait that fails to produce viable pollens, and has been reported in a large number of plant species. CMS is encoded in the mitochondrial genome and can arise spontaneously due to mutation in the genome (autoplasmy) or can be expressed following cytoplasmic substitutions due to nuclear-mitochondrial incompatibility (alloplasmy) (Prakash et al., 2009). A CMS fertility restoration system has been used as a suitable pollination control system to produce commercial hybrid seed for many crops, because it is easy to maintain. Cytoplasmic male sterility phenotypes encompass a wide range of reproductive abnormalities, including degenerate anthers, aborted pollen, carpelloid and petaloid stamens (for review see Chase, 2007). Many CMS genes result from mtDNA rearrangements, as at least 10 essential mitochondrial genes are involved in the origination of CMS genes among 28 types of CMS from 13 crop species (Chen and Liu, 2014). Intriguingly, sequences of the known functional mitochondrial genes are not necessarily included in some CMS genes, for several rice CMS-associated ORFs consist of sequences of putative mitochondrial ORFs (Chen and Liu, 2014). Additionally, some CMS genes (such as orf125 in radish CMS-Kos and its variant orf138 in Brassica CMS-Ogu) are non-chimeric genes with the sequences from single source (Bonhomme et al., 1992;Iwabuchi et al., 1999). CMS genes have been shown to cause mitochondrial dysfunction by disrupting the mitochondrial membranes, reducing the mitochondrial ATP level and increasing the reactive oxygen species content (Hanson and Bentolila, 2004;Chen and Liu, 2014;Horn et al., 2014). Up to now, nine Rf (restorer of fertility) genes have been isolated in seven plant species, such as Rfo (Rfk1) in radish and Brassica (Chen and Liu, 2014). Most of the identified Rf genes encode PPR (pentatricopeptide repeat) proteins, but Rf genes are also highly multifarious and the respective restoration of fertility in CMS/Rf systems may be realized by various mechanisms at genomic, post-transcriptional, translational, or post-translational, and metabolic levels (Chen and Liu, 2014). Stable CMS lines of crops could be produced by introducing the alien cytoplasm from the relatives through interspecific-/intergeneric hybridizations, or by mediating the mitochondrial rearrangement via somatic fusions (Prakash et al., 2009). For Brassica crops, since a sterility inducing cytoplasm identified in a wild population of Raphanus sativus (Ogura CMS) was introduced into B. napus and B. oleracea (Bannerot et al., 1974), a spectrum of alloplasmic CMS lines of diverse origins have been obtained by combining the cytoplasm of Brassica coenospecies with crop nuclei, particularly in B. juncea (see reviews by Delourme and Budar, 1999;Budar et al., 2004;Prakash et al., 2009;Yamagishi and Bhat, 2014). More recently, a novel CMS system in B. juncea incorporating the cytoplasm of B. fruticulosa was developed (Atri et al., 2016). Somatic hybridization via protoplast fusion is a possible alternative for gene transfer from wild relatives to crops, by combining both nuclear and cytoplasmic genomes of two distantly related species, genera or even tribes. In somatic hybrids from protoplast fusion, the chloroplasts are usually inherited from one of the parents, whereas the mitochondrial DNA (mtDNA) is rearranged and may include DNA from both parents (Prakash et al., 1999). Intergenomic mitochondrial recombination with high frequency was found to occur in various somatic hybrids from the combinations of different species within Brassicaceae (Landgren and Glimelius, 1994;Dieterich et al., 2003;Leino et al., 2003;Oshima et al., 2010;Liu et al., 2015), leading to novel CMS genes. In the B. napus CMS Tournefortii-Stiewe produced by protoplast fusion with B. tournefortii (Dieterich et al., 2003), the mitochondrial rearrangement at upstream of the gene atp9 generated a chimeric orf193 that was co-transcribed with atp9. But another B. napus alloplasmic CMS Tournefortii by sexual hybridization with the same species had a chimeric orf263 at the vicinity of the atp6 (Landgren et al., 1996). Furthermore, recombinant mitochondrial genomes were also generated when protoplasts of fertile and sterile cytoplasm were fused (Sakai and Imamura, 1990;Wang et al., 1995;Meur et al., 2006). From the mitochondrial genomes of nap, pol, ogu, and hau CMS sequenced (Handa, 2003;Chen et al., 2011;Tanaka et al., 2012;Heng et al., 2014), comparative analysis revealed that the CMS associated genes were localized close to the edge of syntenic sequence blocks, together with short repeats and overlapped repeats. These repeats were responsible for reorganization of mitochondrial genomes via homologous recombination. Overlapping homologous sequences from the fusion parents were produced when the parental mitochondria fused and the recombination occurred. In the cybrid derived from the somatic hybrid between Nicotiana tabacum and Hyoscyamus niger after being backcrossed with N. tabacum, the mitochondrial genome was highly recombinant with at least 35 intergenomic rearrangement events detected between two parents (Sanchez-Puerta et al., 2015). Recombination occurred via homologous mechanisms involving the double-strand break repair and/or break-induced replication pathways. Currently, the autoplasmic pol CMS system from natural mutations in B. napus is still predominant for rapeseed hybrid production in China, but the alloplasmic Ogura CMSfertility restoration system is widely used in Europe and Canada. Although several CMS systems have been reported, they are not used commercially due to undesirable CMS phenotype and lack of fertility restoration genes. In our previous study, the intertribal somatic hybrids between B. napus and Isatis indigotica Fort. (Chinese woad) of the Isatideae tribe within the Brassicaceae family were obtained and backcrossed successively to B. napus (Du et al., 2009;Kang et al., 2014), resulting in the development of one novel B. napus CMS line with carpelloid stamens (named as inap CMS). Herein, the histological and genetical characterizations of inap CMS were described, together with the mitochondrial rearrangements. Plant Materials The intertribal somatic hybrid (As1, 2n = 52, AACCII) between B. napus L. cv. Huashuang 3 (2n = 38, AACC) and I. indigotica Fort. (Chinese woad, 2n = 14, II) was previously produced by protoplast fusion (Du et al., 2009). The backcrossed progenies with the same chromosome number and complement as B. napus were male sterile due to the development of carpelloid stamens (Kang et al., 2014). The phenotype of the carpelloid stamens was stably maintained after backcrosses with Huashuang 3 (Figure 1), which showed the characteristic of the cytoplasmic inheritance. Then, the novel CMS line of B. napus was developed and named as inap CMS. Its stability of male sterility was observed for 3 years in two locations and two seasons: Qinghai University Nuclear and Organelle DNA Analysis For nuclear analysis, total DNA was extracted and purified from young leaves of the parents and CMS line according to the method by Dellaporta et al. (1983). AFLP analysis was carried out according to the procedures of Vos et al. (1995) with some modifications. Genomic DNA (75 ng) of each sample was digested by using the restriction endonucleases EcoR I and Mse I for 6 h at 37 • C and denaturalized for 1 h at 65 • C. Then two adapters were ligated to the sticky ends of the digested DNA at 22 • C overnight, and the resulting ligation products were amplified by PCR with primers matching the adapters. After being diluted to 15 times, resultant PCR products were amplified using 23 pairs of randomly selected primers. Finally, the PCR products were separated on 6% polyacrylamide gels and DNA bands were visualized by silver staining. For mtDNA analysis, 24 mitochondrial genes were examined in parents and CMS line (Supplementary Table S1). A total of 20 µl reaction contained 4 µl 5Phusion HF Buffer, 0.4 µl 10 mM dNTPs, 0.5 µl forward and reverse primer (10 µM), 0.6 µl 100% DMSO, 1 µl template DNA (50 ng/µl) and 0.2 µl Phusion Hot Start II High-Fidelity DNA Polymerase (2 U/µl). PCR amplification was carried out with an initial denaturation step at 98 • C for 30 s followed by 35 cycles of 94 • C for 5 s, 65 •

C for 20 s (optimal annealing temperature was determined by Tm calculator on website: www.thermoscientific.com/pcrwebtools), 72 • C for 45 s, and a final10-min extension at 72 • C. PCR products were separated on 0.8% agarose gels and purified using the TIANgel Midi Purification Kit (TIANGEN BIOTECH (BEIJING) CO., LTD., Beijing, China). The 3 -A overhang of blunt-end DNA fragments was added using DNA A-Tailing Kit (TaKaRa BIO INC, Dalian, China). Then the products were cloned into the pMD18-T vector and sequenced. The primers for amplifying complete sequence of cox2-1 and cox2-2 were designed based on the rapeseed mtDNA sequence (accession number: AP006444, Table 1). Reverse Transcription-PCR Total RNA was extracted from flower buds of inap CMS line, B. napus cv. Huashuang 3, I. indigotica and the male fertile B. napus monosomic addition line Me (Kang et al., 2014) using the Eastep R Super Total RNA Extraction Kit (Shanghai Promega Biological Products, Ltd, Shanghai, China). cDNA was synthesized from total RNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Inc., America). RT-PCR amplification was carried out with an initial denaturation step at 94 • C for 5 min followed by 35 cycles of 94 • C for 30 s, 57 • C for 30 s, and 72 • C for 90 s, and a final 10-min extension at 72 • C. Histological Studies The inflorescences of B. napus and CMS line were first fixed by 50% FAA solution and stained with Ehrlich's haematoxylin solution. After rinsed with distilled water, the samples were dehydrated with a graded ethanol series (15,30,50,70,85,95, and 100%) for 4 h in each. Then samples were gradually cleared in 20, 40, 60, 80, and 100% chloroform for 2 h, and added paraffin fragments into 100% chloroform twice and kept in an incubator at 37 • C for 2 days. The samples were embedded in Paraffin with ceresin (Sinopharm Chemical Reagent Co., Ltd, Shanghai, China). The tissues were sectioned into 8 µm sections, floated on drops of distilled water and dried overnight onto Poly-L-Lysine-coated slides at 37 • C. After deparaffinized with xylene, the tissues sections were mounted by neutral balsam. All micrographs were taken with a Nikon Digital DS-Ri1. For scanning electron microscopy, the young flower buds were fixed with 2.5% glutaraldehyde in 0.05 M phosphate buffer (pH 6.8) at room temperature for 2 h, then post-fixed in 1% osmium tetroxide in the same buffer at 4 • C for 1 h. The samples were dehydrated with a graded ethanol series, then critical Restorer Search for inap CMS The restorer lines of different CMS systems were used to test their ability to restore the fertility of inap CMS ( Table 2). The seeds of Ogura and pol CMS lines and restorer lines were kindly provided by Prof. Jiangsheng Wu from our University. B. napus cv. Yudal was reported to restore Tournefortii-Stiewe CMS with recombinant mtDNA of both B. tournefortii and B. napus (Dieterich et al., 2003). The disomic rape-radish addition line (DAL-f) (2n = 40, AACC+1II R ) was selected, for it acted as the restorer for a novel CMS of B. napus with Raphanus cytoplasm and also carpel-like stamens (Budahn et al., 2008). However, the addition with the chromosome F did not restore Ogura CMS of B. napus. One B. napus monosomic addition line Me carrying one chromosome of Chinese woad and the same cytoplasm of inap CMS was used, as it produced anthers with viable pollen grains (Kang et al., 2014). After inap CMS line was pollinated by these different lines by hand, the seeds were harvested and F 1 plants were assessed for the degree of male sterility. Phenotype of inap CMS in B. napus The inap CMS of B. napus was selected from the somatic hybrids with Chinese woad by successive backcrossing to recover the B. napus chromosome complement (Figure 1). The male sterility was attributed to the developmental conversion of its tetradynamous stamens into the carpelloid structures with stigmatoid tissues at the tips and ovule-like tissues in the margins, and two shorter stamens into filaments (Figures 2A-E). The two carpelloid structures at the same side fused together in some flowers, and still remained after the flower faded and as the pods developed ( Figure 2F). The size and developmental complex of carpelloid structures diminished after successive backcrosses. The petals and sepals of its flowers were significantly smaller than the donor B. napus cv. Huashuang 3 (Table 3), and the number of nectaries reduced to 2.4 from 4 of Huashuang 3 due to the fused carpelloid stamens. Although the pistil length of the flowers was close to Huashuang 3, the seeds per pod (16.1) of inap CMS after pollination by Huashuang 3 were fewer than those in Huashuang 3 (23.5) by selfing (Table 3), likely some ovules were inviable (24.9/30.9). The negative effect of the cytoplasm from inap CMS on the defect of female development might cause the yield penalty at certain degrees, and optimal combinations needed to be screened for its correction. The male sterility phenotype was expressed consistently during recurrent backcrosses, and in Wuhan and Xining for 3 years, indicating its genetic and environmental stability. Then the male sterility was also easily maintained by Huashuang 3 and all other B. napus genotypes used (see below), as inap CMS had good seed-sets and still produced the male sterile progenies after pollinated by them. The plants of inap CMS grew more slowly and flowered later almost 2 weeks than Huashuang 3 in Wuhan, but they show no leaf chlorosis, likely for they maintained the chloroplast genetic component from Huashuang 3 (Du et al., 2009;Kang et al., 2014). But death of flower buds occurred at the initial stage of flowering, and later the flower buds developed and opened normally. Stamen Development of inap CMS The stages of floral development in Arabidopsis thaliana were defined by Smyth et al. (1990) using scanning electron microscopy. With this standard, the early flower developmental processes of inap CMS line and its maintainer line Huanshuang 3 were observed and compared. They showed no differences until the stamen primordium appeared (Figures 3A,B,E,F). Subsequently, the stamen primordium developed into stamen in the maintainer (Figures 3C,D), while the stamen converted into the carpelloid structure in the CMS (Figures 3G,H). The ovulelike structure appeared on the carpelloid tissues in the CMS line at the stage 9 (data not shown). By histological studies, disordered cell division occurred at the sites of the stamen primordium in the flower buds at stage 4, and petal primordium would arise (Figures 3I-P). The stamen primordium developed into the carpelloid structure as a result of disorganized cell proliferation. So the male sterility of inap CMS initiated at the stage of stamen primordium differentiation. Nuclear and Cytoplasmic Composition of inap CMS With randomly selected 23 pairs of AFLP selective primers, no I. indigotica genomic bands were found in inap CMS line, showing no woad chromosomal segments introgressed after successive backcrosses with B. napus. It has been shown that the somatic hybrid contained the recombinant mitochondrial genome from B. napus and I. indigotica (Kang et al., 2014). As most of mitochondrial genes had no fragment length differences between parents and the hybrid from PCR amplification, 25 mitochondrial genes in the CMS line were sequenced and found that 20 of them originated from I. indigotica (Table 4). Only the genes atp9, cox3, rps7, and rrn26 showed no sequence differences between parents and CMS line. Several SNPs were identified at the first exon and the intron of cox2-2 between CMS and B. napus, while one same band was detected by PCR amplification using its primers. Two copies of cox2 genes were reported to exist in Brassica species, except for B. nigra and B. carinata (Handa, 2003;, which showed a consistent sequence of the first exon and intron and diverged from each other at the second exon. The cox2-1 was homologous to the mitochondrial cox2 genes of other plants, but cox2-2 had an extension with no homology to any other sequences. To obtain the full sequence of cox2-1 and cox2-2 of inap CMS and parents, the primers were redesigned (Table 1 and Figure 4A). Only one copy of the cox2 (GenBank: KY656165) was obtained from I. indigotica, having 99.77% sequence identity with cox2-1 of B. napus (Figure 4B and Supplementary Figure S1). Multiple sequence alignment indicated that the first exon and intron of the cox2-2 of the CMS was from the cox2 of I. indigotica, but the second exon was derived from the cox2-2 of B. napus (Figure 4 and Supplementary Figure S1). RT-PCR analysis showed that this cox2-2 was transcribed in inap CMS flower buds, but the expression level was very low (Supplementary Figure S2). To examine the effect of nuclear restorer gene on the expression of the cox2-2, the expression level was compared between the fertile B. napus addition line Me and inap CMS, but showed no significant difference (Supplementary Figure S2). To determine whether inap CMS was different from Ogura and pol CMS in sterility genes, the primers of their respective CMS-associated genes orf138 and orf224 were used for PCR amplification, but did not produce the products, showing that inap CMS did not share these genes (Supplementary Figure S3). Different Fertility Restoration Relationship of inap CMS from Other CMS Systems As shown in Table 2, all the F 1 hybrids between inap CMS and these restorers of various CMS systems were male sterile and still produced the carpelloid stamens, except for those with the B. napus addition line harboring the woad chromosome E. Although the two alloplamic Ogura CMS lines had the similar phenotype of inap CMS, their restorers with the introgression of restoration gene from one specific radish chromosome F failed to restore our inap CMS. Furthermore, the hybrids between inap CMS and the disomic rape-radish chromosome addition line f (DAL-f) (2n = 39, AACC+1R) with the radish chromosome carrying the restoration gene still displayed male sterility. To identify restorers of inap CMS in B. napus, 112 cultivars of wide origins were used as pollinators (data not shown) and the pods with good seed-sets were produced, but none was found to function as the restorer. DISCUSSION The novel inap CMS in B. napus selected from the intertribal somatic hybrids contained the recombinant mtDNA of both parents but biased to woad, and had the stamens converted into carpelloid structures. The sterile phenotype was stable during years and under different ecological environment. The developmental conversion of stamens occurred at the stage of stamen primordium differentiation. It was different from other CMS lines in phenotype, CMS associated genes, and restorer and maintainer relationships. Homeotic conversion of stamens into carpelloid structures were described in several CMS lines of wheat (Ogihara et al., 1997), tobacco (Farbos et al., 2001), carrot (Linke et al., 2003), rapeseed (Leino et al., 2003), cauliflower (Kamiñski et al., 2012), and broccoli (Shu et al., 2015). Stamens of the CMS lines were replaced by carpelloid organs, thus resembled the B-class genes APETALA3 (AP3) and PISTILLATA (PI) mutants. In B. napus CMS line with the rearranged Arabidopsis mitochondria, downregulation of AP3 in the stamen primordia and subsequent repression of PI resulted in the conversion of stamens into carpelloid organ (Carlsson et al., 2007a). Similarly, the B-class genes were transcriptionally downregulated in carpelloid CMS flowers of tobacco, carrot and wheat (Farbos et al., 2001;Linke et al., 2003;Hama et al., 2004). The mitochondrial background had a distinct influence on nuclear gene expression (Carlsson et al., 2007b). This indicated that CMS-associated mitochondria genes regulated the expression of these nuclear MADS-box genes by mitochondrial retrograde signaling (MRS). Wheat Calmodulin-Binding Protein 1 (WCBP1) significantly upregulated in young spikes of the pistillody line (Yamamoto et al., 2013), indicating that CMS-associated genes played a role in development of pistillike stamens by MRS involved Ca 2+ signaling pathway in wheat CMS. The most mitochondrial genes of inap CMS were found to originate from I. indigotica (Table 4), which was likely caused by the asymmetric fusion to produce the hybrids (Du et al., 2009). The protoplasts of B. napus were treated by 3 mM iodoacetate (IOA), and those of I. indigotica were irradiated with UV (Du et al., 2009). IOA could inactivate the glyceraldehyde-3-phosphate dehydrogenase, and resulted in the prevention of the glycolytic pathway (Zhao et al., 2008). Only when B. napus protoplasts for fusion contained

the normal cytoplasm, the fused cells could divide, formed calli and regenerated plants. Thus, the mitochondria from I. indigotica were retained, and those from B. napus eliminated but several segments were introgressed into the mitochondria of the hybrid by recombination (Table 4). Such situation was also reported in other asymmetric somatic hybrids (Morgan and Maliga, 1987;Akagi et al., 1995). MtDNA sequencing of an Ogura-CMS cybrid derived from somatic fusions between B. napus and sterile radish showed that the mtDNA component mainly inherited from radish (Wang et al., 2012). The B. napus origin of the chloroplast DNA of inap CMS was in accord with the conclusion that the chloroplasts were generally from the iodoacetate-treated parent (Morgan and Maliga, 1987;Leino et al., 2003). This might be due to the better compatibility of recipient chloroplasts with the recipient nucleus, or the loss of the irradiated chloroplasts. Two copies of cox2 gene in Brassica species (Handa, 2003; arose from duplication, for it was involved in a large repeat sequence. Via this large repeat, the Brassica mitochondrial genome could be recombined into two subgenomic circles. I. indigotica contained only one copy of this gene. Of two copies of cox2 gene in inap CMS, cox2-1 was from I. indigotica, but novel cox2-2 was the recombinant between cox2 of woad and cox2-2 of rapeseed (Figure 4 and Supplementary Figure S1). Rearrangements in the cox2 region were also reported in somatic hybrids between B. napus and A. thaliana (Landgren and Glimelius, 1994;Leino et al., 2003), B. napus and R. sativus Kosena CMS (Sakai and Imamura, 1990). CMS-associated genes identified in many CMS systems (Wang et al., 1995;Chen et al., 2011;Tanaka et al., 2012;Heng et al., 2014) were located in flanking sequence of the known mitochondrial genes, and co-transcribed with it. Protoplast fusion offered an opportunity for mitochondrial genome recombination, leading to appearance of novel CMS genes. Three atp9 genes were present in the alloplasmic Tournefortii-Stiewe CMS of B. napus, which was generated by protoplast fusion with B. tournefortii (Dieterich et al., 2003). MtDNA rearrangements happened in the upstream of one of these genes and generated a chimeric gene orf193 that was co-transcribed with atp9. An alloplasmic B. juncea CMS line derived from somatic hybridization with Diplotaxis catholica had two copies of coxI gene, and the recombinant coxI-2 gene was suggested to cause CMS (Pathania et al., 2007). For the novel cox2-2 gene was expressed in the inap CMS flowers weakly and insignificantly from restorer plant, it was likely not responsible for male sterility. The identification of the restorer and maintainer relationship is one of the most classical methods to distinguish different types of sterile cytoplasm (Wan et al., 2008). The test results showed that the fertility of inap CMS failed to be restored by restorer lines from Ogura, pol, nap and Tournefortii-Stiewe CMS systems. Budahn et al. (2008) found that rapeseed plants with the alien radish cytoplasm showed pistilloid stamens with ovules and stigmatoid tips, and the radish chromosome F could eliminate the disadvantageous effect of alien cytoplasm in B. napus. Additionally, the chromosome F did not restore Ogura CMS. In spite of the similar phenotype with inap CMS, the radish chromosome F was unable to restore its fertility, indicating that these two CMS lines had different CMS-associated genes and mechanisms of male sterility. No restorers were screened among 112 B. napus cultivars. This further testified that the fertility restoration gene(s) of alloplasmic CMS lines generally existed in nuclear genome of cytoplasm donor species. The restoration gene of a B. napus CMS line with rearranged A. thaliana mtDNA was located on chromosome III of A. thaliana (Leino et al., 2004). The restorer line for Nsa CMS of B. napus developed by the somatic hybridization with Sinapis arvensis was one disomic addition line with the restoration gene on the alien chromosome (Wei et al., 2010). Similarly, the restoration gene(s) for our inap CMS also existed on one specific woad chromosome, because only the monosomic alien addition line of B. napus with woad chromosome e among the complete set of such lines and also the same cytoplasm as inap CMS produced the normal flowers with fertile pollen grains (Kang et al., 2014). To breed the restorer line for inap CMS, the chromosomal segment carrying the fertility restoration gene(s) should be introgressed into one of rapeseed chromosome and the lines (2n = 38) with high restoration capability selected, as excellently exemplified by the long-term development of the radish introgression carrying the Rfo restorer gene for the Ogu-INRA CMS in B. napus (Delourme et al., 1991(Delourme et al., , 1998Primard-Brisset et al., 2005). The identification of some progeny plants (2n = 38) with the restored flower development and pollen fertility from the addition line gave the expectation that the new CMS and fertility restoration system should be developed for rapeseed hybrid production in near future. AUTHOR CONTRIBUTIONS ZL and LK conceived and designed the experiments, analyzed the data and wrote the paper. LK performed the experiments. PL participated in mtDNA analysis. AW performed the histological section experiments. XG provided technical expertise for molecular analysis and scientific discussions. Defining and Treating Older Adults with Acute Myeloid Leukemia Who Are Ineligible for Intensive Therapies Although acute myeloid leukemia (AML) is primarily a disease of older adults (age ≥60 years), the optimal treatment for older adults remains largely undefined. Intensive chemotherapy is rarely beneficial for frail older adults or those with poor-risk disease, but criteria that define fitness and/or appropriateness for intensive chemotherapy remain to be standardized. Evaluation of disease-related and patient-specific factors in the context of clinical decision making has therefore been largely subjective. A uniform approach to identify those patients most likely to benefit from intensive therapies is needed. Here, we review currently available objective measures to define older adults with AML who are ineligible for intensive chemotherapy, and discuss promising investigational approaches. Patient-specific factors also contribute to outcomes independent of AML characteristics. For example, worse performance status (10,17,18) and the presence of comorbid conditions have been associated with increased mortality and decreased response rates in this population (19,20). The tendency to manage older adults with less intensive measures may contribute to worse outcomes. Several studies have demonstrated improved survival for older patients receiving intensive induction chemotherapy compared to those receiving supportive care alone (2,21). In the United States, however, <40% of older adults with AML receive chemotherapy for their disease (3). These data suggest a need for an improved understanding of factors that define ineligibility for an intensive treatment approach. Defining this subset of patients who are not eligible for intensive therapy involves a great deal of subjectivity, and criteria have yet to be standardized across or within institutions. This review will focus on factors that should be taken into consideration to determine eligibility for an intensive treatment approach in AML and evolving treatment strategies, including investigational approaches, for older adults considered less fit for intensive induction therapy. FACTORS THAT DeTeRMiNe eLiGiBiLiTY FOR iNTeNSive iNDUCTiON CHeMOTHeRAPY Physical Performance Physical performance can be used to help predict outcomes in older patients with AML who are treated with induction chemotherapy. Methods available to quantitatively assess physical performance include the Eastern Cooperative Oncology Group performance status (ECOG PS), the Karnofsky performance status (KPS), and the short physical performance battery (SPPB). Retrospective analysis of data from clinical trials of patients treated with an intensive induction chemotherapy approach showed that in patients older than 65 years with poor ECOG PS of 2 or 3, outcomes declined drastically with age. For example, among patients with an ECOG PS of 3, the likelihood of early death increased from 0% in those <56 years to 29% in patients 56-65 years, and 82% in patients >75 years. However, for those with ECOG PS of 0-1, age appeared to have only a modest effect on the incidence of early death after induction chemotherapy (10). Another retrospective analysis of 998 patients age 65 years or older who underwent induction chemotherapy reported 8-week mortality rates of 23, 40, and 72% for patients with ECOG PS of 0-1, 2, and 3-4, respectively. The same groups had 1-year overall survival rates of 35, 25, and 7%, respectively (22). Similarly, the KPS has been shown to help predict outcomes in older patients (17,23). The SPPB ( Table 1) is another objective measure of physical performance and has been shown to predict future disability, hospitalizations, and mortality among elderly patients in general, with or without a malignancy. The test is relatively simple to perform in the clinic in only a few minutes' time and includes measures of balance, gait speed, and time to rise from a chair. Scores range from 0 through 12, with a score of 12 representing the most physically fit patient. A single-center study showed an association between lower SPPB score and increased risk of death specifically in patients older than 60 years with newly diagnosed AML undergoing intensive induction therapy. All evaluated patients had a reported EGOG PS of 0-1 at the time of evaluation. Those with SPPB scores <9 had a shorter median survival than those with scores >9 (6 versus 16.8 months, respectively). When analyzed as a continuous variable, each 2-point increase in SPPB score was associated with a 15% decrease in hazard ratio for death. This study showed that the SPPB is a valuable tool to further risk-stratify those with good ECOG PS who may have a lower functional reserve (20). Comorbid Conditions Comorbid conditions should also be taken into account when discussing AML management in older adults, as they portend worse prognoses and increased toxicity for patients undergoing intensive induction chemotherapy. Either the Charlson comorbidity index (CCI) or the hematopoietic cell transplantation-specific comorbidity index (HCT-CI) can be used to measure comorbid conditions quantitatively. Neither of these indices was initially designed for use in older patients with AML, but both have been studied in this population with varying results. The CCI assigns point values for certain comorbid conditions, some of which are stratified for severity. The original CCI has been revised slightly for use in older adults with AML. A single-center retrospective study showed that patients with a CCI score >1 had a significantly lower chance of attaining a complete remission (CR) than those with a score of 0 or 1 (35 versus 63%). The group with higher scores also showed a trend toward higher 8-week mortality and lower 2-year survival (19,20). The HCT-CI ( Table 2) was developed to improve the sensitivity of the CCI in the stem cell transplant setting, but has been evaluated as a tool to predict outcomes with intensive induction chemotherapy for AML as well. A retrospective study of 177 patients over the age of 60 years receiving induction chemotherapy for AML showed that HCT-CI scores of 0, 1-2, and >2 corresponded to early death rates of 3, 11, and 29%, respectively. The same groups had median overall survival times of 45, 31, and 19 weeks (24). A single-center study demonstrated that HCT-CI score ≥3 in older patients with AML was the single most significant predictor of overall survival and early death, even outweighing karyotype in that study (25). Cognitive Function Cognitive function should not be overlooked when considering treatment options for older adults with AML, as pretreatment cognitive impairment may increase the risk of complications during and after intensive therapy for AML (20). Data in this area are limited, but a few small studies have shown that cognitive impairment is common in this population and is an independent predictor of outcome. One study with a mean age of 70.8 years found that 31.5% of their patients had cognitive impairment at the time of diagnosis of AML (17). Another study from the same group showed that older patients with AML receiving induction chemotherapy with a modified mini-mental state exam score of <77 out of 100 had a median overall survival of 5.2 months compared to 15.6 months in those with a score ≥77 (20). Prognostic Models Several prognostic models have been developed to risk-stratify and predict outcomes of patients undergoing induction chemotherapy based on patient and disease characteristics. An analysis of 2483 patients age 60 years or older enrolled in two UK trials showed that cytogenetic group, age, white blood cell count, performance status, and type of AML (de novo or secondary) were all associated with outcome in patients treated with either intensive or

non-intensive regimens. When these factors were used to stratify patients into good, standard, and poor-risk groups, the 1-year survival rates were 53, 43, and 16%, respectively (26). In a single-institution study of 998 patients age 65 years or older with AML treated with intensive chemotherapy, significant predictors of outcome were age ≥75 years, unfavorable cytogenetics, ECOG PS >2, antecedent hematologic disorder, lactate dehydrogenase (LDH) >600 IU/L, elevated creatinine, and treatment outside of a laminar air flow room. They went on to devise a scoring system based on the number of poor prognostic factors present. Those with none of the above risk factors had >60% CR rates, induction mortality of 10% within 8 weeks of treatment, and 1-year survival over 50%. This favorable group accounted for 20% of their study population. However, those with three or more risk factors had CR in <20%, induction mortality of >50%, and 1-year survival of <10%. This high-risk group accounted for 25-30% of their sample size (22). Another prognostic model comes from a study of over 1400 older patients with AML who were otherwise healthy and were treated on a clinical trial with standard induction chemotherapy. This tool uses a formula including variables such as body temperature, hemoglobin, platelet count, LDH, age, type of AML (de novo or secondary), fibrinogen level, and molecular and cytogenetic features of the disease to predict probabilities for response and early mortality (27). In a study of over 900 patients over the age of 60 years with AML who received standard induction chemotherapy followed by one cycle of consolidation, independent predictors of survival included karyotype, CD34 expression, white blood cell count at diagnosis, age, LDH, and nucleophosmin 1 (NPM-1) status. Karyotype was, by far, the most significant predictor of survival. Those with favorable risk cytogenetics fared the best, regardless of other factors, with 3-year overall survival rates of about 40%, while those with poor-risk cytogenetics had a dismal 3-year overall survival of only 3%. With this in mind, the authors devised a prognostic score to better define the risk for those in the intermediate cytogenetic category. Those with (28). At the moment, there is no consensus regarding a uniform set of guidelines that affirm fitness for intensive induction chemotherapy. The aforementioned prognostic scoring models, physical performance evaluation, comorbidity indices, and cognitive assessments can guide decision making in a more objective manner. However, validated guidelines are needed to standardize our treatment approaches globally. Several proposed guidelines have arisen from expert opinion and objective data, but no one algorithm has emerged as the standard in patient care. One such guideline developed by the Italian Society of Hematology (SIE), Italian Society of Experimental Hematology (SIES), and Italian Group for Bone Marrow Transplantation (GITMO) uses age, performance status, and comorbid burden to define fitness for intensive or non-intensive therapies (29). Table 3 demonstrates another evolving set of criteria for fitness, vulnerability, and frailty based on performance status, comorbidity assessment, and cognitive assessment that was recently proposed based on review of available evidence (30). Preliminary results of a separate consensus guideline, based on several patient-specific criteria and validated in a retrospective evaluation of 362 patients diagnosed and treated at multiple centers, were recently presented. This study demonstrated that the proposed criteria were able to predict for overall survival, regardless of the treatment modality. When combined with European LeukemiaNet (ELN) risk criteria (31), this model was able to identify a subgroup of fit, low/ intermediate-I risk patients who did relatively well with a median overall survival of 20 months. Fit patients with intermediate-II risk or higher fared significantly worse, with a median overall survival of 8.5 months (32). This underscores the fact that these proposed tools still require the clinician to consider the patient's fitness in the context of the disease biology. Some fit older patients with the highest risk disease may not derive sufficient benefit from standard induction chemotherapy to outweigh the risks, and these patients may be best served by consideration of alternative novel therapeutic strategies. A more uniform stratification of both fitness of the older patient for chemotherapy and appropriateness of that therapy in the context of disease biology would help inform clinical decision making as well as facilitate clinical trial design. THeRAPeUTiC STRATeGieS FOR PATieNTS wHO ARe UNFiT FOR STANDARD iNDUCTiON CHeMOTHeRAPY Treatment options for patients deemed ineligible for intensive induction chemotherapy are few. Possible approaches may involve clinical trial participation, lower-intensity chemotherapeutics such as DNA hypomethylating agents, or supportive measures alone. DNA Hypomethylating Agents The DNA hypomethylating agents decitabine and azacitidine are commonly used to treat this population. Both are approved by the US Food and Drug Administration (FDA) for the treatment of myelodysplastic syndrome (MDS). Decitabine is also approved by the European Medicines Agency (EMA) for AML, and azaciditine is approved for AML with 20-30% bone marrow blasts that arose from MDS. Both drugs are generally well tolerated and can provide some benefit in certain older patients with AML. Decitabine has been investigated in the frontline setting in older adults with AML. A phase III, multicenter study was performed for patients >65 years with newly diagnosed AML comparing decitabine administered on a 5-day schedule in 28-day cycles to conventional care, which consisted of either best supportive care or low-dose cytarabine. The decitabine cohort demonstrated significantly higher CR rates (17.8 versus 7.8%). There was a trend toward increased median overall survival (7.7 versus 5 months, p = 0.11) that did not reach statistical significance. After another year of follow-up, the survival difference between the two groups did reach significance (p = 0.037) (33). The EMA approved decitabine for older adults with AML based on these data. However, the FDA declined a decision that has been criticized by some as representative of overly stringent statistical analysis (34). A phase II single-institution study of 53 older patients with AML not eligible for intensive therapy suggested that a higher CR rate can be obtained when decitabine is given for 10 consecutive days as opposed to 5-day MDS-like regimens. In fact, an impressive 47% of their patients attained a CR, and an additional 17% had no morphologic evidence of disease but had incomplete count recovery. Of those who achieved a CR, the median time to response was three cycles. One-year survival of even the poor-risk patients in this study was 30% (35). The results from this trial have led to the hypothesis that a 10-day schedule of administration may be more active for this agent in patients with AML and has led to further investigation of this schedule of decitabine in a recent cooperative group trial conducted in patients with AML >60 years of age (36). Azacitidine has shown clinical activity in older patients with AML. In a phase III study (CALGB 9221) patients with MDS [45 had AML by current World Health Organization (WHO) criteria] were randomized to either azacitidine or best supportive care, with crossover to azacitidine permitted at the time of disease progression. The azacitidine arm demonstrated significantly improved response rates (60% overall response in the azacitidine group including 7% CR versus 0% in the supportive care arm). The response rates were similar between patients with MDS and AML in a subgroup analysis. The azacitidine group also reported significantly better quality of life measures, including fatigue, dyspnea, physical functioning, and psychosocial stress (37). In a landmark phase III study, AZA-001, patients were randomized to either azacitidine daily for 7 days of a 28-day cycle or a predefined, investigator's choice conventional care regimen, which included best supportive care, low-dose cytarabine, or intensive induction chemotherapy. Most of the enrolled patients had MDS, but about one-third met WHO criteria for AML, with 20-30% blasts. A survival advantage was demonstrated for the azacitidine arm of the trial overall including the subgroup with WHO-defined AML. In that subgroup, median overall survival in the azacitidine arm was 24.5 versus 16 months in the conventional care regimen arm (38). Preliminary results of the AZA-AML-001 study were recently presented. This phase III, multi-institution study compared azacitidine to conventional care regimens including intensive induction chemotherapy, low-dose cytarabine, or best supportive care in patients ≥65 years with AML and blast count >30%. The primary endpoint of median overall survival was 10.4 months in the azacitidine arm versus 6.5 months in the conventional care arm, which did not quite achieve statistical significance (p = 0.0829). There was a trend toward an improvement in the 1-year overall survival rates in the azacitidine arm as well (47 versus 34%) (39). Emerging data suggest that certain subsets of patients may be more likely to respond to hypomethylating therapy. In an updated subgroup analysis of the AZA-AML-001 study, patients with morphologic dysplastic changes treated with azacitidine had twice the median overall survival than their morphologically similar counterparts treated with a conventional care regimen (12.7 versus 6.3 months, p = 0.0357) (40). There was some initial evidence that hypomethylating agents may be more effective in AML characterized by DNMT3A mutations; however, follow up studies were conflicting (41,42). Recent reports also demonstrated that patients with TET2 mutations are more sensitive to treatment with hypomethylating agents (43,44). Further studies examining biological factors predicting response to epigenetic therapies are necessary and are ongoing. Low-Dose Cytarabine Low-dose cytarabine represents another available option outside of a clinical trial for patients unfit for intensive therapy and remains a frequently used comparator or combination partner in clinical studies in this patient population (45). In a multicenter phase III trial, 217 patients with AML or high-risk MDS deemed unfit for intensive therapy were randomized to receive either lowdose cytarabine 20 mg twice daily for 10 days or hydroxyurea. The low-dose cytarabine group had a higher CR rate (18 versus 1%) and improved overall survival with an odds ratio of 0.60. Gemtuzumab Ozogamicin In recent years, there has been a concerted effort to develop novel agents with better efficacy and toxicity profiles particularly for those patients who are considered unfit for standard induction approaches. Gemtuzumab ozogamicin (GO) is an antibody-drug conjugate that consists of an anti-CD33 antibody linked to calicheamicin. GO was granted accelerated FDA approval for patients with CD33 + AML in first relapse who were not candidates for cytotoxic chemotherapy based on several open-label studies showing improved outcomes in this population (46)(47)(48). However, the confirmatory SWOG 106 study, which involved the addition of GO to standard induction therapy in a separate population, namely untreated adults ≤60 years old, found an increase in 30-day mortality in this population which prompted the voluntary withdrawal of GO from the market, and thus it is no longer routinely available to older patients with AML (49). Several subsequent studies focusing on older patients with AML have shown improved outcomes when GO is added to conventional therapy. In one randomized study comparing induction chemotherapy alone versus induction chemotherapy plus GO in older patients ranging from 51 to 84 years old with AML, there was no difference in response rates, early mortality, or toxicities between the two groups, but at 3-year follow up, there was a decreased relapse rate (68 versus 76%, p = 0.007) and improved survival (25 versus 20%, p = 0.05) in the group who received GO (50). Another trial randomized 495 older patients with AML ranging from 54 to 90 years old who were deemed inappropriate candidates for intensive therapy to low-dose cytarabine with or without GO. The addition of GO resulted in significantly improved remission rates (30 versus 17%, p = 0.006), but no improvement in mortality at 12 months (51). The role of GO as post-remission therapy for older patients has also been investigated. A phase III multicenter study randomized patients over the age of 60 years in remission after intensive therapy to receive either three cycles of GO or no post-remission therapy. They found no difference in relapse rates, disease-free survival, or overall survival (52). The ultimate fate of this agent has yet to be determined, but many experts have vocally advocated for its reinstatement in our treatment armamentarium (53)(54)(55). NOveL AGeNTS UNDeR iNveSTiGATiON In recent years, there has been an increasing focus on molecularly targeted therapies in oncology, and AML is no exception. Several targeted small molecule inhibitors are under investigation for older patients with AML who are not fit for standard induction therapy. In general, these agents are hypothesized to be less toxic than traditional chemotherapy, and as such could be

useful in specific molecular subsets of AML in less fit older adults either as single agents or in rationally designed combinations in the near future ( Table 4). Agents That Target the Microenvironment/ Leukemia Stem Cell PF-04449913 is an oral agent designed to inhibit the hedgehog signaling pathway, which has been shown to be aberrantly activated in AML cells. It is currently undergoing phase I/ II trials in combination with induction chemotherapy in fit patients, and with either low-dose cytarabine or decitabine in unfit patients (56). Agents That Target Dysregulated Kinases or Signaling Pathways Quizartinib, an FMS-like tyrosine kinase 3 (FLT3) receptor tyrosine kinase inhibitor, has shown promising results in patients with relapsed refractory AML, particularly those patients harboring an FLT3-internal tandem duplication (ITD). In a phase I study, responses [including CR, CR with incomplete platelet recovery (CRp), CR with incomplete hematologic recovery (CRi) (74), and partial remission (PR)] were seen in 30% of all patients and 53% of FLT3-ITD-positive patients treated with quizartinib (75). Preliminary results of a phase II trial of patients ≥18 years old with relapsed or refractory AML showed an encouraging composite CR rates (CR + CRp + CRi) of 44%, with nearly one-third of patients successfully bridged to stem cell transplant (60). Phase I/II studies combining quizartinib with azacitidine or low-dose cytarabine are ongoing (61). Quizartinib is an oral agent and has been generally well tolerated, making it an exciting prospect for less fit patients with AML. Another multi-targeted tyrosine kinase inhibitor, ASP2215, is also in ongoing phase I trials (62). Volasertib, a small molecule inhibitor of the polo-like kinase 1 (PLK1) protein, was combined with low-dose cytarabine and compared to low-dose cytarabine alone in a randomized, phase II trial for unfit patients with AML. The combination arm showed improved CR + CRi (31 versus 13.3%, p = 0.052) (57). The phase III trial is ongoing (58). GSK2141795, a novel agent that blocks the phosphoinositide 3-kinase (PI3K)/AKT signaling pathway, is being studied in a phase II trial of unfit patients with RAS-mutated AML in combination with the MEK inhibitor trametinib (59). Studies with Pevonedistat, a first-in-class inhibitor of NEDD-8 activating enzymes (NAE) in combination with azacitidine are ongoing. In phase I, this combination was safe and generally well tolerated. Among the 18 patients evaluable for response, a 56% ORR was reported (63). Phase II is underway. epigenetic Therapies Besides the DNA hypomethylating agents discussed above, other epigenetic therapies such as histone deacetylase inhibitors are also of interest in this population, and combination approaches are being evaluated (76). Phase II studies combining pracinostat with azacitidine in older adults with AML are accruing (65), and preliminary results are encouraging (77). The role of valproic acid, perhaps in combination with all-trans retinoic acid and other epigenetic modifiers, has yet to be fully elucidated (66). Other combinations of epigenetic therapy with different modalities continue to be studied. Examples include the topoisomerase-II inhibitor vosaroxin plus decitabine (68,78) and bortezomib plus decitabine (79). Mutations in IDH1 and IDH2, which inhibit TET2 enzymatic function and thereby result in DNA hypermethylation, also represent novel epigenetic targets. Phase I studies of oral IDH1 and 2 inhibitors in patients with the respective mutations are ongoing (69,70). DOT1L, a histone methyltransferase that plays a critical role in leukemic transformation induced by MLL rearrangements, is also a promising therapeutic target that is being investigated in early phase trials in MLL-rearranged leukemias (71,80). Novel analogs and/or novel formulations of existing DNA methyltransferase inhibitors are also under active investigation in AML (72). Managing AML in Unfit Patients Frontiers in Oncology | www.frontiersin.org Other Novel Agents Selinexor, an oral selective inhibitor of nuclear export (SINE), is currently being investigated in a phase II randomized trial for unfit patients as a single agent versus physician's choice, which includes hypomethylating agent or supportive care (73). There are also ongoing early phase trials investigating this agent in combination with chemotherapy or hypomethylating agent therapy. Of course, the importance of supportive measures for patients undergoing less intensive therapy or no therapy should not be overlooked. Prophylactic antimicrobials, transfusion of blood products as needed, and hydroxyurea if needed for cytoreduction can all be utilized in an effort to reduce hospitalization rates in older patients with AML with the hope of improving quality of life. CONCLUSiON As the world's population continues to age, the number of people diagnosed with AML each year can be expected to rise, adding urgency to the need for more effective and less toxic therapies for older, less fit adults. Available therapies outside of the realm of clinical trials are few. Clinical trial participation should be considered the standard of care for unfit patients and patients with high-risk disease whenever possible. Developing and validating uniform definitions for risk stratification according to fitness and integrating this within the context of disease biology are of utmost importance with regard to the design, implementation, and interpretation of clinical trial data in this patient population. The criteria that define patients unfit for intensive induction chemotherapy are currently evolving and require validation. Therefore, at the present time, we recommend that clinicians incorporate the currently available tools described herein plus patient preferences in the development of treatment strategies for the individual patient. AUTHOR NOTeS KP is a fellow in the section of hematology and oncology as well as clinical pharmacology at the University of Chicago. OO is an Associate Professor of Medicine in the section of Hematology and Oncology at the University of Chicago. AUTHOR CONTRiBUTiONS KP and TO both performed the literature review and authored the manuscript. The authors meet criteria for authorship as recommended by ICMJE. The authors received no direct compensation related to the development of the manuscript. All authors have read and approved the final manuscript. FUNDiNG KP was supported by grants from the National Institutes of Health/National Institute of General Medical Sciences Clinical Therapeutics grant (T32 GM007019) and the Basic Research Training in Medical Oncology grant (T32 CA009566). Editorial assistance provided by GeoMed, an Ashfield business, part of UDG Healthcare plc, was contracted and funded by Boehringer Ingelheim Pharmaceuticals, Inc. (BIPI). BIPI was given the opportunity to check the data for medical and scientific accuracy as well as intellectual property considerations. Elevated levels of inflammatory plasma biomarkers are associated with risk of HIV infection Background To determine if individuals, from HIV-1 serodiscordant couple cohorts from Rwanda and Zambia, who become HIV-positive have a distinct inflammatory biomarker profile compared to individuals who remain HIV-negative, we compared levels of biomarkers in plasma of HIV-negative individuals who either seroconverted (pre-infection) and became HIV-positive or remained HIV-negative (uninfected). Results We observed that individuals in the combined cohort, as well as those in the individual country cohorts, who later became HIV-1 infected had significantly higher baseline levels of multiple inflammatory cytokines/chemokines compared to individuals who remained HIV-negative. Genital inflammation/ulceration or schistosome infections were not associated with this elevated profile. Defined levels of ITAC and IL-7 were significant predictors of later HIV acquisition in ROC predictive analyses, whereas the classical Th1 and Th2 inflammatory cytokines such as IL-12 and interferon-γ or IL-4, IL-5 and Il-13 were not. Conclusions Overall, the data show a significant association between increased plasma biomarkers linked to inflammation and immune activation and HIV acquisition and suggests that pre-existing conditions that increase systemic biomarkers represent a factor for increased risk of HIV infection. Supplementary Information The online version contains supplementary material available at 10.1186/s12977-021-00552-6. heterosexual HIV-discordant couple cohort in Lusaka, Zambia, it was found that Schistosome infections were linked to increased HIV-1 transmission in both sexes, increased acquisition of HIV-1 in women, and increased progression to death in HIV-positive women [18]. In a large study of at-risk HIV-negative women in South Africa it was observed that higher levels of inflammatory cytokines in the female genital tract of individuals prior to infection was associated with increased HIV susceptibility [6], allowing a more defined approach to quantitating this risk factor for acquisition. It is clear that in women the composition of the genital microbiome can also greatly influence the genital tract inflammatory state. Consistent with this notion, a study of the FRESH cohort in Durban, South Africa, showed that women who presented with Lactobacillus-deficient microbiota in their reproductive tract produced higher levels of inflammatory cytokines [19]. A follow-up study showed that these women had a greater risk of HIV acquisition compared to women with Lactobacillus crispatus-rich genital microbiota [20]. However, the genital cytokine levels observed were not correlated with those in the plasma [20]. In this study, we examined systemic plasma biomarker levels in individuals who would eventually seroconvert (pre-infection) and those who remained HIV-negative (uninfected) in a cohort of serodiscordant couples from both Zambia and Rwanda. We compared these two groups to determine if there was any association between systemic plasma biomarker levels and increased acquisition of HIV in HIV-negative individuals, and whether those biomarkers were associated with any pre-existing urogenital infections. Elevated systemic plasma cytokine and chemokine levels characterize individuals prior to infection A major goal of this study was to determine whether prior to infection seroconverting partners in the two discordant couple cohorts under study exhibited a different inflammatory cytokine or chemokine profile compared to those who remained seronegative. Plasma from a total of 38 Zambian participants (19 uninfected, 19 preinfection) and 30 Rwandan participants (17 uninfected, 13 pre-infection) were analyzed in a Luminex multiplex assay. All of the individuals included in this study were negative partners in an HIV-1 serodiscordant couple, and were analyzed a median of > 1000 days following enrolment (Zambia) and > 450 days (Rwanda) (see Additional file 1: Table S1). Initial analyzes showed that the levels of the biomarkers were similar in the two countries, allowing an initial analysis of the combined Zambia and Rwanda cohorts (see Additional file 2: Figure S1). We observed that 18 of the 21 biomarkers measured were significantly increased in the pre-infection group compared to the uninfected group (Fractalkine, GMCSF, ITAC, IL-1ß, IL-2, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12, IL-17a, IL-21, IL-23, MIP-1α, MIP-1ß, MIP-3α, and TNFα) ( Fig. 1 and Additional file 3: Table S2). All p-values have undergone FDR-correction to account for the multiplicity of the assay. A majority of these upregulated cytokines and chemokines are involved in the inflammatory response. Because many of these cytokines are upregulated in coordinated pathways, we analyzed the data using a random forest model to establish those exerting most influence on the phenotype. Five biomarkers (ITAC, IL-8, IL-7, TNFα, and Fractalkine) were identified as the most significant contributors to the signature associated with future HIV-1 infection (Fig. 2a). We also analyzed the data using a Partial Least Squares (PLS) analysis. The PLS showed that the levels of GMCSF, Fractalkine, IFNg, ITAC, IL-1ß, IL-2, IL-5, IL-7, IL-8, IL-10, IL-12, IL-17α, IL-21, IL-23, MIP-1ß, and TNFα had higher impact on the separation between the uninfected and the pre-infection cohort ( Fig. 2b and Additional file 4: Table S3 A). This result confirms our initial analysis that multiple inflammatory biomarkers are significantly elevated in individuals who later become infected. To further interrogate these findings, we analyzed the data from the Zambia and Rwanda combined cohorts to try and determine if genital inflammation contributed to the elevated profile observed in the pre-infection cohort. Presence of genital inflammation or ulceration in the combined cohort does not explain elevated biomarker levels in pre-infection group Past studies of the Zambian and Rwandan cohorts, as well as other studies, have shown that the presence of genital inflammation/ulceration increases an individual's susceptibility to infection by HIV [3, 4, 6-9, 19, 20]. To test the possible role of genital inflammation or ulceration in the elevated biomarker profile, we compared the levels in individuals from the pre-infection group for whom any form of genital inflammation or ulceration had been reported in the 6 months prior to sample collection with those for whom no genital inflammation or ulceration was reported. There was no significant difference in the levels of any of the 21 analytes analyzed. Similarly, genital inflammation/ulceration did not significantly impact systemic biomarkers in the uninfected group. Moreover, after removing individuals with genital inflammation from the analysis, 11 of the biomarkers remained significantly higher in the pre-infection group compared to the uninfected group (Fractalkine, GMCSF, ITAC, IL-1ß, IL-6, IL-7, IL-8, IL-21, MIP-1α, MIP-3α, and TNFα) (Fig. 3). We next analyzed the Rwandan and Zambian cohorts separately

to determine whether any differences existed between the two cohorts when analyzed in a similar manner. Elevated systemic plasma cytokine and chemokine levels associated with pre-infection individuals in both Zambian and Rwandan cohorts Despite the smaller numbers analyzed when the data was separated by country, we still observed that several inflammatory biomarkers were significantly higher in the pre-infection group than the uninfected group. In Rwanda, Fractalkine, GMCSF, ITAC, IL-1ß, IL-6, IL-7, IL-8, MIP-1α, and TNFα concentrations were significantly increased in the pre-infection group compared to the uninfected group ( Fig. 4a and Additional file 5: Table S4A). The median, interquartile range, and p-values for this analysis is shown in Additional file 5: Table S4A. When analyzed in a PLS model, Fractalkine, GMCSF, IFNg, ITAC, IL-1ß, IL-2, IL-5, IL-6, IL-7, IL-8, IL-10, IL-17α, IL-21, MIP-1α, MIP-1ß, and TNFα were found to be the major contributors to the separation of the groups ( Fig. 4b and Additional file 4: Table S3B). This supports our initial univariate analysis of the Rwanda cohort since there were a number of additional biomarkers that were trending towards significance (see Additional file 6: Figure S2). In Zambia, Fractalkine, ITAC, IL-7, IL-8, IL-23, and TNFα concentrations were significantly increased in the pre-infection group compared to the uninfected group in univariate comparisons ( Fig. 5a and Additional file 7: Figure S3). The median, interquartile range, and p-values for this analysis are shown in Additional file 5: Table S4B. While elevated levels of Fractalkine, ITAC, IL-7, IL-8, and TNFα are observed for both countries in their respective pre-infection groups, they differ in that they exhibit additional biomarkers unique to each country. The PLS analysis showed that levels of ITAC, IL-5, IL-7, IL-8, and MIP-1α were associated with pre-infection ( Fig. 5b and Additional file 4: Table S3C). Taken together, we observed that the pre-infection group had elevated biomarker levels in both Rwanda and Zambia even when the cohorts are analyzed separately. Presence of Schistosomiasis antibody titers does not appear to be a major contributor to the elevated biomarker profile in pre-infected individuals We have recently reported that infection with Schistosoma haematobium is common even in urban settings in Zambia and that infection was associated with increased susceptibility to HIV-1 infection [18]. Because this parasite colonizes the venous plexus in the bladder, we investigated prior infection as evidenced by antibodies resulted in elevated cytokines. However, screening of the plasma from both the uninfected and pre-infection groups for antibodies to the schistosome showed that a similar proportion of individuals in both had detectable antibody titers (9/19 uninfected; 11/19 pre-infection). The Zambian cohort was therefore divided into four groups: uninfected individuals with negative Schistosomiasis antibody titers, uninfected individuals with positive antibody titers, pre-infection individuals with negative titers, and pre-infection individuals with positive titers (see Additional file 8: Figure S4). We did not observe any significant differences in biomarker levels between seropositive and seronegative individuals in either the uninfected group or the pre-infection group. With the proviso that we are analyzing very small groups of individuals, this suggests that schistosome infection is not associated with the elevated biomarker levels observed in the preinfection group. Predictive analyses identify elevated levels of several biomarkers as markers of pre-infection Based on preliminary results from partial least squares analysis (Figs. 2b, 4b and 5b), we wanted to determine which biomarkers included in the study were associated with future HIV-1 acquisition. To achieve this, we analyzed the combined Zambian and Rwandan cohort using Receiver Operating Characteristics (ROC) curves and established a cutoff for the area under the curve (AUC) of 0.8. Using a cutoff of 0.8 means that any biomarker identified in the analysis has over an 80 % chance of distinguishing pre-infection individuals from those that remain uninfected. ITAC, Fractalkine, IL-7, IL-8, and TNFα were identified as markers predictive for pre-infection (Fig. 6). ROC curves for individual Zambia and Rwanda cohorts are shown in Additional files 9 and 10: Figures S5 and S6. This finding is supported by our random forest analysis (Fig. 2a). Using a kfold validation model, we found that concentrations of ITAC above 30.98 pg/ml was highly predictive of a seronegative partner who would become infected (probability of 0.91; sensitivity of 0.75; specificity 0.94). If an IL-7 value above 6.99 pg/ml was added to the model, the probability of identifying pre-infection individuals increased to 0.97 (Fig. 7). Discussion This study found that individuals who will eventually seroconvert have higher levels of proinflammatory biomarkers in their plasma compared to individuals who remain HIV-negative, suggesting a novel link between predisposition to HIV infection and systemic biomarker levels. Our findings suggest that pre-existing conditions which induce systemic inflammation can represent a risk factor for HIV acquisition. An example of a viral infection capable of inducing this is hepatitis C virus. Plasma levels of IL-10 and MIP1ß were positively correlated with HCV RNA levels and may be involved in HCV immunopathogenesis [21], HIV/HCV co-infected women had higher levels of several proinflammatory biomarkers -biomarkers elevated in each infection or disease discussed in this section are summarized in Additional file 11: Table S5. In addition, HCV-positive HIV-negative women had higher levels of IFNg and IL-17 compared to other groups [22]. However, the prevalence of HCV infections in both Zambia and Rwanda has been reported to be low [23,24]. Elevated systemic levels of cytokines and chemokines are also seen in multiple infections common in sub-Saharan Africa. Multiple researchers have found that malaria infection has been linked to elevated systemic cytokine responses. HIV and P. falciparum co-infected individuals had increased HIV viral loads, a steeper decline in CD4 + T cell counts, and exhibited higher levels of several acute phase proinflammatory cytokines [25]. Similarly, symptomatic malaria infections had increased levels of multiple cytokines compared to uninfected controls [26]. It has also been found that individuals with greater malaria disease severity had higher levels of CRP, TNFα, and IFNg [27]. P. vivax infections also are associated with higher levels of proinflammatory cytokines [28]. Elevated systemic biomarkers have also been found in Tuberculosis (TB). Multiple studies found that individuals with acute and latent TB infections had higher levels of cytokines and chemokines compared to healthy controls [29][30][31][32]. These results show that TB infections can increase several proinflammatory cytokines regardless of the stage of infection. A separate study looked at proinflammatory cytokine levels in TB-infected individuals co-infected with S. stercoralis. S. stercoralis is a soil-transmitted helminth that infects about 50-100 million people worldwide [33]. Individuals co-infected with TB and S. stercoralis have increased levels of several type 2 cytokines compared to individuals only infected with TB [34]. One possible explanation for this is an association with microbial translocation. S. stecoralis infected individuals exhibited significantly higher plasma levels of microbial translocation markers (i.e. LPS) and this was associated with increased levels of several proinflammatory cytokines [35]. Researchers looking at microbial translocation in HIV infected individuals found that LPS was also positively correlated with plasma levels of IL-6, TNFα, and hsCRP [36]. This finding is particularly interesting because it is well known that microbial translocation is associated with disease severity in HIV infections [37]. These studies show that several common infections in sub-Saharan Africa are capable of elevating systemic levels of several pro-inflammatory cytokines, which we found as a risk factor for HIV acquisition. However, it is also possible that higher levels of biomarkers prior to infection may have additional impact once an individual becomes HIV-infected. High levels of several cytokines appear to enhance HIV replication and disease progression [38]. Additional studies are needed not only to determine the cause of the elevated levels of biomarkers in individuals prior to HIV acquisition, but also how these elevated levels may impact the disease progression of these patients once infected. While these common infections in Africa are associated with elevated inflammatory cytokines, the observed biomarker profile of the pre-infection individuals reported on here is not indicative of a classic innate immune response or T cell response. Indeed, the elevated cytokines and chemokines observed play roles in many different arms of the immune system. Our analysis showed that three biomarkers, Fractalkine, ITAC, and IL-7, were highly predictive as risk factors for HIV acquisition if elevated above certain levels. These biomarkers can be elevated from infection, but also from non-communicable conditions. Compared to healthy controls, elevated levels of Fractalkine were found in individuals with type 2 diabetes, systemic lupus erythematosus, and systemic sclerosis, an autoimmune disorder that affects the vascular system and leads to early defective angiogenesis. [39][40][41][42]. In systemic sclerosis, the higher levels of soluble Fractalkine in the blood were associated with vascular activation and increased disease severity [42]. That same study found higher levels of ITAC compared to healthy controls [42]. ITAC, also known as IFN-inducible T cell α chemoattractant, was found to be at higher levels in patients with inflammatory bowel disease (Crohn's disease and ulcerative colitis), fibromyalgia, and sarcoidosis [43][44][45]. Sarcoidosis is a systemic inflammatory granulomatous disease that affects lungs of its patients and higher systemic levels of ITAC were found to be predictive of future pulmonary function test decline [43]. IL-7 is not normally considered an inflammatory cytokine and is more associated with T cell homeostasis. However, higher levels of IL-7 were found to be associated with several diseases, such was fibromyalgia [46] and also in patients with colorectal and esophageal cancers compared to healthy controls [47]. Unfortunately, health data on such conditions were not recorded as part of the Heterosexual Transmission Protocol into which discordant couples were enrolled. Nevertheless, the studies reported here indicate that additional studies of cytokine/chemokine levels in at risk individuals is warranted. There are a number of limitations to the current study. We were only able to analyze the inflammatory biomarkers in a limited number of individuals since the availability of pre-HIV infection plasma samples from the discordant couples studied here are also limited (only 3-7 % of individuals in this cohort seroconverted per year depending on the cohort). In addition, we have limited information on other infections and health conditions of individuals in study. This is compounded by the fact that only a small number of studies have reported on systemic biomarkers in other infections or diseases in the context of an African population. As a result, it is difficult for us to identify specific infections or diseases that may be causing the elevated biomarker profile that we observed. Analogous studies of cohorts in the USA or Europe, where additional biomarker data on other infections/diseases may be available, would represent a valuable follow-up study to this one. The identification of key biomarkers associated with HIV acquisition has important clinical ramifications. Our study identified elevated levels of both ITAC and IL-7 as highly predictive of HIV acquisition risk. If individuals in a HIV high-risk region are found to have elevated levels of one or both of these biomarkers, more intensive HIV-1 prevention approaches could be taken in order to protect those individuals. Additional studies are needed not only to determine the cause of the elevated levels of biomarkers in individuals prior to HIV acquisition and how they facilitate the acquisition event, but also how these elevated levels may impact the disease progression of these patients once infected. Overall, our findings suggest that individuals at risk for infection could be identified by testing for elevated levels of a very limited number of biomarkers. Conclusions We show that in a cohort of seronegative partners from serodiscordant couples, individuals who eventually acquire HIV from their partner have higher levels of inflammatory cytokines and chemokines compared to individuals that remain HIV-negative. This was observed in both Rwanda and Zambia where subtype A and subtype C are the predominant HIV-1 serotypes. A Receiver Operator Characteristics analysis showed that the levels of just two cytokines, ITAC and IL-7, were highly predictive of future infection. This suggests high systemic biomarker levels are both a risk factor and a quantitative predictor for HIV acquisition. Study Subjects All participants were enrolled in the Rwanda Zambia HIV Research Group (RZHRG) discordant couple cohorts in Lusaka, Zambia and Kigali, Rwanda. Subjects from both cohorts were enrolled in human subjects protocols approved by the Emory Institutional Review Board, the Rwanda National Ethics Committee and the University of Zambia Research Ethic Committee and provided written consent. When the participants enrolled in the cohort, they were provided couples counseling and testing, treatment for sexually

transmitted infections (STIs), and condoms to reduce transmission of HIV-1. All the subjects tested in this study were the seronegative partner within a serodiscordant cohabitating heterosexual couple. Subjects were selected on their seronegative status, availability of plasma samples, and selected from both Zambian and Rwandan cohorts. The negative partner was tested for HIV every 1 to 3 months. In the Zambia cohort, the median days from enrollment to when the sample was taken was 1234 for the uninfected group and 1087 for the pre-infection group. Samples for the pre-infection group was collected a median of 46 days before the estimated date of infection (EDI). In the Rwanda cohort, the median days from enrollment to when the samples were taken was 494 for the uninfected group and 457 for the pre-infection group. Samples for the pre-infection group was collected a median of 45 days before the EDI. The algorithm used to determine the EDI has been previously described [3]. Evaluation of plasma cytokines The plasma cytokine and chemokine levels were measured using a Milliplex Map Human High Sensitivity T Cell Panel (HSTCMAG-28SK). This kit measures the levels of 21 inflammatory cytokines and chemokines. The samples were run in duplicate. In order to eliminate batch to batch variation in the assay, all tests were carried out on the same batch of plates and approximately equal numbers of pre-infection and uninfected plasma were run on the same plate. The plates were quantified and standardized on a Bioplex 2000 at the Yerkes Virology Core and final concentrations were extrapolated from a standard curve and expressed in pg/ml. All plasma samples were stored at -80 °C and had undergone a single freeze-thaw for aliquoting prior to use. Genital inflammation and ulceration data collection As described in Haaland et al. [3], medical and laboratory signs and symptoms of inflammatory or ulcerative STI, candida, and bacterial vaginosis were recorded systematically at routine study visits and at interim sick visits, with full physical and/or genital exams conducted annually and as clinically indicated; physical and genital exams were routinely conducted on the visit date when lab test results indicated HIV-1 seroconversion. A self-reported symptom was considered present whether or not the patient sought medical treatment and included treatment administered at external clinics. The generation of the composite variables were described in Wall et al. [4]. Briefly, for each 3-monthly interval, composite variables were created. The genital inflammation composite included inflammatory STIs (clinical or laboratory diagnosis or treatment of gonorrhea, chlamydia or trichomonas) and non-inflammatory STIs (reported discharge, dysuria, dyspareunia; observed discharge or inflammation of external or internal genitalia; and/or laboratory diagnosis of candida or BV). The composite for genital ulcer included observed or reported genital ulcers and/ or incident positive RPR. A subject was considered having positive genital inflammation or ulceration if they had presence of either in the 6 months prior to biomarker sample collection. Schistosomiasis antibody titer data collection As described in Wall et al. [18], plasma samples were collected at baseline analyzed in an ELISA assay for antibodies to schistosome soluble worm antigen preparation (SWAP). A 4-parameter curve fitting model was used to assign units based on the standard curve to each unknown plasma. The positive cutoff value was set at three standard deviations above the average anti-SWAP IgG in serum from egg negative controls from the US and Europe. A positive schistosomiasis result was defined as having a positive SWAP antibody response. Data analysis Comparison between the uninfected and the pre-infection groups were done with nonparametric Kolmogorov-Smirnov tests in Prism 9. We addressed the multiple testing issue by using the Benjamini-Hochberg False Discovery Rate (FDR) correction [48]. The Random Forest model was performed in R 4.0.0 at the default setting. The details were described previously [49]. Partial Least Squares (PLS) analysis was performed using the JMP Pro 15 statistical package. PLS analysis had a variable importance cutoff of 0.8. The Receiver Operating Characteristic (ROC) curve analysis used a cutoff of 0.8 for area under the curve (AUC). The predictive results were generated with a k-fold validation which included two splits in the model (k = 5). Best clinical practice guidance for treating deep carious lesions in primary teeth: an EAPD policy document Purpose The European Academy of Paediatric Dentistry (EAPD) has developed this best clinical practice guidance to help clinicians manage deep carious lesions in primary teeth. Methods Three expert groups conducted systematic reviews of the relevant literature. The topics were: (1) conventional techniques (2) Minimal Intervention Dentistry (MID) and (3) materials. Workshops were held during the corresponding EAPD interim seminar in Oslo in April 2021. Several clinical based recommendations and statements were agreed upon, and gaps in our knowledge were identified. Results There is strong evidence that indirect pulp capping and pulpotomy techniques, and 38% Silver Diamine Fluoride are shown to be effective for the management of caries in the primary dentition. Due to the strict criteria, it is not possible to give clear recommendations on which materials are most appropriate for restoring primary teeth with deep carious lesions. Atraumatic Restorative Technique (ART) is not suitable for multi-surface caries, and Pre-formed Metal Crowns (PMCs) using the Hall technique reduce patient discomfort. GIC and RMGIC seem to be more favourable given the lower annual failure rate compared to HVGIC and MRGIC. Glass carbomer cannot be recommended due to inferior marginal adaptation and fractures. Compomers, hybrid composite resins and bulk-fill composite resins demonstrated similar values for annual failure rates. Conclusion The management of deep carious lesions in primary teeth can be challenging and must consider the patient’s compliance, operator skills, materials and costs. There is a clear need to increase the use of MID techniques in managing carious primary teeth as a mainstream rather than a compromise option. Aim The European Academy of Paediatric Dentistry (EAPD) proposes this best-practice guidance to help practitioners manage deep caries in children during the delivery of oral health care. A similar statement has been published by a Joint ORCA and EFCD Expert Delphi Consensus Statement (Spleith et al. 2020). Treatment options and materials for permanent teeth are excluded from this document. Selection of the topic guide Dental caries is a common, yet preventable disease that affects 20-90% of 6 year-old children in Europe (WHO 2018). The management of dental caries in children has shifted towards controlling caries according to an individual treatment plan including risk estimation, early diagnosis and prevention plan to keep dentition healthy and arrest initial lesions if needed (Pitts et al 2014). This was investigated by the EAPD best clinical guidance management for early caries lesions in children and young adults (Kühnisch et al. 2016). Unfortunately, many children may present with deep carious lesions which require restorative management, either by conventional techniques or by implementing the concept of Minimal Intervention Dentistry (Ericson et al. 2003;Frecken et al. 2012;Dorri et al. 2015;Schwendicke et al. 2019). Conventional approaches to deep carious lesions have focussed on pulpal interventions to avoid extraction and keep the tooth asymptomatic and functional until exfoliation, whereas Minimal Intervention Dentistry (MID) techniques aim to maintain teeth vital, asymptomatic and functional for as long as possible, preferably until exfoliation. The type of treatment provided should follow biological evidence-based caries management concepts, which emphasise preserving as much tooth structure as feasible, and in case of primary teeth, until these exfoliate naturally (Frencken et al. 2012). For all these techniques, the clinician must also consider the most suitable material to use. When a clinician is presented with a child patient with deep carious lesions in the primary teeth, there are many factors to be considered before an appropriate management plan can be reached. These need to consider both the needs of the patient, parent and dentist: Methods Three expert groups were invited by EAPD to undertake systematic reviews of the available literature for the management of deep carious lesions in primary teeth, in particular focussed on: This new guideline is based on the reviews presented by the invited experts in the 12th EAPD virtual interim meeting in Oslo in April 2021. The discussions were carried out by those attending the three working groups consisting of invited speakers and nominated delegates from the EAPD member countries. Each working group was moderated by two members of the EAPD Clinical Affairs Committee (CAC). Discussions were carried out and conclusions were reached by agreement and consensus, and the recommendations from the workshops were presented on the final day of the interim meeting by the CAC moderators. This was used as a basis by the CAC members to develop the guidance. The selection criteria for the three groups is summarised in (Table 1). Due to the different selection criteria and approaches used in the three reviews, it was not possible to determine recommendations using GRADE (Guyatt et al. 2008). This implies that some of the recommendations are based on low-grade evidence and expert opinion. Workshop 1: conventional management of deep caries in primary molars The systematic review and meta-analysis by Stratigaki et al. 2022, concentrated on the following techniques: The evidence demonstrated that pulp reaction to the treatment and applied medicament rely on the status of the pulp before the intervention, and the conditions under which the pulp is being treated (patient's compliance, effective use of Local Anaesthesia (LA), and Rubber Dam Isolation). Recommendations: • Use the least invasive technique for the best predictable clinical outcome. • There was a unanimous agreement that a restoration providing a good coronal seal is essential for the management of vital pulp in primary teeth. • Indirect Pulp Capping (IPC) and Pulpotomy (PP) have high success rates and can be recommended as effective treatment modalities for the management of deep caries in primary teeth. • Direct pulp capping has limited use in daily clinical practice in the event of pulp exposure, except in very restricted non-infectious conditions and on asymptomatic teeth. • Calcium hydroxide has the poorest success rate of all commonly used pulpotomy medicaments, and therefore Recommendations: • The use of 38% SDF once or twice per year can be advantageous for caries arrest, with better outcomes for two applications per year. It is recommended that clinicians should consider the use of 38% SDF in children with a high caries risk, to avoid/delay the need for more invasive treatments in very young children. • The use of pre-formed metal crowns (PMC) using the Hall technique (HT) for the management of dentinal caries in primary teeth can reduce the risk of pain and restoration failure for caries in the primary teeth. The Hall technique (HT) reduced discomfort and was preferred by patients and parents. • Selective (one-step) or step-wise caries removal offer some advantage over complete caries removal for the avoidance of pulp exposure for lesions extending to inner third or quarter of dentine. In the absence of other signs and symptoms indicating irreversible pulpitis, these techniques should be considered to avoid pulp exposure and the need for pulp therapy. • The failure rates for ART when used to restore multisurface caries is unacceptably high. Therefore, this technique is not recommended for the restoration of multisurface carious lesions. ART could be considered as an adequate management option for single surface (occlusal) in certain instances for primary teeth. Gaps in knowledge: • Further investigation is needed into the effectiveness and safety of the HT, as there has only been one systematic review to date. • Comparison studies are needed into the cost effectiveness of different MID treatments modalities Workshop 3: materials The systematic reviews by Amend et al. 2022 Ionomer Cements (RMGIC) are recommended with caution for occluso-proximal (class II) restorations of primary dentition. These materials are not recommended in multi-surface reconstructions. • Metal Reinforced GIC (MRGIC) are not recommended in the restoration of primary molars. • Glass Carbomer is not recommended for both occlusal (class I) and occluso-proximal (class II) restorations of primary carious molars due to the high failure rate. • Compomers are recommended for both occlusal (class I) and occluso-proximal (class II) restorations of primary carious molars. • Hybrid and bulk-fill composite resins are recommended for both occlusal (class I) and occluso-proximal (class II) restorations of primary carious molars. • It is recommended to use a calibrated polymerization lamp and ensure adequate polymerization, omitting the monomers at the surface. • Due to lack of evidence, it was not possible to consider dentine etching times

and margin cavity preparation. • Due to the selection criteria, only one RCT with Preformed Metal Crowns (PMC) was included in the review, using the Hall Technique (HT) in vital primary molars (Santamaria et al. 2017). The study was found to have a low annual failure rate, but a high risk of bias, therefore clear recommendations could not be given due to lack of evidence. • There is a lack of RCTs evaluating restoration techniques in primary anterior teeth. In the one included trial with high risk of bias (Alaki et al. 2020) zirconia crowns and composite strip crowns were compared in the reconstruction of carious primary anterior teeth, but a recommendation could not be given for lack of evidence. • Due to the low levels of evidence, no recommendations for the use of specific isolation techniques could be made for all dental materials. Gaps in knowledge: • More RCTs with power calculations and parallel group design are needed comparing restorative interventions • Narrow age range for included children and longer follow ups • A description of the caries experience among the included participants • Detailed descriptions of the interventions (availability of preoperative radiographs, assessment of carious lesion depth, administration of local anaesthesia, isolation technique, extent of carious tissue removal, restorative materials and application mode, adhesive protocol etc.) • Operator experience should be clearly stated • A precise report of the numbers of patients lost to followup is essential Clinical recommendations Recommendations for management of deep carious lesions in primary teeth were developed in line with the strength of the evidence (Fig. 1). Strong • It is recommended that application of 38% SDF can be advantageous for caries arrest, with better outcomes for biannual application. • Indirect Pulp Capping (IPC) or selective and step-wise caries removal, and Pulpotomy (PP) have high success rates and can be recommended as effective treatment modalities for the management of deep caries in primary teeth. • The use of formocresol for pulpotomy is no longer recommended, due to the availability of more biocompatible medicaments. • ART technique is not recommended for the restoration of multi-surface carious lesions. • Glass carbomer cannot be recommended due to inferior marginal adaptation and fractures. • MRGIC cannot be recommended due to loss of anatomical form and marginal intergrity. • Pre-formed metal crowns (PMCs) using the Hall technique are recommended as a treatment option for the management of dentinal caries. 1 3 • The use of PMCs for endodontically treated primary molar teeth is recommended. Weak • Compomers are recommended for both occlusal (class I) and occluso-proximal (class II) restorations of carious primary molars. • Hybrid and bulk-fill composite resins are recommended for both occlusal (class I) and occluso-proximal (class II) restorations of carious primary molars. • GIC, RMGIC and HVGIC are more favourable given the lower annual failure rate compared to MRGIC. Research recommendations The EAPD interim seminar identified further research needs, to improve comparability of studies to include: • Focus conducting trials with more appropriate study designs and standardised methodology, particularly in relation to use of randomisation and allocation sequence concealment diagnostic and outcome measures o Studies should record the use of radiographic assessment p The depth of caries should be specified using an objective classification such as that proposed by the ICDAS • Quality of life, patient preference, cost effectiveness and burden of care and impact of different treatments modalities on future compliance Conclusion The management of deep carious lesions in primary teeth can be challenging and must consider the patient's compliance, operator skills, materials and costs. The lack of high quality RCTs meant that for some consensus statements only a low level of evidence was available. One of the important outcomes of this review was that Minimal Intervention Dentistry (MID) techniques appear to Fig. 1 Flowchart of treatment protocol of dentinal caries lesions in primary dentition be effective in arresting the progression of dentinal caries in primary teeth when compared to no treatment and conventional restorations. There is some evidence of improved patients reported outcomes with such techniques, however further research is required. A major advantage of MID for the management of dentine carious lesions is that many of these techniques can be carried out without aerosol generation. There is a clear need to increase the emphasis on utilising MID techniques in managing carious primary teeth as a mainstream rather than compromise option in circumstances where the conventional approach is prohibited due to cost or co-operation (Splieth et al, 2020). Due to the heterogenicity of the studies and the reviews, it was not possible to develop guidance using best-practice methods, such as GRADE. Detailed and explicit criteria for ratings of quality and grading of strength, as well as consensus protocols, and input from patients and parents will make judgments more transparent for future guideline development and recommendations. Differential Diagnosis of a Left Atrial Mass after Surgical Excision of Myxoma: a Remnant or a Thrombus? Echocardiographic diagnosis of atrial myxoma may not always be straightforward, and the distinction between myxoma and thrombi is not easy, especially when we observe a mass after successful surgery. Our report describes a 72-year-old woman who presented with right upper limb hemiparesis and was subsequently diagnosed as having transient ischemic attack due to a left atrial myxoma. One month after successful surgical resection of the tumor, the patient developed left-sided weakness. Echocardiography revealed a left atrial mass attached to the interatrial septum. Intravenous heparin was administered as a therapeutic trial for postoperative thrombi, which resulted in a decrease in mass size within a week. Anticoagulation with warfarin was continued, and complete resolution was demonstrated on a 4-month follow-up transesophageal echocardiography. This case highlights the fact that thrombus formation at the surgical site should be considered an unusual but potential complication after surgical resection of left atrial myxomas. Introduction Myxomas are the most common type of primary cardiac tumors, representing 30-50% of all cardiac tumors. 1) Typically, myxomas manifest as a solitary, pedunculated, mobile mass in the left atrium (LA) with a stalk that is attached to the interatrial septum. Echocardiography is the most commonly used diagnostic modality. However, the echocardiographic diagnosis of myxomas is not always straightforward, and the differential diagnosis includes other benign and malignant primary heart tumors, metastatic tumors, and intracardiac thrombi. The distinction between myxomas and thrombi may pose considerable diagnostic challenges, and difficult cases have been reported. 2)3) A precise diagnosis is crucial as the treatment approach and the accompanying risks differ in these two conditions. We describe a case of a LA mass that developed 1 month after surgical resection of a LA myxoma, presenting the difficulty in differential diagnosis between a recurrent myxoma and a postoperative thrombus. Case A 72-year-old Korean woman presented to our emergency department with right upper limb hemiparesis without speech impairment. Her medical history was significant for hypertension, hyperlipidemia, and diabetes. She reported no previous history of cigarette smoking and alcohol drinking. On admission to the hospital, her right upper limb hemiparesis resolved rapidly and completely. The results of laboratory examinations, including complete blood count, serum electrolyte levels, and coagulation studies, were within normal limits, except for slight hypertriglyceridemia. An electrocardiogram showed a normal sinus rhythm, and a brain computed tomography scan appeared normal. Urgent magnetic resonance imaging (MRI) of the brain revealed no definite evidence of acute infarction. Neither significant steno-occlusive lesions nor cerebral aneurysms in the intra/extracranial vessels were demonstrated by magnetic resonance angiography (MRA). Transthoracic echocardiography (TTE) showed a well-defined echogenic mass in the LA with a broad-based attachment to the interatrial septum (Fig. 1). The LA was not enlarged, and left ventricular (LV) systolic function was normal, with an ejection fraction (EF) of 60%. Transesophageal echocardiography (TEE) confirmed the presence of a heterogeneous mobile mass (33x25 mm) with internal echo-free spaces, and color Doppler flow mapping showed flow signal within the mass suggesting hypervascularity (Fig. 2A). The mass did not involve the heart valves directly, and Doppler echocardiography revealed no impairment of flow across the mitral valve. The diagnosis of an embolic transient ischemic attack (TIA) caused by a LA myxoma was made, and the ABCD2 score was 6. She was urgently referred for surgical intervention to prevent further embolic strokes. A median sternotomy was performed and right atriotomy was created with a trans-septal approach. The LA mass was then excised en bloc for complete resection and the iatrogenic atrial septal defect was directly closed without using a patch. Histopathological examination confirmed a cardiac myxoma with a clear resection margin, and without an overlying thrombus. The postoperative course was uneventful and free of major complications. On postoperative day 4, TTE demonstrated no evidence of any remnant mass or thrombus in the LA (Fig. 1). The patient was discharged without antiplatelet therapy or anticoagulants. One month after surgery, the patient was readmitted to our hospital because of a transient left-sided weakness in her arm and leg without speech impairment. Her neurological examination was normal, and brain MRI performed using the stroke protocol demonstrated no acute lesion with diffusion restriction. MRA of the cranial vessels showed neither steno-occlusive lesions nor aneurysmal dilatations, which was the same as before. TTE and TEE were performed to identify the source of the TIA. Scanning of the LA revealed a polypoid mass (13×7 mm) attached to the mid portion of the interatrial septum, with an irregular lobulated surface (Fig. 2B). Lipomatous hypertrophy of the interatrial septum was also noted, and no thrombus was seen in the left atrial appendage. Although LV systolic function was slightly decreased (EF 47%) without regional wall motion abnormalities, the echocardiographic findings of the mass in question were more suggestive of a remnant myxoma. However, given the short period of time after the surgery, thrombus formation at the site of direct closure of the interatrial septum was a reasonable possibility. Therefore, after discussion, intravenous heparin administration was started as a therapeutic trial and it was continued to help differentiate a remnant mass from a thrombus. One week after the initiation of anticoagulation, a follow-up TEE examination revealed a relevant reduction in the burden of the mass in question from 13×7 mm to 8×3 mm, suggesting that the mass was most likely to be a thrombus (Fig. 2C). Therefore, longterm anticoagulation with warfarin was continued to maintain an international normalized ratio of 2-3. At the 4-month followup after warfarinization, the mass attached to the interatrial septum had resolved completely on TEE examination, confirming the diagnosis of a thrombus (Fig. 2D). The patient is now being maintained on warfarin and kept under regular follow-up. Discussion Myxomas are the most common primary cardiac tumors, although their diagnosis is not always straightforward. Most often, the typical morphological characteristics of atrial myxomas can be readily described using echocardiography, but the imaging appearance of myxomas sometimes mimics the appearance of thrombi. 2) Differential diagnosis is more difficult when a recurrent mass is observed after successful surgery of a LA myxoma. Myxomas are larger, pedunculated, mobile, and most frequently attached by a stalk to the fossa ovalis of the LA septum. Less commonly, they are located in the atrial dome, mitral, pulmonic, or aortic valves, or the right atrial septum and free wall. 4) On the other hand, thrombi that are generally attached to the posterior left atrial wall have a broad base, and are mostly immobile. The left atrial appendage is a useful landmark for differentiating tumor from thrombus. The formation of thrombi usually occurs in patients with organic heart disease, and in association with low cardiac output or atrial fibrillation. 1) These characteristics, however, are not specific enough to always reliably distinguish left atrial thrombi from myxomas. 3) In our present case, the questionable finding was a LA mass that developed 1 month after primary surgical resection of the original tumor. It was important to differentiate a remnant myxoma from a postoperative thrombus because prompt surgical excision would have been required to prevent further embolic events in the case of a remnant tumor. Our present case had undergone uncomplicated en bloc resection. The absence of atrial fibrillation, enlarged atrial chamber, valvular heart diseases, and spontaneous atrial contrast echoes were the features weighing against the diagnosis of a thrombus. However, considering the rarity of remnants after successful surgical resection of non-familial sporadic myxomas and the rapid growth rate of the mass in question, a

treatment test with anticoagulation was warranted in order to avoid an unnecessary surgical intervention. Embolization is the second most common initial presentation of a myxoma, occurring in 30-40% of the patients. Prompt surgical excision must be performed as soon as the diagnosis is confirmed because of the high risk of new embolic complications or sudden death. 1) However, postoperative thromboembolic events occur infrequently, and atrial arrhythmias are the most common complications after surgery for myxoma. [5][6][7][8] Although myxomatous fragments of a tumor or thrombi at the tumor surface may dislodge during surgery, the occurrence of atrial fibrillation is the main reason for postoperative thromboembolic complications. 9) Thrombus formation at the closure site is an unusual but potential complication contributing to recurrent neurologic events. However, an intracardiac thrombus after surgery for myxoma has rarely been reported, but thrombus formation after repair of an atrial septal defect or a patent foramen ovale has been reported. [10][11][12] One interesting point is that all these case reports showed thrombus formation in the right atrium along the primary suture line: one case showed a pedunculated thrombus along the right atrial wall, 11) whereas the other two cases showed a thrombus attached to the right atrial side of the interatrial septum. 10)12) Considering low blood flow velocity in the right atrium, short-term anticoagulation has been suggested to prevent this complication after primary surgical or device closure of the atrial septal defect or a patent foramen ovale. 11) Our case is unique as it showed that thrombus formation can occur in the left atrium after primary closure of the atrial septum without using an artificial atrial patch. Removal of an atrial myxoma carries an operative mortality rate of 5% or less, and the mortality of patients waiting for surgery is 8%. 1)13) Atrial thrombi misdiagnosed as myxomas can lead to an unnecessary surgical resection. 3)14)15) Although a trial of anticoagulation with intravenous heparin is advised for making the differential diagnosis, it may not be applicable in case of postoperative thrombi, which can be more or less resistant to anticoagulation therapy. 11) An organized thrombus associated with underlying structural heart diseases may also be less likely to respond to lytic or anticoagulation therapy. 16) In summary, the echocardiographic distinction between myxomas and thrombi is not always straightforward due to their morphological similarities. Our present case highlights 1) the importance of clinical suspicion of intracardiac thrombi after successful primary closure of the atrial septum; 2) the advantage of a clinical therapeutic trial of short-term anticoagulation with serial follow-up TEE, which could discriminate between a postoperative thrombus and a remnant cardiac tumor, resulting in avoidance of an unnecessary secondlook operation. The management of an atrial mass should be based on the clinical situation (accompanying heart disease, postoperative period, etc.) and the echocardiography findings (well-limited mass, echogenicity, etc.). A treatment test with anticoagulation is worthy of attempting in cases in which there is a diagnostic challenge in differentiating myxomas from thrombi. NRDC report trashes recycling critics. doi: 10.1289/ehp.10217 In their otherwise informative and concise review of the current state of evidence concerning risk factors for acute childhood leukemia, Belson et al. (2007) did not correctly address nonionizing radiation and, in particular, power frequency magnetic fields as a possible risk factor for childhood leukemia. This failure may be due to a widespread misconception about the evidence concerning nonionizing electromagnetic fields (EMFs) as a health hazard. It is also apparent in the Churchill County leukemia cluster study published in the same issue, in which Rubin et al. (2007) investigated a multitude of factors, many with sparse or ambiguous previous evidence of an association with childhood leukemia. Although power frequency magnetic fields have been classified as a possible human carcinogen (group 2B) by the International Agency for Research on Cancer (IARC 2002) and by a National Institute of Environmental Health Sciences (NIEHS) working group (NIEHS 1998), based on the evidence of an association with childhood leukemia, these were apparently not considered by Rubin et al. (2007). In their review of nonionizing radiation, Belson et al. (2007) inappropriately mixed original research and pooled analyses, further contributing to the prevailing confusion. Both Ahlbom et al. (2000) and Greenland et al. (2000) presented pooled analyses that included the important study of Linet et al. (1997). Hence, it is inappropriate to present results of the latter as an independent source. Almost all epidemiologic studies of residential exposure to power frequency magnetic fields published before 1999 are included in either the pooled analyses of Ahlbom et al. (2000) or Greenland et al. (2000). Only the study of Myers et al. (1990) was not included because authors refused to provide requested data. Although the study of Linet et al. (1997) is often cited as failing to support the hypothesis of an association between residential exposure to magnetic fields and childhood leukemia [it was also cited by Belson et al. (2007)], it actually was one of the most important supporters of an association in the pooled analyses and contributed the greatest number of highly exposed children. Two large and well-conducted studies published after the pooled analyses (Kabuto et al. 2006; Schüz et al. 2001) lend further support to the results of the pooled analyses of an increased risk from high average levels of magnetic field exposure. It is also incorrect to characterize the evidence as “some have found a small association ... while others have not ....” First of all, the association is not small, but is comparable or larger than that for all other factors considered by Belson et al. (2007). Second, the evidence is consistent across different continents, study types, measurement methods, and other factors. Of course, there are potential sources of bias, in particular selection bias. However, thorough investigations of these potential biases have rendered it unlikely that they can completely explain the association. Up to now, there is no other risk factor of childhood leukemia that has been as comprehensively studied concerning possible biases and confounding factors. It is high time that exposure to power frequency EMFs is recognized as a potential risk factor for childhood leukemia and is properly included in the protocols of cluster studies and in epidemiologic studies of other risk factors as a potential confounder. The author declares he has no competing financial interests. In their otherwise informative and concise review of the current state of evidence concerning risk factors for acute childhood leukemia, did not correctly address nonionizing radiation and, in particular, power frequency magnetic fields as a possible risk factor for childhood leukemia. This failure may be due to a widespread misconception about the evidence concerning nonionizing electromagnetic fields (EMFs) as a health hazard. It is also apparent in the Churchill County leukemia cluster study published in the same issue, in which Rubin et al. (2007) investigated a multitude of factors, many with sparse or ambiguous previous evidence of an association with childhood leukemia. Although power frequency magnetic fields have been classified as a possible human carcinogen (group 2B) by the International Agency for Research on Cancer (IARC 2002) and by a National Institute of Environmental Health Sciences (NIEHS) working group (NIEHS 1998), based on the evidence of an association with childhood leukemia, these were apparently not considered by Rubin et al. (2007). In their review of nonionizing radiation, inappropriately mixed original research and pooled analyses, further contributing to the prevailing confusion. Both Ahlbom et al. (2000) and Greenland et al. (2000) presented pooled analyses that included the important study of Linet et al. (1997). Hence, it is inappropriate to present results of the latter as an independent source. Almost all epidemiologic studies of residential exposure to power frequency magnetic fields published before 1999 are included in either the pooled analyses of Ahlbom et al. (2000) or Greenland et al. (2000). Only the study of Myers et al. (1990) was not included because authors refused to provide requested data. Although the study of Linet et al. (1997) is often cited as failing to support the hypothesis of an association between residential exposure to magnetic fields and childhood leukemia [it was also cited by ], it actually was one of the most important supporters of an association in the pooled analyses and contributed the greatest number of highly exposed children. Two large and well-conducted studies published after the pooled analyses (Kabuto et al. 2006;Schüz et al. 2001) lend further support to the results of the pooled analyses of an increased risk from high average levels of magnetic field exposure. It is also incorrect to characterize the evidence as "some have found a small association … while others have not …." First of all, the association is not small, but is comparable or larger than that for all other factors considered by . Second, the evidence is consistent across different continents, study types, measurement methods, and other factors. Of course, there are potential sources of bias, in particular selection bias. However, thorough investigations of these potential biases have rendered it unlikely that they can completely explain the association. Up to now, there is no other risk factor of childhood leukemia that has been as comprehensively studied concerning possible biases and confounding factors. It is high time that exposure to power frequency EMFs is recognized as a potential risk factor for childhood leukemia and is properly included in the protocols of cluster studies and in epidemiologic studies of other risk factors as a potential confounder. doi: 10.1289/ehp.10080 I read with interest the recent review by on childhood leukemia, particularly the sections dealing with radiation exposure. Like the authors, I believe that ionizing radiation is strongly associated with childhood acute leukemia. I would like to point out that several critical pieces of information were overlooked; these support stronger and more meaningful conclusions. Ionizing Radiation and Childhood Leukemia Although atomic bomb survivors offer the clearest evidence of leukemia risk after childhood exposures to ionizing radiation, studies of children exposed to fallout in other contexts should not be downplayed. stated that "radiation exposure secondary to the Chernobyl accident has not been shown to increase the risk of leukemia in children who were exposed after birth …," but they failed to mention the case-control study of Noshchenko et al. (2002), which found significant increases in childhood and acute leukemias in association with estimated childhood exposures. Children living downwind of the Nevada Test Site have also shown a significant increase in leukemia related to estimated fallout exposure (Stevens et al. 1990). In utero exposure to ionizing radiation has been a known causal factor for childhood cancer for > 50 years. Although stated that the lack of evidence for a childhood leukemia risk among atomic bomb survivors constitutes the "most notable reason for doubt of a true association," they overlooked the reviews of Little (2002, 2003); these authors demonstrated that the highly uncertain atomic bomb survivor data are statistically compatible with the robust set of data found in the Oxford Survey of Childhood Cancers and related X-ray exposure cohorts. There is no valid reason to doubt this association at present. Perspectives Correspondence The correspondence section is a public forum and, as such, is not peer-reviewed. EHP is not responsible for the accuracy, currency, or reliability of personal opinion expressed herein; it is the sole responsibility of the authors. EHP neither endorses nor disputes their published commentary. The association between preconception paternal irradiation (PPI) and childhood leukemia has always been controversial. Two of the major objections to the "Gardner hypothesis," as pointed out, have been mixed evidence from studies of radiation-exposed fathers and a lack of positive evidence in the children of the atomic bomb survivors. Regarding the first objection, Belson et al. overlooked the two largest studies of the children of radiation workers. Draper et al. (1997) conducted a UK-wide case-control study of childhood cancers in relation to paternal radiation exposure. This study showed, based on > 13,000 cases not included in the study of Gardner et al. (1990), that children with leukemia or non-Hodgkin lymphoma were significantly more likely than controls to have fathers who were radiation workers. Dickinson and Parker (2002) conducted a cohort study of > 250,000 births in Cumbria, England, including the cases of Gardner et al. (1990), and found a significant 2-fold increase in the risk of leukemia and non-Hodgkin lymphoma among the children of radiation workers. These and other studies, taken together,

give statistical support to the idea that paternal radiation work is a risk factor for childhood leukemia. When interpreting the evidence for a PPI effect in atomic bomb survivors, it is important to consider what is known about potential mechanisms. As reviewed by Niwa (2003), Nomura (2003), and others, animal studies have consistently demonstrated that PPI can cause or increase the susceptibility to leukemia in offspring. In addition to fascinating evidence of postconception genomic instability after preconception exposure, many studies suggest that there may a window of sensitivity corresponding to postmeiotic stages of spermatogenesis; in humans, this would mean the few months leading up to conception (Adler 1996). Of the roughly 30,000 children of atomic bomb survivors, only about 2% were conceived in the 6 months after the bombings. Based on the spontaneous leukemia rate reported by Yoshimoto (1990), the expected number of spontaneous cases in this subcohort would be < 1, and an excess on the order suggested by the radiation worker studies would not be statistically apparent. For this and other reasons, the atomic bomb survivors may not be an appropriate comparison group. To summarize, it is not unreasonable to observe that the weight of evidence generated to date supports the idea that preconception, prenatal, and postnatal exposures to ionizing radiation are all risk factors for childhood leukemia. The author declares he has no competing financial interests. We reported that prenatal exposure to methylmercury causes cognitive impairment in an estimated 316,588 children born in the United States each year, costing this nation $8.7 billion annually in lost productivity . Each year, this exposure also causes an estimated 1,566 cases of mental retardation (Trasande et al. 2006). The principal (70%) source of the mercury that enters the bodies of American children is combustion of coal in electricitygenerating plants. Abel Russ In their reanalysis of our data, made a series of incorrect judgments and poorly considered assumptions, each of which diminishes the import of our findings. We note the following errors in their analysis: First, incorrectly used a linear model to relate cognitive function to prenatal methylmercury exposure, despite the National Research Council's (NRC) clear finding that a logarithmic model provides a better statistical fit. The NRC, in their examination of the Faroe Islands cohort study, the study on which they place greatest reliance, stated that "[b]ecause these calculations necessitate extrapolating to estimate the mean response at zero exposure level," logarithmic models "lead to lower estimates of the Benchmark Dose (BMD) than linear or K-power models" (NRC 2000, p. 294). Recent analyses of early childhood lead exposure further corroborate the validity of logarithmic models in representing subclinical dose-response relationships of neurodevelopmental injury . Second, unwisely based their analysis on potentially biased data from the Seychelles cohort study. By contrast, our model , like that of the NRC, is based primarily on Faroes data. We chose not to use Seychelles data because of concern that the tests of neurobehavioral function used there were not well validated for a non-American population and therefore may not have been sensitive to detect cognitive impairment (Landrigan and Goldman 2003). Another major potential source of bias in the Seychelles study, not acknowledged by , is that it fails to consider the potentially beneficial nutrients found in the fish-based diet of the Seychelles. These nutrients, omega-3 fatty acids in particular, may partially offset the toxicity of methylmercury. Indeed, if maternal fish intake is taken into account in the Seychelles cohort, as recently was done, the estimate of methylmercury toxicity increases (Budtz-Jorgensen et al. 2007). cited previous meta-analyses of the Faroes, Seychelles, and New Zealand studies by in applying IQ decrements of 0.13-0.18 points/ppm hair mercury, but these are likely underestimates, and further invalidate the analysis of Griffiths et al. Third, in attributing mercury deposition to sources of emission, relied inexplicably and without justification on a mathematical model that posits that only 16% of deposits are attributable to American sources. They ignored empiric data from the U.S. Environmental Protection Agency (EPA)-sponsored Steubenville study, which found that 80-90% of mercury emissions deposit within 30-50 miles of the source (U.S. EPA 2007); and from the A 397 Correspondence Electric Power Research Institute, which estimated that 30% of mercury deposits are attributable to American sources (Seigneur et al. 2004). Fourth, incorrectly assumed that reductions in mercury emissions from power plants do not result in any reduced levels of fish contamination until after 15 years. This is not correct. Reductions in power-plant emissions in 2008 will, in fact, begin immediately to minimize methylmercury body burden among children born to women in 2008, and the degree of reduction will increase further in subsequent years, perhaps through 2038, thus reducing the number of children damaged, the severity of the prenatal brain damage in these children, and the resulting economic burden. Finally, incorrectly based their estimate of the economic value of a child's social productivity on the 1992 Current Population Survey rather than on the currently available 2005 data set. This miscalculation substantially underestimates the economic impact of methylmercury on the developing brain. estimated that this value is currently on the order of $4-9 million/child, a value far greater than that used by Griffiths et al., and greater even than our estimate. doi: 10.1289/ehp.10302R In our review of the article by , we used their published linear model to evaluate the monetized impact of IQ decrements associated with prenatal mercury exposure to methylmercury (MeHg) under different assumptions ). First, we used a corrected dose-response slope to address the error that the authors made in the conversion of the relationship between cord blood and neurodevelopment effects. We then introduced the assumptions that the U.S. Environmental Protection Agency (EPA) used in its Clean Air Mercury Rule (CAMR). Introducing the U.S. EPA assumptions decreased the undiscounted monetized impact of global anthropogenic mercury emissions in the corrected Trasande et al. model by 81% and decreased the estimated impact of U.S. sources by almost 97%. When discounting is included, the U.S. EPA assumptions decreased the monetized estimate of global impacts by 88% and the impact of U.S. power plants by 98%. Methylmercury and the Brain: Griffiths et al. Respond The choice of a linear model (i.e., a K-power model, with K = 1) was based on the recommendation of the National Research Council (NRC 2000): After extensive discussion, the committee concluded that the most reliable and defensible results for the purpose of risk assessment are those based on the K-power model. Trasande et al. choose to emphasize the results of their logarithmic model, which produces their highest estimates of monetized impacts. We do not dispute that there may be cases in which a logarithmic model might be appropriate, but in the case of methylmercury, the NRC (2000) was unequivocal: For MeHg, the committee believes that a good argument can be made for the use of a K-power model with K constrained to be greater than or equal to 1. That rules out square-root (K = 0.5) and log models (the limiting case as K approaches 0). For the U.S. EPA dose-response slope, we used the results of an integrated statistical analysis by , which has been recently updated (Axelrad et al. 2007). The analysis of Axelrad et al. includes results from the Seychelles study and also those of the Faroe Islands study (which was used by , as well as the New Zealand study. All three of these studies were used by the NRC (2000) and are described as being "well designed and carefully conducted, and each examined prenatal MeHg exposures within the range of the general U.S. population exposures." We will concede that controlling for maternal fish intake when assessing the impact of mercury on neurodevelopment is an important consideration that can be addressed in the future. The assumption that, on average, 16% of the total mercury deposition in the United States is from American and Canadian sources comes straight from the U.S. EPA model used for the CAMR. As discussed in our article ), the U.S. EPA used a spatially explicit air quality model to simulate the location of mercury deposition, but we used the average value to compare it to Trasande et al.'s (2005) assumption that 60% of the mercury content in all domestically caught fish is due to American sources. It is true that in the study of Steubenville, Ohio, published after the CAMR was promulgated, Keeler (2006) found a much higher percentage of local and regional deposition (70% of the mercury wet deposition, not 80-90%), but this is an estimate of deposition at a single point and cannot be extrapolated to the entire country. Furthermore, the same U.S. EPA model that produced the 16% average value predicts comparatively high values for the Steubenville region of Ohio (U.S. EPA 2006). With regard to the charge that we assumed there will be no reductions in fish contamination until after 15 years, Transande et al. are wrong. In our article we are clear in our position that benefits build over time during the transition path from the current conditions to the new equilibrium. The choice of 15 years is an average period over which to discount the benefits, reflecting the 5-30 years for freshwater systems and the 30-200 years for ocean systems to reach equilibrium. Furthermore, we reported the undiscounted monetized results, which could be compared to implicit assumption of the instantaneous elimination of all anthropogenic mercury from the environment. Finally, Trasande et al.'s reference to is truly baffling. That article is a review and evaluation of dozens of studies on the value of a statistical life (VSL). A VSL is derived from the tradeoffs witnessed in the market and elsewhere between income and small changes in risk of death. The value for a small change in mortality risk is aggregated to statistical lives in order to be comparable to risk assessment estimates. Because mortality risk and IQ decrements are vastly different items, there is no expected relationship between these two values. doi: 10.1289/ehp.10274 In their article, "Exposures to Environmental Toxicants and Attention Deficit Hyperactivity Disorder [ADHD] in U.S. Children," advanced our knowledge of the effects of environmental tobacco smoke (ETS) and lead on the central nervous system of children. With respect to lead exposure, the study, importantly, focused on an older age group (4-15 years) than is generally studied (< 6 years) because of the greater sensitivity of the developing central nervous system to environmental insult early in life [Centers for Disease Control and Prevention (CDC) 1997]. Environmental Exposures and ADHD In the logistic model used by , the association of ADHD with lead exposure was statistically significant in the highest exposure quintile; however, it was also tenuous. Although not unheard of, the cutoff (p < 0.2) for inclusion of factors and variables associated with ADHD on univariate analysis was generous compared with the commonly used 0.1 or 0.05, and very close to the p-value of the lead-ADHD association of 0.19. The lead-ADHD relationship also exhibited a significant monotonic dose response, so it would have been helpful to know how the authors developed their exposure metric. Why, for example, were quintiles selected rather than another interval scheme, and why were they not of uniform size? Was the reported dose response the only model considered, or did the authors investigate other models, as some have done in studying the relationship of lead exposure and cognition ? noted that their analyses were limited by the cross-sectional nature of the National Health and Nutrition Examination Survey data they used, precluding adjustment of their model for certain covariates and potential confounders (e.g., parental psychopathology). Based on data from multiple studies, ADHD heritability has been estimated to be about 75% (Biederman and Faraone 2005). Inability to adjust for parental psychopathology is therefore an important limitation, because adjustment would likely reduce-and might eliminate-the associations of ADHD with ETS and lead. In studies of lead exposure and cognition, some of which cited as being consistent with their findings, the strength of the IQ-lead relationship can be dwarfed by the relationship of IQ to other factors such as parenting and socioeconomic status (Koller et al. 2004). When reporting associations of environmental contaminants and pathology, it seems prudent to maintain a broader perspective, as well as an environmental health perspective. The author declares he has no competing financial interests. Traumatic aorto-pulmonary artery fistula: a

case report Abstract Background Aorta-pulmonary (A-P) artery fistula following a stab wound to the chest with superimposed infective endocarditis (IE) is a rare, often unrecognized presentation. Herein, we report a case of A-P fistula due to stab chest assessed by two- and three-dimensional (3D) imaging. Case summary A 30-year-old man presented with a history of being stabbed in the chest with a screwdriver. The chest wall laceration was sutured, an intercostal drain inserted for a haemopneumothorax, and he was subsequently discharged. He presented 3 weeks later with exertional dyspnoea, fever, rigours, and loss of weight. On examination, he had a wide pulse pressure and a harsh continuous murmur in the 2nd left intercostal space associated with a palpable thrill. Blood tests revealed raised infective markers and anaemia. All blood cultures were sterile. On echocardiography, the aortic and pulmonary valve was severely damaged, with suspicion of superimposed vegetations secondary to IE. There was severe aortic and pulmonary valve regurgitation. A fistulous connection was noted between the aorta and main pulmonary artery, just below the commissure adjoining the right and left coronary sinus of the aortic valve. On 3D imaging, the defect was quantified. The patient was subsequently referred for aortic and pulmonary valve replacement and closure of the A-P fistula. The presence of multiple vegetations was confirmed intraoperatively. He also received a 6-week course of intravenous antibiotics. Discussion We have described a rare case of an A-P fistula due to a stab wound to the chest complicated by IE. In a patient with stab wound to the chest, a high index of suspicion of cardiac involvement must be maintained, and a careful search for intracardiac shunts must be made on echocardiography, prior to discharge. Furthermore, in addition to two-dimensional imaging, 3D imaging proved useful in providing a comprehensive assessment of the morphology of the lesion prior to surgery. Introduction Aorta-right ventricular fistulas are a rare presentation after penetrating chest trauma. 1 The majority are well-tolerated, but over time, which may result in heart failure if left untreated. Once detected prompt surgical intervention is recommended to prevent sequelae such as heart failure and infective endocarditis (IE). 1 There have been no cases of aorta-pulmonary (A-P) artery fistulas, with concurrent aortic and pulmonary valve regurgitation, reported in the recent literature. One case report, published in 2005, stated that aorta-right ventricular fistula with aortic insufficiency is a rare lesion and only 17 such cases had been previously reported. 2 Some patients may present early due to haemodynamic instability, whereas others may have a more protracted course resulting in a delayed presentation. In a South African case study, a patient presented 6 years after a penetrating chest injury with exertional dyspnoea and a continuous murmur due to aorta-right ventricular fistula complicated by aortic regurgitation. 3 The delayed presentation was explained by gradual fibrosis of the aortic valve cusp and the resultant larger defect. In a case series review of aorta-right ventricular fistula by Samuels et al., 4 it was concluded that due to unfavourable natural history of these lesions, a definitive repair of this structural defect must be performed at the index hospitalization. In 2015, Dieng et al. 5 reported penetrating trauma to the heart with a screwdriver, in this case, the patient sustained injury to the right ventricle and right coronary vein and underwent surgical repair. In the last decade, case reports pertaining to penetrating wounds to the chest resulting in aorta-right ventricular fistula have been scarce. We present a unique case of an A-P artery fistula complicated by IE and double valve involvement, and its assessment by three-dimensional (3D) echocardiography. Case presentation A 30-year-old man with no comorbidities presented to the emergency department with a history of being stabbed in the chest with a screwdriver a few hours prior to arrival. The entry wound was located in left parasternal region in the second intercostal space ( Figure 1). The screwdriver had already been removed. No other injuries were noted. The patient was normotensive (112/74 mmHg), had a tachycardia (112/min), and was tachypnoeic (34 breaths per minute and saturation of 92%). Reduced air entry was noted on the left side of the chest but no murmurs were noted. No blood tests could be traced from the original admission. A chest X-ray confirmed a left haemopneumothorax. The haemothorax occupying about a third of the left hemithorax and the pneumothorax compressed about a third of the left lung. The volume of blood drained was not recorded. No blood transfusion was required. A transthoracic echocardiogram performed in the emergency department seemingly did not detect cardiac pathology. A computed tomography (CT) scan of the chest was not performed. The chest wall laceration was sutured and an intercostal drain inserted on the left side. He was subsequently discharged after removal of the intercostal drain. He presented 3 weeks later with exertional dyspnoea, fever, rigours, and loss of weight. On examination his blood pressure was 102/58 mmHg, heart rate was 115 beats/min, respiratory rate of 30 breaths/min, and temperature of 36.8 C. There was a harsh continuous murmur in the second left intercostal space associated with a palpable thrill. All other systems were clinically unremarkable. A 12-lead electrocardiogram showed sinus tachycardia. Blood tests revealed raised infective markers with a white cell count of 20 Â 10 9 cells/mL (Normal 3.92-10.4 Â 10 9 cells/mL) and an elevated Creactive protein of 118 mg/L (Normal <10 mg/L). He was also anaemic with a haemoglobin of 8.9 g/dL (Normal 13.4-17.5 g/dL). All blood cultures were sterile. The chest X-ray showed a small pleural effusion on the left. The transthoracic echocardiogram showed good left ventricular function, with an ejection fraction of 58%. A small pericardial effusion (<1 cm) was noted anterior to the right ventricle. The aortic valve and pulmonary valves were severely damaged with suspicion of superimposed vegetations secondary to IE ( Figure 2). There was severe aortic regurgitation primarily due to destruction of right coronary cusp. Additionally, there was moderate eccentric pulmonary regurgitation. There was a fistula between the aorta and main pulmonary artery, just below the commissure between the right and left coronary sinus of the aortic valve (Figures 2 and 3). On 3D transoesophageal echocardiography, the defect measured 1.4 Â 0.6 cm in diameter with area of 0.9 cm 2 (Figures 4-6). There was moderate, functional, mitral regurgitation, and moderate tricuspid regurgitation. The pulmonary artery systolic pressure was 40 mmHg and preserved right ventricular systolic function was noted. The patient was subsequently referred for surgical repair. Surgical debridement of all necrotic and infected material was performed. Closure of the A-P artery fistula was performed with a bovine pericardial patch. The aortic valve was replaced with a 25-mm CardiaMed bileaflet mechanical valve, and the pulmonary valve was replaced with a 27mm Mitroflow bioprosthetic valve. The cardiopulmonary bypass time was 196 min. The presence of multiple vegetations was confirmed intraoperatively (Figure 7). No organism was identified on cultures of the resected valves. He was given 2 weeks of antibiotics prior to surgery and 4 weeks after surgery. He received ceftriaxone 2 g daily intravenously for a total of 6 weeks, gentamycin 240 mg daily intravenously for the first 2 weeks and vancomycin 1 g 12 hourly intravenously (drug levels for the latter two drugs were monitored). At follow-up, 2 weeks after discharge by the cardiothoracic surgeons, he was asymptomatic and had no murmurs clinically. On echocardiography, he had a trivial aortic valve paravalvular leak, normal pulmonary valve, and an intact pericardial patch repair. Discussion Aorta-pulmonary artery fistula following a stab wound to the chest with superimposed IE is a rare presentation. There are several points of interest in this case that need to be highlighted. Firstly, this patient was stabbed with a screwdriver, an unusual weapon. Unfortunately, the screwdriver was removed prior to the patient's arrival, so the angle of entry could not be ascertained on clinical examination. Furthermore, there was a delay in the diagnosis of the fistula due to two possible reasons. Firstly, the patient may have been initially asymptomatic of cardiac symptoms and the continuous murmur missed during the first hospital admission. Alternatively, during the index presentation, there was no fistula, and the patient presented later due to superimposed IE, secondary to infection due to the penetrating wound, eventually resulting in tissue necrosis and fistula formation. Thus, a high index of suspicion must be maintained, and a careful search for intracardiac shunts must be made on echocardiography, prior to discharge. Furthermore, the initial cardiac echocardiogram was performed in the emergency department by a junior doctor and not by an expert in echocardiography. We suspect the cardiac lesion may have been missed at the index admission. Thus, we recommend that all cases where there is suspicion of cardiac trauma an expert echocardiographer must be consulted. Furthermore, a CT scan of the chest should have been performed which may possibly have identified the pericardial effusion and the fistula. A multidisciplinary team approach these patients has been shown to improve patient outcomes. 6 Additionally, such patients must be followed-up closely after discharge as these lesions may only manifest at a later stage due to poor healing or fibrosis. 3 It is known that healthy endothelium cannot be easily colonized by bacteria. 7 Bacteria require endothelial disruption to adhere, multiply and cause infection in the form of IE. The mechanism of endothelial disruption may be related to direct trauma, amongst others, from invasive procedures or penetrating trauma as represented in this case. There was a likely injury to the endothelium of the semilunar valves and concomitant introduction of bacteria by the contaminated screwdriver. The endothelial disruption resulted in exposure of the tissue matrix, leading to platelet adherence, deposition of fibrin, formation of microthrombi and promotion of bacterial adhesion and proliferation. This resulted in the formation of vegetations and enlargement of the fistula. The fistula, in turn, resulted in further damage to the endothelium due to a high-velocity jet caused by the left to right shunting of blood from the aorta to the pulmonary artery, causing further bacterial seeding. Physical examination alone is not sufficient for detection of the direction and extent of cardiac injury from penetrating trauma. 8 It is neither sensitive nor specific for recognizing significant pathology. Usually, the shape and size of the weapon and the angle of entry of the instrument provides clues to the potential injuries. The size of the external injury may grossly underestimate the extent of internal injury. Penetrating trauma to the so-called 'box' region should raise a red flag, as there is a high risk of injury to the heart and surrounding mediastinal structures. 8 The box is superiorly defined by the clavicles and sternal notch, laterally by vertical lines through the nipples, and inferiorly by the costal margins. The incidence and the natural history of penetrating cardiac injury are not well studied. Some authors have reported secondary cardiac injuries at follow-up, such as fistulas and ventricular septal defects. Based on the aforementioned observations it is suggested that these patients should undergo a follow-up transthoracic echocardiogram, as haemodynamically insignificant or small cardiac lesions may be missed at the initial encounter. 3,9 A baseline CT scan of the chest should also be performed to exclude a haemopericardium and a haemopneumothorax, fistula and intracardiac septal defects. Concurrent involvement of the aortic and pulmonary valve resulting in haemodynamically significant lesions is unique. The aforementioned combination of lesions secondary to a penetrating chest wound has not been reported in the recent literature. Furthermore, in this case, 3D echocardiography allowed a comprehensive assessment of the morphology of the A-P artery fistula, aortic and pulmonary valve, as well as the superimposed vegetations. Three-dimensional transoesophageal echocardiography allowed acquisition of volumetric data, from which the multiplane reconstruction of the orifice of the fistula allowed us to measure the diameter and the area of the asymmetrical orifice. [10][11][12] This information proved useful prior to surgical repair. Additionally, 3D echocardiography allowed enface view visualization of the defect. Threedimensional imaging has certain inherent limitations which must be taken into account when calculating orifice area in general. 11 The main limitations of 3D echocardiography include: (i) its low temporal and spatial resolution which may cause errors in area measurement of the orifice and can miss small vegetations; (ii) reliance on adequate two-dimensional image quality limits its use in patients with poor echo windows; and (iii) presence

of artefacts such as due to blooming, can result in underestimation of the orifice area. 12 Conclusion We have described a rare case of an A-P artery fistula with concurrent aortic and pulmonary valve regurgitation due to a stab wound to the chest complicated by IE. In patients with stab wounds to the chest, careful physical examination, a chest X-ray, a thorough echocardiogram and a CT scan of the chest should be performed. Furthermore, in addition to two-dimensional echocardiography, 3D echocardiography is useful in providing a comprehensive morphological assessment prior to surgery. Traumatic aorto-pulmonary artery fistula Methemoglobinemia during the Use of Glyceryl Trinitrate Patches in Neonates: Two Case Reports Methemoglobinemia can result in severe hypoxia. It has been frequently reported during the use of inhaled nitric oxide, but can occur where nitrate containing medications are used. Glyceryl trinitrate (GTN) patches have been used in the treatment of digital and limb ischemia in prematurely born infants. Little is known about the pharmacokinetics of GTN when incorporated into patches. Studies of other topical forms of nitroglycerine have shown a wide range of absorption. It is likely that the increased permeability of the prematurely born infant's skin would facilitate absorption. We describe the use of GTN patches in two very prematurely born infants used to treat limb/digit ischemia. This resulted in methemoglobinemia and resultant increase in their supplementary oxygen requirements. Removal of the patches was associated with a reduction in their methemoglobin levels and the supplementary oxygen requirements back to baseline levels. In conclusion, routine monitoring of methemoglobin levels should be undertaken when GTN patches are used in very prematurely born infants. Methemoglobinemia is formed as a result of the oxidization of iron from its ferrous (Fe 2þ ) to ferric (Fe 3þ ) state and can occur following exposure not only to inhaled nitric oxide (iNO) but also nitrite containing medications. Methemoglobinemia has been frequently reported in infants treated with iNO. 1 The levels of methemoglobinemia relate to the iNO levels used. 1 Methemoglobin has a higher oxygen-binding capacity than hemoglobin and thus affected infants can suffer severe hypoxia. Ischemia can be caused by a combination of vascular damage and concomitant vasoconstriction. 2 Glyceryl trinitrate (GTN) patches are used in the treatment of ischemic digits and limbs in neonates 2,3 and in adults. 4 GTN is converted to NO in the vascular smooth muscle, and this activates guanylate cyclase and increases the levels of cyclic guanosine monophosphate (cGMP); 5 cGMP causes relaxation of the vascular smooth muscle in the affected artery and Abstract Methemoglobinemia can result in severe hypoxia. It has been frequently reported during the use of inhaled nitric oxide, but can occur where nitrate containing medications are used. Glyceryl trinitrate (GTN) patches have been used in the treatment of digital and limb ischemia in prematurely born infants. Little is known about the pharmacokinetics of GTN when incorporated into patches. Studies of other topical forms of nitroglycerine have shown a wide range of absorption. It is likely that the increased permeability of the prematurely born infant's skin would facilitate absorption. We describe the use of GTN patches in two very prematurely born infants used to treat limb/digit ischemia. This resulted in methemoglobinemia and resultant increase in their supplementary oxygen requirements. Removal of the patches was associated with a reduction in their methemoglobin levels and the supplementary oxygen requirements back to baseline levels. In conclusion, routine monitoring of methemoglobin levels should be undertaken when GTN patches are used in very prematurely born infants. dilation of the collateral circulation. 2 GTN patches, however, remain unlicensed for use in neonates and little is known about the pharmacokinetics of GTN when used in patch formation in prematurely born neonates. Methemoglobinemia is not listed as a side effect of the use of GTN patches. The absorption of nitroglycerin from topical ointments has been shown to be variable in infants. GTN patches vary in size and dose. A 10-cm 2 patch typically contains 40 mg of nitroglycerin and is designed to release this at 0.2 mg/h; however, in one study of seven infants, an application of 1 mg in an ointment resulted in blood levels of between 0.03 and 3.36 ng/mL. 6 The skin of prematurely born infants has increased permeability compared with term born infants which might lead to greater absorption of topical treatments. 7 We present two cases of methemoglobinemia which occurred during the use of GTN patches to treat ischemic digits and limbs of very prematurely born infants. In both cases, 9 cm 2 patches were used which contained 18.7 mg GTN. Case 1 An infant with a birth weight of 650 g was born at 24 weeks of gestational age following the onset of spontaneous preterm labor. During initial resuscitation, it was noted that the lower limbs of the infant were dusky and this did not resolve despite improving the blood pressure levels by administering volume replacement and inotropes. One GTN patch was applied to each lower limb over the area of the posterior tibial arteries. These were left in place as the infant was stabilized and transferred to the local tertiary unit. A radial arterial line was inserted the following day, day 2 after birth. The thumb and forefinger distal to the radial arterial line became poorly perfused shortly after insertion and a further GTN patch was applied to the hand. This was in addition to the two patches in situ on the feet. The oxygen requirement was noted to rise from an inspired oxygen fraction (FiO 2 ) of 0.21 to 0.4 within 6 hours of application of the third patch. The infant's inspired oxygen concentration was altered as necessary to keep the oxygen saturation levels between 92 and 95%. Serial capillary blood gases showed an increase in methemoglobinemia from 1.1% prearterial line insertion to 8.4% over 6 hours. In total, three patches had been applied over the 48-hour period. All patches were immediately removed and the methemoglobinemia level fell to less than 1% within 2 hours. There was a corresponding reduction in the oxygen requirement to the level prior to the application of the patch. Perfusion of the hand and lower limbs completely recovered. Case 2 An infant with a birth weight of 617 g was born at 24 weeks of gestational age via a vaginal delivery following premature rupture of membranes 1 week previously. After initial resuscitation, the infant was transferred to a tertiary neonatal unit. Umbilical arterial access was not obtained and a right posterior tibial arterial line was inserted on day 2 after birth. The following day, it was noted that it was difficult to sample from the arterial line and to obtain an adequate blood pressure trace. The toes and foot were cool to touch and blanched white on examination. The arterial line was removed and a GTN patch immediately applied and replaced daily. The toes on the right foot became necrotic on day 7 and, following consultation with the vascular team, two GTN patches were applied to the foot, one proximally and one posteriorly to the point where the arterial line had been inserted. The patches were changed twice daily. Over the following 24 hours, the infant's supplementary oxygen requirement increased from an FiO 2 of 0.23 to an FiO 2 of 0.70. The infant's inspired oxygen concentration was altered as necessary to keep the oxygen saturation levels between 92 and 95%. At that time, it was noted that the methemoglobin level was 14.5%. The GTN patches were immediately removed. A total of 10 patches had been used during the previous 5 days and 5 of those patches had been sited in 24 hours prior to removal of the patches. Over the following 24 hours, the methemoglobin level fell from the highest level of 23.3 to 2.5%, with a corresponding reduction in the oxygenation index (mean airway pressure  FiO 2  100 divided by PaO 2 , kPa), from 24 to 8. The tips of the toes that were necrotic sloughed off after several weeks, but no amputation was required. Discussion We have demonstrated GTN patch use in very prematurely born infants can result in methemoglobinemia. In both infants, this was associated with an increase in their inspired oxygen requirement. In the first case, the ischemia completely resolved, but in the second resolution was incomplete and as a consequence, the toes became necrotic. In both cases due to the development of the methemoglobinemia, the patches were removed prematurely, which may have reduced their efficacy. Previous case reports have suggested a treatment duration of 28 days. 2 It is not clear whether the methemoglobinemia was the cumulative effect of all the patches applied or after a certain number of patches a threshold was reached. It is known, however, that prematurely born infants have lower levels of methemoglobin reductase which makes them more susceptible to methemoglobinemia. 8,9 Once the patches were removed, the methemoglobin levels decreased and the supplementary oxygen requirement returned to the level prior to treatment with the patches. Thus, we suggest the use of the GTN patches resulted in the methemoglobinemia. As a consequence, we recommend if GTN patches are used in a very prematurely born infant, frequent assessment of methemoglobin levels should be undertaken. Melanoma whole exome sequencing identifies V600EB-RAF amplification-mediated acquired B-RAF inhibitor resistance The development of acquired drug resistance hampers the long-term success of B-RAF inhibitor (B-RAFi) therapy for melanoma patients. Here we show V600EB-RAF copy number gain as a mechanism of acquired B-RAFi resistance in four out of twenty (20%) patients treated with B-RAFi. In cell lines, V600EB-RAF over-expression and knockdown conferred B-RAFi resistance and sensitivity, respectively. In V600EB-RAF amplification-driven (vs. mutant N-RAS-driven) B-RAFi resistance, ERK reactivation is saturable, with higher doses of vemurafenib down-regulating pERK and re-sensitizing melanoma cells to B-RAFi. These two mechanisms of ERK reactivation are sensitive to the MEK1/2 inhibitor AZD6244/selumetinib or its combination with the B-RAFi vemurafenib. In contrast to mutant N-RAS-mediated V600EB-RAF bypass, which is sensitive to C-RAF knockdown, V600EB-RAF amplification-mediated resistance functions largely independently of C-RAF. Thus, alternative clinical strategies may potentially overcome distinct modes of ERK reactivation underlying acquired B-RAFi resistance in melanoma. Introduction Activating B-RAF V600 kinase mutations occur in ~50% of melanomas 1 , and the ATPcompetitive type I RAF inhibitors, PLX4032/vemurafenib and GSK2118436, display remarkable activity leading to overall survival advantage in patients with V600 B-RAF mutant melanomas [2][3][4][5][6] . Acquisition of drug resistance leading to clinical relapse, however, develops in virtually all patients treated with B-RAF inhibitors (B-RAFi) 4,5 . Heterogeneous mechanisms of acquired B-RAFi resistance hitherto uncovered fall into general MAPKredundant, AKT-dependent 7,8 or MAPK-reactivating 9,10 pathways, indicating specific translatable therapeutic strategies to prevent or overcome resistance. Contrary to expectation, V600E B-RAF secondary mutations have not been found to account for acquired B-RAFi resistance 10 , suggesting V600E B-RAF-bypass mechanisms as the principal means to ERK reactivation. Here we observed an alteration in V600E B-RAF, namely genomic copy number gain, in tumors of melanoma patients whose cancer progressed after initial responses to B-RAF inhibitors. We demonstrated that this V600E B-RAF amplification results in V600E B-RAF over-expression, which is necessary and sufficient for acquired resistance to B-RAF inhibitor. This finding, along with a recent study reporting N-terminal truncation of V600E B-RAF causing acquired B-RAFi resistance in melanoma 11 , underscores key molecular alterations in the drug target itself. We further suggest that V600E B-RAF-instrinsic (amplification, truncation) vs. V600E B-RAF-bypass (N-RAS mutations) mechanisms, both reactivating the MAPK pathway, may offer insights into distinct therapeutic strategies to overcome acquired B-RAFi resistance in melanoma. Whole exome sequencing identifies V600E B-RAF amplification We assembled twenty sets of patient-matched baseline (prior to B-RAFi therapy) and disease progression (DP) (i.e., acquired B-RAFi resistance) melanoma tissues and analyzed them to identify the proposed mechanisms of acquired B-RAFi resistance in melanoma. These reported mechanisms include N-RAS 10 and MEK1 12 mutations, alternativespliced V600E B-RAF variants 11 , and over-expression of RTKs (PDGFRβ 7,10 , IGF1-R 8 ) and COT 9 (Tables 1 and Supplementary Table S1; Supplementary Fig. S1). For DP samples negative for these mechanisms and where there was sufficient frozen and patient-matched normal tissues (from patients #4, 5, 8, 14, 16, 17 & 18), we subjected triads of genomic DNAs (gDNAs) from normal, baseline, and DP tissues to whole exome sequencing. In two available data sets, we searched for somatic DP-specific

non-synonymous single nucleotide variants (nsSNVs) and small insertion-deletion (indels), which were exceedingly few in number or absent, respectively, using our bioinformatic workflow (Supplementary Tables S2 and S3). We also analyzed for DP-specific copy number variations (CNVs) from the exome sequence data (Supplementary Table S2). This identified V600E B-RAF copy number gains in these two patients' DP tissues (2.2 and 12.8 fold in patients #5 and 8, respectively) relative to their respective baseline tissues ( Fig. 1a; Table 1). Gain in V600E B-RAF copy number was reflected in corresponding increased gene expression at the protein level (Fig. 1b). V600E B-RAF amplification was validated by gDNA Q-PCR, producing consistent fold increases in DP-specific V600E B-RAF copy number gain (relative to baseline) (2.0 and 14 fold increase in patient #5 and 8 respectively) (Fig. 1c). We then expanded the analysis of V600E B-RAF amplification to all twenty paired melanoma tissues and detected V600E B-RAF copy number gains in DP samples from two additional patients (2.3 and 3 fold for DP2 of patient #9 & DP of patient #13, respectively) ( Fig. 1c; Table 1). We note that these copy number fold increases are likely underestimates of the true changes due to non-tumor diploid cell contents and tumor heterogeneity, as most disease progressive tumors occur from stable residual tumors as a result of partial responses seen in the vast majority of patients treated with B-RAF inhibitors. An increase in the mutant B-RAF to WT B-RAF ratio was also noted in all four cases of DP harboring B-RAF copy number gain when compared to their respective baseline tissues (Fig. 1d), consistent with selection for V600E B-RAF (vs. the WT B-RAF allele) copy number gain during acquisition of B-RAFi resistance. V600E B-RAF amplification was largely mutually exclusive with N-RAS mutations (no enrichment in MEK1 exon 3 mutation was detected in DP vs. baseline tumors), RTK over-expression (no COT over-expression detected), as well as a novel mechanism involving V600E B-RAF alternative splicing 11 (Table 1; Supplementary Fig. S1). B-RAFi selects for V600E B-RAF gain and over-expression We have derived vemurafenib/PLX4032-resistant (R) sub-lines by providing continuous vemurafenib exposure to seven human melanoma-derived V600E BRAF-positive parental (P) cell lines sensitive to vemurafenib-mediated growth inhibition. Four resistant sub-lines, including M229 R5 and M238 R1 7,10 , over-expressed PDGFRβ compared to their parental counterpart. One sub-line (M249 R4 10 ) gained a mutation in N-RAS, and another (M397 R) an alternatively spliced variant of V600E B-RAF resulting in in-frame fusion of exons 1 and 11 ( Supplementary Fig. S2). As in our tissue analysis, these mechanisms were identified in a mutually exclusive manner. Another vemurafenib-resistant sub-line, M395 R, was derived from a V600E B-RAF-homozygous parental line, M395 P ( Supplementary Fig. S3a). Compared to M395 P, M395 R harbors increased copy numbers of V600E B-RAF gDNA and cDNA, consistent with a dramatic V600E B-RAF protein over-expression ( Supplementary Fig. S3b, c, and d). M395 R displays growth highly resistant to vemurafenib treatment ( Supplementary Fig. S4a), and titration of M395 R with vemurafenib (1 h) after a 24 h of drug withdrawal revealed pERK levels to be highly resistant to acute V600E B-RAF inhibition ( Supplementary Fig. S4b). This pattern of MAPK reactivation was similar to that seen in a mutant N-RAS-driven, vemurafenib-resistant sub-line, M249 R4, and contrasted with that in the RTK-driven vemurafenib-resistant sub-line, M229 R5 ( Supplementary Fig. S4b) 7,10 . Expectedly, the levels of p-AKT are unchanged (Fig. 2b) comparing M395 P vs. M395 R, consistent with a lack of RTK over-expression leading to MAPK-redundant, PI3K-AKT signaling 7 . Accordingly, M395 R does not over-express either PDGFRβ or IGF-1R, in contrast to M229 R5, which has been shown to over-express the RTK PDGFRβ ( Supplementary Fig. S4c) 7,8 . Additionally, M395 R is WT for N-, H-and K-RAS and MEK1, harbors no secondary mutations in V600E B-RAF or an alternatively spliced variant of V600E B-RAF which results in a N-terminally truncated V600E B-RAF protein. Modest V600E B-RAF over-expression leads to B-RAFi resistance Three different but uniformly modest levels of V600E B-RAF over-expression were achieved by infecting M395 P with varying viral titers and subsequent puromycin selection. This resulted in relatively low (1.9 fold over empty vector virus control), medium (2.4 fold) and high (2.8 fold) levels of V600E B-RAF RNA/cDNA over-expression ( Supplementary Fig. S5), with the corresponding protein over-expression levels shown in Figure 2a. In comparison, in two sets of tissues (from patients #8 and #13) where flash-frozen tissues were available, the RNA/cDNA levels of V600E B-RAF in the DP tumors were 9.5 and 1.4 fold relative to those in their patient-matched baseline tumors. Notably, the DP tumor from patient #13 was obtained by an intervention radiology-guided needle biopsy of a pelvic mass (Supplementary Table S1) and contained a high admixture of normal and tumor contents (latter indicated by S100), which likely contributed to an underestimation of the true change in the V600E B-RAF RNA/cDNA levels. V600E B-RAF gain leads to drug-saturable resistance The modest and incremental over-expression of V600E B-RAF at the RNA and protein levels in M395 P conferred similar degrees of vemurafenib resistance (Fig. 2b). Interestingly, further V600E B-RAF over-expression at a much greater level, as in the case of M395 R relative to M395 P (increase in RNA/cDNA level shown in Supplementary Fig. S3c and S5; increase in protein level shown in Fig. 2c) conferred enhanced drug resistance mainly at 1 μM vemurafenib but not 10 μM vemurafenib (Fig. 2d). Thus, a modest V600E B-RAF copy number gain and over-expression can confer vemurafenib resistance, and even high amplitude V600E B-RAF amplification and over-expression can be readily saturable by micromolar concentrations of vemurafenib. Moreover, V600E B-RAF knockdown in M395 R confers vemurafenib sensitivity ( Fig. 2c and d). Consistently, V600E B-RAF over-expression in M395 P (at a level titrated to be comparable to M395 R) and its knockdown in M395 R resulted in pERK resistance and sensitivity, respectively, to acute vemurafenib treatment after a 24 h drug withdrawal (Fig. 2e). We predicted that, regardless of the cellular genetic context, MAPK reactivation due to drug target (i.e., V600E B-RAF) over-expression would be saturable by higher doses of vemurafenib, in contrast to mutant N-RAS-mediated MAPK reactivation where V600E B-RAF may be bypassed by the alternative use of C-RAF 13 . Indeed, dosing of vemurafenib from 1 to 50 μM revealed a significant difference in drug sensitivity of M249 R4 ( Q61K N-RAS) vs. M395 R (amplified V600E B-RAF) (Fig. 3a) (where the latter was highly sensitive to vemurafenib at this drug concentration range), suggesting a potential therapeutic opportunity. To rule out that these results were not due to a difference in genetic backgrounds, we artificially rendered the V600E B-RAF melanoma cell line, M229, vemurafenib-resistant by either Q61K N-RAS or V600E B-RAF viral transduction (Fig. 3b). Again, high dose vemurafenib treatment was more effective at overcoming drug resistance in V600E B-RAF-transduced M229 than in the same cell line transduced with Q61K N-RAS. MEK inhibition restores vemurafenib sensitivity Since both N-RAS mutation and V600E B-RAF amplification-driven acquired resistance mechanisms would be anticipated to result in MEK reactivation, we tested the allosteric MEKi, AZD6244/selumetinib, on the Q61K N-RAS-driven M249 R4 and the V600E B-RAF amplification-driven M395 R sub-lines. MEKi treatment resulted in decreased proliferation in both cases, but the activity was noted at lower concentrations for the Q61K N-RAS-driven resistance mechanism (Fig. 3c). This differential pattern was reproducible by exposing AZD6244/selumetinib to V600E B-RAF melanoma cell lines M229 and M238 transduced with high levels of V600E B-RAF vs. a short-term culture, Pt55 R 10 , with Q61K N-RAS-driven acquired B-RAFi resistance (Fig. 3d). We also tested the combination of B-RAFi with MEKi, which is currently in clinical testing 14 , in three-day survival assays. A calculation of combination index (CI) values using equal ratios of vemurafenib and selumetinib was performed. The results were consistent with a highly synergistic effect of these two agents combined in overcoming both mutant N-RAS-driven (M249 R4) and V600E B-RAF amplification-driven B-RAFi resistance (M395 R) (Fig 3e and 3f), although the combination tended to be more potent against mutant N-RAS-driven acquired resistance to vemurafenib. This B-RAFi and MEKi combinatorial synergy was further corroborated in longer-term clonogenic assays (Fig. 3g). Differential C-RAF dependency of ERK-reactivating mechanisms We also predicted that MAPK reactivation due to V600E B-RAF over-expression would be C-RAF-independent, in contrast to mutant N-RAS-mediated MAPK reactivation where V600E B-RAF may be bypassed by the alternative use of C-RAF. Indeed, C-RAF knockdown by shRNA sensitized the mutant N-RAS sub-line, M249 R4, but not the V600E B-RAF amplified sub-line, M395 R, to vemurafenib in three-day survival assays (Fig. 3h). C-RAF knockdown restored vemurafenib sensitivity to M249 R4 ( Q61K N-RAS/ V600E B-RAF) even more strikingly in a longer-term clonogenic assays which afforded fresh drug replacement every two days (Fig. 3i). An independent C-RAF shRNA also restored vemurafenib sensitivity to M249 R4 (Supplementary Table S4). Additionally, B-RAFi and MEKi synergy and C-RAF-dependence in mutant N-RAS-driven acquired B-RAFi resistance was confirmed in a short-term culture derived from a tumor with clinical acquired vemurafenib resistance (Supplementary Fig. S6). Discussion Identification of V600E B-RAF amplification as a mechanism of acquired resistance in B-RAFi treated patients provides evidence for alterations in the drug target causing clinical relapse. Based on these studies, therapeutic stratification of MAPK reactivation underlying B-RAFi resistance into drug-saturable or C-RAF-dependent pathways may be translatable into the design of next-generation clinical trials aimed at preventing or overcoming B-RAFi resistance (Fig. 4). These findings also provide pre-clinical rationale for dose escalation studies in selected patients with B-RAFi-resistant V600E/K B-RAF metastatic melanomas, particularly given the wide range of effective dosing and the fact that the maximum tolerated dose of GSK2118436 has not been determined. The combination of current B-RAF inhibitors (or next-generation RAF inhibitors that enhance B-RAF potency or feature pan-RAF inhibition) with MEK1/2 inhibitors may potentially broadly block MAPK reactivation. Emerging evidence points to B-RAF mutant cancers of other tissue origin or lineage being less responsive to specific B-RAF inhibition than B-RAF mutant melanomas. Mechanisms of acquired B-RAF inhibitor resistance may turn out to be instructive for understanding primary resistance of B-RAF mutant cancer types to B-RAF inhibitors, as primary (de novo) and secondary (or acquired) drug resistance may be clinical manifestations from a spectrum of molecular alterations that are mechanistically linked. Thus, multiple modes (e.g., mutation, copy number gain) of up-regulating oncogene activity, which may pre-exist in the same tumor and/or patient, may help explain the range of heterogeneous responses of B-RAF mutant cancers to direct B-RAF, MEK or ERK inhibition. Cell culture experiments Cells were maintained in DMEM with 10 or 20% fetal bovine serum and glutamine. shRNAs (Supplementary Table S4) for B-RAF and C-RAF were sub-cloned into the lentiviral vector pLL3.7; pBabe B-RAF (V600E) was purchase (plasmid 17544, Addgene); viral supernatants generated by co-transfection with three packaging plasmids into HEK293T cells; and infections carried out with protamine sulfate. Stocks and dilutions of PLX4032 (Plexxikon, Berkeley, CA) and AZD6244 (commercially available) were made in DMSO. Cells were quantified using CellTiter-GLO Luminescence (Promega) or crystal violet staining followed by NIH Image J quantification. Whole exome sequencing Human tissues were obtained with patient-informed consent under UCLA Institutional Review Board (#10-001089) approval. For each sample, 3ug of high molecular weight genomic DNA was used as the starting material to generate the sequencing library. Exome captures were performed using Agilent SureSelect Human All Exon 50mb and Agilent SureSelect Human All Exon 50mb XT for PT #5 and Pt #8, respectively, per manufacturers' recommendation, to create a mean 200bp insert library. For Pt #5, sequencing was performed on Illumina GenomeAnalyzerII (GAII) as 76+76bp paired-end run. The normal sample was run on 1 flowcell lane and the tumor samples were run on 2 flowcell lanes each. For Pt #8, sequencing was performed on Illumina HiSeq2000 as 50+50bp paired-end run and 100+100bp paired-end run. The three samples (normal, baseline and DP) were initially mixed with 9 other samples and run across 5 flowcell lanes for the 50+50bp run. For the 100+100bp run, they were mixed with 3 other samples to be run across 5 flowcell lanes with barcoding of each individual genomic sample library. For Pt #5, approximately 62 million, 137 million, 147 million reads were generated for normal tissue (skin), baseline melanoma and DP melanoma, respectively, with 75.2%, 78.1%, and 74.7% of the

reads mapping to capture targets. Based on an analysis of reads that uniquely aligned to the reference genome and for which the potential PCR duplicates were removed, an average coverage of 52X, 88X, and 114X was achieved with 87%, 92% and 93% of the targeted bases being covered at 10X or greater read depth for normal, baseline and DP, respectively. For Pt #8, approximately 198 million, 270 million, 256 million reads were generated for normal tissue (skin), baseline melanoma and DP melanoma, respectively with 43.2%, 44.1% and 42.3% of the reads mapping to capture targets. Based on an analysis of reads that uniquely aligned to the reference genome and for which the potential PCR duplicates were removed, an average read depth of 107X, 132X and 123X was achieved with 89%, 90% and 90% of the targeted bases being covered at 10X or greater for normal, baseline and DP, respectively. Sequencing data analysis For Pt #8 where the samples were indexed and pooled before the sequencing, Novobarcode from Novocraft was used to demultiplex the data. The sequence reads were aligned to the human reference genome using Novoalign V2.07.13 from Novocraft (http:// www.novocraft.com). For Pt #5, hg18 downloaded from UCSC genome database was used and for Pt #8, b37 downloaded from GATK (Genome analysis toolkit) resources website was used for the reference genome. SAMtools v.0.1.16 16 was used to sort and merge the data and Picard (http://picard.sourceforge.net/) was used to mark PCR duplicates. To correct the misalignments due to the presence of indels, local realignment was performed using RealignerTargetCreator and IndelRealigner of GATK 17 . Indel calls in dbSNP132 were used as known indel input. Then, GATK CountCovariates and TableRecalibration were used to recalibrate the originally reported quality score by using the position of the nucleotide within the read and the preceding and current nucleotide information. Finally, to call the single nucleotide variants (SNVs), the GATK UnifiedGenotyper was used to the realigned and recalibrated bam file while GATK IndelGenotyperV2 was used to call small insertion/ deletions (Indels). To generate a list of somatic variants for DP tumor, the difference in allele distribution was calculated using one-sided Fisher's exact test using normal sample or the baseline sample. Variants with p-value<0.05 were included in the "somatic variant list". Low coverage (<10X) SNVs and SNVs with more than one variant allele in normal tissue and baseline melanoma were filtered out during the process. These somatic variants were further annotated with SeattleSeqSNPannotation (http://gvs.gs.washington.edu/ SeattleSeqAnnotation/). For DP-specific, non-synonymous SNVs that result in missense mutations, we assessed the level of amino acid conservation using PhyloP score (provided in UCSC genome database) where a score > 2 implies high conservation and the nature of amino substitution using Polyphen-2 analysis 18 . CNV analysis was performed using an R package, ExomeCNV 15 . ExomeCNV uses the ratio of read depth between two samples at each capture interval. Here, the read depth data between baseline and DP melanomas were compared. Briefly, the read depth information was extracted through the PILEUP file generated from the BAM file after removing PCR duplicates using SAMtools. The average read depth at each capture interval was calculatedand the classify.eCNV module of ExomeCNV was run with the default parameters to calculate the copy number estimate for each interval. Subsequently, another R package commonly used to segment the copy number intervals, DNAcopy 19 , was called through ExomeCNV multi.CNV.analyze module with default parameters to do segmentation and sequential merging. The genomic regions with copy number 1 were called deletion and any regions with copy number >2 were called amplification. Circos 20 was used to visualize the CNV data. Genomic DNA and RNA quantifications For real-time quantitative PCR, total RNA was extracted and cDNA quantified by the iCycler iQ Real Time PCR Detection System (BioRad). Data were normalized to TUBULIN and GAPDH levels. Relative expression is calculated using the delta-Ct method. gDNAs were extracted using the FlexiGene DNA Kit (Qiagen) (Human Genomic DNA-Female, Promega). B-RAF relative copy number was determined by quantitative PCR (cycle conditions available upon request) using the MyiQ single color Real-Time PCR Detection System (Bio-Rad). Total DNA content was estimated by assaying β-globin for each sample, and 20 ng of gDNA was mixed with the SYBR Green QPCR Master Mix (Bio-Rad) and 2 pmol/L of each primer. All primer sequences are provided in Supplementary Table S4. Supplementary Material Refer to Web version on PubMed Central for supplementary material. Growth curve of did not alter the pERK level in the absence of vemurafenib/PLX4032 but conferred growth resistance to the parental line, M395 P when exposed to indicated concentrations of PLX4032 for 72 h (relative to DMSO-treated controls; mean ± SEM, n = 5). Dashed line, 50% inhibition. (c, d) Transduction of shRNA to knockdown BRAF V600E in the drug-resistant sub-line, M395 R, did not alter the pERK level in the absence of PLX4032 but restored growth sensitivity to PLX4032 (72 h; mean ± SEM, n = 5). (e) Increasing (in M395 P) or decreasing (in M395 R) BRAF V600E levels decreased or increased pERK sensitivity to PLX4032 (0, 0.1, 1, 10 μM) treatments for 1 h, respectively. Table 1 Clinical characteristics and acquired resistance mechanisms in patients with matched baseline and disease progression (DP) melanomas tissues. High throughput transcriptome analysis of lipid metabolism in Syrian hamster liver in absence of an annotated genome Background Whole transcriptome analyses are an essential tool for understanding disease mechanisms. Approaches based on next-generation sequencing provide fast and affordable data but rely on the availability of annotated genomes. However, there are many areas in biomedical research that require non-standard animal models for which genome information is not available. This includes the Syrian hamster Mesocricetus auratus as an important model for dyslipidaemia because it mirrors many aspects of human disease and pharmacological responses. We show that complementary use of two independent next generation sequencing technologies combined with mapping to multiple genome databases allows unambiguous transcript annotation and quantitative transcript imaging. We refer to this approach as “triple match sequencing” (TMS). Results Contigs assembled from a normalized Roche 454 hamster liver library comprising 1.2 million long reads were used to identify 10’800 unique transcripts based on homology to RefSeq database entries from human, mouse, and rat. For mRNA quantification we mapped 82 million SAGE tags (SOLiD) from the same RNA source to the annotated hamster liver transcriptome contigs. We compared the liver transcriptome of hamster with equivalent data from human, rat, minipig, and cynomolgus monkeys to highlight differential gene expression with focus on lipid metabolism. We identify a cluster of five genes functionally related to HDL metabolism that is expressed in human, cynomolgus, minipig, and hamster but lacking in rat as a non-responder species for lipid lowering drugs. Conclusions The TMS approach is suited for fast and inexpensive transcript profiling in cells or tissues of species where a fully annotated genome is not available. The continuously growing number of well annotated reference genomes will further empower reliable transcript identification and thereby raise the utility of the method for any species of interest. Background Clinical dyslipidaemia is defined as an abnormal concentration of lipids in the blood circulation caused by cholesterolrich diet or exposure to elevated insulin levels. It is long known that elevated High-density lipoprotein-cholesterol (HDL-c) levels are associated with reduction of the risk of atherosclerotic cardiovascular diseases [1]. This effect is mechanistically related to reverse cholesterol transport (RCT), a process that mediates excretion of excess HDL cholesterol from peripheral tissues to the liver ending ultimately in the faeces [2]. Cholesteryl ester transfer protein (CETP) plays a central role in this process and it was recognized that pharmacological inhibition of CETP mediated cholesterol transport would raise HDL-c levels resulting in protection from coronary heart disease. The Syrian hamster Mesocricetus auratus has become an important animal model for pre-clinical dyslipidaemia research because HDL-c levels are raised in response to drugs designed for human target proteins [3]. In contrast to mice and rats, CETP is expressed in hamster liver which explains raise of serum HDL levels in response to the CETP inhibitor Anacetrabip [4]. The Illumina human body map 2.0 project provides high-quality RNAseq based gene expression data for major human organs using pooled samples from healthy donors including liver (http://www.ncbi.nlm.nih.gov/geo/). This database allows a comparative analysis of liver gene expression between humans and model organisms such as the primate Macaca fascicularis where a microarray based expression database of 36 liver samples became recently available as a part of a genome sequencing project [5]. Currently, lack of a fully annotated genome for M. auratus prevents development of reliable genome-based tools for transcript imaging such as microarrays or qPCR panels. Recently, the genome of the Chinese hamster Cricetulus griseus derived cell line (CHO)-K1 was published because it is the preferred host cell line for industrial production of recombinant proteins such as therapeutic antibodies [6]. For this reason, genome analysis and annotation was focused on the analysis of protein modifying systems such as fucosylation or glycosylation pathways. A partial transcriptome analysis of an antibody producing CHO line by RNA sequencing had been performed earlier [7]. The RNA source, the low transcript annotation coverage of approximately 40% combined with the absence of data in the public domain limits utility of this study for our project. Furthermore, a recent mitochondrial genome based study of rodent phylogeny shows that evolutionary relationship between M. auratus and C. griseus is comparable to that between R. norvegicus and M.musculus [8]. These circumstances suggested use of well annotated genomes such as rat, mouse or human as template for mapping and reliable annotation of Syrian hamster liver transcripts. We show here, that simultaneous application of two independent deep sequencing technologies combined with searches across multiple database allows indeed unambiguous annotation and quantification of transcripts expressed in liver of the Syrian hamster. Triple match sequencing (TMS) can be applied for transcript profiling of any organism provided that annotated genomes of related species are available. Experimental strategy The underlying principle and the workflow of TMS are summarized in Figure 1 (A and B). From total RNA, two libraries are constructed for NGS using a long-read (Roche 454) and a short-read (ABI SOLiD) technology. For the 454 library construction, we used polyT-primed cDNA because the 3 0 sequence upstream of the poly-A tail is required for quantification of short-read tags. This approach allows identification of genes belonging to homologous families because the untranslated region has in general higher sequence diversity than the coding region. To achieve broad representation of transcripts independent of their expression levels it was crucial to normalize RNA abundance in the cDNA samples by competitive hybridization. 1.2 million 454 reads with an average length of 350 bases were assembled into contigs. Unique contigs together with unassembled reads were subsequently used as query for homology searches across well annotated transcriptome databases such as mouse, rat and human (RefSeq). This approach led to unambiguous annotation of 10'800 transcripts corresponding to 57.4 per cent of the contigs assembled from the normalized 454 long-read library. Probing of human, mouse and rat RefSeq databases with unique query contigs typically produced matches to the same transcript in each library thereby improving annotation fidelity of hamster transcripts. Our primary objective is to quantify hamster liver mRNA abundance for interspecies comparisons (see below). Using total liver RNA from the same samples sequenced by 454 technology, we generated SAGE short read libraries for processing on an ABI-SOLiD instrument. We relied on SAGE based bead libraries because the frequency of unique tags is directly proportional to transcript abundance [9]. We performed four SOLiD runs yielding altogether 82 million usable reads. 93% of this pool mapped to the 454-contig assembly and the remainder mapped either to human, mouse, rat, or dog RefSeq databases ( Figure 1B). The high mapping efficiency supports the TMS approach and highlights the requirement of a long-read library for reliable mapping and annotation. Quantitative comparison of transcripts involved in lipid metabolism in liver of H.sapiens, M.fascicularis, S.scrofa, M.auratus and R.norwegicus The primary motivation for our study was to compare liver gene gene expression in hamster with humans and model organisms used in metabolic disease research. For this study we constructed SAGE-based liver RNA libraries from H. sapiens, the minipig S. scrufa and the rat R. norwegicus. Human and rat sequence-tags were annotated and quantified by mapping to public domain genomes or

to a minipig draft genome assembly [R. Schmucki, unpublished data]. M. fascicularis liver expression data came from a previous study [5]. The general concordance of the normalized datasets for each species is illustrated by the comparable expression levels of housekeeping genes across all species included in this deep sequencing based analysis (Figure 2, right panel). To compare expression of genes involved in lipid digestion, mobilization and transport we selected a set of 48 genes from the REACTOME database (http://www. reactome.org/), because their functional interactions are well documented. Hierarchical clustering generated six distinct clusters of genes with similar gene expression patterns across all species (Figure 2; left panel; CL1-CL6). Cluster 1 contains apolipoprotein isoforms that are abundantly expressed in all species included here. ApoA2 was assigned to cluster five which contains transcripts involved in HDL-biosynthesis with very low expression in the rat. The cholesteryl-ester-transfer-protein (CETP) is the target of HDL-modifying drugs explaining lack of pharmacological responses in the rat model [10]. Cluster 2 contains highly expressed transcripts in all species adjacent to the most abundant liver specific transcript albumin. Clusters 3 and 4 are compiled of genes in the intermediate expression range, with notable interspecies variability. Finally, the mRNA expression levels in cluster 6 are in general low because these genes mostly belong to a network operating in the intestinal lumen. The protein kinase PRKACG as a testis-specific enzyme was properly assigned to cluster 6 [11]. Altogether we failed to identify transcripts for five genes (Figure 2, left panel, grey boxes). The ApoA1 gene of M. fascicularis lacks a CATG motif required for SAGE-based sequencing using SOLiD technology. Three S. scrofa genes (PRKACG, CETP, LPA) are missing in the draft genomes of the pig [12] or the minipig (R. Schmucki, unpublished data). The M. auratus AMN receptor gene (cluster 5) was not present in our long read library. Discussion We show that triple match sequencing (TMS) allows robust deep sequencing based transcript imaging in absence of an annotated genome. Using a SAGE based comparison of liver gene expression across five species we show that TMS derived data are comparable to mRNA quantification data using conventional genome based mapping. Similar to other vertebrates about To ensure representation of low-abundance transcripts it is essential to normalize the 454 library by competitive hybridization. The long 454 reads are assembled into contigs and mapped to human, rat, mouse and dog RefSeq databases yielding about 10'800 annotated genes (hamster liver transcriptome). 93% of 82 million short SAGE tags mapped to the hamster transcriptome and the remaining 7% to human RefSeq entries allowing quantification over 4 orders of magnitude. All 454 and SOLiD sequencing data generated here are available on request. 10'800 genes are expressed in hamster liver at comparable levels (Additional file 1: Table S1). The value of transcript identification in absence of an annotated genome has been recognized by others. Fletcher et al. have generated a partial transcriptome of the Woodchuck (Marmota monax) an animal model for experimental hepatitis B infection and hepatocellular carcinoma [7]. Randomly primed cDNA libraries were sequenced using 454-long read technology followed by contig assembly and annotation. This yielded a pool of 13'448 unique transcripts which was used to build a custom microarray to analyze the transcriptional responses to virus infection and cure. Compared to TMS this process is more time consuming due to the need of microarray production and validation. A method termed TRINITY relies on graphics based assembly of high coverage short read or paired end cDNA libraries without reference genome [13]. In contrast to the Woodchuck approach, the three software modules constituting TRINITY can detect novel splice variants as demonstrated for fission yeast, mouse, and whitefly whose genomes were not available at the time of release. This 48 genes from the public domain network REACTOME "Lipid digestion, mobilization, and transport" were selected to highlight species specific differences in gene expression relevant for HDL biosynthesis (left panel). Gene clusters generated by hierarchical clustering are labelled as CL1 to CL6. The highly abundant albumin transcript is marked. The right panel shows mRNA levels of a standard reference set of liver housekeeping genes compiled from public domain data (L. Badi, unpublished). The log2 values of normalized read counts are presented in a standard heat map as indicated. Grey fields indicate genes lacking valid expression data. SAGE tags of cluster 6 transcripts were quantified using the Chinese hamster genome draft as resource because they had no matches in the 454 library due to low abundance. The RNA source for each species is given in the material section of the paper. All SAGE libraries were built using commercial kits and the source of tissue is given under Methods. Abbreviations of organisms included are indicated at the bottom. approach is particularly suited for the analysis of tumor samples where gene re-arrangements and structural deviations can lead to novel variants of any gene [14]. Although TRINITY has powerful transcript assembly capabilities for short and long read data the current version of the program cannot perform transcript quantification. In our approach contig assembly of long-reads is solely performed for annotation followed by quantification using frequencies of short-read SAGE tags ( Figure 1A). Therefore, TMS is inherently unable to quantify or discover splice variants of the same gene. This requires construction of tissue specific cDNA libraries by random priming which was used to build the human "bodymap" gene expression database (http://www.ensembl.info/blog/2011/05/24/ human-bodymap-2-0-data-from-illumina). TMS does not attempt to annotate species-specific transcripts because the approach relies on mapping to known mRNAs in other species. We note that any unmapped long reads are candidates for species specific transcripts of hamster. Normally, unique 454-reads or contigs matched the same RefSeq transcript from more than one species ( Figure 1B) resulting in a significant improvement of annotation confidence. As it turned out, the genome sequence or the partial transcriptome analysis of the C. griseus derived cell line CHO(K1) had only limited value for our project [6,15], because the annotation efforts in C. griseus focused mainly on pathways relevant for protein production and modification. Furthermore, the evolutionary distance between C. griseus and M.auratus is comparable to that between mouse and rat [8]. Finally, the partial transcriptome of CHO is derived from an ovary cell line which expresses probably different genes than liver tissue from another species. Conclusions Triple match sequencing is a robust and rapid method for annotation and quantification of transcripts in the absence of a reference genome. The method is especially useful in projects where reliable mRNA quantification rather than primary sequence information is important. TMS does not require implementation of specialized hard-or software which makes it fast, affordable and user friendly. It might even contribute to the identification and selection of novel animal models prior to a genome sequencing effort. Animals Male Syrian hamsters, weighing 125 ± 5 g were obtained from Charles River Germany (Sulzfeld). The animals were kept at standard housing conditions (22±2°C, 50-60% humidity with a range of 40-80% and 12/12 light/dark cycle). They were supplied with standard laboratory chow and water ad libitum, and left to acclimatize for 1 week before the experiments. The harvested liver was cut in small pieces (approximately 3 × 3 mm) and snap frozen immediately for further processing. All experimental procedures were carried out in accordance with international guidelines for care and use of laboratory animals. RNA isolation RNA was isolated from Syrian hamster liver samples using RNeasy Mini Kit from Qiagen (No.74104). RNA integrity was analyzed using an Agilent Bioanalyzer (RNA 6000 Nano Kit) and all samples reached a quality score (RIN) above 9.7. Total Wistar rat liver RNA was a gift from Dr T. Heckel and snap-frozen minipig liver tissue was a gift from Dr Niels-Christian Ganderup (www.ellegaard. com). Normal human liver RNA was provided by the Human Tissue and Cell Research foundation (project-nr: 2012-13 approved 06/25/2012). Construction of SAGE based libraries and NGS procedures The Applied Biosystems SOLID 3 System SAGE kit (No.4443756AB) was used for library construction. Presence of the expected 100-base pair SAGE templates was confirmed using an Agilent Bioanalyzer (High Sensitivity DNA Assay). The emulsion PCR (ePCR) library was constructed using DNA at a concentration 0.75 pM. The ePCR was carried out on a SOLID EZ Bead amplifier and the enriched beads were quantitated using a NanoDrop 2000 spectrophotometer. Sequencing was performed on a SOLiD4 instrument with DOLF-TOP sequencing chemistry Frag-Lib F3 Tag MM35 (No. 445352). cDNA normalisation for GS FLX (Roche 454) libraries To normalize the levels of cDNA prior to library construction we used a competitive hybridization protocol provided with commercial kits (Evrogen MINT cDNA synthesis kit No. SK001; TRIMMER cDNA Normalization kit (No. NK001). For 454 sequencing the use of special cDNA primers is required. The 3 0 -kit primer is replaced by polTdeg (5 0 -AAG CAG TGG TAT CAA CGC AGA GTA CTT TTG TTT TTT TTT CTT TTT TTT TTV N -3 0 ) The kit PCR primer M1 is substituted by polTM1 (5 0 -AAG CAG TGG TAT CAA CGC AGA GTA CGG-3 0 ). The product of the full-size preparation of ds cDNA is used for normalization with the TRIMMER kit using 700-1300 ng of purified cDNA. The normalization is performed according to the supplied protocol without any changes. emPCR was carried out with 5cpb input on SOLID EZ Bead Amplifier. Typically a GS FLX+ Sequencing Run generated about 800'000 reads with an average length of 410 bases. Raw sequencing data processing, read mapping, contig assembly and gene annotation Following depletion of ribosomal sequences, 454-sequencing reads were assembled de novo using the software The Impact of Launching Surgery at the District Level in Niger Background In 2005, the Ministry of Health in association with the Faculty of Medicine of Niamey decided to launch surgery at the district hospital (DH) level as part of the health strategy for the country. Surgical procedures were provided by general practitioners who received 12 months of training in basic surgery. Methods Whereas the initiative was launched nationwide, we chose randomly to study the region of Dosso during a 1-year time period of January 2007 to December 2007 in the three district hospitals as well as the regional hospital of Dosso. Results During the course of 1 year, 544 patients received operations in the three DHs, of which 37.9% (n = 206) were emergent and 62.1% (n = 338) were elective. The most common emergent interventions were cesarean sections (70%) and uterine ruptures (7.8%). For elective surgeries, hernia repairs comprised 80.8% of the cases. The mortality rate of emergent surgeries was 7.3 and 0% in the cases of elective surgeries. Of note, there was a large reduction in transfers to the regional hospital: 52% compared to 2006 and 82% compared to 2005. In 66.1% of the transfers, the cases consisted of fractures, and in 10.4% of abdominal trauma and critical thoracic emergencies. Further study of this initiative has highlighted other challenges, including that of human resources, equipment maintenance, provision of consumables, and the need for continued training. Conclusions Results from this governmental initiative to provide surgery in rural district hospitals by general practitioners are promising and encouraging. In the rural district of Dosso, there have been no deaths from elective surgery, and the number of surgical transfers to the regional hospital has drastically diminished. Introduction The inability to provide adequately for the surgical needs in many low-income countries has received increasing attention in the global health community [1,2]. One of the main barriers has been the shortage of surgical workforce [3], and there have been numerous efforts to provide different types of surgical training to various healthcare providers, from generalist physicians, clinical officers, and nurses [4][5][6]. Little literature has documented the specifics of these training initiatives in sub-Saharan Africa, and few studies have provided more outcomes-oriented evaluations of these programs. In Niger, the ability to offer emergency obstetrical and surgical care has been very limited in rural areas, where 72% of the population lives. In 2005, the Ministry of Health in association with the Faculty of Medicine of Niamey decided to launch surgery at the district hospital (DH) level as part of the health strategy for the country. Surgical procedures were provided by general practitioners (GPs) who received 12 months of training in basic surgery, who received a certificate designated ''Capacity of District Surgery'' (CDS). This study describes the implementation and

effects of this training program for generalist physicians to provide surgical care in district hospitals rural areas after these physicians were deployed at the end of their training in 2006. Description of surgical training program Students chosen to be in the program were government physicians (generalists) who were already working in the District Hospital and had previously practiced in rural areas. Interested parties were required to take a written examination that covered both medical and surgical aspects of medicine, and the top students were chosen to participate in the program. The training occurred in two stages as resident of surgery. A 3-month theoretical and practical training took place in the university hospitals. This was a defined curriculum designed by the Ministry of Health, covering aspects of general surgery (e.g., appendectomies, laparotomies, hernia repairs, splenectomy), trauma and orthopedics (e.g., reductions of open and closed fractures, amputations), obstetrics and gynecology (e.g., cesarean section, uterine rupture repair, dilatation, and curettage), and urology (e.g., testicular torsion, hydrocele repair). This training was then followed by 9 months of essential and practical training in the regional hospitals, requiring a certain number of each procedure under each category to be performed independently. The follow-up of the training was performed by the Faculty of Medicine, who named a coordinator to survey each site multiple times. The first two cohorts of trained practitioners consisted of 41 physicians from rural areas who graduated in 2006 and 2007. To train ancillary staff, there was a concomitant program to train nurse anesthetists and surgical aides during a 3-year period, which began in 1995 by the University of Niamey. This was to provide enough technical support for the provision of these services. These two programs were sponsored by the special program of the president and partners (Belgian Technical Cooperation, Italian Cooperation). The cost for the first year of the program was approximately $100,000 USD, or $4,762 per student. Methods For the purposes of studying the volume and outcomes of this initiative, the region of Dosso was randomly selected. Dosso is 139 km south of the capital and located in the southwest region of the country, which receives relatively more rain than the east of the country, and thus, more agricultural, with comparatively fewer desert areas. It is therefore relatively more populated than the east (with a greater population density compared with the arid desert in the northeast and central portions of the country); distances between district hospitals and referral hospitals in Dosso are shorter than those in the northeast and central portion of the country. This is important because our estimates of referral for more mortality and referrals will be more conservative due to even poorer accessibility in other regions. This region has one regional hospital (RH) and four district hospitals (DH), which serve a catchment area of 1,890,000 inhabitants over a surface area of 31.000 km 2 . The districts areas were: Gaya (318,000 inhabitants), Doutchi (620,000 inhabitants), Loga (168,000 inhabitants), and Boboye (339,300 inhabitants). We wanted to evaluate the impact of this strategy beginning with the deployment of these GPs with a CDS certificate at the end of 2006 over three DHs (Gaya, Loga, and Doutchi) where these surgical services were provided. We studied a 1-year period from January 2007 to December 2007 in the three DHs and the regional hospital of Dosso. We reviewed patient charts that had been listed in the operating room registers and hospitalization records in these four hospitals; we also analyzed the regional level reports of reported activities in addition to supervisor reports from the Ministry of Health and Faculty of Medicine. Furthermore, we obtained qualitative information regarding these initiatives through discussions with the operating teams, service chiefs, and responsible administrative personnel at the district hospitals. We studied the following measures: the trend in the number of emergency transfers from DH to RH, the prevalence and type of surgical operations performed in DHs, morbidity and mortality rates of surgical procedures, and overall working conditions. Results In 2007, 544 patients received operations in the 3 DHs, of whom 37.9% (n = 206) were emergent and 62.1% (n = 338) were elective. Table 1 illustrates this data by district hospital. Emergent surgeries For emergent surgeries, cesarean sections were the most common surgical procedure and accounted for 70% (n = 144). As shown in Fig. 1, this procedure was followed by laparotomy for acute peritonitis accounting for 9.2% (n = 19). The etiology of peritonitis was appendicitis in five cases and perforation of ileum secondary to typhoid fever in 14 cases. These two procedures were followed by hysterectomy for uterine rupture for 7.8% (n = 16), repair of strangulated hernia for 4.4% (n = 9) without bowel necrosis, and laparotomy for intestinal obstruction for 1.9% (n = 4) from adhesions. The ''other'' interventions included amputation, wound debridement, salpingectomy for ectopic pregnancy, and splenectomy. The mortality of all procedures combined amounted to 7.3% of cases (n = 15), split between 10 cases for the obstetrical procedures of cesarean section and uterine rupture (6.25% mortality rate), and 5 cases for other interventions (10.9% mortality rate). For comparison, the mortality rates of the regional hospital in 2007 (where surgical and obstetrical activities were performed by two general surgeons and one gynecologist) were only slightly lower: 5.7% (18/318) for obstetric and gynecologic procedures, and 9.25% for other interventions (21/227). As shown in Fig. 2, overall the rates of emergent referral to the regional hospital decreased steadily during the 3 years. Elective surgeries For elective cases (Table 2), hernia repair was the most performed operation (80.8%), followed by hydrocelectomy (13.3%). We documented 15 cases (4.4%) of postoperative complications, the majority of which were wound infection (66.7%, n = 10). There were no cases that resulted in the death of the patient for elective procedures. Effects on referrals to regional hospital With respect to referrals to the regional hospital, in 2005 before the advent of surgery at the district level, there were 635 patient transferred to the regional hospital of Dosso. In 2006, some surgical activities were introduced to the district in the form of 2-week surgical camps provided by surgical teams from the National Hospital of Niamey, who came twice during the 6-month period. During this period, there were more than 240 patients transferred to a higher-level facility. In 2007, the number of transfers drastically reduced to 115 patients, which was 52% of the numbers transferred or thoracic trauma, and the rest (13.1%, or n = 15) were due to iatrogenic causes, mainly equipment malfunction (e.g., rupture of an oxygen tank or shortage of anesthetic drugs precluding further surgery). Supervision Supervision of these training programs in the different hospitals was led by one surgeon, one gynecologist, and one anesthesiologist-all faculty of the University of Niamey (Table 3). Each hospital was to receive six supervisory visits lasting 2 days each, although one hospital received only five due to transportation problems. Supervision included active monitoring of procedures as well as further in-service training. After a total of 17 visits after the implementation of this district surgery program, certain problems were noted. The most common were management of human resources, in particular the poor use of specialized nurses in the operating room, and lack of technical personnel to repair and maintain surgical equipment. Providers and administrators also complained of difficulty in procuring surgical consumables (including those for orthopedic and trauma cases) and anesthetic medication. Imaging, such as ultrasound, was not available; the only imaging modality was conventional roentography or X-ray. In addition, only one operating room was functional in two of the three district hospitals. Finally, providers discussed their need for continued training in surgical techniques and management of postoperative complications. Discussion Reliable access to surgery for the population in rural areas has been a political preoccupation in Niger. Since 1974, mobile surgical teams had been dispatched to provide for surgical needs in these areas, but this arrangement was limited due to its sporadic and temporary nature. Niger is more than four times the size of the UK and consists mainly of desert, where the distances between health structures are significant, and poor road conditions make transfers formidable and difficult. Effective coverage of surgical needs, including that of obstetrical emergencies, requires that surgical capacity be available within the area. In Niger, the lack of surgeons is largely attributable to the prolonged training (5 years); the current system has only graduated three to five surgeons annually since 2004. Based on this reality, a decision was made to train general practitioners who were going to be working in rural areas in surgical skills during 12 months; In Ethiopia, the length of training consists of 6 months [7]. In Canada and Australia, in certain rural locations, surgical activities are performed by general practitioners, whereas in the United States, many procedures are provided by general surgeons [8][9][10][11]. The results of this program in Dosso since the implementation of district surgeons (or general practitioners trained in surgery) are encouraging and promising. The division of 62.1% elective and 37.9% emergent cases can be compared with the results of Humber et al. in British Columbia, Canada, of a respective split of 81 and 19% [11]. The higher percentage of emergent cases in Niger suggests that the needs for these types of programs are even greater in sub-Saharan Africa. For elective surgeries, we found that the largest needs were in the areas of hernias and hydrocele. In fact, the mobile surgical camps that began in 1974 were specifically developed to address these pathologies in an active population (e.g., cultivators and breeders). In Niger, hernia repair is the most common gastrointestinal surgery, in nearly one of two patients [12]. The findings from a comparative study done in 2003 regarding the management of inguinal hernia at the University Hospital of Niamey and the district hospital of Gaweye showed that the pathology was treated similarly in both institutions with equivalent surgical outcomes and, in fact, the length of stay and cost of treatment was significant less in the district hospital [13]. In emergent surgeries, the obstetric interventions of cesarean and ruptured uterus represented the majority of activities (77.8%) with a mortality rate of 6.25% and minimal morbidity, compared with 5.7% in the regional hospital where fully trained surgeons provided care. This is similar to a Canadian study where results from cesarean surgeries done by generalists was compared with that of specialists; the most surprising observation was the low morbidity rate from these major surgeries was similar for both groups. With 4-month training, on average, the generalists were practicing cesarean sections with a level of acceptable risk and safety [14]. The referrals to the regional hospital were mainly due to orthopedic trauma (66.1%) and abdominal-thoracic trauma (10.4%). Although these generalists are trained in the basic treatment and management of orthopedic fractures, almost half of these cases were referred to the regional hospital simply because of a lack of materials necessary for RH regional hospital; MOH Ministry of Health treatment of these orthopedic trauma cases (e.g., external fixators) and even orthopedic beds for traction. Another contributing factor to the higher referral rate for these injuries was the need for further training in these injuries. For the management of abdominal trauma requiring splenectomy (0.9%) at the district hospital, the transfers for serious cases required the care of a specialist and possibly further imaging, such as ultrasound. To improve the treatment of abdominal trauma, a budget would need to include the purchase of materials as well as continued education of staff in primary trauma care, as well as an additional training at the University of Niamey. Two district hospitals (Gaya, Doutchi) of the three studied in the region of Dosso are on the route of well-frequented roadways and are the main institutions that receive the patients who are victims from these road traffic injuries. According to the World Health Organization, the burden of mortality and morbidity from these types of trauma are particularly elevated in low-and middle-income countries. In fact, 90% of the burden of road traffic injuries is borne by these countries [15]. Regarding the implications of this type of training program as a method for increasing surgical workforce, the World Health Organization (WHO) has calculated that Africa bears approximately 25% of the burden of the world's diseases but only 1.3% of the world's health work force [16]. The

lack of surgeons is partly related to the ''brain drain'' where physicians from developing countries go to more developed regions, such as Europe or the United States. The other factor contributing to the lack of surgeons is the insufficient training mechanisms, which is common across many countries in sub-Saharan Africa and has thus spawned the adoption of different political responses. In Malawi and Mozambique, for example, the health system has trained a cadre of nonphysicians called ''assistant medical officers'' who can perform common surgical procedures, including laparotomies and cesarean sections [16]. This particular program in Niger of equipping generalists through a dedicated surgical training is based on training physicians who have already been working in district hospitals for 2 years or more and are familiar with life in rural and more isolated areas. The other reason to choose these physicians for this type of training is the reality that all of the physicians in the 34 districts across the country are generalists. This particular training program was of interest to these physicians because this certificate of completion also provides a bonus to their annual salary, in addition to the indemnity they receive as an incentive for working in rural areas. In terms of retention, all physicians trained from this program have remained in their posts, which local officials contribute to the initial selection of trainees for this program (those already working in rural areas who originated from these areas and have settled their families there) as well as the fact that the CDS is not recognized outside of Niger. The problem of human resources is the common denominator across all developing countries. In Niger, certain operating rooms were not functional because of lack of anesthetists, which raises the question of whether it may be necessary to train generalists in surgery and anesthesia, as in Canada [17]. According to Chen et al. [18], during the last 20 years, economic reforms have diminished public expenditure, freezing any new recruitment and weakening the salaries in the public sector. The authors also measured ''health worker density'' and demonstrated its correlation with survival rates, calculating that sub-Saharan Africa has only ''a tenth of the nurses and doctors for its population'' compared with Europe and, even more starkly, ''Ethiopia has a fiftieth of the professionals for its population than Italy'' [18]. The enormous problems must be resolved with political engagement over the short-term and long-term based on recognition of the reality that these countries face and the importance of the provision of basic life-saving surgical procedures for the population. In Niger, the engagement of the political leaders to relieve the burden of surgical disease has been led by Ministry of Health, the University of Niamey, and development partners (e.g., the Belgian Technical Cooperation, WHO, World Bank, Italian Cooperation) to establish a short-term emphasis on the creation of the General Practitioner for District Surgery and, in the long-term, for the training of surgical specialists. The goals of these welltargeted efforts address the lack of human resources and materials. Awareness of the human resource issues in district hospitals has encouraged us to plan a strategy that, in additional to financial incentives, could enhance the satisfaction of surgeons in the practice of their profession. Their responsibility would be to follow-up outcomes from those generalists who they have trained and also provide continuous education and supervision during short stays in the district hospitals. Conclusions The results from this study of rural surgery performed by generalists trained in surgical procedures are promising and encouraging. Mortality and morbidity is low for both emergent and elective procedures, and referrals to the regional hospital have been reduced drastically, except for cases of fractures and abdominal-thoracic trauma. Evaluation of this program has highlighted the need for increased human resources, provision of materials, maintenance of equipment, and continuous education, especially in the field of trauma. Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited. A Comparison of Weight-Related Behaviors of Hispanic Mothers and Children by Acculturation Level Hispanic mothers and children in the United States experience a high prevalence of obesity, which may be affected by maternal acculturation level. Little is known about the association of acculturation on weight-related behaviors. This study describes differences in weight-related behaviors by acculturation level of Hispanic mothers residing in the U.S. and compares them to behaviors of White mothers. Acculturation level was determined using personal acculturation and acculturation environment variables. Cluster analysis of acculturation variables identified three groups of Hispanic mothers: low personal and environmental acculturation (n = 46), high personal and low environmental acculturation (n = 65), and high personal and environmental acculturation (n = 38). Results indicate that, compared to White mothers (n = 340), the least acculturated cluster of Hispanic mothers tended to model physical activity less often and the most acculturated exerted more pressure on children to eat. Mothers in the least acculturated cluster tended to rate children’s health status lower, indicate that children had greater fruit juice and less milk intakes, have more meals in locations associated with less healthy meals, and have the least space and supports for physical activity. Findings highlight relationships between maternal acculturation level and weight-related behaviors and suggest strategies for helping acculturating Hispanic mothers create healthier lifestyles and home environments. Introduction Individuals of Hispanic origin make up the largest ethnic group in the U.S. [1]accounting for 17% of the population. These individuals experience disproportionate health disparities compared to most other racial and ethnic groups [2]. Of the top leading causes of deaths for Hispanic persons in the U.S., most are diet related (e.g., cancer, heart disease, diabetes, liver and kidney disease) [3]. Many of these health conditions have strong links with obesity [4][5][6][7][8]. Obesity prevalence in the U.S. is rising, affecting 42% of the population; racial and ethnic disparities are seen in obesity, with prevalence being highest in Hispanic and non-Hispanic Black adults (i.e., 45% and 50%, respectively) [9]. Hispanic children in the U.S. have the highest rates of overweight among all racial and ethnic groups, with 26% being obese versus 22%, 14%, and 11% of non-Hispanic Black, White, and Asian children, respectively [10]. This weight disparity emerges at a young age-22% of preschool-aged Hispanic children in the U.S. are obese compared to 8% of all other U.S. preschool-aged children and 5% of preschoolers living in Mexico classified as moderately to severely overweight [11,12]. A factor complicating the study of Hispanic individuals is the range of acculturation levels or extent of cultural and psychological transition in this population. Approximately one-third of all Hispanic individuals living in the U.S. are foreign born, originating predominately from Mexico, Puerto Rico, and Cuba [13][14][15]. Foreign-born individuals undergo the Materials and Methods The study protocol was approved by the Rutgers University Institutional Review Board. This study uses the baseline data from the HomeStyles randomized control trial, which used the Social Cognitive Theory and a social-ecological framework to investigate environmental, social, and personal characteristics of home environments associated with health and body weight [38,41]. Outcomes have been reported elsewhere for a subset of study participants [38,39,41,54]. Participant recruitment occurred using a variety of methods, in both English and Spanish, including word of mouth, outreach at community events, and electronic postings that invited parents to join a program to help them "build even happier, healthier, safer families". Participants To be eligible to complete the baseline survey, individuals had to have at least one child aged 2 to <9 years, be between the ages of 20 and 45 years, make all or most household food purchasing and preparation decisions, live in the study target areas (New Jersey or Arizona), give informed consent, and complete the baseline survey. Participants received a $15 stipend. Participants who replied to recruitment announcements (n = 5495) were not included in the data set analyzed in this study if they did not complete the study screener (n = 217), did not consent (n = 405), did not meet eligibility criteria (n = 3343), failed to complete the baseline survey (n = 862), or gave implausible responses (e.g., answered many questions similarly; n = 34). In addition, those not pertinent to the purpose of this secondary analysis (i.e., males [n = 49] and those who were of a race/ethnicity that was neither White nor Hispanic [n = 95]) were culled from the data set. The final analytic sample was 489. Survey Instrument The "Home Obesogenicity Measure of EnvironmentS" (HOMES) baseline survey was used to collect data. To summarize the previously reported details of this survey [38,40], it was based on a comprehensive literature review that identified behavior and environmental variables associated with weight status in parents and young children along with validated, reliable, relevant scales for assessing these variables [38,40]. When multiple scales assessing a variable were located, a panel of experts in nutrition and survey methods reviewed them to identify the scale that had the greatest reliability, validity, and relevance to the study and was easy for participants to complete and researchers to administer and score. In cases where a suitable scale for assessing a variable was not available, a new scale was developed using standard practices for scale development; that is, items were developed and reviewed by experts for content validity, tested cognitively with parents of young children, pilot tested, pretested, and field tested. At each stage of development, the findings were subjected to expert review and refined [40,[55][56][57]. All participants in the survey development stages had the same characteristics as those who completed the HOMES baseline study but did not participate in the baseline study. The HOMES baseline survey gathered data on demographic characteristics (e.g., age, education, race/ethnicity, total parents in the household, maternal employment, family affluence), food insecurity risk, maternal health status, maternal weight-related behaviors, and parenting practices. Mothers also reported the health status and weight-related behaviors for one of their children. Mothers with 2 or more children between the ages of 2 and <9 years were directed to provide data on one child born that was closest to a randomly chosen time and date (i.e., noon on 1 June). Assessments of the home environment were family mealtime importance, frequency, and location; household food availability; physical activity space and supports; and neighborhood safety. Race/ethnicity was reported as White, Hispanic, Black, Asian, Native American, or American Indian, and/or Alaskan Native or Pacific Islander; this item was used to identify the participants for this secondary analysis. The Family Affluence Scale, a reliable marker of socioeconomic status [58,59], has 4 items that generate a family affluence score of 0 to 9, with higher scores indicating more affluence [58,59]. Food insecurity risk was evaluated with the 2-item scale developed by Hager et al. [60]. Maternal and child health status were appraised using the 5-point (poor to excellent) health rating item from the Health-Related Quality of Life Scale developed by the Centers for Disease Control and Prevention [61]. Mother and child physical activity levels were evaluated using the 3-item HOMES Physical Activity Questionnaire level [38,62,63], which measured days per week spent walking, doing moderate activity, and doing vigorous activity. An indicator of sedentary behavior of mothers and children was determined using total minutes spent watching television, movies, and videos, and using computers each day. The Block Fruit/Vegetable Screener determined mothers' consumption of fruits and vegetables daily and children's intake of fruit/vegetable juice each day [64][65][66]. The HOMES Drinks Intake Screener estimated daily intake of sugar-sweetened beverages, such as soft drinks, fruit drinks, tea, and coffee by mothers and children [38,67,68]. The drinks screener also determined children's daily milk intake [38,67,68]. Weight-related parenting practices included maternal modeling of behaviors and their child feeding practices. The Modeling of Physical Activity scale assessed how frequently mothers actively played with their children [38,40,62]. The Modeling of Healthy Eating scale evaluated how important mothers felt it was to engage in this behavior, with answers being a 5-point Likert agreement scale ranging from strongly disagree to strongly agree [38,40,62]. Child feeding practices evaluated included using food to reward children for eating healthy foods, pressuring children to eat, and controlling food intake amounts of children [62,[69][70][71][72][73][74][75]. These child feeding practices scales used a 5-point Likert agreement scale ranging from strongly disagree to strongly agree. Higher

scores on all weight-related parenting practices scales indicate greater use of the practice. Characteristics of the home environment assessed included the value mothers placed on family meals using a 5-point Likert agreement scale ranging from strongly disagree to strongly agree. Mothers also reported the total number of family meals eaten each week as well as how many days per week family meals were consumed in locations associated with healthier meals (i.e., dining table) and less healthy meals (i.e., car, fast food restaurant, in from of a television) [76][77][78]. The amount of fruits/vegetables and sugar-sweetened beverages available in the home was determined with a household food supply frequency questionnaire [38]. The Hop-Up questionnaire appraised the space and supports available for physical activity inside the home, in the outdoor/yard area right outside the home, and in the neighborhood as well as overall neighborhood safety [79,80]. As acculturation is an abstract construct that is fluid in nature, it can be difficult to define and measure [47]. Proxy variables, such as language use or nativity, can be used to estimate acculturation [81]. The strongest single indicator of acculturation tends to be language use or preference [82]. In this study, personal and environmental acculturation were assessed for those reporting Hispanic as their race/ethnicity. Personal acculturation was assessed using these 3 variables: language chosen to complete the survey, language most commonly spoken at home, and participant's country of birth [34][35][36][37][83][84][85]. Each personal variable was dichotomously coded with those who completed the survey in English, spoke English at home, and/or were born in the U.S. being considered more acculturated while the opposite was considered less acculturated. Acculturation environment variable information were derived from statistics geocoded as census tract data using the U.S. postal zip code of each participant's home residence [37]. Acculturation environment variables in each participant's home census tract were: percentage of foreign-born individuals, percentage of foreign-born individuals arriving within 5 years prior to the census, and percentage of Spanish-speaking households reporting speaking English less than very well [37]. Acculturation environment variables were standardized by state (New Jersey or Arizona) by dichotomously coding each variable as more acculturated if the percentage was at or above the median threshold for the participant's state of residence and less acculturated if the percentage value was below the median threshold. For example, a person who lived in a neighborhood with a higher percentage of foreign-born individuals than the state's median was coded as less acculturated for this variable. Data Analysis Each of the dichotomously coded personal acculturation and acculturation environment measures were included in Ward's Hierarchical Cluster Analysis to partition Hispanic participants into meaningful acculturation groups. Cluster analysis allows the relationships among the personal and environment measures to be explored while accounting for the complex, latent interactions among these measures. The goal of cluster analysis is to merge individuals into similar groups that maximize within-group homogeneity and between-group heterogeneity [86]. Ward's hierarchical method was selected because it determines the total number of clusters in circumstances where the number of clusters is unknown, as was the case in this study [86]. Descriptive statistics (e.g., means, standard deviations, confidence intervals) were computed to describe the participants and scale scores. Differences in weight-related behaviors were compared among and between study groups (i.e., Hispanic acculturation cluster groups and non-Hispanic White participants) using analysis of covariance (ANCOVA) and Bonferroni post-hoc tests, with family affluence score serving as the covariate because family socioeconomic status is thought to be associated with differences related to acculturation [25,[87][88][89][90][91][92][93][94][95]. Family affluence was compared among and between study groups using Analysis of Variance (ANOVA) and Tukey post-hoc tests. Effect size was calculated as partial eta-squared values, with 0.01, 0.06, and 0.14 serving as the thresholds for small, medium, and large effect sizes, respectively [96]. Due to the multiple comparisons, the probability level for the main effects was set at p ≤ 0.01 to reduce the risk of type I errors. Post-hoc probability was set to p < 0.05. Analyses were conducted with SPSS software version 27.0 (IBM Corporation, Chicago, IL, USA). Results The final analytic sample of 489 mothers was comprised of 340 who were non-Hispanic White and 149 who were Hispanic. The scree plot and agglomeration schedule generated from Ward's hierarchical cluster methodology indicated that Hispanic mothers grouped into three acculturation groups. An examination of the personal acculturation and acculturation environment scores indicated that Hispanic mothers assigned to Cluster 1 (n = 46) were the least acculturated, achieving a total acculturation score of 1.7 on a 6-point scale ( Table 1). As shown in Table 2, most completed the survey in Spanish, spoke Spanish at home, and were born outside the United States. Cluster 1 mothers also lived in areas where the percent of foreign-born individuals, percent of foreign-born individuals arriving between 2010 and 2015, and percent of households speaking English less than very well were at or above the median. Cluster 2 (n = 65) was somewhat acculturated, having a total acculturation score of 3.5 out of 6. Cluster 2 Hispanic mothers were similar to Cluster 1 regarding environmental acculturation measures but differed on personal acculturation measures in that all or nearly all completed the survey in English, spoke English at homes, and were born in the United States. That is, Cluster 2 lived in low acculturation environments but had higher personal acculturation. Cluster 3 (n = 48) Hispanic mothers were the most acculturated, achieving a mean acculturation score of 5.37 on the 6-point scale. Most Cluster 3 mothers lived in areas where percent of foreign-born individuals, percent of foreign-born individuals arriving between 2010 and 2015, and percent of households speaking English less than very well were below the median. Like the somewhat acculturated mothers in Cluster 2, all mothers in Cluster 3 completed the survey in English, spoke English in their homes and were born in the U.S. Thus, Cluster 3 lived in high acculturation environments and were personally acculturated. ANOVA and follow-up procedures comparing total acculturation scores by cluster revealed significant differences for all pairwise comparisons of clusters, indicating that each cluster uniquely represented a different level of acculturation. 2 Acculturation level scale contains six items; scores range from 0 to 6; higher scores indicate higher acculturation level. 3 Personal Acculturation scale contains three items; scores range from 0 to 3; higher scores indicate higher personal acculturation. 4 Acculturation environment contains three items; scores range from 0 to 3; higher scores indicate higher acculturation environments. 5 Family Affluence Scale contains four items; scores range from 0 to 9; higher scores indicate greater family affluence. Mothers were of moderate family affluence. ANOVA and post-hoc test results indicated that White mothers and Cluster 3 mothers did not differ; however, White mothers had significantly greater family affluence than both Clusters 1 and 2, while Cluster 3 mothers had significantly greater family affluence than Cluster 1 mothers. These differences in family affluence confirmed the need for family affluence to serve as a covariate in subsequent analyses. As shown in Table 3, results from ANCOVA and post-hoc pairwise comparisons of mothers indicated that they tended to be similar regarding age, education level, number of parents in the household, and food insecurity risk. Employment status differed significantly among study groups, with mothers in the least acculturated cluster (Cluster 1) being less likely to have paid employment than all other groups of mothers. However, effect sizes for all demographic characteristic differences were small. Mothers reported good to very good health status, with no groups differing significantly (Table 4). Mothers' weight-related behaviors were similar across groups. Few significant among-and between-group differences were noted in mothers' weight-related parenting practices and child feeding practices. White mothers tended to model physical activity more often than Hispanic mothers in Cluster 1, and the most acculturated Hispanic mothers (Cluster 3) tended to pressure children to eat more than White mothers. Children's health status was very good to excellent, with Hispanic mothers in the two least acculturated groups (Clusters 1 and 2) rating child health significantly lower than White mothers. ANCOVA of children's weight-related behaviors indicated that physical activity and screentime behaviors did not significantly differ among groups. However, beverage intake of children was significantly different among study groups. That is, children of Hispanic mothers in Cluster 2 drank significantly more sugar-sweetened beverages than the least acculturated Hispanic mothers (Cluster 1) and White mothers. Children of Hispanic mothers in Cluster 1 tended to drink significantly less milk and more fruit juice than children in other groups. However, the effect sizes for all child assessments were small. The home environment characteristics described in Table 5 indicate no significant differences among study groups in the importance placed on family meals or frequency of family meals. However, family meal locations differed significantly among study groups, with the least acculturated Hispanic mothers in Cluster 1 tending to more meals eaten in locations associated with less healthy meals, such as cars and in front of the TV than other groups and fewer meals at dining tables. Availability of fruits/vegetables and sugarsweetened beverages in the household did not differ significantly among and between groups. Hispanic mothers in Cluster 1 tended to have less home and outdoor/yard space and supports for physical activity, with these differences being significantly lower in comparison to White mothers. The two less acculturated clusters of Hispanic mothers (Clusters 1 and 2) reported significantly lower neighborhood safety than the most acculturated Hispanic mothers as well as White mothers. Effect sizes for home environment characteristics were small, except for frequency of eating family meals in cars, which had a large effect size. [1]. 4 Parents in household: possible score range = 1 to 2. 5 Employment: possible score range = 1 to 3; 1 = does not work, 2 = works part time, and 3 = works full time. 2 The 5-point agreement rating: poor, fair, good, very good, excellent; scored 1 to 5, respectively; higher score indicates better health [2,3]. 3 Days/week engaged in walking, moderate activity, and vigorous activity weighted by exercise intensity (weights of 1, 2, 3, respectively) and summed to create a scale score; higher scale score indicates greater activity level. Possible score range = 0 to 42 [4][5][6]. 4 Higher score indicates greater servings consumed daily [6][7][8][9][10][11]. 5 Days/week mother engages in physical activity with child. Possible score range = 0 to 7. 6 The 5-point agreement rating: strongly disagree, disagree, neither agree nor disagree, agree, strongly agree; scored 1 to 5, respectively; scale score equals average of item scores; higher scale scores indicate greater expression of the trait. Possible score range = 1 to 5. Cronbach alphas for the 4-item Models Healthy Eating, 3-item Uses Food to Reward Child's Healthy Eating, 4-item Pressures Child to Eat, and 4-item Controls Child Food Intake Amounts scales are 0.71, 0.75, 0.66, and 0.66, respectively. Discussion The findings of this study indicate that using personal acculturation and acculturation environment variables in a cluster analysis generated three distinctly different acculturation clusters of Hispanic mothers. Cluster 1 had the lowest personal acculturation. Both Clusters 1 and 2 lived in less acculturated neighborhoods in contrast to Cluster 3 mothers who lived in acculturated neighborhoods. Clusters 2 and 3 had high personal acculturation. It is interesting to note there was not a cluster with low personal acculturation and acculturation environment was not found-this is likely a reflection of the impact a high acculturation environment has on the progression of personal acculturation characteristics [97]. The observation that White and the most acculturated cluster of Hispanic mothers had similar family affluence, coupled with the significant differences between these two groups and the two less acculturated clusters for family affluence, supports previous research reporting that socioeconomic status increases with acculturation [25,[87][88][89][90][91][92][93][94][95]. The lower rate of employment of the least acculturated Hispanic mothers may be due to lack of facility with the English language, lack of cultural capital (e.g., knowledge of societal customs and valued attitudes and behaviors), lack of social capital (i.e., inclusion in social networks that provide instrumental relationships and offer access to resources, such as employment opportunities), and neighborhood environments with limited employment opportunities [87,[98][99][100][101][102][103][104][105][106][107]. However, given that the least acculturated Hispanic mothers (Cluster 1) differed significantly from all other groups, including Cluster 2 which had a similar acculturation environment, the findings appear to indicate that personal acculturation factors, such as ability to use the prevailing

language, play a central role in employment status. Study findings suggest that mothers are more alike than different regardless of their acculturation level and health status. However, Hispanic mothers in the two less acculturated clusters reported a significantly lower health status for their children than White mothers. These data suggest that acculturation level is positively linked to child health status. This supports previous research indicating Hispanic mothers with lower acculturation tend to have children with poorer health [108,109], which could be related to lack of health insurance or access to health care facilities in less acculturated neighborhoods [110,111]. Mothers nearly met the recommended intake of fruits and vegetables (five or more servings per day) [112] and reported low intake of sugar-sweetened beverages. Although mothers had similar dietary intake, they reported significant differences in their children's intake. Children of the least acculturated Hispanic mothers, like those of White mothers, had the lowest intakes of sugar-sweetened beverages, yet these children of Cluster 1 Hispanic mothers consumed more 100% fruit juice and less milk than White and more acculturated clusters. Although it is not clear why these differences occurred, it is possible that the least acculturated mothers were not able unable to discern the difference between 100% fruit juice and fruit-flavored sugar-sweetened beverages because sugar-sweetened beverages often feature pictures of fruit on the label, yet contain no real fruit juice. These images can lead to confusion regarding the fruit content of the product, especially among those with less English language facility [113,114]. The least acculturated Hispanic mothers had a lower family affluence score, and thus, were more likely to participate in the Special Supplemental Feeding Program for Women, Infants, and Children (WIC) which would increase their access to juice [115]. Indeed, WIC participants tend to report higher intakes of juice and sugar-sweetened beverages than non-WIC participants, but an intake of milk similar to non-WIC families was not observed in this study [116]. The lower milk intake by children of the least acculturated Hispanic mothers could be due to the relatively lower intake of fluid milk in many Hispanic groups, such as those from Mexico and Central America, due to the prevalence of lactose intolerance and the higher prices of milk seen in less acculturated neighborhoods in the U.S. [117,118]. Study findings suggest that acculturation was not associated with mothers' own diets, but mothers' acculturation played a role in children's intake. These findings contrast with previous research indicating that acculturation is linked with adult diets [24,30,31], but the acculturation level of children's caregivers contributes little to the dietary intake of young children in their care [53]. Although the reason for these differences is not known, it is important to consider that previous studies tend to consider only personal acculturation characteristics, unlike this study which included both personal and environment acculturation characteristics. Mothers and children in all groups had low physical activity scores and exceeded the recommendations for screentime (<1 h daily of high-quality programming for children older than 2 years) [119]. Although differences among groups were not significant, the least acculturated mothers and children tended to have the lowest physical activity and modeled physical through co-play with children the least. This likely is because they also had less access to space and supports for physical activity within their homes, yards, and neighborhoods and lower neighborhood safety, suggesting that acculturation environment limited access to amenities that promote physical activity [120][121][122][123][124][125][126]. Similarly, Cluster 2 who also lived in a low acculturation environment reported lower neighborhood safety. The frequency of family meals eaten at less healthy locations tended to decrease and meals at healthy locations rise as acculturation level increased. The frequent consumption of family meals in cars by the least acculturated cluster may be a time management response to the shift work schedules and unskilled labor job types common to this population [31,44,106,107]. In addition, the frequent meals in cars may reflect limited space and equipment within homes for preparing, storing, and/or eating foods thereby making it convenient and necessary to frequently eat meals outside of the home [127]. Living space also may be constrained to the point that it is not possible to accommodate a dining table [127,128], so meals are eaten in front of the television. Despite the frequent, on-the-go eating of less acculturated families, mothers reported adequate (at least five servings/person/day) availability of fruits/vegetables in their homes [112]. This study supports previous findings that acculturation is associated with behaviors [24,25,27,89,90,[92][93][94] and home environment characteristics [120][121][122][123][124][125]127] that impact weight status and demonstrates that both personal and environmental factors contribute to links among acculturation [37], independent of family affluence, and lifestyles and home environments of mothers with young children. The results highlight key similarities and differences in health status and weight-related behaviors of Hispanic mothers and their young children based on maternal acculturation level. Findings suggest that interventions for reducing obesity and overweight risk among both Hispanic and White adults and children in the U.S., regardless of Hispanic mothers' acculturation level, should focus on increasing physical activity, reducing screentime, and increasing the frequency of maternal modeling of healthy behaviors. For audiences who have lower personal acculturation, strategies for differentiating between sugar-sweetened fruit drinks and 100% fruit juice may be needed. Strategies for successfully addressing the barriers in low acculturation environments to healthy behaviors, such as eating at a dining table rather than locations associated with less healthy meals, easing access to health care facilities, increasing availability of physical activity space and supports, improving safety perceptions of neighborhoods, and possibly changing labeling of sugar-sweetened fruit drinks to make it easier to determine these are not fruit juice equivalents, also are needed. This study is among the first to examine the weight-related behaviors of Hispanic mothers of young children by acculturation level. It is important to note that the findings of this study differ from the narrative that acculturation unequivocally leads to negative health behaviors and health consequences among Hispanic populations, even finding some benefit to acculturation in this population. In addition, it is the first known to include acculturation environment variables in the determination of mothers' acculturation level. The strengths of this study include the use of valid, reliable scales in the survey and overall large sample size, although the number of participants in each cluster was more modest. This study is limited by the potential for social desirability bias and self-reporting of data. The cross-sectional nature of this study also negates the possibility of inferring causal relationships. Additionally, the race/ethnicity of spouses/partners was not available and should be considered in future studies. Conclusions Overall, findings suggest weight-related behaviors that educational interventions for this population should target and highlight the important roles of considering maternal acculturation when developing interventions aiming to help acculturating Hispanic mothers of young children create healthier lifestyles and home environments. For instance, interventions could focus on encouraging families to share meals at home together or develop strategies for accessing safe locations to engage in physical activity with their children. Clarifying how acculturation is linked to health behaviors can lead to health promotion materials that are tailored to and perhaps resonate better with acculturating Hispanic audiences, thereby contributing to the reduction in health disparities that are so prevalent in this population. Future research should investigate whether the findings of this study hold true for other immigrant populations undergoing acculturation in the United States and other nations. UNDERSTANDING STRESS AND AGGRESSIVE BEHAVIOR OF UNDERGRADUATE STUDENTS AT (UUM) Students are people who are pursuing higher education in a college. Students are known as intelligent and critical thinkers in acting, but students often experience stress stemming from academic activities and sometimes students behave aggressively. This study aimed to understand the stress and aggressive behavior of undergraduate students at Universiti Utara Malaysia. This study was qualitative by using data collection techniques with observations and interviews. The sample that the researcher selected was ten undergraduate students at UUM based on the knowledge, the researcher's consideration appropriate to the purpose of the study. The results show that the stress experienced by undergraduate students at UUM is the stress of Daily Hassless & Personal Stressor. While the aggressive behavior of undergraduate students at UUM is direct passive verbal behavior and indirect passive verbal behavior. Emotional maturity in self-control is an effort undertaken by undergraduate students at UUM in dealing with stress and aggressive behavior. ISSN: 2320-5407 Int. J. Adv. Res. 8(01), 323-333 324 (WHO) is ranked 4th in the world. The UK Health and Safety Executive's review of stress prevalence covered 487,000 people in the UK who were still active from 2013-2014. Statistics have been collected that the prevalence of stress in women (54.62%) is greater than in men (45.38%).Many universities around the world have conducted research on stress levels in students according to their preferred faculty. In a stressed environment, the prevalence of students is 38-71 %, while in Asia it is 39.6-61.3% (Habeeb 2010, Koochaki 2009). Each student can experience stress in school, the stress that college students face when there are too many demands and assignments a student has to deal with, can also be stressed because of the pressure to show their academic achievement and excellence, which can result in students feeling overwhelmed (Olejnik & Holschuh, 2007; Fernández-González, González-Hernández & Trianes-Torres, 2015). This anxiety can contribute to stress, which can interfere with the student's mind's success in carrying out all of its activities. Another sentiment that goes with stress is anger. In one's life aggressive attitudes and actions can emerge from frustration because it is a comprehensive response to the body, stress often influences the actions of the person experiencing it, which can decide if the behaviour is expressed or not. Behaviors, however, can also be unreliable when stress accompanied by frustration, anxiety, low self-esteem, aggressive behavior and destructive behavior can increase (Buss & Perry (1992)). This is consistent with Burton, Hafetz, & Henninger (2007)) that stress has to do with actions of aggression because when a person is stressed they are anxious, irritable, violent and concentrated on work. Students suffering from stress rarely act violently, which is due to their lack of emotional maturity. Emotional maturity, according to Cole, Cole., & Dean (1980), is the individual's capacity to be compassionate, relaxed, selfcontrolled, the feeling of embracing himself and others, in addition to expressing his feelings constructively and creatively. Emotional maturity is important to the development of positive ability in relationships with other people. Individuals who have achieved emotional maturity can be described as individuals who can objectively assess the situation before they act, who can no longer respond without thought in advance like children or people who are not emotionally mature.Having good self-control, being able to express one's feelings correctly or in compliance with the conditions he faces so that he can adapt better because he can embrace a variety of people and situations and respond to the demands he faces (Hurlock, 2004). Aggression behavior is a type of negative behavior that occurs from stimulus, particularly environmental stimulation that often results in a greater impact. Aggression activity can be physical or verbal, and may occur in other people or objects that are the focus of actions of aggression. According to Koeswara (1988), many figures which explain the notion of aggression behavior is the action of individuals aimed at injuring or harming other people. Meanwhile, according to Dill and Dill (1998), aggressive behavior is viewed as behavior based on experience and the presence of certain stimuli such as pressure which causes somebody to take aggressive actions. Typically, this behavior is done in a planned, instantaneous, or by stimulation of certain conditions. The goals of the research are to: (1), stresscausing factors and aggressive behavior in undergraduate students at Universiti Utara Malaysia. (2), types of stress and aggressive behavior of University Utara Malaysia undergraduate students. (3), the efforts of University Utara Malaysia undergraduate students to conquer tension and aggressive behaviour. Definition of Students: According to Yusuf (2012), students are youths between the ages of 18 and 25 and pursuing higher education at a university or college. Students are considered to have a high level of intellect, skill in reasoning, and action planning. Critical thinking and behaving quickly and correctly is a characteristic that appears to be inherent in each pupil, which is a complementary concept (Dwi Siswoyo, 2007). Based on the description above it can

be concluded that a student is a term for someone who is currently studying or undergoing higher education in a college and has a high intellectual level. Definition of Stress: Stress is an internal disorder, according to Lazarus & Folkman (1986), which may arise from physical demands of the body or environmental and social factors that are regarded as potentially harmful, uncontrolled or beyond the capacity of the person to resolve. Stress is also a physical and psychological stress-state (Chapplin, 1999). Rice & Dolgin (2002) argue that stress is an event or stimulus to the environment which causes people to feel stressed. Atkinson (2000) suggests that stress refers to things that are perceived to be harmful to one's physical and psychological well-being, which in response to stress is referred to as the cause of stress and the reaction of the 325 person to that stress situation. It can be inferred on the basis of the above theories that stress is a state of selfsuppression. Stress is a complex mechanism that creates physiologically, mentally, and behavioral responses to the individual that experiences it, by which the processes are individual in nature and vary from one person to another. Factors causing stress: Stressors may come from different sources, from both physical, psychological, and social conditions, and may also occur in work situations, at home, in social life, and other external environments. Stressors are, according to Lazarus & Cohen (1977): 1. Daily hassles are small incidents that occur on a daily basis such as work, school and so on. 2. Personal stressor is a greater threat or disturbance or failure to something that occurs at an individual level, such as loss of a loved one, job loss, financial problems and other personal problems. Stress signs and symptoms: Stress symptoms can be classified into three groups according to Andrew Goliszek (2005), namely physical, mental, and behavioral symptoms as follows: 1. Physical signs: fatigue, muscle aches, back pain, exhaustion, indigestion, nausea or vomiting, abdominal pain, loss of appetite or appetite, heart palpitation, frequent urination, high blood pressure, inability to sleep or heavy sleep, excessive sweating, and a number of other symptoms. 2. Emotional symptoms: irritable, upset by little things, mood swings, hallucinations, anxieties, fear, sometimes weeping, feeling helpless, feelings of loss of control, suicidal thoughts, confused thoughts, decision-making inability, etc. 3. Behavioral symptoms: smoking, taking drugs, wandering around, losing interest in physical appearance, hair pulling or twisting, changing social habits, and more. As for mental stress symptoms, including: uncontrollable anger or rage / aggressiveness, concern about small things, inability to develop, focus and determine what to do,difficult mood or inappropriate behaviour, intense anxiety or phobia, loss of self-esteem, appear to isolate oneself, talk too much or become too communicative, distracted and, in serious or actual cases-really messed up (Walia, 2005). Definition of aggressive: Saad (2003) describes that aggressive is actions intended to hurt, attack people, damage the things around them to protect themselves as a result of a sense of frustration.Sarwono (1988) states that violence is an outlet for feelings of frustration. Whereas, according to Berkowitz (1987), Koeswara (1988) & Dill and Dill (1998) are violent behaviors which have some intentions to harm others physically or psychologically. It can be inferred from the various violent formulas described above that aggressive behavior is a relieving act of aggression to overcome strong resistance or to punish others, intended to harm others physically or psychologically. Factors that affect aggressive behavior: According to Taylor, Peplau & Sears (2009) and the emergence of aggressive behaviour, this is closely related to rage in an individual. Rage can grow due to stress and frustration, stress and frustration occurring in an individual. Since goals are not attainable.One of the concepts of psychology, may continue to arouse feelings of hostility among people who experience stress and frustration. The circumstance might have arisen because human beings are unable to endure the pain that had befallen him.Whereas, according to Aryani (2006), Guswani, and Karyuan (2011), hostility can may be triggered by dealing with situations or adverse conditions within their environment. Aggressive Shapes: In the schedule, the following can be seen in offensive ways according to Morgan Relationship of Stress and Aggressive Behavior: Some reports indicate that stress is related to aggressive behaviour. Anger is emotional or affective like awakening and psychological readiness to be violent, according to Buss (2007), saying rage is an emotion that has characteristics of high parasympathetic nervous system activity and the existence of a very strong feeling of resentment that is usually caused by a mistake that can be clearly wrong or perhaps not, and when angry there is a feeling of wanting to attack, hit, kill or throw something and a cruel thought emerges. Horowitz (2002) notes that, when a person experiences stress, angry behavior is one of the emotional symptoms. According to Strutcher, Perr & Menec (2000), individuals tend to become more irritable and have reduced mental capacity, with high stress levels. Tobin, Graziano, Vanman & Tassinar (2000) state that angry behavior is a negative emotion due to an unfavorable situation which causes a propensity towards aggressive behavior such as struggle with verbal movements or expressions. Method:- In this study, the researcher chose the approach most appropriate for undertaking the project and according to the research questions and objectives to be examined.This work uses a qualitative approach, as it can give a real image of a phenomenon (Bogdan & Biklen, 2003;Creswell, 2013). The qualitative approach consists of investigating, explaining or defining the real phenomenon from the research participants perspective.However, qualitative research helps the researcher to get better, more comprehensive information about what is being studied (Merriam, 1998;Creswell, 2013). The analysis has a real, concise context that explains the state of affairs at the time of the study, and the researchers analyze data based on the events of the observations and interviews used as a basis for drawing a conclusion. This work makes use of techniques for collecting data through observation and interviews.The aim of the observation and interview was to assess the stress and aggressive behavior of UUM undergraduate students.It is in line with the view expressed by Grieshaber (2006) that observation is the most reliable way to obtain knowledge on social symptoms. The observations made by the researcher are direct, ongoing, and full-time involvement in the field, whether physical or behavioral, during the course of the investigation. In this way the researcher discovers all the attitudes, actions, utterances that occur naturally and inevitably without the resea rcher's development or design. Looking at the sample size of Patton (1990), it does not give the right number of cases for use in qualitative research.He recommends however that the sample be chosen on purpose, where the subject of the analysis is the basis of data acquisition (Fraenkel & Wallen, 2007). The sample the researcher selected on the basis of the information, the decision of the researcher according to the purpose of the study. Individuals within this community are called components. The sample that researchers set is only part of the chosen population for analysis, because the data that researchers have collected should meet the standard needed for the theme construction. Sampling in qualitative research aims at accurately reflecting both the population and the method of continuous data collection (Creswell, 2013). Generally speaking, this study does not examine a single person, but rather chooses a sample that will provide data based on the knowledge it has, which may justify the researcher's conclusions, which is the view of the degree to which the sample represents the population under investigation. Sampling is a process in which research objects are chosen from a group that represents a large selected group of people. The sample used in this analysis was the selection of undergraduate students at UUM and where the researchers wanted to investigate the phenomenon. The sampling technique used in this analysis was the purposeful sampling which was previously described by researchers. The researchers used their own judgment in this sampling 327 to engage the study participants which best suited the intent of this study (Grieshaber, 2006). The sample used in this analysis was the selection of undergraduate students at UUM and where the researchers wanted to investigate the phenomenon. Researchers then brought ten graduate students to UUM. The justification for picking 10 UUM undergraduates is; 1. The pupils are both male and female 2. Undergraduate students ' readiness to engage in the study. 3. Students from Indonesia, Malaysia, Thailand, Nigeria and China are foreign undergraduates at the UUM. Scientists protect the identity of participants by marking their names according to their code number, based on adherence to research ethics. The number of code used in the abbreviation is 1 through 10. Assignment from lecturer Ten participants expressed stress daily hassles on the job demands of the lecturer, following the findings of the researcher and participant interviews: Results and Discussion "I or another student must have felt the same stress as me ... the reason I became stressed while I was a UUM student such as the demands of the assignments from the lecturer, the demands were large, the assignments made sometimes had to be according to what was ordered and if it wasn't appropriate it had to be fixed while we still had a lot of assignments another". (Participants 1) There are lessons which can be hard to understand The findings of the investigative and participant meetings were the following ten participants who emphasized everyday hassles in the field of the life of lessons that were hard to understand: "The thing that makes learning stressful for me is that sometimes it's hard to understand and understand the lessons given by the lecturer ... sometimes I have to go ask a smart friend to help me understand the lesson".(Participants 3) Number of assignments per person & group Ten participants who, following the outcomes of the investigative and participant meetings, expressed daily stress on aspects of many individual & group tasks: "What definitely stressed us was several assignments from lecturers, both individual and group assignments ... So many of us have had to do such tasks poorly, as long as we can collect them on time".( Participants 9) The results of Personal Stressor interpersonal aspects among participants are as follow Financial Problem Seven participants who, following the findings of the investigator and the participant meetings, reported personal stressor on aspects of financial problems: 328 "Of course a financial problem is one that stresses me because we are far away and sometimes remittances fro m family and scholarships are slow to enter into our accounts". (Participants 6) Family problem: Ten participants who stated personal stressors on aspects of family problems, the following were the results of the investigative and participant meetings: "Family problems are also one that makes us stressed, sometimes we miss, the slow delivery of money from parents, not to mention if our parents are sick, of course we are worried". (Participants 5) In general, through this research, the causes of undergraduate student stress are triggered by Daily Hassless factors such as the demands of lecturers ' assignments, the nature of hard-to-understand lessons and the many individual & group assignments.While what triggers Personal Stressor exists in UUM undergraduate students, financial and family problems do exist.This was also discussed in the discovery (Lazarus & Cohen, 1977), an examination of a situation that can cause stress induced by two factors, namely personal factors and situational factors. His finding is consistent with Strutcher, Perr & Menec (2000), who notes that the origin of student stress lies in the learning process, similar to that noted by Saklofske, Austin, Mastoras, Beaton & Osborne (2012) that learning stress occurs when students have too many demands and tasks to perform.We also struggle with personal issues such as family and financial issues, in addition to the stressors faced by undergraduate students. This is also consistent with the statement by Hasida & Moshe (2012) that the stressor of a student arises from within, for example, the physical and emotional state, as well as the effects or demands of similar contexts such as family and social culture.It is noted that the cause of undergraduate student stress is new learning demands and obligations, such as stresses to boost academic achievement, self-reliance and financial stability, according to a study conducted by Zeidner & Schwarzer (1996). College

time creates a lot of problems for students because of the requirements that need to be met as a student and have to obey a program that has been set up in the university or the instructor who offers university education to students. Here are the researchers ' observations and interviews about the physical stress felt by the participants Back pain: Ten participants who claimed the type of stress encountered in the physical aspect being back pain, the results of the study meetings and the participants were as follows: Stress Forms and Aggressive Forms of Undergraduate Students at UUM: Based on the researchers' interviews and findings about the type of stress in "Physically I always get back pain while studying, it makes me stressed because when I want to go back to study I sometimes can't sit anymore so I have to lie down". (Participants 7) Headache: Ten participants who stated the form of stress experienced in the physical aspect were headache, following the results of the investigative and participant meetings: "Stress is a headache, often we experience if we cannot understand the learning provided by the lecturer". (Participants 2) 329 Sleeplessness: Six participants stated the form of stress experienced in the physical aspect, namely insomnia, following the results of the investigative and participant meetings: "Sleeplessness also makes you stressed out because we are worried if the task is not finished ... we sometimes sleep in class and do not pay attention to the teaching lecturer as a result our stress increases". (Participants 9) Always Sleepy and Want to Sleep: The results of the investigative and participant meetings are the following four participants who reported the type of stress encountered in the physical aspect that is always sleepy and who want to sleep: "If we are always sleepy and want to sleep if we want to learn, this is because of the many tasks, seeing it creates stress". (Participants 5) Always hungry and want to eat: Ten participants who stated the form of stress experienced in the physical aspect of being always hungry and want to eat, the following were the results of the investigative and participant meetings: "When I want to learn and do my assignments, I am always hungry and want to continue eating, stressful learning makes me want to eat all the time". (Participants 1) The following are the findings and outcomes of the investigator's gibberish about participants sources of emotional stress. Easily angry & easily offended Seven participants who stated that the stressed form was emotionally charged were irritable and offended, along with the results of the investigative and participant meetings: "When I was stressed I learned to be sensitive to being angry and easily offended, like when I was stressed then my friend made a fuss immediately when I was angry and angry ... I thought that I was not valued and no one wanted to help me" (Participants 6) Restless: Ten participants who stated the form of stress experienced in the emotional aspect were, restless following the results of the investigative and participant meetings: "I feel restless when learning stress, the thing that makes me nervous when learning stress is when a deadline is chased to collect assignments while I don't understand the assignment". (Participants 3) Distracted mind: Ten participants who stated the form of stress experienced in the emotional aspects were chaotic thoughts, the following were the results of the investigative and participant meetings: "Stress is sure to cloud my mind, because many items are thought that assignments have not been done, fees for the semester have to be made, and others". (Participants 8) 330 It's hard to make a decision Ten participants who stated the form of stress that was absorbed in the emotional aspect that is difficult to make decisions, the following are the results of the investigative and participant meetings: "Stress makes it difficult for me to make decisions, small things like eating are difficult to determine when I experience the stress of learning, even when I'm hungry I am lazy to move to buy food so I have heartburn". The findings of observations and gibberish from researchers concerning the types of stress activity encountered by participants are as follows: Lazy Learning & Not going to class Ten participants who stated the form of stress experienced in the aspects of stress behavior were lazy learning and not entering the classroom, following the results of the investigative and participant meetings: "When I'm stressed I become lazy to study and don't even attend class for one to three days". (Participants 2) Smoke Five participants who stated the form of stress experienced in the aspects of stress behavior are smoking, the following are the results of the investigative and participant meetings: "I am a smoker, but if I am not stressed I will smoke as much as one pack a day or even two days, but when I am stressed I will smoke as much as 2 packs a day". (Participants 10) Based on observational and theme data related to the stress of undergraduate students, the UUM shows physical stress such as back pain, headaches, insomnia, always sleepy, want to sleep, always hungry and want to eat. When UUM undergraduate students undergo types of emotional stress, they are irritable and irritable, anxious, noisy thinking, and hard to make decisions.Further, UUM undergraduate students ' type of stress behavior is lazy studying and not attending class and smoking.Stress is the state of physical or mental stress in a person, besides stress is an issue or a demand for adaptation, which is because the individual demands to disrupt his or her life and cause physical, emotional and behavioral problems (Lazarus & Folkman, 1986;Morgan et al, 1986 Verbal, passive, direct: Refuse to talk to other people: Ten participants expressed verbal, passive, direct forms of aggression by refusing to talk to others, following the findings of the researcher and participant interviews: "When I refuse to answer my questions I also refuse to talk to others because I don't want to add to my stress problems anymore". (Participants 6) Verbal, passive, indirect: Be quiet and then leave the person while talking: Ten participants who expressed aggressive form experienced in verbal, passive, direct aspects by refusing to answer other people's questions, following the results of the investigative and participant meetings: "Usually when I am stressed learning sometimes I just stay quiet and leave people while talking, this is because my mind is chaotic". (Participants 10) Verbal, passive, indirect: Avoiding interactions with others: Ten participants who expressed aggressive forms experienced in verbal, passive, indirect aspects by avoiding interaction with others, following the results of the investigative and participant meetings: "So that I don't get stressed, I usually avoid interactions with friends ... I do this so I don't get distracted by things that damage my mind". (Participants 3) According to Burton, Hafetz, & Henninger (2007), aggressive behavior types may occur to any individual, that aggressive behavior towards a person is an individual act or action that is offensive or destructive with the intent to injure or hurt others, both verbal and non-verbal.Aggressive behavior, however, is not always violent or hurtful in killing people, and aggressive acts are divided into several categories, ranging from actions that can hurt others to actions that simply refuse to speak to others (Morgan et al, 1986).The aggressive form of undergraduate UUM aggression, which is only passive verbal in nature, refuses to answer questions from other people and refuses to talk to others, while indirect forms of passive verbal aggression are silent and vanish while talking and avoiding contact with others. Attempts to Address Undergradute Students' Strees and Aggressive Behavior at UUM: Attempts by UUM undergraduate students to overcome stress and aggressive behavior are as follows: Attempts by undergraduate students at UUM to overcome stress and aggressive behavior Emotional maturity Emotional Control Positive thinking Participants 1'2'3'4'5'6'7'8'9'10 Participants 1'2'3'4'5'6'7'8'9'10 Source: face-to-face interviews and observations The following are the results of observations and exploratory meetings related to aggressive forms experienced by participants: Emotional Control Ten participants who stated the efforts of undergraduate UUM students overcoming stress and aggressive behavior by controlling emotions, the following were the results of the investigative and participant meetings: "My efforts to deal with stress and my aggressive form by controlling emotions to stay calm, not angry and not cause fights between me and my friends". (Participants 1) 332 Positive thinking: Ten participants stated the efforts of UUM undergraduate students to deal with stress and aggressive behavior with positive thinking, following the results of the investigative and participant meetings: "In addition to controlling my emotions, of course I must keep thinking positively that many assignments also have good effects for me to be smarter and more critical in my own field of science". (Participants 10) Emotional maturity is an attempt that undergraduates make to deal with stress and aggressive behaviour.Emotional maturity is an important thing individuals need to learn because emotional maturity is a psychological concept that describes how people manage their emotions while experiencing stress.People who are emotionally mature are people who are able to manage feelings and positive thinking, so that individuals can communicate them in an adaptive way. According to Cole, Cole., & Dean (1980) individuals ' ability to respond to stimuli that affect their environment can be demonstrated by a healthy person, guided and specifically in line with the stimulation and responsibility for all environmental decisions and actions.If this is done then it is said that the person matures his emotions. According to Hurlock (2004) as individuals objectively assess the situation before responding emotionally, emotional maturity is no longer behaving without feeling like children or inexperienced people beforehand. Conclusion:- Based on the results of the data analysis and the results of the studies addressed, it can be inferred that the Daily Hassless and Personal Stressor are causing stress and aggressive behavior of undergraduate students.In general, undergraduate students ' types of stress and aggressive behavior include physical, mental, interpersonal, direct passive verbal and indirect passive verbal, this can be seen as the undergraduate students ' stress and aggressive behavior have no great potential to harm themselves and others.In general, the types of stress and aggressive behavior of undergraduate students include physical, emotional, interpersonal, direct passive verbal and indirect passive verbal, which can be seen as the stress and aggressive behavior of undergraduate students has no great potential for harming themselves and others. Neural correlates of theory of mind in children and adults using CAToon: Introducing an open-source child-friendly neuroimaging task Theory of Mind (ToM) or mentalizing is a basic social skill which is characterized by our ability of perspective-taking and the understanding of cognitive and emotional states of others. ToM development is essential to successfully navigate in various social contexts. The neural basis of mentalizing is well-studied in adults, however, less evidence exists in children. Potential reasons are methodological challenges, including a lack of age-appropriate fMRI paradigms. We introduce a novel child-friendly and open-source ToM fMRI task, for which accuracy and performance were evaluated behaviorally in 60 children ages three to nine (32♂). Furthermore, 27 healthy young adults (14♂; mean = 25.41 years) and 33 children ages seven to thirteen (17♂; mean = 9.06 years) completed the Cognitive and Affective Theory of Mind Cartoon task (CAToon;www.jacobscenter.uzh.ch/en/research/developmental_neuroscience/downloads/catoon.html) during a fMRI session. Behavioral results indicate that children of all ages can solve the CAToon task above chance level, though reliable performance is reached around five years. Neurally, activation increases were observed for adults and children in brain regions previously associated with mentalizing, including bilateral temporoparietal junction, temporal gyri, precuneus and medial prefrontal/orbitofrontal cortices. We conclude that CAToon is suitable for developmental neuroimaging studies within an fMRI environment starting around preschool and up. The foundations of mentalizing are laid during the first few years of life, though they become more refined throughout childhood and adolescence. Early conceptualizations of ToM tasks have particularly focused on explicit measures (e.g., Sally and Anne Task ( Baron-Cohen et al., 1985)), which are mastered around the age of 4 ( Baron-Cohen et al., 1985;Wimmer and Perner, 1983). However, studies employing implicit ToM assessments during infancy, as for example through the investigation of an infant's anticipatory looks, have suggested that ToM may develop as early as 13-15 months (Onishi and Baillargeon, 2005;Southgate et al., 2007;Surian et al., 2007). Consequently,

the type of task employed to measure ToM, or mentalizing, may have a significant influence on the interpretation of the reported skill levels. Studies assessing the neural correlates of ToM in adults (Bzdok et al., 2012;Molenberghs et al., 2016;Van Overwalle and Baetens, 2009) have consistently linked mentalizing to brain areas within the frontal (e.g., anterior dorsal medial and ventromedial PFC, inferior frontal-and precentral gyri and the anterior cingulate cortex), temporal and parietal cortices (e.g., bilateral temporoparietal junction, middle temporal gyri, posterior superior temporal sulci and the precuneus (Molenberghs et al., 2016)). In accordance with the conceptual separation of affective and cognitive ToM, distinct networks can be identified (Schlaffke et al., 2015;Sebastian et al., 2012a;Shamay-Tsoory and Aharon-Peretz, 2007). While affective and cognitive mentalizing are controlled by a shared network comprising bilateral temporal poles, superior temporal sulci and the dorsomedial prefrontal cortex, the specific role of orbitofrontal and ventromedial prefrontal cortices in affective mentalizing has been highlighted based on research in clinical and healthy populations (Hynes et al., 2006;Sebastian et al., 2012a;Shamay-Tsoory and Aharon-Peretz, 2007). Affective ToM has also particularly been associated with basal ganglia functioning (Schlaffke et al., 2015;Bodden et al., 2013). On the other hand, cognitive mentalizing processes are more specifically linked to activation in the dorsomedial prefrontal cortex, precuneus, cuneus, bilateral temporoparietal junction, and the middle of the superior temporal gyri (Molenberghs et al., 2016;Van Overwalle and Baetens, 2009;Schlaffke et al., 2015). While various reports describe the neural correlates of ToM in adults, less is known for younger children, with or without neurodevelopmental disorders. Potential reasons may include practical and technical challenges as well as a lack of age-adequate scanner tasks (Raschle et al., 2012(Raschle et al., , 2009Thieba et al., 2018). However, in recent years new studies have emerged investigating mentalizing in young populations through task-based functional magnetic resonance imaging (fMRI) or functional near infrared spectroscopy (Gweon et al., 2012;Hyde et al., 2018;Moraczewski et al., 2018;Richardson et al., 2018;Richardson and Saxe, 2020a). Such studies implement auditory paradigms, false belief tasks or incorporate more naturalistic settings such as passive movie viewing tasks. Alternatively, task-free functional (e.g., resting state fMRI (Xiao et al., 2019)) or structural measures, (e.g., white matter measures (Wiesmann et al., 2017)) can be further substantiated through the use of additional behavioral ToM measures. Here we present three experimental studies conducted to assess the development and implementation of the Cognitive and Affective Theory of Mind Cartoon task (in short CAToon; available at: www.jacobscenter. uzh.ch/en/research/developmental_neuroscience/downloads/catoon. html), a novel, open-source, engaging and child-friendly fMRI mentalizing task. Study 1 introduces development and behavioral assessment of CAToon in children. Specifically, we aimed to assess the age at which children were able to complete CAToon behaviorally. We hypothesized that CAToon may be completed starting around preschool/kindergarten (around 4 years of age). Study 3 aimed to assess the neural correlates during CAToon performance in a first group of typically developing children. We hypothesized, that CAToon may elicit similar, though still developing neuronal activation patterns in children (Richardson et al., 2018;Richardson and Saxe, 2020b). CAToon task Task creation of the Cognitive and Affective Theory of Mind Cartoon task (CAToon) included the following steps: (1) standardized literature review on ToM fMRI studies (as described in (Fehlbaum et al., 2021), (2) evaluation for child-appropriateness according to the following requirements: the task should be feasible for young children, including non-readers (no text), has to be engaging and fun, and should ideally be visually entertaining since this has previously been reported to reduce head motion (Huijbers et al., 2017). As a result, we decided on the use of cartoon stories, which are commonly used in the literature, successfully evoke distinct neuronal activation associated with ToM (e.g., (Schlaffke et al., 2015;Sebastian et al., 2012a;Völlm et al., 2006;Walter et al., 2009)) and which adhere to the aforementioned requirements. In Study 1 (behavioral) and 3 (fMRI) these requirements were re-evaluated. CAToon consists of a total of 30 hand-drawn stories, including two experimental conditions targeting affective ToM (AT) and cognitive ToM (CT) and a control condition (physical causality (PC); Fig. 1). Each condition comprises 10 stories of similar visual complexity. Ten backgrounds were prepared for the task and each background occurs only once in each condition. The two experimental conditions were designed to differentially motivate affective versus cognitive aspects of ToM reasoning. That means participants should have to infer how a character would react to a fellow character's expressed or expected emotions during AT trials, whereas during CT trials participants should assume how characters would act based on another character's intentions or beliefs. PC trials serve as a control condition, requiring a basic understanding of cause and effect and basic physical laws. All trials start with three consecutively presented images, followed by a single image displaying three possible endings. CT trial endings consist of one possible, one improbable and one highly improbable/ impossible solution. AT trial endings consist of two possible solutions (negative expectancy/positive expectancy) and one impossible solution. In positive expectancy endings a character's emotional needs are met with care or reassurance, whereas in negative expectancy outcomes the character is scolded, ridiculed or ignored. This manipulation allows the investigation of differences in positive or negative outcome expectancy. PC trial endings consist of one possible and two impossible solutions. As physical causality and cognitive ToM conditions have only one correct answer the chance of getting a correct answer is 33 % (1/3). For affective ToM (AT) we present two correct (negative and positive expectancy) answers resulting in a higher chance of 66 % (2/3). Therefore, for the overall task the chance level is 44 %. In other words, in the PC and CT conditions participants have to get at least 3.3 tasks correctly to reach the level of chance (that adds up to 6.6) and in AT condition they have to get 6.6 answers correctly to reach level of chance. Across the 30 trials that adds up to 13.2, which is 44 % of 30 tasks. CAToon task evaluation (Study 1: behavioral) CAToon was evaluated behaviorally in a group setting (first and second grade school classes) or in an individualized manner (preschoolers/kindergarteners). Each participant was asked to look at three images presented in a row and then indicate their choice of the most likely story ending out of three options by either pointing to it (preschoolers/kindergarteners) or by crossing off their choice in a booklet (school-aged children/group setting; details in Supplement 1). CAToon task evaluation (Studies 2 & 3: fMRI) A total of 30 cartoon stories were rear-projected onto a screen behind the scanner, viewed by participants via a prism attached to the head coil and displayed using Presentation® software (V16.5, Neurobehavioral Systems, Inc., Berkeley, CA, www.neurobs.com). Adults completed the fMRI task in one single run (trial order in Supplement 2), while children completed the task over the course of two runs (15 trials each) as suggested for fMRI in developmental population (Raschle et al., 2009). Both runs included 5 AT, 5 CT, and 5 PC stories. The run order for children completing the fMRI experiment was alternated (starting with either run1 or run2). The task had a rapid event-related design with fixed inter-trial intervals of 3 s. Before the start of each run, a fixation cross was present for 2 s. Each trial started with the consecutive presentation of three images (3 s each), followed by a decision phase of 7 s for adults. Based on adult feedback from Study 2 (i.e., challenged by the relatively short answer time), the decision time was extended to 10 s for children to assure age appropriateness. Participants had to choose one out of three possible endings through use of a button box (task design in Fig. 1). Task duration was 8 min 36 s for adults, and 11 min 4 s for children (two runs of 5 min 32 s). Before solving the CAToon neuroimaging task inside the MRI scanner, participants completed three practice trials on a laptop and by use of a cardboard model button box. After these practice trials it was verbally assured that participants understood the task and key points were repeated prior to the start of neuroimaging (further info in Supplement 2). Participants Study 1: CAToon task evaluation in children (behavioral) For Study 1 60 children ages three to nine years (mean age: 5.77 years; 32 boys; group characteristics in Supplementary Table S1) completed the CAToon task behaviorally. All children were recruited through daycare, kindergarten or schools. Answer choices, age and gender, but no identifiable personal data, was collected. In line with approval by the local ethics board (Ethikkomission Nordwest-und Zentralschweiz), informed assent to participate was provided by the daycare teachers, school principals or parents. Families were informed about the participation and had the option to withhold contribution. Analyses Study 1 The mean percentage of correct answers for overall task performance was calculated for each participant. A one-way analysis of variance (ANOVA) was employed to inspect overall performance difference across age groups (by year). The group of nine-year old children was excluded from this analysis due to small sample size (n = 2). Prior to conducting the ANOVA, assumption of normality (kurtosis and skewness) and homogeneity of variance (Levene's test) was tested. The results from the ANOVA were followed up with Bonferroni-corrected post-hoc group comparisons. In addition to the categorical investigation of change in performance based on age in years, changes based on age in months were assessed (dimensionally) using partial F-tests to select the best-fitting regression model. Projected changes in performance based on age were calculated using the CurveExpert Professional Software (https://www.curveexpert.net/) by displaying the instantaneous rate of change (the slope of the tangent line at a given point on the curve). All behavioral analyses were performed in SPSS (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.) or R version 4.0.3 (R Core Team, 2020; https://www.R-project.org/). Participants Studies 2 & 3: CAToon task evaluation in adults and children (fMRI) 28 healthy young adults and 37 typically developing children took part in the fMRI experiments assessing the neural correlates of mentalizing using CAToon. Participants included in the fMRI studies had normal or corrected-to-normal vision, sufficient German skills, no previous neuropsychological disorder, and average to above average IQ based on their level of education (for adults) or an IQ ≥ 70 (for children; verbal and non-verbal subtests of the German version of the Wechsler Intelligence Scale for Children (WISC-IV; (Daseking et al., 2007)). One adult was excluded from the study due to strong visual prescription glasses that could not be used or adjusted for within the scanner. Four children were excluded due artefacts caused by braces (n = 2), and claustrophobia (n = 2). The final sample therefore included 27 adults (20-39 years), and 33 children (7-13 years; group characteristics are listed in Table 1). Adult participants further completed standardized questionnaires assessing callous-unemotional traits (callous-unemotional dimension of the Youth Psychopathic traits Inventory (YPI; (Andershed et al., 2002)) and empathy (Interpersonal Reactivity Index (IRI; (Davis, 1980)). This allowed the investigation of the association between behavioral scores of callous-unemotional traits or empathy and neural activation during mentalizing using post-hoc assessments. All participants and in case of the children their legal caretakers provided verbal and written consent for taking part in the study. fMRI data acquisition (Studies 2 & 3) For the fMRI task whole-brain T2-weighted echo-planar images were collected on a Siemens 3 T Prisma MR scanner using a 20-channel head coil (transverse slice orientation, interleaved acquisition) and the following specifics: field of view = 220 mm, TR = 2000 ms, TE = 30 ms, 42 slices, slice thickness = 2 mm, voxel size = 2.0 × 2.0 × 2.0 mm, 333 volumes. One additionally structural image was acquired for coregistration during image preprocessing, using the following specifics: voxel size: 1.0 × 1.0 × 1.0 mm; TR = 1900 ms; TE = 3.42 ms; TA = 4.26; flip angle = 9 degrees; field of view = 256 × 256 mm, 192 slices with a slice thickness of 1.00 mm. For fMRI acquisition, the first twelve seconds prior to the start of the first stimulus included simultaneous multislice acquisition and dummy scans (discarded), which allowed accounting for T1 equilibration effects. The ToM task lasted 8 min and 38 s for adults, and 11 min and 4 s (5 min 32 s per run) for children.

The structural image acquisition lasted 4 min and 26 s. Analyses of in-scanner data (Studies 2 & 3) In-scanner performance (correct versus incorrect answers) and differences between conditions were investigated employing one-way ANOVAs in adults and children. Bonferroni-corrected post-hoc tests were employed to further investigate significant differences between trials. Whole brain fMRI analyses (Study 2 & 3) fMRI data was analyzed using the Statistical Parametric Mapping software (SPM12; http://www.fil.ion.ucl.ac.uk/spm/) in MATLAB 2016a (Mathworks, Natick, MA). Regressors of interest were created using a boxcar-function for experimental and control condition and contrasts of interest were calculated for affective ToM (AT > PC), cognitive ToM (CT > PC) and mentalizing (CT|AT > PT). For adults three additional contrasts were calculated. To detect shared activation of affective and cognitive ToM we conducted a conjunction analysis, testing areas activated in both, AT and CT conditions ((CT > PC) & (AT > PC)). Finally, contrasts of distinct activation representing affective ToM (AT > CT) and cognitive ToM (CT > AT) were calculated. For all contrasts the statistical parametric maps were cluster-level FWE-corrected at p < 0.05, with an initial cluster-building threshold of p < 0.001, uncorrected. Regressors of interest were implemented for the full trial duration of 16s (adults) or 19s (children), including story presentation and decision time. To assess whether regressor length significantly impacted neural activation, post-hoc analyses were also conducted implementing a reduced regressor (i.e., only considering the story phase of the trials in adults, excluding the decision phase). Standard fMRI preprocessing included realignment and unwarping, co-registration to each participant's structural image and segmentation prior to normalization into standard space (ICBM152 template). All images were smoothed using an 8-mm full width at half maximum isotropic kernel. Using the art imaging toolbox (https://www.nitrc.org/ projects/artifact_detect/) seven additional regressors accounting for motion and variations in mean signal intensity as well as a high-pass filter of 0.01 Hz (128 s) were added to the first-level model of each participant. Performance across conditions Summarizing the ratio of correct answers in all children for each condition showed that AT trials were solved correctly in 90.7 %, CT trials in 60 % and PC trials in 73.3 % of all trials. Overall task performance results in a ratio of correct answers ranging from 40 to 97%. All children reached an accuracy above chance; Table 2. Considering only the incorrect solutions within the cognitive ToM conditions, children selected improbable scenarios in 54.9 % and highly improbable/ impossible solution in 45.1 % of the cases. One-way analysis of variance showed a significant effect of condition for correct answers (F(2,177) = 43.214, p<0.001). Bonferroni post-hoc tests for pairwise comparisons indicated that the ratio of correct answers differed significantly between all three conditions (p < 0.001). It has to be noted, that a direct comparison may not be warranted and has to be interpreted with great caution, since AT conditions consisted of two possibly correct endings unlike CT and PC trials (one possible ending). Performance across age A one-way ANOVA investigating the effect of age on accuracy rate was conducted after Levene's test indicated equal variances (F(5, 52) = 0.420, p = 0.832) and normal distribution within the age groups (Supplementary Table S2; (Field, 2013;Ghasemi and Zahediasl, 2012)). The analysis of variances showed a significant effect of age on performance, F(552) = 39.215, p < 0.001. Bonferroni post-hoc test for pairwise comparisons revealed that 3 and 4-year-olds scored Table 1 Group characteristics and behavioral scores of adult and child participants of Studies 2 and 3 (fMRI). Notes. ISCED = International Standard Classification of Education, IRI = Interpersonal Reactivity Index (sum scores), YPI = Youth Psychopathic Traits Inventory (mean scores), WISC-IV = Wechsler Intelligence Scale for Children, Fourth Edition, SD = standard deviation, n.a. = not applicable. Table 2 Ratio of correct answers (in %) in the different conditions across age groups (in years). significantly less correct compared to all other age groups, with no further difference between the two youngest age groups (p = 1.000). Also, while variations in performance remained, no significant differences between 5, 6, 7 and 8-year-olds were observed (Supplementary Table S3). Assessing a dimensional age-performance model revealed a significant improvement of fit comparing the linear and quadratic models (F(1) = 4.918, p = 0.031). The age-performance relationship was best described by a quadratic model (Fig. 2), no further improvement was observed when using a cubic model. The regression model indicates that the instantaneous rate of change (f'(x)) is higher in younger ages and becomes lower in older children. This implicates bigger steps of improvement taking place in younger children. More specifically, in the youngest participants the performance is predicted to improve by 0.97 % with each passing month (f'(37) = 0.97; equaling one additional correctly solved trial every 3 months). In contrast, the rate of change drops to an improvement of 0.57 % at 77 months (f'(77) = 0.57), and decreases even further with progressing age (f'(117) = 0.17). This means, that the oldest participant within our sample is predicted to improve by one additional correct answer when aging 17 months, reflecting a deceleration of improvement in the older participants. In-scanner task performance Overall, adult participants scored above chance with 87.90 % correct answers across the 30 trials (AT: 97.80 %, CT:71.90, PC: 94.10), while children scored 81.92 % correct across all trials (AT: 95.15 %, CT: 72.73 %, and PC: 80.91 %). The range of correctly answered trials was between 73-100 % for adults. Children scored between 63-93 % corrects overall. The analysis of variance showed a significant effect of condition on the ratio of correct answers in both, adults (F(2,78) = 45.373, p < 0.001) and children (F(2,96)=24.35, p < 0.001). In adults the ratio of correct answers within CT was significantly lower as compared to AT and PC (both p < 0.001), however there was no significant difference between AT and PC trials (p = 0.636) according to the Bonferroni post-hoc test for pairwise comparisons. In children, the post-hoc Bonferroni pairwise comparison yielded significant differences between AT and CT (p < 0.001), AT and PC (p < 0.001), and also PC and CT (p = 0.041). It has to be noted though, that a direct comparison between AT, CT and PC conditions may not be warranted and has to be interpreted with great caution, since AT conditions consisted of two possibly correct endings unlike CT and PC trials (one possible ending). When looking at the incorrect solutions within the cognitive ToM trials, adults selected improbable scenarios in 87.1 % and highly improbable/impossible solutions in 12.9 % of the cases. Children selected improbable scenarios in 62.7 % and highly improbable/impossible solution in 37.3 % of the cases. Whole brain fMRI analyses Mentalizing ((CT | AT)>PT) yielded a significant increase in activation in adults and children in frontal brain regions, including medial prefrontal, and orbitofrontal cortices, and inferior frontal gyrus. Activation increase was further observed in temporal regions, such as bilateral temporoparietal junctions, temporal poles and superior temporal sulcus. Parietal regions with heightened activity during mentalizing included inferior parietal lobule, precuneus and supramarginal and angular gyri. Further areas with an increased activation included limbic regions (e.g., cingulate cortex, insula, hippocampus), and basal nuclei (e.g., right thalamus). Affective (AT > PC) and cognitive (CT > PC) ToMrelated activation increases were within expected areas, such as medial PFC, temporoparietal junction and precuneus; see The repeated analyses of the three main contrasts in the adult group with a shortened regressor including only the story-phase yielded similar activation pattern when compared to analyses implementing the full regressor (data provided through NeuroVault: https://identifiers.org/ne urovault.collection:9698). Shared and distinct activation for AT and CT. The conjunction of affective and cognitive trials revealed areas of shared activation in bilateral temporal poles and temporoparietal junctions, right superior temporal sulcus, anterior cingulum, precuneus, bilateral inferior frontal gyri and dorsomedial prefrontal cortex (Fig. 4, Table 5). An increase in activation was observed for affective versus cognitive ToM (AT > CT) trials in anterior precuneus, middle and posterior cingulate cortex bilaterally, inferior temporal gyrus, ventromedial prefrontal and orbitofrontal cortices. Significantly greater activation in cognitive versus affective ToM (CT > AT) trials was observed in left middle and superior frontal gyri, right insula, left inferior and middle temporal, bilateral angular and right supramarginal gyri, hippocampus, and posterior precuneus (Fig. 4, Table 5). In order to assess whether the analysis timeframe had any effect on the neural activation obtained during cognitive and affective ToM, we re-analyzed contrasts of interest (e.g., 'AT > CT'; 'CT > AT') using a shorter regressor, which resulted in an overall similar activation pattern. However, for the cognitive trials a relative increase in activation in right insular and inferior frontal gyrus was no longer observed employing the Fig. 3. Statistical parametric maps displaying neural activation during affective ToM, cognitive ToM, and mentalizing in the adult (red) and child (blue) groups (cluster-level FWE-corrected p < 0.05, using a cluster-building threshold of p < 0.001, uncorrected) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.). Note. Hem = hemisphere, ACC = anterior cingulate cortex, IFG = inferior frontal gyrus, OFC = orbitofrontal cortex, inf. = inferior, mid. = middle, sup. = superior, L/R = left/right, T-scores, k = cluster size and xyz co-ordinates of peak voxel according to Montreal Neurological Institute (MNI). Note. Hem = hemisphere, ACC = anterior cingulate cortex, OFC = orbitofrontal cortex, g. = gyrus, inf. = inferior, mid. = middle, sup. = superior, L/R = left/right, Tscores, k = cluster size and xyz co-ordinates of peak voxel according to Montreal Neurological Institute (MNI). shorter regressor (data provided through NeuroVault: https://identifie rs.org/neurovault.collection:9698). Callous-unemotional traits/empathy and neuronal activation during mentalizing. Two post-hoc partial correlation analyses were conducted in the adult group to investigate the association of callousunemotional traits and empathy with activation in right temporoparietal junction during mentalizing ((AT | CT)>PC). The analyses were controlled for age and gender and Bonferroni correction for multiple comparison testing was used to adjust for the number of tests conducted (p<(0.05/2)). Mean parameter estimates from the right temporoparietal junction as defined according Dufour et al. (2013; retrieved from: http ://saxelab.mit.edu/use-our-theory-mind-group-maps) and extracted through the MarsBar toolbox [Brett et al., 2002; http://marsbar.sourcef orge.net/]. The right temporoparietal junction was chosen as an area of interest because of its key implication across a wide range of mentalizing tasks and studies (Döhnel et al., 2012;Mahy et al., 2014;Powell et al., 2017;Saxe, 2010). Partial correlation analyses revealed a significant negative correlation between CU-traits and neural activation during mentalizing in right temporoparietal junction (r(23) = -0.533, p = 0.006, but no significant association between empathy levels and activation in right temporoparietal junction (r(23) = 0.366, p = 0.072). Further partial correlation tests revealed no significant relationship between callous-unemotional traits and (1) motion during scan (as measured by the average head motion during task or the number of outliers over 1.5mm), and (2) performance on the task, in order to test confounds (Supplementary Table S5). Discussion Here we evaluate feasibility and neural activation patterns evoked by CAToon, a newly developed child-friendly and open-source fMRI Theory of Mind cartoon task. Evaluation included one behavioral study (Study 1; behavioral assessment of 60 children; 3-9 years) and two neural evaluations (Study 2: fMRI in 27 adults and Study 3: fMRI conducted in 33 children). Behavioral results support task feasibility as early as three years of age. However, reliable performance skills are reached around 5 years, which we suggest as an ideal age for fMRI task implementation. fMRI evidence in children and adults confirmed that CAToon is associated with significant activation increases in brain regions associated with mentalizing (e.g., dorsomedial PFC, ventromedial PFC, bilateral temporoparietal junction, middle temporal gyrus, posterior superior temporal sulcus, precuneus, inferior frontal gyrus, precentral gyrus, anterior cingulate cortex and temporal pole). Affective and cognitive ToM trials led to brain activation increases of shared (e.g., bilateral temporal pole, temporoparietal junction, superior temporal sulcus, precuneus and parts of the dorsomedial prefrontal cortex) and distinct brain regions (e.g., AT-specific: orbitofrontal cortex, anterior parts of the precuneus, posterior cingulate cortex, CT-specific: right insula, parahippocampal and fusiform gyrus and posterior potions of the precuneus). Moreover, activation increases in the right temporoparietal junction were negatively correlated with levels of callous-unemotional traits, but not empathy, in adults. Feasibility of the CAToon task for

children Behavioral data (Study 1) and fMRI data acquisition (Study 3) revealed that children of all ages tested were able to complete CAToon above chance level. More specifically, Study 1 indicated that while all children were able to complete CAToon, children aged five years and up performed significantly better than three and four-year-olds. While children of five years and older still displayed variations in performance, no further significant change in task performance was observed, indicating reliable task performance. Behaviorally, children were most accurate in the affective ToM condition, followed by physical causality and cognitive trials. These findings have to be considered with caution, however, since outcome options were not identical for all conditions (e. g., two possible correct endings for AT compared to CT and PC Table 5 Peak activation reports for distinct (AT > CT), (CT > AT) and shared activation (conjunction analysis in adults ((AT > PC) & (CT > PC)). Note. Hem. = hemisphere, OFC = orbitofrontal cortex, inf. = inferior, mid. = middle, sup. = superior, post. = posterior, L/R = left/right, T-scores, k = cluster size and xyz co-ordinates of peak voxel according to Montreal Neurological Institute (MNI). conditions). An increasing performance accuracy of children ages five and up as reported here is in line with previous evidence of children performing reliably on explicit ToM tasks starting around four to six years of age Wellman et al., 2001). Notably, implicit ToM tasks reveal false belief understanding in infants already (Southgate et al., 2007;Surian et al., 2007). However, demands posited by an explicit and/or fMRI task require complementary skills to basic false belief understanding (Lillard and Kavanaugh, 2014). Younger children have been reported to be more challenged by or fail mentalizing tasks that require inhibitory control and working memory (Carlson et al., 2002;Müller et al., 2012;Rakoczy, 2010;Scott and Baillargeon, 2017). The observed performance improvements may result from individual improvements in ToM skills and/or maturation of executive functions typically observed around this age (Roebers et al., 2011;Röthlisberger et al., 2010). Such skill improvements have been linked to the start of formal schooling (e.g., (Brod et al., 2017;Roebers et al., 2011)). For fMRI purposes we therefore recommend the use of CAToon starting around the age of five years and up, which considers increased challenges posed by an MRI environment (Raschle et al., 2012(Raschle et al., , 2009. The use of implicit ToM fMRI tasks by passive movie has shown to be possible in children as young as three years of age (Richardson et al., 2018). Here we additionally evaluated neural activation associated with the story-phase of the CAToon trials only (as compared to the implementation of regressors that include the story and explicit answer phase) in adults, with comparable outcome. While this may be viewed as a first step towards testing CAToon's suitability as a potential passive viewing task, future investigations in younger children are warranted. Through use of conjunction analyses we observed shared and distinct activation patterns when further investigating affective and cognitive ToM trials, which is in line with past evidence (Bodden et al., 2013;Hynes et al., 2006;Schlaffke et al., 2015;Sebastian et al., 2012a). Brain regions that were implicated during both affective and cognitive ToM included bilateral temporal pole, temporoparietal junction, right superior temporal sulcus, anterior cingulum, precuneus, bilateral inferior frontal gyri and parts of the dorsomedial prefrontal cortex. Activation was greater for affective as compared to cognitive trials within the anterior part of the precuneus extending into the posterior cingulate cortex, as well as within the cuneus and orbitofrontal cortex. This pattern remained when analyzing only the story portion of the trials, indicating that passive viewing of the CAToon stories may be sufficient to induce affective mentalizing. A distinct activation of the orbitofrontal and ventromedial prefrontal cortex in affective ToM is in line with literature emphasizing its role for affective processing (Hynes et al., 2006;Molenberghs et al., 2016;Schlaffke et al., 2015;Shamay-Tsoory et al., 2010). Similarly, the posterior cingulate, has been linked to empathetic perspective taking (Schlaffke et al., 2015;Völlm et al., 2006). In contrasts to previous findings (Schlaffke et al., 2015), we have not detected distinct activation in basal ganglia for affective compared to cognitive ToM. This may result from different task designs, as Schlaffke et al. (2015) measured affective and cognitive ToM through the use of the same set of images, but different questions, while CAToon included distinct trials for each condition. Areas with increased activation during cognitive versus affective trials included posterior parts of the precuneus, parahippocampal gyrus, hippocampus and right insula. However, removing the explicit decision phase from the model, the right insula and inferior frontal gyrus did not remain significant. Both regions have been linked to decision making processes (Hartwright et al., 2016;Paulus et al., 2005), which may explain why a shortened model, excluding the decision phase, no longer results in activation increases of these areas. Callous-unemotional traits, empathy and mentalizing We observed a negative association between callous-unemotional traits and neural activation within the right temporoparietal junction in adults. The right temporoparietal junction is most commonly implicated when inferring about thoughts, beliefs and emotional states (Molenberghs et al., 2016). Within the limited literature investigating the relationship between callous-unemotional traits and neural correlates of mentalizing, our findings support those establishing a negative link (Lockwood et al., 2013;Sebastian et al., 2012b) between callous-unemotional traits and mentalizing skills. They may thus be in line with evidence suggesting that adults with higher levels of callous-unemotional traits are more likely to disregard others' feelings (Scheepers et al., 2011) or more likely to display deficient affective perspective taking (Lui et al., 2016). However, past findings are inconclusive, with some reporting a positive association (Gao et al., 2019) or those missing to find a significant connection (O'Nions et al., 2014). Notably, levels of callous-unemotional traits displayed within our adult group did not correlate with an increase in motion during fMRI task performance. Neural correlates of mentalizing using CAToon in children After successfully evaluating the neural correlates associated with CAToon task performance in adults (Study 2), Study 3 further assessed the neural correlates in a group of children (ages 7-13 years), thus testing feasibility in an initial fMRI study of children using CAToon. We observed activation of the mentalizing network comparable to findings in our adult study. Activation clusters in the child group were similar, though seemed slightly less pronounced as reported in adults, which is in line with studies investigating adult and developmental populations during mentalizing (Fehlbaum et al., 2021;Richardson et al., 2018). An increase in neural activation was observed in areas including bilateral temporoparietal junction, medial prefrontal cortex and precuneus (in line with (Gweon et al., 2012;Richardson and Saxe, 2020b)). Our findings provide first evidence for the feasibility of employing CAToon as an fMRI task in children. An effort in developing and subsequently sharing age-appropriate neuroimaging tasks may further replicability and reproducibility of findings (Klapwijk et al., 2019). To provide opportunity for others using CAToon or our findings further, the task and all T-maps reported in the manuscript are made openly available at https://www.jacobscenter.uzh.ch/en/research/developmental_neuro science/downloads/catoon.html. CAToon adds to previous tasks used in children by measuring affective and cognitive aspects of mentalizing. However, it has to be highlighted that the two conditions are not as well isolated (e.g., Sebastian et al., 2012a had distinct cognitive and affective trials) or matched (e.g., Schlaffke et al., 2015 implementing the same images for cognitive and affective trials, but different questions asked), since for CAToon cognitive and affective ToM both include people and affective elements. While naturalistic (e.g., cognitive and affective ToM are social processes, rarely isolated from humans or fully free of affect in real life), this implementation results in a certain confound rendering it challenging to fully isolate individual aspects. Future studies could further test individual stimuli ratings by content, which may then be associated with neural activation in each condition to test the influence of different stimulus characteristics. CAToon was employed as an explicit task, including in-scanner responses. However, first analyses reveal its potential as a passive viewing task, which might be more appropriate for very young participants. Study limitations Several limitations have to be noted. In Study 1 (behavioral evaluation in children), CAToon was presented either in a one-on-one setting or, for older children, in groups. This procedure ensured appropriate understanding and task conduction for younger children, but limits comparability across all age groups. It is notable that younger children performed still lower than children ages five and up. It is also mentionable that there was no strict timing or time limit for the image presentation and decision phase within the behavioral study, which is not possible when running CAToon within an fMRI setting. Additionally, since characters' emotional expressions were included in cognitive ToM scenarios, the affective and cognitive scenarios presented in CAToon are less clearly distinct as compared to previous paradigms (e.g., Sebastian et al., 2012a). While the correct solving of the CT and AT trials is designed to rely on the inference about the targeted mental states (i.e., intentions in cognitive and emotions in affective trials), a direct investigation about the cognitive process underlying participants' answers (i.e., understanding what inferences they make during certain trials) should be investigated in future studies (e.g., collecting subjective responses of participants' reasoning for the answer selection). However, neural activation in adults provides initial evidence of the conditions eliciting activation in established areas (e.g., increased recruitment of ventromedial prefrontal and orbitofrontal cortex during affective trials)." A deliberate choice to include two correct answers within the affective trials was made in order to be able to investigate positive/ negative expectancy in later studies (e.g., of children with and without disruptive behaviors or prior maltreatment). This might be considered as a caveat as it makes the direct comparison of behavioral performance between the different trial types challenging. From a neural perspective, it might be hypothesized that the inclusion of two possible correct answers may require children to evaluate their response even more, thus increasing the need for mentalizing. However, such an effect will have to be further evaluated. We would also like to note, that while there are variations in behavioral task performance across all ages, there are no significant improvements after the age of 5 years when measured outside the fMRI environment (Study 1), or after the age of 7 years when using CAToon inside the fMRI environment (Study 3). In future studies, the use of an additional, established behavioral measure is recommended in order to establish whether children's performance and neural activation is clearly associated. Also, CAToon task stimuli were not tested in different cultures, limiting generalizability and highlighting opportunities for future investigation. Due to the small sample size for brain-behavioral correlations (e.g., comparison of neural activation with callousunemotional traits/empathy) such findings must be interpreted with caution (Cremers et al., 2017). Funding This study was funded through a Jacobs Foundation Early Career Research Grant (Nr. 2016201713) and an early career research grant by the University of Basel (both to NMR). Declaration of Competing Interest The authors report no declarations of interest. The Detection and Characterization of Herpes Simplex Virus Type 1 in Confirmed Measles Cases Based on measles surveillance in Shanghai, People’s Republic of China, from 2006 to 2015, we found that measles virus isolates from 40 throat swab samples exhibited atypical cytopathic effects in Vero/hSLAM cells, which was found to be a result of coinfection with measles virus (MeV) and human herpes simplex virus type 1 (HSV-1). Serological and molecular approaches were used to confirm and characterize the coinfections in these patients. Among the 40 measles cases, measles-specific IgM was detected in 37 cases, while measles-specific IgG was detected in 27 cases. HSV-1-specific IgM and IgG were detected in 7 and 34 cases, respectively, suggesting that most of the MeV infections were primary, but that HSV-1 infection was due to the reactivation of latent virus in most cases. The titers of HSV-1 IgG in patients with either measles or measles-HSV-1 coinfection were significantly higher than those in the healthy group (P = 0.0026 and P < 0.0001, respectively); however, there was no significant difference in the titers of HSV-1 IgG in the MeV and MeV-HSV-1 coinfection patients (P = 0.105). Nucleic acids from MeV and HSV-1 were detected in 40 and 39 throat swabs, respectively. Twenty five MeV RNA sequences were genotyped, and all represented genotype H1, which

is the endemic genotype in China. Sequences from the glycoprotein G gene of HSV-1 were used to classify the isolates into two distinct phylogenetic groups: 34 belonged to group A and 3 belonged to group B. www.nature.com/scientificreports www.nature.com/scientificreports/ from clinical specimens and/or viral isolates. MeV infection can cause cell membrane fusion and induce the formation of multinucleated giant cells (syncytia) in susceptible cells 11 . Such a pronounced cytopathic effect (CPE) is often used to indicate the presence of MeV in cell culture. During serial passages of cultures from a throat swab specimen from a suspected measles case, we observed changes in the types of CPE observed in infected Vero/ hSLAM cells. Syncytia were observed in cells inoculated with the first two passages of virus; however, beginning with the third passage, the infected cells exhibited cell rounding with a decrease in syncytia. Human herpes simplex virus type 1 (HSV-1) infection was confirmed in this culture by real-time PCR. A retrospective study was conducted to investigate coinfection of MeV and HSV-1 based on samples collected from 2006 to 2015. Forty coinfected cases were identified from a total of 4921 cell cultures obtained from throat swabs. were confirmed in Shanghai. The measles incidence was between 0.35/100,000 and 6.79/100,000 annually. The ratio of male to female cases was 1.28:1. Most cases occurred in populations aged 8-11 months, 20-29 years and 30-39 years. Adults aged >20 years accounted for 50% of all cases. Measles immunization has been documented since 2009 when the MSS was established. The majority of measles cases occurred in individuals without an immunization history or who had an unknown history. According to the MSS, a total of 186 cases were identified in individuals who had received at least one dose of measles-containing vaccine (MCV), who accounted for 2.48% to 13.25% of the cases between 2009 and 2015. Among these cases, 70.43% had received one dose of MCV, 24.73% had received two doses of MCV, and 4.84% had received more than two doses of MCV. Measles surveillance in From 1253 MeV isolates during this period, 40 cases of MeV and HSV-1 coinfection were identified. The average age was 23 (the age ranged from 6 months to 53 years old). Twenty-four were adults older than 20 years old. Three had documented MCV (SH11300, SH15370, and SH15185), whereas the remaining had an unknown or no history of MCV or disease (Table 1). All 40 patients reported fever and rash, 33 (82.5%) experienced coughing, and 24 (60%) experienced catarrhal symptoms. Detection of viral nucleic acids. Viral nucleic acids were detected by real-time PCR/RT-PCR from throat swab samples or infected Vero/hSLAM cultures, which were divided in to 3 groups, the coinfection group (n = 40), measles group (n = 463) and healthy group (n = 192). Among the 40 MeV and HSV-1 coinfection cases, MeV RNA was detected in 40 (100.0%) throat swabs by RT-qPCR. There were no significant differences in the MeV RNA copy numbers between the coinfection group and the measles group (P = 0.5033). HSV-1 nucleic acids were detected in 39 of the 40 throat swabs by qPCR, and the copy number of HSV-1 DNA in the coinfection group was significantly higher than that in the measles group (P = 0.0028). Of the 463 throat swabs from MeV cases, 17 (3.67%) were positive for HSV-1 DNA, whereas 3 (1.56%) swabs from the healthy group were positive. There was no significant difference in the positive HSV-1 DNA rate between these two groups (P = 0.1316). Serological results. Serum samples from the 40 MeV and HSV-1 coinfection cases were tested for IgM and IgG antibodies (Table 1), the OD value of IgM and IgG and judgement were supplied in detail as Supplementary Table S1. Thirty-seven (92.5%) were positive for MeV IgM, and 27 (67.5%) were positive for MeV IgG. Although three (SH11300, SH12649, and SH15370) were negative for MeV IgM, MeV RNA was detected in the corresponding throat swabs. Of the 40 cases, seven (17.5%) and 34 (85.0%) were positive for HSV-1 IgM and IgG, respectively. Five (SH06064, SH08035, SH11300, SH12538 and SH151187) were negative based on both HSV-1 IgG and IgM. Among these five patients, four were children under 5, while SH151187 represented a patient older than 50. Table 2 shows the titers of HSV-1 IgG antibody in the healthy group, coinfection group, and measles group. Statistically significant differences were found between the measles group and the healthy group (P = 0.0026), the healthy group and the coinfection group (P < 0.0001), but no significant difference was found between the measles group and the coinfection group (p = 0.105). Virus isolation and laser scanning confocal microscopy results. A total of 1253 MeV isolates were obtained from 4921 throat swabs between 2006 and 2015. The proportion of MeV and HSV-1 coinfection ranged from 0.0-11.1% each year. All MeV isolates produced syncytium formation in infected Vero/hSLAM cells during the first three passages. Forty MeV isolates exhibited another type of CPE characterized by larger and rounder cells in cultures as early as the 3 rd passage, which became more prominent in the subsequent passages. These cultures were confirmed to be coinfected with MeV and HSV-1 by the detection of viral nucleic acids. Laser scanning confocal microscopy was used to demonstrate the dynamics of MeV and HSV-1 infection in coinfected Vero/hSLAM cells (Fig. 1). In cells infected with SH13390 at passage 3, MeV infection, detected by antibody of anti-MeV matrix (M) protein, as indicated by green fluorescence, was readily observed in some cells, whereas very few HSV-1-infected cells, detected by antibody of anti-HSV-1 ICP0, as indicated by red fluorescence, were found. In the 6 th passage of the infected culture, significant increases in MeV and HSV-1 infection were observed. In HSV-1-infected cells, the red fluorescence appeared as granules in the cytoplasm. As passaging continued, the number of cells stained with red fluorescence increased, while the number of cells stained with green fluorescence decreased. Eventually, at the 17 th passage, only cells with red fluorescence were detected (data not shown). www.nature.com/scientificreports www.nature.com/scientificreports/ Supplementary Table S2), and all were genotype H1a (Fig. 2). The nucleic acid sequences of these 25 MeV exhibited 97.8-100.0% identity, and there was 98.0-100% amino acid sequence identity. The nucleotide sequence identity among the sequences of the MeV viruses and the reference sequence of genotype H1 (GeneBank: AF045212) was 97.3-98.0% (96.0-98.0% for the amino acid sequences). The overall nucleotide sequence identity between We sequenced 37 HSV-1 isolates from the 40 coinfected cultures (GenBank accession numbers: MN166405-MN166441; More sequence information of HSV-1 isolates were provided as Supplementary Table S3). All contained 4-10 guanine-adenine-guanine (GAG) trinucleotide tandem repeats beginning at nt 235 (GenBank: JN555585, using the same numbering scheme) in the gG region. Point mutations in the GAG repeats were detected in some viruses. Based on phylogenetic analysis, these 37 HSV-1 isolates were divided into two lineages. The majority (34/37) of the sequences belonged to group A, and the remaining sequences belonged to group B (Fig. 3). The group A HSV-1 isolates had 98.8-100.0% nucleic acid identity and 97.0-100.0% amino acid sequence identity, while all three group B isolates had identical sequences. The nucleic acid identity between groups A and B was 96.5-97.3%, and the amino acid sequence identity was 94.0-95.7%. In the region between nt 267-703, the viruses in group A had completely different bases than the viruses in group B at 17 loci ( Table 3) www.nature.com/scientificreports www.nature.com/scientificreports/ commonly associated with HSV-2 and HIV infection, and coinfection of Lyme borreliosis and HSV-1 has also been reported [15][16][17] . In addition, since both HSV and VZV can establish latent infections in sensory neurons after primary infection, a rare case of coinfection due to the reactivation of both HSV and VZV has been reported 18 . In this study, nucleic acids from the rubella virus, dengue fever virus, EBV, B19, Adv, HEV, HHV6, CMV or VZV were not detected in any of the clinical throat swabs or cell cultures from the measles or coinfection groups. Sequence analysis of Based on the serological testing and immunization records, the majority of the codetection cases (37/40, 92.5%) were primary measles infections. SH11300, SH15185 and SH15370 had received at least one dose of MCV. SH11300 was negative for MeV IgM and IgG. This case had received a dose of MCV 5 days before the onset of rash. Unfortunately, we were not able to determine whether the infection was a coincidental event or a vaccine-associated measles case due to a lack of sequence information. SH15370, which was only positive for MeV IgG, was classified as secondary vaccine failures. SH15185, which was only positive for IgM and the patient was 1.5 years old, might be a primary vaccine failure 19,20 . SH12649, which was lack of immunization records and positive only to IgG test, couldn't be confirmed as immune failures, suggesting prior exposure to MeV or vaccine. In the same way, there were 25 cases testing positive to both IgM and IgG and it may be due to the window period of immunological test under primary infections 21 . SH11300, SH12649 and SH15370 were MeV IgM negative but were confirmed measles cases based on RT-qPCR testing. Sera were collected from these three cases within 2 days after the onset of rash. The rates of HSV-1 IgG detection in the coinfection (85%) and measles groups (73%) were significantly higher than in the healthy group (57%). The HSV-1 IgG detection rates are lower than the reported seroprevalence (92%) in China 22 . Among the 40 coinfection patients, serological results suggested that 13 of these cases were likely to be primary infections of HSV-1, and 10 of these patients were under the age of 5. Upon comparison of the HSV-1 IgG titers, the titers in the MeV group and the coinfection group were significantly higher than in the healthy Table 3. Analysis of the loci in recombined regions in the HSV-1 gG gene. The nine loci between nt 324 and nt 429 were those where recombination occurred between groups A and B in AF1171219. www.nature.com/scientificreports www.nature.com/scientificreports/ the nucleic acid sequence of the whole gG gene, the phylogenetic analysis showed that most of the HSV-1 isolates identified in this study were in group A and that three isolates were in group B. This distribution is different from that found in European countries 26,27 . The nt 267-703 region in the HSV-1 gG gene has been identified as a recombination hotspot between groups A and B 23 . It has been noticed that the grouping of HSV-1 differs when different segments of the gG gene sequence are used for phylogenic analysis, suggesting that recombination has occurred within this region between the two groups 23,28 . Using sequence analysis and SimPlot, we did not find any evidence of recombination events in this region in the 37 HSV-1 isolates. In this study, we found 17 loci in the gG gene where the bases were completely different between group A and group B. More sequences are needed to determine whether these loci also serve as recombination hotspots. Fatal cases of coinfection with MeV and other pathogens have been reported 12 . It is not clear whether MeV infection increases susceptibility to HSV-1 infection or triggers reactivation. Because of the high prevalence of HSV-1 in the population and the ability of HSV-1 to cause latent infections, the number of coinfection cases may be underestimated. Additional research is needed to investigate coinfections of MeV and other pathogens and to determine whether these co-infections affect the disease severity and/or require special treatment. In this study, the measles group consisted of cases confirmed by either MeV IgM testing or MeV RNA testing by real-time RT-PCR assay in 2015. The MeV and HSV-1 coinfection group included confirmed measles cases with HSV-1 infection that exhibited atypical CPE in infected cells after three passages. In addition, sera and throat swabs were collected from individuals who did not have clinical presentation or signs of measles at the time of sample collection to serve as controls ("healthy group"). The total numbers of sera samples in the measles, coinfection, and healthy groups were 741, 40 and 741, respectively, and the numbers of throat swab specimens collected for these three groups were 463, 40 and 192, respectively. Methods To obey ethical guidelines and ensure data protection, the serum and throat swab

samples used in this study were collected according to the guidelines of the national and Shanghai municipal measles and rubella surveillance programs and did not involve human experimentation. This study strictly followed the relevant ethical requirements and strictly protected the data security and privacy of the participants. Detection of viral nucleic acids. Viral nucleic acid from throat swab samples or infected Vero/hSLAM cultures was extracted using the QIAamp Viral RNA/DNA Mini Kit (QIAGEN; Hilden, Germany) according to the manufacturer's instructions. Viral nucleic acid was detected by real-time PCR/RT-PCR using a detection kit for MeV and rubella virus (RuV) (BioPerfectus Technologies, Taizhou, Jiangsu, China), human herpes virus type 6 (HHV6) and varicella-zoster virus (VZV) (Tianlong Biotechnology Co. Suzhou, Jiangsu, China) according to the manufacturer's instructions. The detection of other viral febrile respiratory illnesses, such as human cytomegalovirus (CMV) 29 , Epstein-Barr virus (EB) 30 , human parvovirus B19 (B19) 31 , human adenovirus (Adv), human enterovirus (HEV) 32 and herpes simplex virus type 1 33 , were carried out using the One Step PrimeScript™ RT-PCR Kit (Perfect Real Time) or Premix Ex Taq™ (Probe qPCR) (TaKaRa; Dalian, China) for RNA or DNA viruses as reported previously. Serological tests. Serological specimens were collected within 5 days after the onset of measles infection. Detection of MeV IgM was carried out using an IgM ELISA test kit (Haitai Biological Pharmaceutical Co.; Zhuhai, Guangdong, China). MeV IgG, HSV-1 IgM, and HSV-1 IgG were detected using reagent kits manufactured by Institute Virion/Serion GmbH (Würzburg, Germany). All procedures and the interpretation of the results were performed according to the manufacturer's instructions. Specifically, if the results of two repeated experiments were consistent for the cases within the cut off value, it would be judged as negative. Virus isolation. The Vero/hSLAM cell line was used for MeV isolation from clinical specimens 11 . Vero/ hSLAM cells were kindly shared by Dr. Yanagi at Kyushu University, Japan, through the WHO Global Measles and Rubella Laboratory Network (GMRLN). Procedures used for the maintenance of cells, sample handling and virus infection were performed as described in the Manual for the Laboratory-based Surveillance of Measles, Rubella, and Congenital Rubella Syndrome (WHO) 34 . Indirect immunofluorescence assay (IFA) and laser scanning confocal microscopy. A throat swab sample, SH13390, from a confirmed MeV and HSV-1 coinfected case was used to infect Vero/hSLAM cells, which was followed by ten cell passages. Infected cells from each passage were subjected to IFA. Briefly, the cells were washed with 0.01 M phosphate-buffered solution (PBS) and fixed with 4% polyoxymethylene Integrating Massage, Chiropractic, and Acupuncture in University Clinics: A Guided Student Observation Background Several studies have reported on the health benefits of applying an integrated complementary health care model. Purpose This paper presents the results of pilot research focusing on the observations massage therapy students made about complementary health care education and integration during massage, chiropractic, and acupuncture treatments at two university clinics. Setting: Observations took place at Northwestern Health Sciences University’s associated clinics that offered massage, chiropractic, and acupuncture. Research Design: Students directly observed how clinicians and interns educated their patients and integrated other forms of complementary health care into their practice. Participants: chiropractors, massage therapists, and acupuncturists, and their patients. All participants were English-speaking and 18–65 years old. Main Outcome Measures: Observations recorded by students in journals about education and integration during massage therapy, chiropractic, and acupuncture treatments were coded and counted. Results Qualitative observations showed that clinicians and interns educated patients to some degree, but the clinicians were less apt to integrate other modalities than the interns. Conclusions Observations support that professional integrity may limit clinicians in their ability to integrate multiple modalities of health care while treating patients. Since it is well established that integration of multiple health care modalities is beneficial to patient health, it is recommended that clinics assist their clinical staff in applying an integrative approach to their practice. introduCtion Background Patient education is instructed health information that empowers a patient with self-care methods to improve their own health. Providers of complementary and alternative medicine (CAM), such as massage, chiropractic, and acupuncture (MCA), improve health by emphasizing patient wellness and not illness while treating and educating patients (1) . For example, persons with chronic low back pain generally benefit from repeated acupuncture and massage treatments (1) , whereas persons with chronic arthritis (2) or diabetes (3) have improved health and lowered health care costs with self-care. Integration of allopathic and complementary health care models into patient education and care have also improved patient health. Many allopathic medical clinics and hospitals have successfully integrated complementary health care practices into their health care philosophy (4)(5)(6)(7)(8) . Primarily, studies have shown that the key to successful integration was provider education (4,5) . Secondarily, when providers knew about CAM integration benefits, they could educate patients about them. A partnership between the Oregon College of Oriental Medicine and Oregon Health and Science University showed that it was important for providers and patients to believe that acupuncture improved health in order for providers to recommend it and patients to choose it (7) . Beyond being helpful for patients, providers at the University of Pennsylvania (6) and Melton Rutland Harborough Primary Care Group (England) (8) believe that integration of CAM modalities could reduce health care costs. All of these examples are important, but may make no difference to some health care providers because of personal opinion. CAM may not be recommended if the opinion is negative (7) . Providers may be biased against the technique because of "lack of evidence" stereotypes some people may have about CAM (9) . Additionally, economic competition between health care modalities could influence care, because one provider may not want lose money to another (10) . CAM providers are not exempt from these limitations. Variation in CAM provider opinions appears to be the main reason why it has been difficult to integrate CAM modalities. Opinions include power (e.g., believing that one modality is superior to another), knowledge, and criticism about the effectiveness of other methods, or that the modalities are simply different and are the result of individual choices along a Alejandra A. Estrin Dashe PhD College of Undergraduate Health Sciences, Northwestern Health Sciences University, Bloomington, MN, USA continuum of health care (e.g., ranging from natural remedy, wellness, accepted, and established modalities) (8)(9)(10)(11)(12) . These limitations are especially problematic because providers separate each modality from the others as independent health care choices. Furthermore, this does not provide the best care for patients as evidenced by the relationships established between CAM and allopathic techniques (1)(2)(3)(4)(5)(6)(7)(8) . Some health sciences universities are working on integrating complementary health care modalities and transmitting the benefits to the students and patients alike. CAM modality integration is the process of combining multiple health care disciplines (such as MCA) into one team that shares one paradigm: to heal and maintain wellness (10,13) . For example, at National University of Health Sciences (NUHS) in Lombard, Illinois, patients, clinicians, and interns reported that they would like to see more education and integration of the modalities (13) . Becoming comfortable with the knowledge may be a factor in the integration process. At least 50% of NUHS clinicians (n=10) and interns (n=89) said they were very familiar with massage therapy, chiropractic, acupuncture, and naturopathic treatment (13) . setting Northwestern Health Sciences University (NWH-SU) in Bloomington, MN is a CAM institution with four colleges: massage therapy, chiropractic, acupuncture, and an undergraduate program. NWHSU also has several teaching clinics that serve students in all colleges. The University has committed itself to improve the integration of MCA into an integrated complementary health care system (14) by offering classes that address the commitment. One of the classes offered in the undergraduate program is a general Sociology class. As a way to fulfill the mission of the University, a requirement of Sociology students is to observe the culture of health care. Based on the needs of the University to fulfill its mission and the general consensus in the academic literature that CAM modality integration is beneficial to patients, an objective of this pilot study is to begin to understand how education and integration of complementary health care is culturally transmitted between patients and providers. Also, a second objective is to show that observing patient-provider relationships (as an example of health care culture) can serve as an experiential learning model. It is hypothesized that providers educated patients about health care without regard to modality boundaries. The data presented are the Sociology students' observations, analysis, and recommendations for future student projects. The author also explores how this pilot study fits into a discussion about patient education and integrative relationships between MCA practices. Methods sociology 1001 Sociology 1001 is an introductory survey course about theory and methods of sociology. The course serves undergraduate students who want to obtain an associate degree in massage therapy or a bachelor degree in health sciences. The course description is as follows: This course will guide students through an introductory study of various elements of human interaction. Historical elements related to the development of social theories will be reviewed. This course will review the concepts of social stratification, class, race and ethnicity, gender, culture, deviance as they relate to a study of Sociology. The topics of science, medicine and health care as they relate to a study of Sociology will provide a focus of this course. Fundamental concepts related to methodology of Sociology research will be addressed. We will examine all topics through the lens of the sociological perspective (15) , which is the ability to see and understand how general society works by observing the particular behaviors of people. Ten students (three males, seven females) participated in the course. Eight were majoring in massage therapy, one was pre-chiropractic, and one was preacupuncture. The students received an explanation of the study on the first day of class. participants Sociology 1001 students observed five appointments at any of three selected clinics: University Health Services (UHS), Pillsbury House (PH), and Winter Family Clinic (WFC). UHS and PH are free clinics that provide MCA with NWHSU clinical professors. UHS serves the NWHSU community and PH serves the local community in south Minneapolis (the Pillsbury neighborhood). WFC is a stand-alone clinic in south Minneapolis offering MCA. Anyone can receive care at the three clinics for low to no cost. Students registered with the clinics to observe appointments. Participants in this study were the patients seeking and providers of MCA practices at the three clinics. Appointments were selected by the receptionists at each clinic based on the participant parameters for patients: aged 18-65, English-speaking, and not a member of a protected class (minority, juvenile under the age of 18, or pregnant). The receptionists were responsible for making sure the participant parameters were met. The participants were not approached before appointments. The study was explained to potential participants when they arrived in their treatment room. human subjects protection This project was approved by the Northwestern Health Sciences University Institutional Review DASHE: INTEGRATING MASSAGE, CHIROPRACTIC, AND ACUPUNCTURE Board (NWHSU-IRB) on October 17, 2011 (AED-95-09-11). Once approval was obtained, students were told to commence data collection. All participants in the study remained anonymous and no identifying information was recorded. The research project in this course has intrinsic worth to each student seeking to develop an understanding of sociology and the methods by which sociological knowledge is generated. As a result, requiring participation of each student in the data collection was considered justified, and seeking an IRB for student participation in the course was not considered necessary. Further, the interdepartmental planning that set up the observational research and the training of students prior to their research activity presented an experience of extremely low-risk to everyone concerned-patients, providers, and students of the Sociology course. However, as a precaution in the future, subsequent studies involving student data collection will include an IRB for students participating in this course. procedure At each appointment, the student observer introduced him/herself to the patients (people seeking treatment) and providers (clinicians and interns), explained the study, provided informed consent forms, secured informed consent, and quietly observed the patient-provider interaction. Students were asked to keep field notes, or write down their observations in a journal. Observations included only what they could see and hear during the appointment, because the IRB committee advised that it

was safer for the patients and students to avoid communication. This meant that students could not collect data on the numbers of eligible participants, the number of participants who agreed or refused, or demographic data. However, students were allowed to speak to providers after patients were out of hearing range. Understandably, these limitations may have encouraged the Hawthorne Effect (16) (observer influence on the social interaction) and made it difficult to tell if there was participant selection bias or demographic influence on the results. Putting these limitations in place was an important risk because they minimized patient risk and simplified the procedure students had to follow. Data collection started on October 17, 2011 and continued through December 2, 2011. Data were analyzed in class starting on December 2, 2011 and at home with guidelines set by the professor. Students presented their findings as a poster or presentation as part of a final grade for the Sociology 1001 class on December 9, 2011. Students were also required to turn in their field notes, informed consent forms, and a copy of the poster or presentation in Microsoft Power Point (Microsoft Corporation, Redmond, WA) for grading and completing this report. data analysis & statistics The data from the students' field notes were coded (by the professor) to create the following variables: number of observations, number of providers (clinician and intern), number of patients, modality (chiropractic, massage, or acupuncture), location of services (UHS, PH, or WFC), education, and modality integration. Number of observations refers to the number of times a student observed a treatment. Number of providers are the number of providers observed, and this was divided up by whether students observed clinicians (professional chiropractors, acupuncturists, or massage therapists who are teachers at the clinics and oversee intern education) and/or interns (students at NWHSU completing their clinical education). Number of patients refers to the number of patients observed. Modality refers to the type of treatment observed, either chiropractic, acupuncture, or massage. Education refers to the instances of observed health care education provided in the treatment. Modality integration refers to the instances of multiple observed treatments from more than one modality, during, in the past, or referred to for future health in the appointment. All quantitative data were recorded and analyzed using Microsoft Excel (Microsoft Corporation, Redmond, WA). results Of ten students in the Sociology 1001 course who performed observations, eight students turned in their field notes. The results reported and analyzed are based on those notes. Referring to Tables 1 and 2, there were 30 total observations, where 31 patients and 41 providers were observed. Students observed treatments in teaching clinics, and frequently saw interns and clinicians working together or, in one instance, there were two patients being treated together. Student observers mostly saw treatments at UHS. Only nine observations took place at PH. No treatments were observed at WFC because it was more convenient for students to stay on campus and observe at UHS or relatively close by at PH (data not reported). The majority of treatments observed were chiropractic (n=27 at UHS and PH, where n=21 at UHS only). Most of the students agreed that they consistently observed patients and providers communicating about health care education and that the providers did their job-they healed and educated their patients. However, upon analyzing the students' field notes, education communication was observed 27.5% of the time (11 out of 30 observations) and modality integration was observed 26.7% of the time (8 out of 30 observations). Further investigation and discussion with the Sociology students uncovered that they observed mostly interns educating, not clinicians. The students noted that it appeared that each clinician (chiropractors especially) wanted to treat their patient and move on to the next one. It is possible that the clinicians were moving quickly to accommodate several appointments. In comparison, interns may have spent more time with patients in order to learn how to become better clinicians, which may have given the impression of providing more education. On the other hand, patients were observed asking their providers to educate them about other forms of health care. For example, one chiropractic patient asked their chiropractic intern about "therapeutic order" (13) , whether to get a massage or chiropractic treatment first. The intern suggested getting a massage first because it loosens up the joints for chiropractic. The patient was surprised by this response, but it is a standard practice. The surprise could have been a response to a cultural assumption that a chiropractor would encourage chiropractic first over massage. In fact, it was observed that some of the clinicians encouraged treatment from their own professions first. There were two instances of professional preference observed, detailed in the next two paragraphs. In one case, a patient asked about seeking a second opinion from an orthopedic surgeon. The chiropractor gave their personal experience with this form of allopathic medicine. The student observer thought that the chiropractor was "pushy" about his opinion regarding getting chiropractic treatment over orthopedic surgery, but in the end said to the patient that they could do what they wanted. In another observation, an acupuncturist suggested to a patient to wait on massage therapy so that the focus of the treatment would be on acupuncture first. It seemed to the students observing that the clinician did not wish for their patients to engage in another modality until their own modality exhausted all health care possibilities, and clinicians tried to talk patients out of going to other providers. These instances, while few, might suggest that clinicians show some discomfort or resistance to integration. The observations of the interns provided a different scenario. During one observation, a chiropractic intern asked the student observer what they would do as a massage therapist to help the patient. This put the student observer in an uncomfortable social situation. It was clear to the observing student that the intern felt comfortable asking the student observer their opinion. However, the student observer could not say anything, even if they wanted, because they were not allowed to treat patients and only observe as stipulated by the IRB. Finally, the Sociology students noted that there was a difference in how patients were seen by providers. Observations at UHS showed that clinicians were modeling independent practices for their interns because of how the appointments were arranged-in rooms big enough for one provider and one patient. Observations at PH showed that integration worked because the appointments were arranged in a larger room with space for several providers to work on one patient. The students noted that education and integration would have been easier if treatments could have been done in larger rooms or with floating massage therapists, chiropractors, and acupuncturists around during each appointment. patient-provider interactions and CaM integration The students' observations generally show that interns and clinicians experience different patientprovider interactions. The interns in this pilot study provided education and promoted integration of multiple MCA modalities whereas the clinicians were not as active. Such findings partially concur with recently published data: NUHS clinicians and interns were familiar with, interested in, and actively promoted education and integration of more than one MCA modality into their practice (13) . Possible reasons why the NWHSU interns seem more active in education and integration than clinicians are presented in this discussion. Interns are invested in their own education. They are learning multiple modalities at once, particularly while in a clinic setting. This is especially true at PH, where several modalities are used on the same patient in the same appointment. Interns at other universities have shown this pattern as well. Acupuncture interns at the Oregon College of Oriental Medicine stated that they learned more about conventional medicine while interning as acupuncturists at a conventional medical clinic (7) . Interns are also under the educational direction of clinicians. They may feel some academic freedom in not knowing what to do in every instance. Interns can call upon a second provider to assist in patient treatment, which may explain why one chiropractic intern asked the student observer to assist him/her with their patient. On the other hand, MCA clinicians may not feel like they have the same flexibility as the interns to recommend or support other forms of health care because of paradigmatic constraints. They may be more accepting of paradigms from their chosen professions and, possibly, are not as accepting of others because that is how they were trained (10) . There also may be a perceived superiority of MCA practices over others because of pride and need to dominate the market. Perceived superiority may also encourage "use of closure or exclusive strategies aimed to limit or regulate the entering of competitors into the market" (10) . In other words, there are some economic incentives to promoting MCA practices over others. The result is a power structure that is difficult to disengage: that one's own MCA practice is more important or better than others. In turn, the health care market is dominated by one type of health care over others and is a disintegrated system. MCA clinicians may promote their own practice, but may also involuntarily disempower their patients. A resulting effect is less patient agency to make informed health care choices when only one type of care is provided. In the instances observed where the clinician suggested acupuncture over massage or chiropractic over orthopedic surgery, the patients' agencies in their health care were muted by the clinicians' opinions. These observations show that it may be important to teach providers how to promote health care integration without offending MCA practice integrity, as well as why integration is beneficial to patients seeking wellness. One way to begin the process of making changes in how education and integration are applied in clinical settings is to model the behavior for the interns. Interns know that they should provide education to their patients. Yet when interns observe clinicians in action, it is important that the clinicians are able to show that they believe in what they do and what other MCA providers do, too. The Sociology students also suggested a few ways to encourage integration. One way would be to improve massage education techniques for chiropractic students so that it is up to the level that massage therapists learn them. Another way to support an integrated model would be to provide a floating MCA provider in case a recommendation is needed. Teaching chiropractors more massage and having extra staff at a clinic would relieve a provider from having to wonder what another provider would do. Finally, clinicians need training opportunities to become comfortable with an integrated model. Educating providers about other health modalities is important for delivering good health care, even if most providers already know about them (13) . It is also equally important to educate providers when to recommend more than one form of health care without professionally offending other providers. This form of training may be necessary to culturally incorporate complementary health care modality integration into professional MCA practice. learning from students: limitations and future study One of the limitations to this pilot study is that the Sociology students' presence in the treatment room influenced the behaviors of the patients and providers and, therefore, their observations, which is also known as the Hawthorne Effect (16) . This psychological variable should be eliminated from future studies. Future studies would focus on collecting data behind a video camera in another room, so that patients and providers are not aware that they are being observed (but have given consent ahead of time). Sociology students also wanted to know more about the patient-provider interaction beyond just observing without interaction. The next group of students should interview patients and providers directly before and/or after the treatment to get more specific data about demographics, how the patients and providers communicate, and details on integration of health care modalities. Interviews would need to be recorded and transcribed to provide more information about patient-provider interactions and CAM modality integration. The study was a pilot study. As such, the entire body of literature on educational intervention about the integration of multiple CAM modalities was not presented. Future studies will include a greater literature review. Also, only three clinics were associated with this study and the study was limited the participants to English-speaking patients only. As a result, the study

had a small sample size and broad conclusions. The next round of data collection would be expanded to include all health care clinics on the NWHSU campus and non-English speaking patients, so that sample size and data collection are increased to confirm the broad conclusions made in this study. A Multilevel Logit Estimation on the Determinants of Utilization of Preventive Health Care and Healthy Lifestyle Practice in China The purpose of this study is to provide policy implications by estimating the individual and community level determinants of preventive health-care utilization in China based upon data from the China Health and Nutrition Survey. Two different frameworks, a human capital model and a psychological-behavioral model, are tested using a multilevel logit estimation. The results demonstrate different patterns for medical and nonmedical preventive activities. There is a strong correlation between having medical insurance and utilizing preventive health services. For the usage of medical-related preventive health care (MP), age, gender, education, urban residence, and medical insurance are strong predictors. High income did not provide much of an increase in the usage level of MP, but the lack of income was a huge obstacle for low-income people to overcome. Community variation in number of facilities accounted for about one third of the total variation in the utilization of MP. The utilization of MP in China remains dependent upon the individual's social-economic conditions. Introduction How do people deal with threats to their health? In the developed world, the answers to this long-standing question tend to be either behaviorally or cognitively oriented (Kirscht, 1983;Prentice-Dunn & Rogers, 1986). However, the utilization of preventive health care in developing countries can be substantially different from that in the developed world (Dupas, 2011). Preventive health-care activities are associated with socioeconomic conditions, and their use depends on the types of activities involved; for example, whether they require medical facilities or can be done by individuals. In this study, we will investigate the determinants of both medical preventive health care and healthy lifestyle practice in China, using a data set from a nationally representative survey. "Prevention first" has been one of the major health-care policies in China since the 1950s. This policy was considered a success in improving the health status of the Chinese people in pre-reform China (Sidel & Sidel, 1982;World Health Organization, 1975). The merit of the pre-reform Chinese medical system was characterized by an emphasis on prevention: "low-cost, locally controlled health services and promoting accessible primary health care in rural areas" (Hillier & Shen, 1996, p. 258). The pre-reform era's health-care system was efficient in terms of the control of infectious diseases; despite pervasive low productivity and shortage of services in almost every area, Chinese people's health status improved throughout the 1980s and 1990s (Fang & Bloom, 2010;Shu & Yao, 1997). During the economic reform, the utilization of preventive health care became challenging for many people, especially people living in rural areas and people without medical insurance, because of decentralization and marketization. Financial decentralization limits the role of the government in providing public health programs. During the reform, the urban medical scheme has been financed mainly through municipal-level risk pooling for employees with basic medical insurance. The rural "New Cooperative Medical Scheme" (NCMS) initiated in 2003 focuses "almost exclusively on medical care and funds only a few preventive services" (Fang & Bloom, 2010, p. 32). This policy may lead individuals to a more curative care focus, since one of the fundamental changes lies in increasing out-of-pocket expense. Thus spending on preventive health care becomes more of a personal choice for future well-being, rather than a public policy arrangement as it was before the reform. Although the ongoing health-care reform initiated in 2009 aims to provide comprehensive universal health coverage for all citizens by 2020, individuals' incentives would remain an important element in the new system (Yip et al., 2012). In this way, it becomes more about the individual's investment in their own human capital, impacted both by the individual's understanding of the importance of preventive care, and their economic ability to carry the associated costs. In recent years, an increasing literature has focused on various key factors affecting the utilization of health care in China (Gao, Raven, & Tang, 2007;Liu, Zhang, Lu, Kwon, & Quan, 2007;Qian, Pong, Yin, Nagarajan, & Meng, 2009;Wagstaff, Lindelow, & Hsiao, 2009;Zhang & Kanbur, 2005;Zhang, Wang, & Zhang, 2014). However, although early studies explored the preventive healthcare system before and after the reform (Hillier, 1986;Hillier & Shen, 1996;Kaufmana & Jing, 2002), if we limit our searching area to the determinants or factors affecting the utilization of preventive health care, we still have the impression that we know little in this area. Among the very few studies on the determinants of preventive health care in China, van Dalen (2006) estimated a few socioeconomic variables, including insurance and wealth, using probitestimation and instrumental variables (IV) based on the data of the China Health and Nutrition Survey 2004. Using probit-estimation he found that possessing health insurance increased the probability of using preventive health care by 3 percent, but this effect was not statistically significant per the results of the IV probit-estimation. He also found that being in the lowest 40 percent of population wealth would reduce the likelihood of using preventive health care by 1 percent. However, being in the top 20 percent wealth group did not have any effect when compared with the middle wealth group. Many issues in the areas of preventive health care in China remain unexplored. One of the reasons that we should pay more attention to preventive health care is that it is an efficient way to improve the well-being of the population, compared with curative strategies. Since much disease is preventable, and "[p]reventable illness makes up approximately 70 percent of the burden of illness and the associated costs" of health care (Fries et al., 1993, p. 322), preventive care is an efficient way to reduce health risk and significantly improve health outcome (Harvey, 2001;Rose, 1992). Also, because preventive health care reduces the incidence rate of disease, it is considered to be far more cost-efficient than a curative strategy (Kaplan, 2000;Rose, 1992). The paucity of literature in this area and the costs and benefits of preventive care motivate us to examine the determinants of utilization of preventive care in the context before the current health-care reform initiated in 2009 using the 2006 wave of the China Health and Nutrition Survey. We are going to estimate the determinants of preventive health care with Grossman's (1972) framework, in which individuals, as the producers of health, allocate their resources to produce health. The reason we test the determinants with Grossman's framework is that individual behaviors in the utilization of preventive health care are at least partly rational and predictable. As such, understanding the determinants of utilization of preventive health care can help policymakers formulate better strategies for improving health outcomes in the current health-care reform. Specifically, this paper attempts to answer the following research questions: (1) What are individuals' determinants for using medical preventive health care and non-medical healthy lifestyles in China? (2) Do community factors make a difference in determining the use of preventive health care? (3) What are the policy implications regarding the utilization of preventive health care? The remainder of this paper is organized as follows: The next section describes the data set, analytic framework, and variables estimated. The third section discusses the results and provides research implications. Finally, the last section will summarize the findings. Theoretical Framework Different theories and models have been used to explain the utilization of preventive health care, including human capital, psychological, and behavioral models. These models are based upon a variety of frameworks and assumptions. The Grossman (1972) model recognizes that health is both a consumption good used for direct satisfaction and an investment good for higher productivity. Using this model, the utilization of preventive health care for individuals can be understood as an investment in human capital. Although individuals cannot "purchase" health from the market, they can prevent the occurrence and reduce the severity of certain diseases by diagnosing them in the latent stage through the use of preventive health care. That is, the return on investment with preventive health care is the improved health status through prevention and reduction in the risks of severe illness. Similar to an investment in physical capital, an individual makes a decision to undergo preventive health care based on vectors of their individual characteristics and social-economic conditions (Glanz, Rimer, & Viswanath, 2008). The Grossman model performs well when the preventive care activities are more medically focused in nature (Cropper, 1981;Kenkel, 1991Kenkel, , 1994Wagstaff, 1986). However, the utilization of preventive health care is not simply the "purchase of health." Preventive health care differs from other commodities. First, uncertainty is intrinsic to health care, both in patient outcomes and in financial concerns. This nature makes health care different from other commodities (Cropper, 1977). Uncertainty always exists and you continue to remain unsure of whether an investment in health will yield a positive result. Second, individuals may not have the knowledge and necessary information in order to estimate the value and possible outcomes of heath care (McGuire, Henderson, & Mooney, 1988). Third, some preventive activities are based on a choice of lifestyle rather than a rational choice of investment. The above discussion implies that the characteristics of different types of preventive health care are different. Preventive health care covers a wide range of activities, "undertaken by a person believing himself to be healthy, for the purpose of preventing disease or detecting it in an asymptomatic state" (Kasl & Cobb, 1966, p. 246). These activities can be classified in different ways (Gordon, 1987; National Institute on Drug Abuse, 1997). For the purposes of this study, we use a classification of the dichotomy of medical-related preventive health care (MP) and nonmedical healthy lifestyle (Lifestyle). MP includes treatments conducted in hospitals for the purpose of early detection and prevention of disease, including regular blood tests, scans, and immunizations. The utilization of Lifestyle reflects an individual's behavior-based activities, which reduce the probability of illness. These behaviors include a healthy diet, physical activity, and not smoking, and have been studied within the frameworks of psychology (Kirscht, 1983;Prentice-Dunn & Rogers, 1986), and classified as alternative health care in medical studies (Kelner & Wellman, 1997). Since health care includes a wide range of activities, the same factors may have different effects on the utilization of medical and nonmedical healthy lifestyle. In this paper, we will estimate the effects of the same set of determinants on MP and Lifestyle separately. The research strategy of this study is to compare the performance of different theoretical frameworks on the utilization of preventive health care activities. The explanatory power of the Grossman model and that of the behavioral and lifestyle frameworks on MP and Lifestyle can be compared using the results of the estimation. Therefore, the complexity of preventive health-care activities can be analyzed in both rational and irrational frameworks. Data The data sets used in this study are from the Adult and Community Files of the China Health and Nutrition Survey (CHNS) 2006. This publicly available survey, overseen by the Health Sciences Institutional Review Board at the University of North Carolina at Chapel Hill, and primarily funded by the National Institutes of Health (United States), is used to evaluate the effects of economic reforms, as well as ongoing government programs addressing public health and nutrition issues. This survey was conducted by the Carolina Population Center (CPC) at the University of North Carolina beginning in 1989. Using a multistage, random cluster sampling design, the data were collected from 4,446 households, with 11,742 individuals, in nine provinces of China, namely Guangxi, Guizhou, Heilongjiang, Henan, Hubei, Hunan, Jiangsu, Liaoning, and Shandong. The advantages of using the Adult and Community files of the CHNS should be emphasized. The CHNS is a major survey done in China, containing a significant amount of information on demographics, health and health care, and socioeconomic indicators of individuals and communities. The Adult File contains a wide range of information about adults (aged 18 and older), including age, gender, education, urban-rural status, occupational activities, individual eating patterns, nutritional knowledge, daily health-related activities, health status, reproductive health, utilization of health care, medical insurance, and individual and household income. The Community File contains information

regarding the socioeconomic situation of each community (such as educational and income profiles), and the medical conditions of the community (such as the number of medical facilities, the number of beds, and the number of doctors serving at these facilities) though there is no community weight variable that can be used to reflect weighted summary statistics based on the communities of the whole county. A community (village or neighborhood) is a government-designated administrative district, not a natural population cluster. There are 218 communities in the 2006 CNHS. Combining the Adult and Community Files provides a rich source of information for an evaluation of public-health policies. It satisfies the purpose of this study for understanding the social, demographic, and community determinants affecting MP utilization and Lifestyle practice. The Model An individual's decision to use preventive health care can be studied as a choice between using preventive care services or engaging in other competing activities, based upon vectors of the individual's demographics, social-economic characteristics, and his/her socioeconomic environment. To account for the community-level clustering in the CHNS data, a multilevel logit model is used. In constructing this model we assume that individuals will try to maximize their return on investment through their decision-making process. Then, following Rabe-Hesketh and Skrondal (2008, p. 247), the probability of an individual i, living in community j, using preventive health care, given his/her demographic characteristics, social-economic status, and community profile is described as follows: where y ij is binomially distributed with p ij being the probability that the ith individual of the jth community utilized preventive health care; x 1 is a vector of the individual's demographic characteristics, such as age and gender; and x 2 is a vector of community variables, such as number of facilities in the community. The u j is the random error at the community level, assumed to be normally distributed, with a mean of 0 and variance of s u 2 (after taking the communitylevel covariates into account). This allows us to calculate the portion of the variation at the community level: r¼ s u 2 /(s u 2 þs k 2 ), where s k 2 denotes the variance at the individual level, and r is known as the inter-class correlation. A higher value of r indicates a higher proportion of the total variance in utilization of preventive health care, derived from community level differences. This multilevel model permits the estimation of two groups of parameters: fixed and random effects. Fixed effects represent the relationship between the observed individual explanatory variables and the outcome variable (the utilization of preventive health care). In contrast, random effects do not come from the observed individual variables, but from the variations among communities, measured in terms of the value of r. The estimation is conducted using the xtlogit command of STATA 10 software package (Rabe-Hesketh & Skrondal, 2008, p. 248). Four models are estimated. Model 1 is an empty model, measuring only the community-level variance, using community ID without any independent variables. Model 2 includes community variables only. Model 3 includes individual variables only. Finally, Model 4 incorporates both individual and community variables. These models produce consistent estimations of fixed and random effects. To conceptualize the determinants of the utilization of preventive health care in China, the remaining part of this section will discuss the specific measurements of the variables used in these models. Dependent Variables Utilization of Medical Preventive Health Care (MP). This is a binomial variable, defined by whether the respondent used any MP services within four weeks preceding the interview date of the CHNS. This includes health examinations, eye examinations, blood tests, blood-pressure screening, tumor screening, prenatal and postnatal examinations, and any other type of preventative examinations. The value is 1 if the respondent used one of these preventive services, otherwise the value is 0. The nature of these services is different from that of curative health care and from a Lifestyle focus. First, the individual's perception of MP is different from that of curative care, since MP is not as urgent as curative care, while people use curative care when they are sick. On the other hand, the use of preventive care can be delayed, and may yield to other competing consumptions and investments, given the constraints of money, time, and other resources. Second, MP differs from Lifestyle, because the former has to be done in a hospital and/or requires the use of medical equipment. Practice of Nonmedical Healthy Lifestyle (Lifestyle). The term "nonmedical healthy lifestyle" covers a wide range of practices dealing with illness and other forms of health care, often known as alternative health care, including physical manipulation, acupuncture and traditional Chinese medicine and naturopathy, vitamin supplements, and healthy diet (Kelner & Wellman, 1997). Given the availability of the information of the survey data and the preventive focus of this study, we use a narrow definition. This binomial variable is defined as whether the respondent performed certain health-related activities or had a healthy lifestyle. A value of 1 is applied if the person was a nonsmoker, who had not consumed alcohol in the past year, maintained healthy food choices, and performed regular physical activity. Otherwise a value of 0 is applied. Here, healthy food choice is an aggregated index calculated from four Likertscale questions within the survey: fp ¼ score of fruits preference þ score of vegetable preference À score of fast food preference À score of salty snack food preference. A healthy food preference is indicative of a fp value greater than 0. Physical activities included one of the following: walking; Tai Chi practice; or participation in sports such as ping-pong, badminton, tennis, soccer, etc. This variable is labeled nonmedical healthy lifestyle, since all of the included activities are nonmedical in nature, and are related to personal lifestyle choices. These nonmedical behavior-related choices contribute to a large proportion of the population's health status (Mokdad, Marks, Stroup, & Gerberding, 2004). We predict that there will be a significant difference between the results based on the same set of determinants when comparing the two dependent variables: MP and Lifestyle. Independent Variables We use of the same set independent variables to estimate both medical and nonmedical dependent variables for comparative purposes. It may be obvious that individuals need stronger financial status to utilize medical preventive health care than to practice nonmedical healthy lifestyles. For example, people do not necessarily need to be rich or have medical insurance in order to practice healthy lifestyles. However, it is still useful to know how these independent variables affect the utilization of preventive heath care empirically. Age. Age is one of the most important characteristics in predicting the utilization of health care in Andersen's (1995) behavior model. This is due to the predisposition of elderly individuals to degenerative diseases, resulting from "biological imperatives." However, the relationship between age and utilization of preventive care, as is a focus of this study, is complicated in terms of investment in human capital. Young people tend to invest more in human capital (e.g., education and health) versus the elderly, simply because they have a longer term in which to earn a return (Cropper, 1977). This propensity however, may be lowered as a result of two other processes. First, young people have several competing investments in human capital, aside from health (e.g., education and physical capital). Quite often young people have to choose to invest in different dimensions of human capital at the expense of health care. An example of this is evident in that a high rate of return on investment in education may limit or prevent an investment in health. Second, younger people in general have better health status than the elderly and thus they are less concerned about their health, and therefore tend to invest in it less. Although the combined results of these competing factors can be complex, we hypothesize that age is positively associated with the utilization of both MP and Lifestyle. Female. There has been a longstanding belief that women in the developed world are more likely than men to use MP (Kirscht, 1983;Lairson & Swint, 1978). However, the gender gap in employment and education can be an important factor affecting the utilization of health care in developing countries. Women traditionally remain out of the formal labor market and are more likely to be involved in unpaid housework and family responsibilities. For this reason, women are less likely to obtain higher education, and thus have decreased opportunities for workplace medical insurance and lower levels of income to use for health-care expenses in developing countries (Fan & Habibov, 2009). For this dummy variable, male is given the value 0 and female is 1. Given the fact that a significant gender gap still exists in China, we are not sure about impact of being a female on the use of MP. However, we can predict that being a female increases the likelihood of practicing Lifestyle. High Education. Education is the most important "environmental variable" affecting the production of both health and home goods, per the Grossman (1972) model. People with higher levels of education have more knowledge about the importance of utilizing preventive health care and will possess more resources with which to invest in health. For these reasons, people with a higher level of education are more likely to use both MP and Lifestyle than those with lower levels of education. This dummy variable takes the value 1 if the respondent received higher than lower-middle school education, and takes the value 0 if otherwise. Urban Residence (Urban¼1/Rural¼0). The huge urban-rural gap in China, also known as the dual economy, was a result of the well-known Chinese Household Registration System that occurred prior to the reform. It remains in place today, not only in terms of income but also in relation to social services, including medical services. Urban Residence is a dummy variable that takes the value of 1 for an individual living in an urban area and 0 for an individual living in a rural area. Place of residence becomes especially relevant to the use of preventive health care services in China, given the huge rural-urban disparities in health care education, accessibility of social welfare, and the availability of health-care facilities. Medical Insurance (Yes¼1/No¼0). Insurance can be an enabling factor in the utilization of health care, according to the Andersen Behavior Model (1995). However, as discussed earlier, studies on the role of insurance in the utilization of health care in China are mixed (Akin, Dow, & Lance, 2004;Wagstaff et al., 2009;Wong, Tang & Lo, 2007;Zhu, Zhu, & Liu 2008;). We predict that people having health insurance are more likely to use MP, but not necessarily more likely to use Lifestyle. Low Household Income PC. This dummy variable takes the value 1 if the individual's per capita household income is in the lowest quartile (the lowest 25 percent of the households in the sample). In China, individuals in the lowest quartile have limited buying power with which to maintain their daily activities. In our sample, the cut-off annual per capita household income is 2050 yuan (approximately $255 U.S. in 2006). People in this quartile have difficulties meeting their basic need for food. If they do not have insurance coverage, preventive health care is considered a luxury. For them, preventive health care may not be their priority and the service charges for a basic general physical examination may be beyond their monthly living spending. We predict that being in the lowest quartile of household per capita income limits the affordability of both MP and Lifestyle. High Household Income PC. This is a dummy variable defined as 1 if the per capita household income is in the highest quartile. High household income increases the ability to financially manage the costs of preventive health care. We predict that being in the top quartile increases the opportunities for an individual to use preventive health care. Distance to the Nearest Facility. Utilization of preventive health care increases with this supply factor. Having to travel to a health-care facility that is far from an individual's residence can seriously discourage the utilization of health care (Khan, Hotchkiss, Berruti, & Hutchinson, 2006). For the purposes of this study, we use Distance to the Nearest Facility as a supply factor, or in other words, the accessibility of the health-care service. It is

expected that with an increase in distance to the nearest facility, the accessibility of MP would decrease. However, this should not affect Lifestyle. Number of Facilities in the Community. This variable is defined as its name suggests. It is a measurement of the availability of health-care facilities at the community level. The study on the role of the community level supply factor (health-care facilities) is well documented for developing countries (Gwatkin et al., 2007). The lack of health-care facilities in their community forces people to travel a long distance in order to seek health care. We expect that the number of facilities available to the community is positively associated with MP. However, Lifestyle should not be significantly affected by it. Per Capita Income of the Community. This is an indicator of the general level of the social-economic condition of the community. The level of per capita income at the community level is associated with many other factors such as: medical infrastructure, transportation, and the community's general social development level, including health education and knowledge of preventive health care. This variable is defined as per capita income of the community in 1000 RMB yuan. We predict that this variable is positively associated with both MP and Lifestyle. For more details about the variables, please see Table 1. Utilization of MP The estimation results from the multilevel regression for MP are reported in Table 2. The intra-class correlation (r) reported in model 1 (the empty model) shows that a significant proportion (36.93 percent), of the total variation in using MP comes from the community level with a statistical significance at the 0.01 level. This indicates that the variations among communities using MP may be the result of different resources such as financial ability, level of development, and the infrastructure of medical facilities. Results from Model 2 partly confirm the variations noted at the community level. The estimation results indicate that the Number of Facilities in the Community is positively associated with MP, and this relationship is statistically significant at the 0.05 level. The odds ratio for this variable is 1.1892, meaning that for each increase of one facility, the probability of using MP increases by 18.92 percent. However, Community per Capita Income had almost no effect on MP. The combined results from Model 2 suggest that community level variation is mainly caused by supply factors rather than economic development level. Therefore, communities with similar economic conditions could benefit through the provision of resources for MP services. Next we will discuss the individual factors reported in Model 3. First, the age of the individual is positively related to MP at the 0.01 significance level. The odds ratio of Age indicates that for every year an individual's age increases, the likelihood of using MP increases by 1.57 percent. This is in line with the Andersen (1995) model, in which older people tend to be less healthy than younger people, and they tend to use more medical preventive services. Second, the variable Female has a strong positive relationship with MP (with a significance level of 0.01). The odds ratio indicates that the probability of a female using MP is about 61 percent higher than it is for a male. This is in line with previous studies focusing on the background of developed countries (Kirscht, 1983). Third, the result for High Education shows that the possession of high education moderately increases the probability of using medical preventive health care by about 55.75 percent. This relation is statistically significant at the 0.01 level. Education is an important variable in the Grossman model, and our study confirms that people with higher education are more likely to use preventive health care. Fourth, living in an urban area radically increases the prospect of using medical preventive services, by about 102 percent. The high odds ratio for this variable reflects the huge urban-rural gap that exists on all dimensions of the social-economic conditions in contemporary China inherited from the household registration system. This gap is not expected to close for a considerably long period of time due to a few fundamental reasons (Chan & Buckingham, 2008). Notes: Odd ratios are reported, standard errors are in parentheses. Ã p < 0.10, ÃÃ p < 0.05, ÃÃÃ p < 0.01. Fifth, the results of our model estimation indicate that having medical insurance is significantly associated with the utilization of preventive health care at the 0.01 level. Having medical insurance improves the individual's likelihood of using MP by as much as 77.68 percent compared to those having no medical insurance. Sixth, the estimation of the two variables of household income, Low Household Income pc and High Household Income pc, demonstrates some interesting results. People in the lowest quartile, well below the World Bank absolute poverty line, only had a 71.04 percent probability of using preventive health care compared with those in the middle two quartiles, with a statistical significance level of 0.1. In other words, if we can lift a person from the lowest quartile, up to the middle two quartile levels (the reference group) of earnings of the remaining population, his or her likelihood of using medical preventive health services would increase by 40.76 percent ((100-71.04)/71.04). However, the relationship between being in the highest quartile and the utilization of MP is not statistically significant. The results from these two variables demonstrate that having a high income level does not necessarily increase the utilization of preventive health care as much as would be evident through raising the income level of the individual person from the lowest quartile. Finally, the Distance to the Nearest Facility variable is not statistically significant. From a comparison of the results of all four models in Table 2, we can see a moderate change in the value of r from the empty model to Model 4. A comparison of Model 3 and Model 4 (the full model) generally shows consistent and similar results for the variables estimated, as well as for the random effects. This indicates that multi-collinearity is limited. Utilization of Lifestyle The estimation results for Lifestyle presented in Table 3 show that the intraclass correlation (r) is only about 20 percent (Model 1), compared with about 37 percent for MP. This indicates a moderate intercommunity variation among the total variation of Lifestyle. Results from Model 3 also show that the odds ratios for the two community variables are close to 1, and are statistically insignificant. These results demonstrate that the community level variations have less impact on the utilization of Lifestyle. Six individual variables, Female, High Education, Urban Residence, two household income variables, and Distance to the Nearest Facility, are statistically significant, as reported in Model 3 of Table 3. Two individual variables, Age and Medical Insurance, are not related to the use of Lifestyle statistically, as predicated. First, among the significant variables, Female has a very strong positive relationship with the practice of Lifestyle at the 0.01 level. The odds ratio reveals that the probability of a female using these practices is 3.63 times higher than that of a male. This result is expected, since the dependent variable Lifestyle contains a measure of whether the individual smokes or drinks alcohol, as two of the four measured activities, and smokers and alcohol users are predominantly male. For example, male and female smokers present a prevalence ratio of 22 to 1 in China in 2010 (Li, Hsia, & Yang, 2011). Drinking alcohol is deeply rooted in a maledominated tradition, with the ability to drink being considered a feature of being a man. Second, the results indicate that having high education increases the probability of practicing Lifestyle by 21.62 percent. Education is widely believed to be an important factor influencing preventive health care, as is demonstrated in both the Grossman and behavior models. People with higher education not only have more knowledge on the importance of using preventive health care, but they also tend to choose healthier lifestyles. The possession of high education is not only a means to a good job, but also means more knowledge of the consequences surrounding particular lifestyles, including chosen activities. Thus, it provides a way of changing perceptions and encouraging choice of a healthy style of living. Third, living in an urban area increases the odds of practicing Lifestyle services by about 1.2 times more than living in rural area, similar to that for MP. Notes: Odd ratios are reported, standard errors are in parentheses. Ã p < 0.10, ÃÃ p < 0.05, ÃÃÃ p < 0.01. This means that although the practice of Lifestyle is not dependent upon urban medical facilities, urban environments are still important for practicing Lifestyle. Fourth, the estimation of the two variables of household income, Low Household Income pc and High Household Income pc, demonstrates a similar pattern as that for the MP estimation. People in the lowest quartile had a 79.66 percent probability of using preventive health care, with statistical significance at the 0.01 level. It is understandable that people in economic difficulties also have difficulties in maintaining a healthy life style. The relationship between being in the highest quartile and the utilization of MP is statistically significant at the 0.05 level. The results from these two variables demonstrate that household income does matter in Lifestyle practice. Fifth, Distance to the Nearest Facility has about a 20 percent effect on performing Lifestyle, with a statistical significance at the 0.05 level. Since practicing healthy lifestyle does not have a direct relationship with the distance to the facility, the mechanism for this relationship is empirically unclear. Our results show that the practice of Lifestyle is impacted by the following factors: the individual's gender, education, where they reside, and household income. It is, more importantly, about a person's choice, rather than supply factors. From a comparison of the two estimations reported in Tables 2 and 3, we can see that the same set of independent variables have differing impacts upon the two dependent variables. For MP, Age, Female, High Education, Urban, Medical Insurance, Low Household Income pc (though only at the .1 level), and Number of Facilities are significant variables. For Lifestyle, important factors determining lifestyle and consumption preferences also include being female, high education, living in urban areas, and household income; but unlike in the case of MP, distance to nearest facility also proves to be a significant factor, whereas age and medical insurance do not. Thus, in both cases some factors beyond the individual's control, such as economic conditions and availability of facilities, are important. Conclusions and Policy Implications Through an analysis of the China Health and Nutrition Survey 2006, this study estimates the demographic, economic, social, and community determinants of the utilization of medical preventive health care (MP) and nonmedical healthy lifestyle (Lifestyle) in China. The determinants are constructed in an effort to test alternative economic and behavior frameworks. The major findings, as well as their policy implications, from our multilevel logit estimation, are summarized below. First, for the usage of MP, age, gender, education, urban residence and medical insurance are strong predictors. Among these predictors, urban residence has a substantial impact on the likelihood of using medical preventive services. Although the urban-rural gap has been longstanding and will remain a hallmark of the Chinese dual-economic system for a considerable length of time, it is still vital that new initiatives in rural health improve upon these current conditions (Eggleston, Li, Meng, Lindelow, & Wagstaff, 2008;Xu, Zhang, & Zhu 2008). The results indicate that there is a strong correlation between having medical insurance and utilizing preventive health services. This implies that there is a great potential for improving the level of utilization of preventive health service through increasing medical insurance coverage, since more than 50 percent of the participants in the sample did not have any type of medical insurance coverage. Second, the model estimation for the two household income variables demonstrates that having a high income did not improve the utilization level of MP as much as was evident in lifting the individual up from the lowest income quartile. In other words, high income did not provide much of an increase in the usage level of MP, but the lack of income was a huge obstacle for low-income people to overcome. In this sense, poverty reduction is an efficient measure that could significantly improve the utilization

of MP and eventually improve health status for the population. Third, the multilevel model estimation shows that community variation accounted for about one third of the total variation in the utilization of MP. The number of facilities, rather than the community per capita income, is a good community predictor. Thus under the same economic conditions, improvements to the infrastructure of medical facilities can make a significant difference. Fourth, in considering the practice of nonmedical healthy lifestyle, gender, education, urban residence and household income are important predictors. Although these lifestyle-related activities are "nonmedical", they are crucial to the health status of the general population. Cancer, heart diseases, and cerebrovascular disease-the top three causes of death in China in 2000-are all related to diet preferences, smoking, alcohol consumption, and a lack of physical activities. Lung cancer is the number one cause of death in China and smoking is a principal risk factor in lung cancer. Although the public has an increased awareness of the dangers of smoking, reducing smoking is a tough task. China is the largest tobacco producer and consumer in the world and this current pattern of economic development for a higher level of GDP could remain. The Chinese state-owned tobacco companies produce over 1.7 trillion cigarettes annually, which contributed 7.4 percent to the total government revenue in 2003 (Hu et al., 2005). Unless the government yields to health concerns, by changing the current GDP-oriented policy, there will not be a big move in reducing smoking behavior. Fifth, the results of this study indicate that the Grossman framework performs very well for the use of MP. However, a behavior-based framework provides a more reasonable explanation for the practice of Lifestyle. The results of this study support the assertion that preventive health actions appear to represent several independent domains (Kirscht, 1983). Since social-economic factors have differing impacts on preventive health care activities, classifying these activities and choosing proper research frameworks are important in preventive health-care research. Further studies may consider classifications and theoretical frameworks other than the MP and Lifestyle dichotomy used in this study. It is essential to do so, since this would allow health-care policymakers to apply different strategies, targeting specific preventive activities, in order to improve the health status of their population. Finally, the previously discussed results of our study indicate that the utilization of MP in China remains dependent upon the individual's socialeconomic conditions. However, MP should not simply rely upon the individual's ability and effort, since some infectious diseases could eventually become a public issue if no effective preventive care plan is in place. For example, currently there are about 120 million individuals infected with the hepatitis B virus (HBV) in China. This accounts for about 9 percent of the Chinese population and one-third of the total number of the HBV infections in the world (Custer et al., 2004;Liu & Fan, 2007). High-quality HBV vaccines are still too costly for low-income Chinese people. If the government were to provide this vaccination to all citizens, especially those without health insurance, a significant improvement could be achieved. Although the results of this study clearly indicate that the utilization of Lifestyle is partly a matter of an individual's personal characteristics such as education and gender, the government could still bring about significant change. MP and Lifestyle are different and not wholly substitutable for each other, thus they do not compete for the same resources of the government. For improvement of Lifestyle, the government could refrain from seeking profits from the tobacco and alcohol industry, and set new regulations for cigarette and liquor production and consumption, while smoking in public could be prohibited. In conclusion, all of these suggestions are based upon the assumption that the government would forego GDP growth as their sole objective. One obvious limitation of this study is that the available information on medical preventive health care only accounts for limited items of preventive health care within four weeks preceding the interview date of the CHNS, which does not capture the comprehensive nature of preventive health care. In addition, it should be noted that the results of this study reflect the situation of health care before the health-care reform beginning in 2009. This reform aims to provide nearly universal health-care coverage by 2020. With the progress of the reform, more people will have medical insurance, and their health-related behaviors will change. The results of this study might be useful for new policy arrangements in providing information on behaviors in the case that most rural people did not have medical insurance. Preventive health care could be a more important component in the new health-care system if the government takes more responsibilities in the reformed system. Finally, it would be interesting to make comparisons on the changes of health-care behaviors during the reform in further study. Those comparisons would provide useful information for improving the efficiency of the new medical system. Lida Fan is associate professor in the School of Social Work at Lakehead University in Thunder Bay, Ontario, Canada. Jianye Liu is associate professor in the Department of Sociology at Lakehead University, Thunder Bay, Ontario, Canada. Nazim N. Habibov is associate professor in the School of Social Work at the University of Windsor in Windsor, Ontario, Canada. Two-stage closed sinus lift: a new surgical technique for maxillary sinus floor augmentation Bone tissue atrophy may constitute a relative contraindication for implantation. The methods used in reconstruction of the alveolar ridge within the lateral section of the maxilla have been well known but not perfect. Presentation of the two-stage, closed sinus lift technique as well as efficacy evaluation of reconstruction of the alveolar ridge in the maxilla within its vertical dimension with the use of this technique. The total procedure was performed in 26 out of 28 patients qualified for the study. The height of the alveolar ridge at the site of the planned implantation was no <3 mm, the width of the ridge was no <5 mm. During the treatment stage 1 the sinus lift was performed for the first time. The created hollow was filled with allogeneic granulate. After 3–6 months stage 2 was performed consisting in another sinus lift with simultaneous implantation. The treatment was completed with prosthetic restoration after 6 months of osteointegration. In 24 out of 26 cases stage 1 was completed with the average ridge height of 7.2 mm. In stage 2, simultaneously with the second sinus lift, 26 implants were placed and no cases of sinusitis were found. In the follow-up period none of the implants were lost. The presented method is efficient and combines the benefits of the open technique—allowing treatment in cases of larger reduction of the vertical dimension and the closed technique—as it does not require opening of the maxillary sinus. Introduction The height of the alveolar ridge in the maxilla is the resultant of masticatory forces transferred by the periodontal ligament system to the bone and pneumatisation of maxillary sinuses beginning with eruption of the third molars (Misch 1999). Bone atrophy in the maxilla is a physiological process, which accelerates in case of tooth extractions (Sornié t al. 2005). In females higher post-extraction bone resorption is observed compared to males (Saglam 2002), which may be related to density of the bone tissue and hormonal balance of the body. The research proves that more severe atrophy may be expected when molars are extracted rather than premolars (Wehrbein and Diedrich 1992) and when a greater number of adjacent teeth are extracted (Sharan and Madjar 2008). Prolonged healing time resulting from numerous and traumatic extractions also promotes more severe atrophy of bone tissue (Sharan and Madjar 2008). Unskilful tooth extraction may be associated with the damage of the thin lamina dividing the maxillary sinus and the alveolus as well as rupture of the sinus-lining membrane, which hence exacerbates the extraction-related, physiological atrophy of the ridge hard tissue. Insufficient vertical dimension of the alveolar ridge is a relative contraindication for implantation. Owing to techniques of alveolar ridge reconstruction introduced in surgery in 1970s the optimal size of the ridge bone may be restored (Sorní et al. 2005;Schwartz-Arad et al. 2004) and implantation may be successfully performed (Levin et al. 2004). The first to be described was the open technique, which allowed performing the procedure in patients with a ridge of at least 4 mm (Balaji 2013); however, successful attempts were made in more severe reduction of the vertical dimension (Chaushu et al. 2009). If the atrophy of the vertical dimension of the alveolar ridge is less severe, closed techniques are used. Their advantages include lower invasivity and single-stage sinus lift combined with implant embedment. However, in view of limited visibility within the operative field and greater initial dimension of the alveolar ridge (7 mm), the planned range of augmentation must be smaller (Pal et al. 2012). Limitations of the techniques mentioned above inspire clinicians to seek new methods of ridge reconstruction in the lateral segment of the maxilla before implantation, which would allow combination of the advantages of open sinus lift with the low risk of the closed sinus lift. Objective of the study Presentation of two-stage closed sinus lift and evaluation of this new technique in maxillary alveolar ridge reconstruction within its vertical dimension. Materials and methods The technique of two-stage sinus lift was used in 28 subjects aged 29-66 (mean age: 44) who had reported to have a dental defect restored with implant insertion. Before treatment computed tomography of the maxilla ( Fig. 1) was performed in all the patients. Inclusion criteria comprised no inflammation within the sinus on the side of the dental defect, minimum height of the alveolar ridge within the implantation area: 3 mm, minimum width of the ridge: 5 mm (thus no necessity for widening procedure), lack of general diseases. Stage 1 Under local anaesthesia with 4 % Ubistesin forte an incision was made at the top of the alveolar ridge from the palatal side within the toothless gap. The cut was extended perpendicularly to the ridge, across periodontium of the teeth adjacent to the gap and further on to the oral vestibule. After the mucoperiosteal flap was detached normal bone tissue was found. With a spot drill the optimal Fig. 1 CT scan before grafting place was marked for future intraosseous implantation. Then, with a guide drill a hollow was made 1 mm shallower than the height of the alveolar ridge within this area previously calculated based on CT. Subsequently the sinus floor was augmented for the first time with a chisel kit for sinus lifts and a surgical hammer. Owing to the concave shape of the upper part of the chisel, bone shavings were obtained and the sinus floor was shifted inwards. Additionally, the other working part of the chisel of slightly conical shape caused concentration of bone tissue within the lateral walls of the tunnel. In order to reduce the force needed to push the bone with next chisels, the outer lamina dura was removed with an implant drill 1 number bigger than the next chisel. The last chisel used to elevate the sinus floor was one size wider than the expected diameter of the implant to be embedded. Maxillary sinus floor elevation was done in stages with the use of subsequent chisels. In order to reduce the risk of rupture of the Schneiderian membrane or bone chip dislocation, chiselling was performed very slowly and carefully so dilation of the bone canal progressed gradually. Continuity of the mucous membrane was verified on numerous occasions intraoperatively with a sinus probe ended with a ball. The elevated sinus floor was fixed and filled with the patient's own bone shifted from the alveolar ridge, whereas the bone void of conical shape with a cut apex was filled with frozen, radiation sterilised allogeneic bone obtained from the Tissue Bank (Fig. 2). The procedure was finished by extending the mucoperiosteal flap obtained with cut peritoneum which was then repositioned and fixed with mattress sutures. Postoperatively an antibiotic, anaelgesic, and anti-oedemic treatment was administered and mouth rinsing with an antiseptic and surgical site protection was recommended. The treatment stage 1 was finished when the sutures were removed following 2 and 6 weeks after the procedure healing of the tissues was investigated. Temporary prosthetic restorations were also examined for pressure exerted on the surgical sites. Stage 2 After

3-6 months, before the next stage of surgical treatment, when no inflammation was found within the adjacent tissues and the sinus, a follow-up CT was performed (Fig. 3). Under local anaesthesia an incision of the mucous membrane was done just like in stage 1. After the flap was detached and bone structure was evaluated (Fig. 4), the procedure of the alveolar ridge drilling and gradual sinus floor elevation with a chisel kit was repeated. The only difference was that the last chisel to be used was the same diameter as the planned implant. The resulting bone void was filled with the embedded implant which after obtaining primary stability and covering with the flap was sutured without drainage. The patient received the same recommendations as in stage 1. The sutures were removed after 2 weeks. Treatment stage 2 was finished with a 6-month osteointegration period. Results The two-stage sinus lift was performed in 28 patients. Two subjects initially qualified for the procedure did not report for continuation of treatment after stage 1. In the other two cases stage 1 was unsuccessful (the required increase of the ridge height was not obtained), most probably due to a rupture of the Schneiderian membrane and partial resorption of the graft. In those cases standard closed sinus lift was performed and a shorter implant was embedded during stage 2. Twenty-six implants BIOMET 3I were embedded in regenerated bone tissue. In three subjects partial separation of the wound edges was found, which healed by granulation. As the incision line did not cross the surgical site, the graft was not revealed and no major complications were caused. In one case pressure of the prosthetic restoration exerted on the surgical site was observed which was corrected at a follow-up visit after 2 weeks of the procedure. No symptoms of inflammation in the sinuses were found in any of the cases, including those with unsuccessful stage 1. The mean, maximum, and minimum primary height of the alveolar process, the growth of bone tissue after stage 1 measured in CT as well as the height of the ridge before implantation were presented in Table 1. The augmentation areas and length of the implants are presented in Table 2. During the follow-up period, the visual of which is presented in Fig. 6, no loss of stability was found in any of the implants. In two cases the prosthetic crown had been partially loosened, which was easily corrected by tightening with a torque wrench. Discussion The choice of a technique for bone augmentation depends mostly on the initial height of the ridge at the site of future implantation. If the thickness of the bone does not provide primary stability of the implant and is \5 mm (Valentini et al. 2000), the method providing good and predictable outcomes is the procedure of open sinus lift. However, it constitutes a burden for the patient as it interferes with the sinus and bears a higher risk of infection Balaji 2013), in smokers in particular (Barone et al. 2006). The advantage of this technique is that it allows restoration of a severely reduced ridge (Balaji 2013). However, studies proved that it was better to insert a shorter implant and perform a closed sinus lift than risk an open sinus lift to embed a longer implant (Esposito et al. 2010). Authors of this article had similar experiences; therefore in order to reduce the failure risk the closed sinus lift was performed twice with the doctor's own modification, which despite unfavourable conditions for simultaneous implant embedment, provided good outcomes in two stages. There have been also reports on successful ridge reconstruction with 1 mm of the patient's own bone (Winter and Pollack 2002). The most important disadvantage of the closed sinus lift is the risk of a rupture of the sinus lining membrane; therefore some clinicians use an inflatable device or fill the void with augmentation material before the wall of the sinus is forced into (Stelzle and Benner 2011). Studies by Hernandez-Alfaro demonstrated that implant survival depends on the size of perforation (Hernández-Alfaro et al. 2008). In the discussed study on one hand the two-stage sinus lift prevented excessive straining of the Schneiderian membrane and allowed good primary stability of the implant during its embedment combined with the second sinus lift. On the other hand it allowed regeneration of the bone tissue to a greater extent (the mean of 3.94 mm in stage 1), which in the twostage procedure provided better outcomes than the traditional method suggested by Summers (Summers 1994). Despite multiple verifications of continuity of the sinus-lining membrane as well as better visibility within the operative field, we failed to avoid the membrane rupture in two cases. There have been also reports of no influence of the membrane rupture on the success of the closed sinus lift (Ardekian et al. 2006;Karabuda et al. 2006). An autograft is commonly believed to be the best material for reconstruction. However, this choice is associated with the necessity to harvest the graft from the area of the mentum or the retromolar pad or extraoral locations, when a larger amount of the graft is needed. Another surgical site is associated with increased number of possible complications (Guarnieri et al. 2006;Ewers 2005), which makes the patients dissatisfied. There have been reports on lack of advantage of autografts over allogeneic materials (Valentini et al. 2000;Del Fabbro et al. 2004), and even reports on high susceptibility of autografts to resorption (Wallace 2003), reaching as much as 49.5 % after 6 months of the procedure (Ewers 2005). Similar implant survival following the use of bone substitutes and autografts inspires clinicians to choose the former (Valentini et al. 2000), and osteoinductive and osteoconductive properties of allografts (Hallman et al. 2005) lead to the choice of the material in this study. The latest research reported that the primary stability of the implant depended on the diameter of the implant rather than its length (Maiorana et al. 2014). Good outcomes during a 10-year follow-up were obtained when using 8 mm implants (Mangano et al. 2014). Alternatively, short implants may be used of efficacy comparable to those of standard length (Al-Hashedi et al. 2014), as it happened in two cases when the procedure was modified due to the failure of stage 1. However, the expected time of exploitation of those implants remains unknown (Esposito et al. 2010). The studies reported that in the lateral segment of the maxilla the preferred length of the implant amounted to 6-10 mm (Tutak et al. 2013). In case of severe atrophy an open sinus lift was necessary to embed implants of this size. The suggested technique seemed to be worth considering as there was no interference with the lumen of the sinus. A disadvantage related to the procedure described above is the longer time between commencement of reconstruction and delivery of the prosthetic restoration compared to a standard sinus lift procedure. The minimum of 3 months is needed after stage 1 for the graft to reorganise and replace it with the patient's own bone in order to obtain normal primary stability of the implant during stage 2. However, when using the open technique this period is longer, reaching nearly 1.5 years as graft healing and reorganisation in these cases takes 6-9 months, i.e. it is longer than the time suggested in this paper. Not before this period is finished the implants can be embedded and they may not be weighed down with prostheses until another 6 months pass (Davarpanah et al. 2003). Conclusions The discussed technique of the two-stage sinus lift was an efficient method for reconstruction of atrophied alveolar ridge with the initial ridge height of 3 mm with no necessary opening of the maxillary sinus. The technique had the advantages of the closed sinus lift, i.e. lower risk of infection within the sinus and the advantages of the open technique, i.e. more extended reconstruction of the ridge when its vertical dimension is severely reduced. It allowed ridge expansion up to approximately 5 mm only in stage 1. A smaller initial height of the ridge (3 mm) compared to the conventional closed method, provided good control over the procedure owing to better visibility within the operative field. Two-stage, delayed surgical treatment extended the time of the entire dental defect restoration procedure up to 13 months, which must be clearly explained to the patient before the treatment is started. Employing drills for removal of the lamina dura of the alveolar ridge before the chisels are used for dilating the bone reduced negative sensations of the patient during the sinus lift and did not provoke negative disposition of the patient towards subsequent treatment stages. Efficacy of Conventional, Immune-Botanical and Egg Yolk Mixture on the Management of Diseases and Insect Pests of Tomato Solanum lycopersicum L. in Uganda Vegetables are important for nutrition and health, however, production in sub-Saharan Africa is low, partly due to disease and pest damage. Three integrated pest management (IPM) packages: Conventional, Immune-Botanical and Egg yolk mixture were evaluated for control of tomato diseases and insect pests. During 2017A and 2017B seasons, there were no significant differences in severity of Bacterial wilt (Xanthomonas campestris pv solanacearum), bacterial spot (Xanthomonas campestris pv. Vesicatoria), early blight (Alternaria solani), late blight (Phytophthora infestans) and Tomato yellow leaf curl virus disease (TYLCV) as influenced by the interactive effect of differences in variety and IPM package. Plots treated with conventional package had lower severity scores for TYLCV. Variety MT 56 and Pink-top were the best performers. There were significant differences in insect pest damage due to cutworms and Thrips in 2017A, and Aphids and Thrips in 2017B due to differences in variety and IPM package. Egg yolk significantly reduced thrips damage among different tomato varieties while conventional method significantly reduced aphid damage. Conventional package induced the highest fruit yield for MT 56 in 2017A and Pink-top in season 2017B. Both conventional and Egg yolk methods were cost effective and are therefore recommended for control of tomato insect pests and diseases. Introduction Tomato (Solanum lycopersicum L.) is one of the most important vegetable crops cultivated worldwide [1]. However, productivity in sub Saharan Africa particularly Uganda is among the world's lowest. In Uganda, tomatoes are among the most important and prominent horticultural crops grown for both home consumption and the market [2]. About three million households in Uganda consume tomato in their sauce at every meal [3]. Recently, tomato production has been emphasized as a source of food security and income in Uganda [3]. It is a top priority for production, as the main income crop compared to other vegetables [4]. In Uganda, farmers harvest 1.5 to 14 t ha -1 as compared to the world average of 27.5 t ha -1 [5]. Production has intensified over the years with the introduction of high yielding varieties such as Asira F1 and Tengeru 2010; however, yields continue to be low due to several production constraints such as insect pests, diseases, and other environmental factors [6]. The major economically important insect pest species for tomato include whitefly (Bemisia tabaci Gennadius), leaf miners (Liriomyza sp.), thrips, (Thrips tabaci Lindeman), cotton aphids, (Aphis gossypii Glover), tomato fruit worm (Helicoverpa armigera Hubner. Major diseases include bacterial wilt (Ralstonia solanacearum), early blight (Alternaria solani), Bacterial spot (Xanthomonas campestris pv. vesicatoria), Tomato yellow leaf curl virus and late blight (Phytophthora infestans) [6,7]. Tomato farmers in Uganda rely entirely on the use of pesticides to manage insect pests and diseases. However, the high susceptibility of tomato cultivars to insect pests and diseases has caused farmers to obtain low yields in spite of the increased production cost [8]. According to Kagezi et al. [9], fresh tomato yield losses attributed to thrips are as high as 23.7% without the use of pesticides [10]. Even though insecticides have proven to be highly effective in protecting vegetable crops under extreme pressure from insect pests [11], the indiscriminate and widespread use of synthetic insecticides in vegetable cultivation usually results in development of insecticide resistance [12]. On the other hand, it has been established that farmers' limited knowledge on the appropriateness of pesticides to use, timely application, and the quantity to apply have led to low yield and undesirable accumulation of pesticides in food. Some pesticides are hormone disruptors, cancer-causing agents or neurotoxins that can have harmful effects on

the brain and on the growth and development of babies. Little is known about how these pesticides, which we can't smell, see or taste on our produce, act in combination. Alternative earth-friendly solutions such as cooking oil+egg yolk mixture is urgently needed. Because of the critical role they play in vegetable production, there is a need to evaluate some of the most common pesticides used by farmers as well as test new ones on the market, together with non-harmful concoctions. This will generate useful information for effective management of diseases and insect pests for increased yield of tomato in Uganda. The objectives of this study were to (i) determine the efficacy of three Integrated pest management (IPM) packages i.e., Conventional (Cypermethrin+Mancozeb), Immune-Botanical (Nimbecidine +Bacterimycin) and Egg yolk mixture (Egg yolk+Sunflower cooking Advances in Crop Science and Technology oil) against diseases and insect pests of tomato, and (ii) determine the effects of these IPM packages on the yield of tomato. Materials and Methods The study was conducted at the Horticulture and Oil palm program of the National Crops Resources Research Institute located at Namulonge, Wakiso district, 19 km north of Kampala. Mean daily temperatures were 28.5°C maximum, and 13.0°C minimum. Namulonge lies at an altitude of 1150 meters above sea level, with a bimodal rainfall (1196 mm annually) ( Field layout Four tomato varieties were used for this trial namely: MT 56, Pinktop, Tengeru 97 and Ten-ten. Pink-top and Ten-ten are introductions from South Korea; Tengeru 97 was obtained from the local market but bred by world vegetable center (AVRDC). Variety MT56 was developed by Makerere University (Uganda) and Iowa State University (USA). These varieties were chosen because of their high tolerance to major pests and diseases of tomato in Uganda as confirmed from an adaptability study done earlier. The seeds were nursed, and seedlings transplanted at a spacing of 0.9 m × 0.35 m on ridges 30 cm high. There were two rows of 18 plants in each treatment plot. The experimental field was laid in a randomized complete block design with two factors (Variety and IPM package) in three blocks. Each treatment plot measured 3 m × 3 m, with 0.6 m alley between treatment plot and 1.5 m alley between blocks. Fertilizer (200 g of poultry manure per planting hole) was applied once at planting. Weeds were controlled, and watering was done when necessary. IPM treatments and their application The treatments used were: (i) Conventional, containing a mixture of Cypermethrin-40 mL+Mancozeb-50 g per 20 liters of water, (ii) Immune-botanical: Nimbecidine-90 ml+Bacterimycin-10 g per 20 liters of water and (iii) Egg Yolk mixture: 1 local chicken Egg Yolk +Sunflower cooking oil-60 mL per 20 liters of water. A control plot (water application only) was also maintained. Application of treatments was done using separate knapsack sprayers (CP 20) at weekly intervals, starting four weeks after transplanting, and continued up to end of harvesting. Disease severity and Insect pest's damage assessment A sampling of insect pests began four weeks after transplanting before treatments were applied. Among the 18 plants within each plot, ten were randomly selected and tagged for subsequent sampling and assessment of damage and disease severity. Pest damage assessment was done as according to Nagrare et al. [13] for Aphids, Lopez et al. [14] for leaf miners [15] for thrips, Bardner and Fletcher [16] for cutworms and, whiteflies damage. Disease assessment was as according to different scale: bacterial wilt [17] and bacterial spot [18], early blight [19], late blight [20] and Tomato yellow leaf curl virus disease [21]. Estimation of fruit yield Yield (fruit weight) was taken from the ten tagged plants. Tomato fruits were harvested every three or four days when they reached maturity and then weighed in the field using a Switzerland-made Metler Toledo PB302 electronic weighing scale. The cumulative results obtained for each treatment were then extrapolated to kilograms per hectare (kg ha -1 ). This yield was referred to as harvest yield. Furthermore, the fruits collected were divided into marketable and non-marketable (fruits with blemishes or injuries caused by insect and pathogen) yield. Data Analysis Disease severity and insect pest damage data were entered in an excel sheet before subjecting it to Analysis of Variance (ANOVA) using Genstat statistical software 4 th Edition (2012). Treatment means were separated using Tukey at 5% probability. Diseases in season 1 2017A and season 2 2017B For season 2017A, no significant differences (P>0.05) in the interactive effect of variety and treatment on the severity of Bacterial wilt, Early blight, Late blight, bacterial spot, and Tomato yellow leaf curl virus disease among the different varieties was observed. Significant difference (P <0.05) was observed in the severity of TYLCV among the treated plots. The control performed better than the other treatments with lower severity scores for TYLCV (Figure 1). Tomato variety MT 56 was the best performer with no disease at all ( Table 2). In season 2017B, the interactive effect of variety and treatment followed a similar trend as in season 2017A. Insect pests damage in seasons 2017A and 2017B There were no significant differences (P>0.05) observed for in the interactive effect of variety and treatment except for the level of damage caused by cutworms and thrips (Table 3). Egg yolk mixture induced lower cutworm damage among the different varieties ( Figure 2). Ten-ten variety exhibited lower levels of thrip damage with the use of egg yolk mixture (Figure 3). Pest damage in season 2017B followed the same trend as in season 2017B except for damage caused by Aphids which was found to be significantly different as influenced by the interactive effect of variety and IPM package. There was less pest damage on Pink-top variety and conventional package influence the greatest reduction in Aphid damage in season 2017B (Figure 4). Fruit yield as affected by various treatments in season 2017A and 2017B The interactive effect of variety and treatment did not significantly (P>0.05) influence the total and marketable yield in season 2017A however conventional package influenced higher yields in 2017A among the different tomato varieties ( In the second season (2017B) there was a significant (P<0.05) interactive effect between variety and treatment for the total fruit yield. Plots treated with conventional package yield more than the rest while Pink-top was the best yielder under all treatment expects the control ( Figure 5). The interactive effect of variety and treatment on marketable yield was found not to be significant (P>0.05) however plots treated with conventional package had still higher marketable yield compared to the rest of the treatment (Table 4). Overall the yields in season 2017A were far higher than those obtained in season 2017B (Table 4). Cost estimations for pesticide usage From the cost estimation of the amount of pesticides needed to spray a hectare, the cost for use of immune-botanical package was almost twice the cost of using conventional and egg yolk mixture (Table 5). Discussion Tomato production in Uganda is an important enterprise most especially for the smallholder farmers. It has been shown that farmers choose different tomato varieties to grow depending on the production potential, market demand, regional adaptability, disease resistance and the end use of the fruits [22]. However, the major constraint they face is the crop pests and diseases which may require integrated pest and disease management options [23]. In Uganda, the majority of farmers rely on the use of pesticides to manage insect pests and diseases in order to increase yield. However, the effectiveness of these pesticides depends on several factors namely their stability, physicochemical properties, the nature of the medium into which they are applied, the occurring organisms in the soil as well as the prevailing climatic conditions [24][25][26]. Generally, the severity scores of the different diseases observed within this trial were low. Bacterimycin, Mancozeb and egg yolk mixture were incorporated to control the diseases. There was no significant difference in between these treatments. Bacterimycin consists of binitro-dibromo propane 2-3 diol, an immune modulator that imparts resistance to plants against bacterial diseases such as canker in tomato and other vegetables. Furthermore, it was reported to inhibit the growth of Verticillium lecanii but accelerate the growth of Trichoderma viride [27]. Bacterimycin exhibits a unique mode of action which mimics the natural systemic activated resistance (SAR) response found in most plant species [28]. On the other hand, egg yolk destroys mycelia wall, distributes respiration and lipid metabolism of insects as well as repels insects and prevents fungal spore from germinating on the plant surface [29]. However, for this study, their effect on the insect pests was limited. There was a significant difference in the level of damage caused by thrips among the treatments. The best treatment was conventional method containing Cypermethrin, a pyrethroid that has a greater knockdown effect and great photostability capability. Following in strengthen was Nimbecidine a neem-oil-based botanical insecticide containing Azadirachtin and other limonoids including Meliantriol, Salannin, Nimbin and a host of other terpenoids in the ratio as it occurs naturally in Neem. Azadirachtin has several effects on insect pests as it induces anti-feeding, regulates insect growth and causes sterility [30]. Due to seasonal variations, the effectiveness of the IPM packages varied significantly. In season 2017A there was less rain and temperatures were high. These climatic conditions enhanced the pest population and the severity of the different diseases in presence of the host plant since the pest population and inoculum build up was very fast [31]. However, the yield was not compromised since intensive watering was done. In the second season (2017B) there was enough rainfall in the month of September and November that favored the multiplication of pathogen spores thus more disease and insect damage on the plants resulting into lower yields as was the case observed by Aloysius et al. [32] in Ghana. The different IPM packages significantly influenced the total yield among the different tomato varieties. Conventional package followed by egg yolk mixture influenced higher yields within MT56 and Pinktop varieties. Due to variation in climate within the two seasons, significant variation in yield was realized in the two seasons and this was attributed to changes in several climatic and soil factors: temperature, rainfall pattern humidity and soil fertility respectively [31,33]. Conclusion The study showed that Bacterial wilt, Xanthomonas campestris pv solanacearum, Early blight, Alternaria solani Late blight, Phytophthora infestans and bacterial spot, Xanthomonas campestris pv. vesicatoria and Tomato yellow leaf curl virus disease were the major diseases, while aphids, Aphis gossypii (Glover), Whiteflies, Bemisia tabaci (Gennadius), thrips, Thrips tabaci Lindeman leaf miners, Liriomyza sp, and the tomato cutworm, Agrotis spp were the most important insect pests observed to affect tomato in this study area (Namulonge). There was varying effect of the three IPM package on the severities of the different diseases found in tomato. None of the IPM packages had better control of the observed diseases. Tomato variety MT 56 was the best performer with almost no disease at all. Variety and IPM package significantly influenced the level of pest damage caused by cutworm and thrip damage. Egg yolk mixture had the greatest influence on reducing damage caused by tomato pests. MT 56 and Pink-top variety had the highest yield as well as the lowest level of damage by both pests and diseases. Conventional package influenced higher yields in among the different varieties. Variety difference has a significant influence on the level of damage and severity of disease expressed by the plant. From this study, we can conclusively recommend Egg yolk mixture for effective control of tomato insect pests and conventional method (Cypermethrin+Mancozeb) for managing tomato diseases. Both methods are cost effective as compared to the use of the immunebotanical package. Oxygen Tension Modulates Differentiation and Primary Macrophage Functions in the Human Monocytic THP-1 Cell Line The human THP-1 cell line is widely used as an in vitro model system for studying macrophage differentiation and function. Conventional culture conditions for these cells consist of ambient oxygen pressure (∼20% v/v) and medium supplemented with the thiol 2-mercaptoethanol (2-ME) and serum. In consideration of the redox activities of O2 and 2-ME, and the extensive experimental evidence supporting a role for reactive oxygen species (ROS) in the differentiation and function of macrophages, we addressed the question of whether culturing THP-1 cells under a more physiologically relevant oxygen tension (5% O2) in the absence of 2-ME and serum would alter

THP-1 cell physiology. Comparisons of cultures maintained in 18% O2 versus 5% O2 indicated that reducing oxygen tension had no effect on the proliferation of undifferentiated THP-1 cells. However, decreasing the oxygen tension to 5% O2 significantly increased the rate of phorbol ester-induced differentiation of THP-1 cells into macrophage-like cells as well as the metabolic activity of both undifferentiated and PMA-differentiated THP-1 cells. Removal of both 2-ME and serum from the medium decreased the proliferation of undifferentiated THP-1 cells but increased metabolic activity and the rate of differentiation under either oxygen tension. In differentiated THP-1 cells, lowering the oxygen tension to 5% O2 decreased phagocytic activity, the constitutive release of β-hexosaminidase and LPS-induced NF-κB activation but enhanced LPS-stimulated release of cytokines. Collectively, these data demonstrate that oxygen tension influences THP-1 cell differentiation and primary macrophage functions, and suggest that culturing these cells under tightly regulated oxygen tension in the absence of exogenous reducing agent and serum is likely to provide a physiologically relevant baseline from which to study the role of the local redox environment in regulating THP-1 cell physiology. Introduction While it is widely accepted that immortalized cell lines do not exactly replicate primary human cells, cell lines can be extremely powerful experimental models and are generally more widely accessible to the research community than primary human cells. However, there is increasing awareness that cell culture conditions can significantly influence cellular differentiation and function in vitro, and thus, it is extremely important to determine whether changing specific cell culture parameters influences the fidelity by which cell lines replicate the functions performed by primary cell types. The THP-1 cell line was originally derived from human monocytes approximately 30 years ago [1], and has becoming a widely used in vitro model system for studying the differentiation, physiology and pharmacology of monocytes and macrophages. Like most commonly used cell lines, THP-1 cells are typically maintained in culture at atmospheric oxygen tension ((18-21% O 2 v/v) in medium supplemented with the reducing agent 2mercaptoethanol (2-ME) and serum. While cells in certain microenvironments, such as the alveoli of the mammalian lung, may encounter oxygen tensions approaching atmospheric levels, normoxic levels in most mammalian tissues range from 3 to 12% O 2 (v/v) [2]. Hyperoxia increases intracellular levels of reactive oxygen species (ROS) [3] and, thus, conventional culture conditions may predispose cells to oxidative stress. The supplementation of culture medium with 2-ME and serum likely provides some protection against the oxidative stress generated in cells cultured under atmospheric oxygen tension. Maintaining intracellular reserves of reduced glutathione (GSH) is critical to maintaining intracellular redox homeostasis [4], and as a reducing agent, 2-ME can facilitate the maintenance of reduced levels of thiol-containing proteins and peptides. 2-ME was originally added to media used to culture murine lymphocytes to increase intracellular levels of reduced glutathione and thereby enhance cellular functions [5]; however, ME does not enter the cells freely but does increase uptake of Cys which may result in increased GSH synthesis. This practice has since been adopted and recommended for culturing diverse cell types derived from multiple species, including human THP-1 cells, with little experimental evidence to support its value in enhancing cell viability and/or cell-specific functions. Given the influence of ambient oxygen tension on redox reactions, and the thiol-reducing activity of 2-ME, it seems likely that changing these culture parameters will influence the redox balance in the cell. This in turn is likely to have significant impacts on cellular functions since intracellular ROS levels are tightly regulated not only to prevent oxidative stress-induced cell damage, but also because ROS are crucial signaling molecules in energy production, phagocytosis [6], and cellular differentiation [7]. Moreover, there is evidence that some of the same transcription factors that are activated by oxidative stress, such as NF-kB and AP-1, are also involved in mediating the effects of ROS on other cellular functions, such as cytokine production [8]. Consistent with the proposed role of ROS in normal cell physiology, changes in oxygen tension have been shown to modulate cell proliferation [9], maturation [10], differentiation [2] and cytokine production [11][12][13]. For example, studies have demonstrated that the exceptionally low oxygen tensions associated with the tumor environment are causally linked to upregulation of transcription factors that enhance cytokine production in tumor-associated macrophages [14]. The goal of this study was to determine whether culture conditions, specifically reducing agents and oxygen tension, have a significant influence on the macrophage functions of THP-1 cells. The answer to this question has important implications with respect to optimizing THP-1 cell culture to better replicate primary human macrophages, and for interpreting results obtained with THP-1 cells across different laboratories. In this study, we compared the effects of 5% O 2 , representing a physiologic normoxic level, and 18% O 2 , representing the atmospheric hyperoxic levels used in conventional tissue culture, on the proliferation, differentiation and primary macrophage functions of THP-1 cells grown with and without 2-ME and serum. Our studies indicate that altering the oxygen tension significantly influences THP-1 cell physiology, whereas omitting 2-ME and serum from the culture medium has minimal impact. Results In all experiments, undifferentiated THP-1 cells were synchronized by serum starvation for 48 h prior to exposing cells to varying oxygen tension, 2-ME and serum. Synchronization aligns all cells at the same point in the cell cycle prior to initiating experimental manipulations. While synchronized cell populations are not common in vivo, synchronization of proliferating cell lines is a widely used experimental strategy to minimize variability in the experimental readout since cells in different stages of the cell cycle are well known to be differentially susceptible to and/or respond differently to environmental cues. Many experimental approaches have been described for synchronizing cells at specific phases of the cell cycle [15], and several common methods involve pharmacological agents acting at various points throughout the cell cycle [16,17]. However, because of adverse cellular perturbations that can result from exposure to these pharmacological agents [18], we chose to use serum deprivation as the method for synchronizing undifferentiated THP-1 cells. Oxygen Tension does not Affect Proliferation of Undifferentiated THP-1 cells A principal characteristic of undifferentiated THP-1 cells is their ability to proliferate in culture. Thus, we initially determined the influence of oxygen tension, 2-ME and serum on the proliferation of undifferentiated THP-1 cells. Specifically, we determined the percent increase in cell number at 24 and 48 h after synchronization in cultures maintained under 18% (hyperoxic) versus 5% (normoxic) O 2 in the absence or presence of 2-ME and serum (Fig. 1). Across all treatment groups, the percent increase in cell number was greater at 48 h (Fig. 1B) relative to 24 h (Fig. 1A), suggesting that none of the treatments were overtly toxic to undifferentiated THP-1 cells. At either 24 or 48 h post-synchronization, there was no significant difference in the percent increase in cell number between cultures grown under 18% versus 5% O 2 . Removal of 2-ME from the culture medium had no effect on cell proliferation in cultures grown under either oxygen tension. In contrast, the removal of both 2-ME and serum significantly decreased cell proliferation in cultures grown under 18% O 2 , and a similar trend was observed in cultures grown under 5% O 2 , although the effect did not reach statistical significance. Oxygen Tension, 2-ME and Serum Influence the Metabolic Activity of THP-1 Cells While there was no evidence that removal of 2-ME and serum was overtly toxic to THP-1 cells, it is possible that the absence of these factors decreased cell viability resulting in decreased cell proliferation. To address this question, we next used the MTT assay to determine whether these culture conditions altered the metabolic activity of undifferentiated THP-1 cells. To account for differences in proliferation between cultures exposed to varying culture conditions, results from the MTT assay were normalized to protein concentrations in the same samples. Oxygen tension had no effect on metabolic activity in undifferentiated THP-1 cells cultured in the presence of both 2-ME and serum or with serum alone. Removal of both 2-ME and serum from the culture medium had no significant effect on metabolic activity in THP-1 cells cultured in 18% O 2 , but it significantly increased metabolic activity in cells cultured in 5% O 2 ( Figure 2A). These data suggest that the effect of serum on proliferation ( Figure 1) is not due to effects on cell viability. THP-1 cells can be stimulated to differentiate into macrophages by treatment with phorbol 12-myristate 13-acetate (PMA) [19,20]. PMA induces cell cycle arrest followed by differentiation [21]. Quantification of metabolic activity in PMA-differentiated THP-1 cells indicated that similar to observations in undifferentiated THP-1 cells ( Fig. 2A), culture in 5% O 2 significantly increased metabolic activity (Fig. 2B). In differentiated THP-1 cells, however, this effect of oxygen tension was observed in the presence and absence of 2-ME and serum. Another difference between undifferentiated and PMA-differentiated THP-1 cells is that in the latter, removal of both 2-ME and serum significantly increased metabolic activity relative to cells cultured in the presence of both 2-ME and serum under either oxygen tension (Fig. 2B). Oxygen Tension, 2-ME and Serum Influence PMAstimulated THP-1 Differentiation Differentiation of THP-1 cells can be monitored phenotypically as a switch from a non-adherent to an adherent cell type. Thus, to evaluate the influence of culture conditions on THP-1 differentiation, we quantified cell adhesion at 3 h as a percentage of cell adhesion at 24 h after PMA stimulation in cultures maintained in 18% O 2 versus 5% O 2 in the presence or absence of 2-ME and serum. Relative to cultures exposed to 18% O 2 , differentiation was significantly accelerated at 3 h in cultures exposed to 5% O 2 (Fig. 3). Under either oxygen tension, removal of 2-ME had no effect on cell adhesion at 3 h relative to cultures grown under standard culture conditions; however, removal of both 2-ME and serum significantly increased cell adhesion at 3 h (Fig. 3). Undifferentiated THP-1 cells that were not PMA-stimulated did not adhere when grown in serum free media for extended periods (data not shown). Oxygen Tension, 2-ME and Serum Influence bhexosaminidase Release from Differentiated THP-1 Cells Critical to innate immune function is the constitutive release [22,23] from macrophages of the lysosomal enzyme b-hexosaminidase [24]. To assess the effects of culture conditions on this macrophage function, we quantified both secreted and intracellular amounts of b-hexosaminidase in PMA-differentiated THP-1 cells cultured in 18% O 2 versus 5% O 2 in the absence or presence of 2-ME and serum. PMA-differentiated THP-1 cells released detectable quantities of b-hexosaminidase into the medium during 24 and 48 h of culture (Fig. 4A, B). This release was not dependent on stimulation by lipopolysaccharide (LPS) (data not shown), but it was influenced by oxygen tension, 2-ME and serum. The amount of b-hexosaminidase in the medium at 24 h was significantly decreased in cultures exposed to 5% O 2 relative to 18% O 2 in the presence of both 2-ME and serum or serum alone (Fig. 4A). However, by 48 h, this influence of oxygen tension on bhexosaminidase release was no longer apparent (Fig. 4B). b-Hexosaminidase levels in the medium were also reduced by removal of both 2-ME and serum (Fig. 4A, B). While this effect was observed under both oxygen tension conditions and at both 24 and 48 h, it reached statistical significance only at the 24 h time point and only in cultures exposed to 18% O 2 . To investigate whether the reduced b-hexosaminidase in the medium was due to decreased intracellular levels of b-hexosamin- idase or decreased release from cells, we quantified intracellular levels of the enzyme, normalizing enzyme activity to total protein concentration. Culture in 5% O 2 significantly increased intracellular b-hexosaminidase across all culture conditions, and removal of 2-ME and serum significantly increased intracellular levels of this enzyme in cultures exposed to either oxygen tension (Fig. 4C). These data suggest that reduced levels of b-hexosaminidase in the medium (Fig. 4A, B) reflect decreased release of this enzyme from PMA-differentiated THP-1 cells. While we did not measure the effects of oxygen tension on gene expression of bhexosaminidase subunits, Cowan and collaborators [25] have previously shown that changing oxygen tensions did not alter mRNA levels of this enzyme in cardiomyocytes. Oxygen Tension Significantly Impacts the Phagocytic Activity of Differentiated THP-1

Cells The essential role of macrophages in host-defense mechanisms is mediated in large part by their ability to phagocytose pathogens and cellular debris that contribute to inflammatory reactions and immune responses [26]. To quantify phagocytosis, we used pHrodo TM E.coli fluorescence conjugated BioParticlesH. The fluorescence of these BioParticlesH increases upon lysosomal uptake and subsequent acidification in the lysosomal compartment. Culturing PMA-differentiated THP-1 cells in 5% O 2 significantly decreased phagocytosis of the E. coli BioParticlesH relative to cells cultured in 18% O 2 (Fig. 5). Pretreatment of cultures with cytochalasin-D decreased the mean fluorescence intensity by .75% in cultures exposed to E. coli BioParticlesH under either oxygen tension, confirming that the fluorescence measured in these cultures was the result of phagocytosis of the BioParticlesH. To further evaluate the influence of oxygen tension on phagocytosis, THP-1 cells were PMA-differentiated at low or high oxygen levels for 24 h and switched to high or low oxygen, respectively, for 1 h immediately prior to adding the E. coli BioParticlesH. Consistent with the data shown in Fig. 5 (Table 1). These data suggest that phagocytosis is dependent on the oxygen tension during the phagocytosis assay and not on the oxygen tension during the PMA-induced differentiation, and that phagocytosis is increased at the higher (hyperoxic) oxygen tension, which is consistent with evidence that phagocytosis is dependent on the availability of extracellular oxygen for its respiratory burst [27,28]. Oxygen Tension Influences NF-kB Activation and Cytokine and Chemokine Release in Differentiated THP-1 Cells A key intracellular signaling molecule that links various external stimuli to transcription of target genes in macrophages is NF-kB. NF-kB is a redox-responsive transcriptional factor, and its activation is a key regulator of the cellular response to oxidative stress [29]. NF-kB is also activated by LPS, which induces the expression of multiple genes encoding soluble mediators of inflammation, including cytokines, chemokines and growth factors [30,31]. Thus, we next evaluated the effects of oxygen tension on NF-kB activation using THP-1 XBlue cells, which are stably transfected with an NF-kB-SEAP (secreted embryonic alkaline phosphatase) reporter gene. SEAP expression was measured in differentiated THP-1 XBlue cells cultured in 18% versus 5% O 2 in the absence (baseline) or presence of LPS for 24 h. In the absence of LPS, oxygen tension had no effect on baseline levels of NF-kB activation (Fig. 6A, 6B). NF-kB was significantly activated by LPS relative to baseline levels under either oxygen tension; however, this response was attenuated in cells cultured in 5% O 2 relative to cell cultured in 18% O 2 (Fig. 6A, 6B). A key question raised by these results is whether differential effects of oxygen tension on NF-kB activation translate into altered expression of cytokines and chemokines. To address this question, we used a multiplex cytokine array (specifically, the Milliplex Human Panel) to quantify 14 different cytokines and chemokines at the protein level in differentiated THP-1 cells cultured under different oxygen tensions in the absence or presence of LPS for 24 h. While no clear oxygen tension-related patterns emerged in the expression of individual cytokines or chemokines detected by the multiplex array either in the absence or presence of LPS ( Fig. 6C and D), culturing differentiated THP-1 cells in 5% O 2 caused a general decrease in baseline cytokine/chemokine expression levels and a greater upward shift from baseline with LPS stimulation (Fig. 6D). A second key question raised by the differential effects of 18% versus 5% oxygen tension on LPS-induced NF-kB activation is whether this reflects differences in cellular ROS levels since NF-kB is a redox-responsive transcriptional factor [29]. To address this question, we determined whether by pretreating cells with inhibitors of ROS-generating sources (i.e., NADPH oxidase and lipoxygenase) would attenuate LPS-induced NF-kB activation and whether this attenuation would vary in magnitude between cultures grown under 18% versus 5% O 2 . Differentiated THP-1 cells were cultured under different oxygen tensions in the absence or presence of varying concentrations of the diphenylene iodinium (DPI), an NADPH oxidase inhibitor or nordihydroguaiaretic acid (NGA), an inhibitor of lipoxygenase, for 4 h followed by LPS stimulation for 24 h. Both DPI and NGA significantly inhibited LPS-induced NF-kB activation in a concentration-dependent manner in THP-1 cells grown under either oxygen tension, although a significantly greater inhibition was observed in cultures grown under 18% O 2 relative to cultures grown under 5% O 2 (Fig. 7A and B). Oxygen Uptake Pericellular pO2, which is a primary determinant of oxygendependent cellular responses, is influenced by both atmospheric oxygen and the oxygen consumption of the cells. Thus, we used a Clark-type O 2 electrode to determine whether mitochondrial oxygen consumption varied in differentiated THP-1 cells grown under 18% versus 5% O 2 for 48 h after synchronization. The rate of mitochondrial oxygen consumption in cells grown under 5% O 2 was 0.55 nmol O 2 6(min60 6 cells) 21 ; whereas cells grown under 18% O 2 exhibited a slightly higher rate of oxygen consumption of 0.62 nmol O 2 6(min610 6 cells) 21 . Under both oxygen tensions, rates of oxygen consumption were inhibited by more than 90% by addition of oligomycin (data not shown), indicating that most, if not all, oxygen uptake was linked to oxidative phosphorylation (i.e., mitochondrial ATP production). The ratio of the steady-state concentrations of oxygen around the cells grown at 18% vs. 5% was 3.8 as calculated using the cellular rates of oxygen uptake and the experimental concentration of oxygen in growth media at 20C which is 262.2 mM. This ratio of 3.8 is similar to the ratio of the oxygen solubility at these two pO 2 , which is 4.2. Given that the rates of oxygen uptake of the cells grown at each pO 2 were not dramatically different, the major determinant of the steady-state concentration of oxygen around the cells is the solubility of this gas at each of the pO 2 . Discussion Our data demonstrate that adapting conventional culture conditions to more physiologically relevant conditions significantly alters THP-1 cell physiology. Specifically, we observed that while lowering oxygen tension from 18% O 2 to 5% O 2 had no effect on the proliferation of undifferentiated THP-1 cells, this endpoint was significantly altered by the removal of serum from the culture medium. Changing the oxygen tension from hyperoxic to normoxic did, however, significantly increase metabolic activity in both undifferentiated and differentiated THP-1 cells as well as enhance the differentiation of THP-1 cells and signficantly influence key aspects of macrophage function in differentiated THP-1 cells. Quantification of cellular uptake of oxygen in THP-1 cells grown under 18% O 2 versus 5% O 2 confirmed that the major determinant of the steady-state concentration of oxygen around these cells was the solubility of this gas at each pO 2 and not cellular oxygen consumption. Removing 2-ME from the culture media had negligible effect on these endpoints. In contrast, removing both 2-ME and serum had significant effects on THP-1 metabolism, differentiation and macrophage functions under both conditions of oxygen tension, with more pronounced effects observed in THP-1 cells cultured under 5% O 2 . Serum is commonly used as a supplement in cell culture to improve cell viability; however, there are a number of downsides including cost and the fact that serum is chemically undefined with high variability between batches. Adapting THP-1 cells to serumfree culture conditions would, therefore, significantly decrease costs and potentially increase culture consistency and experimental reproducibility. Removal of serum decreased proliferation of undifferentiated THP-1 cells; however, this effect was largely ameliorated by lowering the oxygen tension from 18% to 5% O 2 . This is consistent with previous studies demonstrating that the proliferation rate of peripheral blood mononuclear cells (PBMC) cultured in medium supplemented with a very low serum concentration was enhanced under normoxic conditions relative to hyperoxic conditions [32]. The decrease in cellular proliferation observed with the removal of serum (and 2-ME) was not due to decreased cell viability as evidenced by the fact that relative to their counterparts cultured in the presence of serum, THP-1 cells cultured in the absence of serum exhibited higher metabolic activity and faster differentiation. PMA-differentiated THP-1 cells cultured in the absence of serum also released approximately 30% less b-hexosaminidase, which correlated with significantly greater retention of b-hexosaminidase in the intracellular compartment. Thus, one of the most striking findings of our study was that the removal of serum from the culture medium for 48-96 hours enhanced THP-1 cell viability and function. Hypoxic conditions have been shown to profoundly affect a broad range of myeloid cell properties in vitro, including expression of chemokine receptors and other cell-surface proteins, cytokine secretion, adhesion, migration, phagocytosis and cell survival [33,34]. Thus, it is perhaps not surprising that lowering the oxygen tension from hyperoxic levels to normoxic levels altered macrophage functions in differentiated THP-1 cells. It has been reported that macrophages constitutively release small amounts of b-hexosaminidase independent of external stimuli [22,23]. This continuous low-level release is important for maintaining the normal turnover of glycosaminoglycan in the tissue matrix; however, the release of hexosaminidases increases significantly during an inflammatory event and this contributes to the degradation of the surrounding tissue [35]. Differentiated THP-1 cells continuously release b-hexosaminidase under all the culture conditions tested, but the amount of enzyme released is significantly influenced by not only the removal of serum and 2-ME, as discussed above, but also by oxygen tension. Decreasing the oxygen tension to 5% O 2 decreased b-hexosaminidase release coincident with increased intracellular levels of the enzyme. These data suggest that under normoxic conditions, the cells are better primed for responding to an inflammatory signal. Phagocytosis by macrophages is an essential component of innate immunity. Previous studies have demonstrated that oxygen tension influences this activity. Pfau and collegues [36] demonstrated that peritoneal macrophages, which are exposed to normoxic conditions, and bone-derived macrophages cultured under low oxygen tension exhibed increased phagocytic activity relative to alveolar macrophages, which are exposed in vivo to hyperoxic conditions, and bone marrow-derived macrophages cultured under high oxgen tension. In contrast, there are numerous reports that alveolar macrophages have greater functional activity related to antimicrobial defense, including phagocytosis, compared to interstitial macrophages which display enhanced capabilities relevant to specific immune responses such as antigen processing as well as in antioxidant defenses [28,37,38]. Additional studies observed that hypoxia, in vivo and in vitro, increase the phagocytic activity of macrophages [39]. These Oxygen Tension Influences THP-1 Cell Physiology PLOS ONE | www.plosone.org apparent discrepancies may be attributed in part to the fact that phagocytosis is positively regulated by hypoxia-inducible factor-1a (HIF1a) [39], which is upregulated by hypoxia [40], and hypoxic conditions are typically associated with inflammatory processes [34]. In contrast, under normoxic conditions, HIF-1a is ubiquitinated and therefore less active [41]. Our data parallel these findings in that we observed that differentiated THP-1 cells exhibited increased phagocytosis when cultured under hyperoxic (18% O 2 ) conditions relative to normoxic (5% O 2 ) conditions. This result is also consistent with evidence that phagocytosis is dependent on the respiratory burst, which requires extracellular oxygen [27]. Collectively, these data implicate O 2 or a metabolite in the acute regulation of phagocytosis. We tested this possibility by switching the oxygen tension under which differentiated THP-1 cells were maintained during the last hour of a 25 h incubation period. Regardless of the oxygen tension under which THP-1 cells were originally cultured, phagocytic activity was predominantly influenced by the oxygen tension during the specific time of phagocytosis, thus linking control of phagocytosis to environmental oxygen tension. It is well documented that NF-kB activation requires an oxidative burst, which releases it from IkB [42]. Therefore, it was not particularly surprising that while LPS stimulated NF-kB in differentiated THP-1 cells grown under either oxygen condition, this response was attenuated in cells exposed to 5% O 2 relative to 18% O 2 . NF-kB activation is closely linked to cytokine release in macrophages [43], so we predicted that LPS-stimulated release of cytokines would similarly be attenuated by culture in 5% O 2 . However, as indicated by multiplex cytokine analyses, differentiated THP-1 cells cultured in 5% O 2 exhibited a more robust increase in cytokine release upon LPS stimulation relative to baseline than cells cultured in 18% O 2 . The concentrations of cytokines released upon LPS stimulation were comparable between cultures exposed to either oxygen tension, so the increased

differential between baseline and LPS-induced levels of cytokine secretion in cells cultured under 5% O 2 reflect the fact that the basal levels of cytokine secretion in the absence of LPS were significantly lower in these cultures relative to those exposed to 18% O 2 . Although the NF-kB activation baseline was not changed at 5% O 2 , these results from the cytokine release studies are consistent with results from the antioxidant studies showing enhanced inhibition of LPS-induced NF-kB activation at 18% O 2 versus 5% O 2 , suggesting a more pronounced role for ROS signaling in THP-1 cells grown under 18% O 2 . Conclusions In response to societal pressures to refine, reduce and replace the use of animals in experimentation, the increasing costs associated with animal models, and the advances in bioinformatics and systems biology, in vitro model systems are an increasingly important tool in biomedical science. While there are limitations associated with cell lines, particularly those that have been immortalized and thus express significant mutations that may alter the physiology of these cells relative to the primary cell type from which they were derived, cell lines, particularly those of human origin such as the THP-1 cell line, are especially useful for pilot projects, drug and toxicity screening, biochemical studies of signal transduction pathways and other types of studies that require large number of cells. Although widely used, standard tissue culture methods expose cells to oxygen levels considerably higher than those encountered by most cells under physiological conditions, and our data corroborate earlier studies in other cell types suggesting that altering oxygen tension impacts cell behavior. Regulating oxygen levels to optimize cell function in vitro is not unprecedented, and extensive and efficient use of this approach has been made in specific tissue culture models, including cultures of placental explant [44], embryos [45] and stem cells [2]. The present study confirms a major role for extracellular oxygen tension in modulating THP-1 cell physiology, which is consistent with literature documenting the regulation of immune function by the cellular redox environment both in vivo and in primary human monocytes and macrophages [46]. Our findings also suggest that THP-1 cells grown under tightly regulated oxygen tension in the absence of exogenous reducing agent are likely to provide a more physiologically relevant baseline from which to study the role of the local redox environment in regulating macrophage differentiation and function. Cell Culture Human monocytic THP-1 cells were purchased from American Type Culture Collection (ATCC) (Manassas, VA) [1]. THP1-XBlue cells, which were derived from the human monocytic THP-1 cell line, were purchased from Invitrogen (San Diego, CA). THP1-XBlue cells are NF-kB-reporter cells in which activation of the transcription factor NF-kB results in the secretion of embryonic alkaline phosphatase (SEAP), which is detected using Quanti-Blue reagent (Invitrogen). Cells were maintained in RPMI 1640 medium containing 11.11 mM glucose in but no phenol red (GIBCO, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (FBS, GIBCO), 1% penicillin (GIBCO), 1% streptomycin (GIBCO) and 50 mM 2-ME (Fisher Scientific, Pittsburgh, PA) at 37uC in a humidified incubator with 5% CO 2 and 95% air (e.g., standard culture conditions). Prior to experimentation, the cells were starved for 48 h and subsequently cultured with or without 2-ME and/or FBS at 37uC in a humidified Thermo Scientific CO 2 tissue culture incubator (NAPCO Series 8000WJ, Thermo Forma, Marietta, OH) equipped with built-in CO 2 and O 2 monitors and attached nitrogen and carbon dioxide gas supplies. Carbon dioxide was set to 5% v/v and oxygen to 5% of 18%. The oxygen and carbon dioxide contents of the incubator atmosphere were periodically verified using a Fyrite gas analyzer (Bacharach Inc., New Kensington, PA). For some experiments, cultures were treated for 24 or 48 h with phorbol 12-myristate 13-acetate (PMA, Sigma-Aldrich, St. Louis, MO) at 20 ng/ml to trigger THP-1 cells to undergo differentiation into macrophages [19,20]. A stock solution of PMA at 40 mg/ml in dimethyl sulfoxide (DMSO, Sigma-Aldrich, Saint Louis, MO) was diluted in tissue culture medium with the final DMSO concentration of 0.1%. Addition of 0.1% DMSO alone did not cause THP-1 cells to undergo macrophage differentiation, nor did it affect their viability as assessed using the MTT assay (data not shown). Proliferation Assays Non-differentiated THP-1 cells were synchronized by serum deprivation for 48 h prior to being plated at an initial density of 0.7610 6 cells/ml in 35 mm tissue culture dishes and cultured under the conditions indicated in Figure 1. At 24 or 48 h after plating, cell density was determined using a hemocytometer. The percent growth was calculated according the following equation: [(final cell density at 24 or 48 h *100)/(0.7610 6 )] -100). Experiments were independently repeated five times with 3 samples per treatment in each experiment. Metabolic Activity Assays The metabolic activity of the cells was evaluated by quantifying the reduction of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, Sigma-Aldrich) to formazan, a reaction catalyzed by cellular reductases and dependent on the availability of reducing equivalents in the cell [47]. After synchronization (serum deprivation for 48 h), THP-1 cells were plated at 1610 5 cells/well in 96-well plates. To compare effects in monocytic THP-1 cells versus THP-1 macrophages, PMA (20 ng/ml) was added to the latter cultures 24 h prior to experimentation. For undifferentiated THP-1, cells were plated onto poly-D-lysine (100 mg/ml, Sigma-Aldrich)-coated wells immediately prior to initiating the experiment. MTT was added to the wells at 500 mg/ml final concentration and cells were incubated at 37uC under the culture conditions indicated in Figure 2 Quantifying the Rate of THP-1 Macrophage Differentiation Undifferentiated THP-1 cells were synchronized by serum deprivation for 48 h prior to re-plating at 1610 5 cells in 96-well plates in media containing 20 ng/ml PMA. Cells were then cultured at 37uC in 18% or 5% O 2 with or without 2-ME and/or FBS. After 3 or 24 h of PMA stimulation, the non-adherent cells were removed with 3 rinses of PBS. The adherent cells were lysed with 50 ml of 1% triton X-100 in PBS, pH 7.4., and the protein content of the cell lysate was measured using the BCA protein assay. Cell adhesion was determined as the protein concentration of cultures at 3 h expressed as a percentage of the protein concentration at 24 h. Measurement of b-hexosaminidase Spontaneous release of lysosomal contents of THP-1 macrophages was determined by measuring the enzyme b-hexosaminidase. Undifferentiated THP-1 cells were plated in 24-well plates at a density of 2610 5 cells/well and stimulated to differentiate by incubating with 20 ng/ml PMA for 24 or 48 h. After differentiation, conditioned medium was collected from each well and saved, and then cells were washed twice and lysed in 1% triton X-100 in PBS, pH 7.4. Triplicate aliquots of each conditioned medium and cell lysate sample (50 ml each) were mixed with an equal amount of substrate, 1.3 mg/ml p-nitrophenyl-N-acetyl-b-D-glucosaminide (Sigma-Aldrich), in 0.1 M citrate, pH 3.5. After incubation for 1 h at 37uC, 50 ml of 0.2 M glycine, pH 10.5, was added to stop the reaction, and the absorbance was measured at 405 nm using a TECAN spectrophotometer. Results were normalized against protein concentration in each sample, which was determined using the BCA protein assay. Experiments were independently repeated four times, and the results were comparable across all four experiments. Phagocytosis Assay Phagocytosis was measured using the pHrodo TM E.coli fluorescence conjugated BioParticlesH (Invitrogen/Molecular Probes, Eugene, OR). The fluorescence of the BioParticlesH increases upon lysosomal uptake and subsequent acidification. Cells were synchronized for 48 h, plated at a density of 10 5 cells/well in a 96well plate and differentiated by PMA treatment for 48 h without 2-ME and FBS. Negative controls were incubated with 2 mM cytochalasin D (Sigma-Aldrich) for 1 h before the addition of the E. coli BioParticlesH to inhibit phagocytosis. For some experiments, THP-1 were PMA-differentiated at low or high oxygen tension for 24 h and then switched to high and low oxygen tension, respectively, 1 h before the addition of the BioParticlesH. Cells were incubated with BioParticlesH for 90 minutes, washed and fluorescence was quantified using the Molecular Devices Spec-traMax plate reader (Molecular Device, Sunnyvale, CA, USA) with the excitation wavelength set at 550 nm and the emission wavelength detection set at 600 nm. Results were normalized against protein concentration as determined using the BCA protein assay. Quantification of Cytokine and Chemokine Release THP-1 cells grown without 2-ME and FBS were synchronized by serum deprivation for 48 h followed by PMA-differentiation for 48 h under 5% or 18% oxygen. Differentiated THP-1 cells were plated at 0.5610 6 cell/ml in 6-well plates and cultured for an additional 24 h at 18% or 5% O 2 in the absence (baseline) or presence of LPS at 20 ng/ml. Conditioned medium was collected from each well at the end of the 24 h incubation. A human Milliplex Kit (Millipore, Billerica, MA) was used to measure chemokine and cytokine concentrations in duplicate aliquots of each conditioned medium sample. This kit simultaneously interrogates 14 human cytokines, chemokines, and growth factors, including: IL-1b, IL-6, MIP-1a, IP-10, TNFa, IFNc, IL-1ra, IL-10, INFc, MCP-1, FKN, G-CSF, GM-CSF and VEGF. Samples were analyzed using the Bio-Plex array system, which includes a fluorescent reader and Bio-Plex Manager Analytic software (Bio-Rad, Hercules, CA). One hundred beads were counted for each analyte per well and cytokine concentrations (pg/ml) were calculated using Bio-Rad software. Oxygen Uptake The oxygen uptake of intact THP-1 cell suspensions (6 to 7610 6 cells/ml) at 20uC was measured using a Clark-type O 2 electrode from Hansatech (King's Lynn, UK) [48]. Cells were incubated in the same RPMI 1640 culture medium used to maintain the cell line (e.g., medium containing 11.11 mM glucose but no phenol red). To evaluate mitochondria-derived oxygen uptake, measurements were repeated in the presence of 3 mM oligomycin (Sigma Chemical Company, Saint Louis, MO). A model for the steadystate concentration of oxygen was used that is based on the flow of oxygen delivered into the chamber and its pO 2 , the solubility of oxygen in the growth media (measured), and the oxygen uptake by cells in growth media (measured). Statistical Analysis All data are presented as the mean 6 SEM. Differences between 2 treatment groups were analyzed by Student's t-test; whereas differences between .2 groups were determined by oneway ANOVA followed by Tukey's post-test using GraphPad Prism 4 software (San Diego, CA). p values ,0.05 were considered statistically significant. These Shoes Are Made for Walking: Sensitivity Performance Evaluation of Commercial Activity Monitors under the Expected Conditions and Circumstances Required to Achieve the International Daily Step Goal of 10,000 Steps Introduction Physical activity is a vitally important part of a healthy lifestyle, and is of major benefit to both physical and mental health. A daily step count of 10,000 steps is recommended globally to achieve an appropriate level of physical activity. Accurate quantification of physical activity during conditions reflecting those needed to achieve the recommended daily step count of 10,000 steps is essential. As such, we aimed to assess four commercial activity monitors for their sensitivity/accuracy in a prescribed walking route that reflects a range of surfaces that would typically be used to achieve the recommended daily step count, in two types of footwear expected to be used throughout the day when aiming to achieve the recommended daily step count, and in a timeframe required to do so. Methods Four commercial activity monitors were worn simultaneously by participants (n = 15) during a prescribed walking route reflective of surfaces typically encountered while achieving the daily recommended 10,000 steps. Activity monitors tested were the Garmin Vivofit ™, New Lifestyles’ NL-2000 ™ pedometer, Withings Smart Activity Monitor Tracker (Pulse O2) ™, and Fitbit One ™. Results All activity monitors tested were accurate in their step detection over the variety of different surfaces tested (natural lawn grass, gravel, ceramic tile, tarmacadam/asphalt, linoleum), when wearing both running shoes and hard-soled dress shoes. Conclusion All activity monitors tested were accurate in their step detection sensitivity and are valid monitors for physical activity quantification over the variety of different surfaces tested, when wearing both running shoes and hard-soled dress shoes, and over a timeframe necessary for accumulating the recommended daily step count of 10,000 steps. However, it is important to consider the accuracy of activity monitors, particularly when physical activity in

the form of stepping activities is prescribed as an intervention in the treatment or prevention of a disease state. Introduction Physical activity is a vitally important part of a healthy lifestyle, and is of major benefit to both physical and mental health. A daily step count of 10,000 steps is recommended globally to achieve an appropriate level of physical activity. Accurate quantification of physical activity during conditions reflecting those needed to achieve the recommended daily step count of 10,000 steps is essential. As such, we aimed to assess four commercial activity monitors for their sensitivity/accuracy in a prescribed walking route that reflects a range of surfaces that would typically be used to achieve the recommended daily step count, in two types of footwear expected to be used throughout the day when aiming to achieve the recommended daily step count, and in a timeframe required to do so. Methods Four commercial activity monitors were worn simultaneously by participants (n = 15) during a prescribed walking route reflective of surfaces typically encountered while achieving the daily recommended 10,000 steps. Activity monitors tested were the Garmin Vivofit ™, New Lifestyles' NL-2000 ™ pedometer, Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, and Fitbit One ™. Introduction Physical activity is an essential part of a healthy lifestyle, playing an important role in improving and maintaining both physical and mental health [1][2][3][4]. Indeed, lack of physical activity has been recognized as the fourth leading risk factor for global mortality, associated with 6% of deaths worldwide [5]. Since the benefits of physical activity have been recognized globally, governments internationally have highlighted to their citizens the need to be physically active on a regular basis. In 1996, the US Surgeon General recommended that as part of a healthy lifestyle, people of all ages should partake in at least 30 minutes of moderate-intensity physical activity, such as brisk walking, on a daily basis [6]. The 2008 'Physical Activity Guidelines for Americans' from the Centers for Disease Control (CDC) recommended that adults need 2 hours and 30 minutes of moderate intensity aerobic physical activity, for example brisk walking, per week [7]. These two recommendations correspond closely with the 2010 WHO recommendation that adults should do at least 150 minutes of moderate intensity aerobic physical activity throughout the week [5]. An important consideration in assisting persons to change their physical activity behavior and to adhere to these public health recommendations is providing them with the ability to easily quantify the extent of physical activity completed in a given day. The development of commercial pedometers in the late 1960s provided a convenient, low cost method of quantifying one form of physical activity, namely walking. These devices provided straightforward feedback to the user of the quantity of physical activity completed, specifically their step count. A series of pedometer-based studies, from 2004 to 2011, evaluated the relationship between step count and adherence to physical activity guidelines and reported that less than 7,500 steps per day represented sedentary or "low-active" behavior [8,9] and taking 10,000 steps per day was consistent with a physical activity level associated with person who is "active" [8,10,11]. Furthermore, a study carried out in an overweight population from the Lower Mississippi Delta of the U.S.A. reported a step count of 9,154 steps per day to equate to 30 minutes of moderate-to-vigorous physical activity. The authors concluded that, in this population, a step count of 8,300 to 9,100 steps a day should be accumulated in order to achieve recommended physical activity guidelines [12]. As part of their physical activity guidelines, the WHO states that physical activity in adults includes recreational or leisure-time physical activity, occupational (i.e. work), household chores, transportation (e.g. walking or cycling), play, games, planned exercise or sports in the context of daily, family, and community activities [5]. Thus the recommended "quantity" of physical activity of 10,000 steps being proposed internationally would only typically be achieved through a combination of these activities and under the varied circumstances associated with these activities. In other words, the recommended step count of 10,000 steps would typically be achieved, not only through a specific programme of exercise, but also through a combination of activities in the home, in the work place, and while getting to and from work. These activities would typically involve the person walking on a wide variety of surfaces from footpaths, indoor floors with varied surface types, and outdoor natural walking surfaces such as grass pathways. Indeed, one group have examined the effect of walking surface on step count sensitivity/accuracy of a pedometer and found surface type significantly affected step count [13]. Additionally, a number of physical activity device manufacturer websites state that surface type can affect step count sensitivity/accuracy of their devices [14,15]. Furthermore, the range of activities conducted in achieving the recommended step count of 10,000 steps would also typically involve the person changing their footwear from more formal "dress" shoes in the workplace to more comfortable runner-type shoes in the home and while exercising. Therefore, a physical activity monitor being used to quantify adherence to the recommended daily step count must be able to deal with: (i) a variety of walking surfaces, (ii) different footwear, and (iii) extended periods of walking, where the accumulated number of steps is in the order of 10,000 steps. These three variables have the potential to affect step count detection accuracy. Thus, it is the view of the authors that activity monitor testing must involve: (i) walking on different surfaces that reflect the range of surfaces that will typically be encountered to achieve the daily step count goal, (ii) walking with different footwear that reflects the different types of footwear expected to be used throughout the day, and (iii) walking for periods of time that properly reflect the time required to achieve the recommended daily step count. In fact, a review of the literature highlights that these considerations have not been widely adopted in a wide range of studies where the sensitivity of activity monitors has been assessed [16][17][18][19][20][21]. Often, the types of surfaces and types of footwear tested are either not reported or not taken into consideration during testing. Furthermore, the timeframe of testing is often far less than would be required to achieve the recommended daily step count of 10,000 steps. For example, timeframes of between 11 and 25 minutes have previously been utilized and would be completely insufficient for accumulating 10,000 steps [16][17][18][19][20]. In this paper we aimed to assess four commercial activity monitors: the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, the NL-2000 pedometer ™, the Garmin Vivofit ™, and the Fitbit One ™. We aimed to assess the step detection sensitivity of each activity monitor, i.e. the ability of the activity monitors to detect and count an actual step as a step, over a prescribed walking route that reflects a range of surfaces that would typically be used to achieve the recommended daily step count and in a timeframe required to do so. In addition, we also aimed to investigate two different types of footwear typically used throughout the day when aiming to achieve the recommended daily step count of 10,000 steps. Participants Fifteen healthy participants (8 female, 7 male) took part in this study, with a mean age of 21.1 ± 1.1 years. Males had a mean BMI of 23.60 ± 2.70kg/m 2 while females had a mean BMI of 21.88 ± 1.81kg/m 2 . None of the participants had any history of cardiovascular disease or neurological disorders. Ethics committee approval was obtained from the Galway University Hospitals Research Ethics Committee, and all participants provided written, informed consent. Study Protocol All 15 participants completed a prescribed walking route. The activity monitors tested in this study were the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™ (Withings, Issy-les-Moulineaux, France), NL-2000 pedometer ™ (New Lifestyles, Missouri, USA), Garmin Vivofit ™ (Garmin, Kansas, USA), and Fitbit One ™ (Fitbit, San Francisco, USA). (See Table 1). All activity monitors were put in place at the manufacturer's recommended body location (Fig 1) by the investigators as per the manufacturers' instructions. All four activity monitors were worn simultaneously on each participant for the duration of testing. The Garmin Vivofit ™ was worn on the non-dominant wrist, with the NL-2000 ™ and Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™ worn on opposite sides of the waist. The Fitbit ™ activity monitor was clipped onto clothing at the level of the chest. Participants were video recorded throughout the study with a hand-held camcorder. The true step count was extracted manually from the recorded video in real time and was then compared to the step count registered by each activity monitor. Additionally, the ActivPAL micro ™ (PAL Technologies Ltd., Glasgow, UK) was worn by each participant as a reference device for measuring the overall total step count for the prescribed walking route. The ActivPAL ™ was attached to the thigh using Tegaderm transparent dressing (3M Health Care, Minnesota, USA). Participants completed a prescribed flat walking route consisting of five different walking surfaces: linoleum (800m), natural lawn grass (900m), gravel (990m), ceramic tile (400m) and tarmacadam/asphalt (880m). In addition, the prescribed walking route included stair walking (49 steps up and 49 steps down, step height 16cm) and ramp walking (240m up and 240m down at an incline of 4.05%) (Fig 2). Each participant completed the walking route twice: the first time wearing standard running shoes, the second time wearing hard-soled dress shoes. At all times the participants were asked to walk at their normal walking pace, in other words, self-selected walking speed. Inter-device Reliability There were between three and five of each device type used during testing. Prior to testing, each unit of each device type was tested over a 400m flat walk of uniform ceramic tile surface while wearing hard-soled shoes. One investigator wore each unit of a device type simultaneously for the 400m walk. The step counts on each unit of that device type were then compared. There was found to be no difference in step counts between the Fitbit ™ units. The largest difference between the units of the remaining activity monitors was 0.19% for the NL-2000 ™ units, 1.55% for the ActivPAL ™ units, and 13.67% for the Garmin ™ units. For the Garmin units, any step counts outside of ±2 SD of the observed count were excluded from analysis. Statistical Analysis All statistical analyses were carried out using SPSS (SPSS for Mac, version 20, IBM Corporation). Sample size was chosen by selecting a type I error rate (α) of 5%, a power (1-β) of 0.80, a sampling ratio of 1 (κ = n A / n B ), and a standard deviation of 250 steps (2.5% of the recommended daily step count of 10,000 steps). The Shapiro-Wilk test was used to analyze normality of data. Data were then fitted to a repeated measures model as this study followed a repeated measures design. Repeated-measures analysis of variance and post-hoc follow-up were used to detect differences in step count between the true step count extracted manually from the video recordings and the step count registered by each activity monitor and to detect step count detection errors associated with each surface type during the walking route. The Greenhouse-Geisser correction was used to correct any violations of the assumption of sphericity. The mean absolute percentage error (MAPE) was calculated for each activity monitor using Eq 1 with N the number of steps as extracted from video analysis and Ñ the number of steps as recorded from each activity monitor. There is no standard for acceptable error in step detection in activity monitors, but we have chosen to select a 5% error zone as being acceptable. Bland-Altman plots were plotted to examine the agreement in step count between the observed step count and the step count recorded on the individual activity monitors. An unpaired t test was carried out to assess any differences in step count between participants classed as 'fast' and 'medium to slow' speed walkers. Results Step counts obtained from video analysis are referred to as the observed step count and were compared to output counts from each activity monitor. Surface type did not affect step count detection on the four activity monitors tested. The mean step counts for the four activity

monitors tested over each section of the walking route are given in Table 2. There was no statistically significant difference in activity monitor detected step count vs. the observed step count for the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, NL-2000 pedometer ™, Garmin Vivofit ™, or Fitbit One ™ activity monitors on the linoleum (P = 0.106), gravel (P = 0.131), natural lawn grass (P = 0.195), tarmacadam/asphalt (P = 0.286), or ceramic tile (P = 0.457) surfaces. There was also no statistically significant difference in activity monitor detected step count versus the observed count for the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, NL-2000 pedometer ™, Garmin Vivofit ™, and Fitbit One ™ activity monitors on the ramp section of the walking route (P = 0.591). Stairs walking did not affect step count, with no statistically significant difference in activity monitor detected step count vs. the observed step count for any activity monitor (P>0.05 for all). Over the entire walking route, independent of surface type, there was no statistically significant difference in step count between the observed count and the four activity monitors tested (P > 0.05; Fig 3). There was also no statistically significant difference in step count between the step count on the ActivPAL ™ reference device and the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, NL-2000 pedometer ™, Garmin Vivofit ™, or Fitbit One ™ activity monitors (P > 0.05 for all). There was no statistically significant difference in step count when the devices were compared to each other on any surface type or overall, independent of surface type (P > 0.05 for all). Mean absolute percentage errors for each of the four activity monitors are given in Table 3. Over the prescribed walking route, eight mean absolute percentage errors were found to be outside of our selected 5% error zone. The Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™ had a step count error for stair walking only, with errors on descending stairs of 6.64% and on ascending stairs of 5.93%. The Garmin ™ had step count errors on the ceramic tile of 5.43%, on ascending stairs of 11.33% and on descending stairs of 6.48%. The NL-2000 pedometer ™ step count error on ceramic tile was 7.66% and on ascending stairs of 8.65%. The Fitbit ™ had step count error on descending stairs of 6.03%. All other mean absolute percentage errors remained within the 5% error zone. The absolute percentage error range is also is in Table 3. Although there was no significant difference in the step counts observed versus the step counts detected by the activity monitors, these values show the variation in the percentage error for the activity monitors (Table 3). For example, when walking down stairs, the Fitbit One ™ activity monitor had a mean absolute percentage error of 6.03% and a mean absolute percentage error range of -16.36% to 99.16%. When all values lying ±2SD from the mean observed step count were removed, only two mean absolute percentage errors were found to be outside the 5% error zone (Table 4). These were both for the Garmin ™ device, with error on ascending stairs of 6.59% and on descending stairs of 5.03%. Bland-Altman plots for the number of steps showed average ± limits of agreement underestimation of 84±238, 44±735, 75±235, 35±232, and 254±488 for the Withings ™, Garmin ™, Fitbit ™, ActivPAL ™, and NL-2000 ™ activity monitors respectively (Fig 4). Footwear type did not affect step detection of any activity monitor; there was no statistically significant difference in step count between the observed count and any activity monitor (P>0.05 for all). Participants took a median time of 4 hours and 4 minutes to complete the entire walking route. Seven participants were classed as being walkers of a 'fast' speed as they completed the walking route in a time of between 2 hours and 6 minutes and 3 hours and 2 minutes, a gap of 1 hour and 2 minutes from the median walking time. Eight participants were classed as being walkers of a 'medium to slow' speed as they completed the walking route in a time between the median time (4 hours and 4 minutes) and 4 hours and 38 minutes. There was a statistically significant difference in step count between the 'medium to slow' speed walkers and 'fast' speed walkers, with the 'medium to slow' speed walkers taking a greater number of steps (P = 0.002). Additionally, the mean walking speed for the 'fast' walkers was 0.84m/s compared to 0.74m/s for the 'medium to slow' speed walkers, a statistically significant difference (P = 0.031). Discussion This study assessed four commercial activity monitors: the Withings Smart Activity Monitor Tracker (Pulse O 2 ) ™, NL-2000 pedometer ™, Garmin Vivofit ™, and Fitbit One ™ for their step count detection accuracy over a prescribed walking route that reflected a range of surfaces that would typically be encountered to achieve the recommended daily step count and in a timeframe required to do so. In addition, we investigated two different types of footwear typically used throughout the day when aiming to achieve the recommended daily step count of 10,000 steps. As stated by the WHO, physical activity includes a variety of activities, such as recreational or leisure-time physical activity, occupational (i.e. work), household chores, and transportation (e.g. walking or cycling) [5]. As such, the recommended daily step count of 10,000 steps would typically be achieved over a variety of different surfaces. Taking this into consideration, we tested a prescribed walking route consisting of various surfaces that would be encountered on a daily basis, namely linoleum, natural lawn grass, gravel, tarmacadam/asphalt, and ceramic tile surfaces. All activity monitors were found to be accurate in step detection, irrespective of surface type. This is an important finding as step count is one of the main parameters utilized in physical activity quantification. It is vital that persons changing their behavior to adhere to public health physical activity recommendations can be confident in the sensitivity of the activity monitors they are employing to achieve their physical activity goals and reach the recommended daily step count of 10,000 steps. Additionally, our results correspond with those obtained by Brown et al. when they investigated the accuracy of the ActiPed ™ activity monitor over two different surface types, namely grass and concrete. Participants walked 1,010m on grass and 1,070m on concrete, with the authors finding the ActiPed ™ to be accurate in step count detection regardless of surface type [36]. When considering step counts, as opposed to surface type, participants in our study exceeded the recommended daily step count of 10,000 steps, taking an average of 10,950 steps to complete the prescribed walking route. Over the course of the entire walking route, independent of surface type, all activity monitors were found to be sensitive in their step detection accuracy, on both flat surfaces and stairs walking. Indeed, it is important to consider the effect of stairs walking, which is an activity of daily living and part of the normal daily step count goal for many adults. We observed similar results to those obtained by Storm et al., who investigated the sensitivity of both the ActivPAL ™ and Movemonitor ™ activity monitors during stairs walking at a self-selected walking speed, finding both activity monitors to be accurate in step detection [19]. Accuracy is an important factor to consider when the quantification of physical activity is paramount, such as when physical activity may be prescribed as an intervention in the treatment or prevention of a disease state. Sensitivity in this case is the probability that a step will be detected when a step is actually performed and a measure of all true positives (TP; i.e. a step was performed and it was detected) divided by the true positive (TP) plus the false negatives (FN; i.e. a step was performed but not detected). Due to the nature of the data output from the activity monitors, calculating FN is not possible thus we presented the MAPE as a valid measure of accuracy. When we investigated the effect of two different types of footwear: running shoes and hardsoled dress shoes, on the sensitivity of the activity monitors, we found no effect on step detection sensitivity. To the best of our knowledge, this is the first study to assess the effects of footwear on the step detection sensitivity of activity monitors and is an important and positive finding, as activities carried out to achieve the recommended step count of 10,000 will most likely be done in various different types of footwear. However, we only investigated two different types of footwear. It would also be of interest to investigate other types of footwear that would often be worn on a daily basis when accumulating the recommended daily step count of 10,000 steps. Examples include slippers, flip-flops, high-heeled shoes, and hiking boots. These different types of footwear differ in terms of their sole type, degree of foot contact with the ground/floor, and the way in which they are worn, i.e. loose fitting or tight fitting on the foot, all of which may play a role in the step detection sensitivity of activity monitors. The authors do recognize that limitations exist with our study, namely we have tested the accuracy of four commercial activity monitors but have not evaluated their specificity. For example, it is also of vital importance to validate activity monitors during a variety of activities of daily living that do not involve direct stepping, such as driving, cycling, and swimming. Furthermore, acceleration signals recorded by activity monitors can differ depending on the location of attachment [37]. Thus, the attachment location for each activity monitor is important to consider. In this study, we assessed each activity monitor in only one location. However, we specifically chose attachment locations for each activity monitor that was specified by the manufacturers. Additionally, it is important to consider walking speed when evaluating the sensitivity of activity monitors. Slow walking speeds, for example, have been shown to reduce the sensitivity of some activity monitors [18,27,[38][39][40]. In this study, we evaluated different distributions of walking speeds within the normal walking self-selected walking speeds and found a greater speed and lesser number of steps in the 'fast' speed walkers. However, evaluation on a range of specific walking speeds, such as on a treadmill, would be of benefit to further elucidate the effect of walking speed on activity monitor step detection sensitivity. Conclusions Our study provides a comprehensive evaluation of four commercial activity monitors when (i) walking on different surfaces that reflect the range of surfaces that will typically be used to achieve the daily step count goal, (ii) walking with different footwear that reflects the different types of footwear expected to be used throughout the day, and (iii) walking for periods of time that properly reflect the time required to achieve the recommended daily step count. All activity monitors tested were accurate in their step detection sensitivity and are valid monitors for physical activity quantification over the variety of different flat surfaces tested and when stairs and ramp walking, when wearing both running shoes and hard-soled dress shoes, and over a timeframe necessary for accumulating the recommended daily step count of 10,000 steps. It is important to consider accuracy for activity monitors, particularly when physical activity in the form of stepping activities is prescribed as an intervention in the treatment or prevention of a disease state. All aspects of daily physical activity that go towards achieving the daily recommended step count of 10,000 steps are necessary to consider in accurate physical activity quantification. Residual cancer burden in locally advanced breast cancer: a superior tool Objectives Locally advanced breast cancer (labc) poses a difficult clinical challenge with an overall poor long-term prognosis. The strength of the association between tumour characteristics, treatment response, and outcome is not well defined. In the present study, we attempted to gain further insight into labc by reviewing tumour characteristics of patients treated with neoadjuvant chemotherapy and by studying the association of those characteristics with outcome. We calculated the residual cancer burden (rcb) score obtained at surgery and attempted to study its correlation with event-free survival (efs) and overall survival (os). Methods We studied patients diagnosed primarily with labc (n

= 45). Pathologic and clinical responses were determined. Pathology slides were reviewed. Results Of the 45 study patients, 9% had stage iib disease; 29%, stage iiia; 51%, stage iiib; and 11%, stage iiic. Inflammatory breast cancer (ibc) was found in 16%. Pathologic complete response (pcr) was achieved in 22% of all patients. None of the patients with ibc achieved pcr. Patients with estrogen receptor–negative (er−)/progesterone receptor–negative (pr−) tumours were more likely to achieve pcr than were those with er+/pr+ tumours. Among patients with tumours that overexpressed human epidermal growth factor receptor 2 (her2/neu), 17% achieved pcr as compared with 25% of patients with non-overexpressing tumours; only 1 patient had received trastuzumab. The rcb scores were calculated in 32 patients and ranged between 0 and 4.6. Conclusions The present study examined practical issues related to the classification and management of labc and ibc. The rcb, defined from routine pathology materials, was easily quantifiable. It appears to be a better predictor than pcr of outcome following neoadjuvant chemotherapy in labc. Higher rcb scores were associated with lower efs and a lower rate of os. A continual quest for reliable predictive and correlative prognostic markers, and for better surrogate endpoints for outcome, is essential to advance our understanding of labc and to improve treatment outcomes. INTRODUCTION Since the late 1970s, tremendous progress has been achieved in the understanding and management of breast cancer. However, locally advanced breast cancer (LABC) remains a difficult clinical challenge, with a long-term survival rate of less than 50% 1 . Treatment of LABC uses a multimodality approach involving chemotherapy, surgery, and radiation therapy 2 . The optimal neoadjuvant chemotherapeutic regimen for LABC continues to evolve. The aim of neoadjuvant chemotherapy is to decrease tumour bulk and ideally to achieve complete clinical and pathologic responses. Pathologic complete response (pCR) has been viewed as a reliable primary endpoint for outcome and survival following neoadjuvant chemotherapy for breast cancer 3,4 . However, residual disease may encompass a range of pathologic responses likely encompassing a variety of prognostic groups from near-complete response to resistance. Therefore, additional surrogate endpoints for outcome and survival are needed. Also, many of the classic biologic predictive and prognostic factors such as hormone receptors and tumour grade may have implications that are different in LABC than in earlier stages of the disease [5][6][7][8][9] . To advance our understanding of this disease, identification of reliable markers that would lead to better disease classification and improved treatment outcomes is desirable. However, few trials studying primary (preoperative) 272 272 272 272 272 chemotherapy have focused exclusively on patients with locally advanced disease. In the present observational study, we attempted to gain further insight into LABC by reviewing tumour characteristics in patients treated with neoadjuvant chemotherapy at a single institution, and by studying the association of those tumour characteristics with outcome. We were specifically interested in determining the practicality of calculating the residual cancer burden (RCB) scores obtained at surgery and in studying the correlation of those scores with eventfree survival (EFS) and overall survival (OS) as compared with pCR. The RCB index was proposed by Symmans et al. 10 as a determinant of the extent of residual disease in the post-treatment surgical resection specimen of patients with breast cancer who received preoperative chemotherapy. The RCB index was found to be an improvement over currently used risk factors for the prediction of distant relapse after neoadjuvant chemotherapy. If independently validated, the RCB index is suggested to provide an accurate surrogate endpoint for patient survival. The RCB index is determined from • the bi-dimensional diameters of the primary tumour bed in the resection specimen (d 1 and d 2 ), • the proportion of the primary tumour bed that contains invasive carcinoma (f in ), • the number of axillary lymph nodes containing metastatic carcinoma (LN), and • the diameter of the largest metastasis in an axillary lymph node (d met ). Largest bi-dimensional measurements of the residual primary tumour bed are recorded from the macroscopic description and are combined as follows: The proportion of invasive carcinoma (f inv ) within the cross-sectional area of the primary tumour bed is estimated from the overall percentage area of carcinoma (%CA) and is then corrected for the component of in situ carcinoma (%CIS): Symmans et al. calculated RCB indexes based on a review of patients who completed neoadjuvant chemotherapy for invasive breast carcinoma (T1-3, N0-1, M0) at the M.D. Anderson Cancer Center 10 . We reviewed pathology slides and reports from 432 patients in two completed neoadjuvant trials: • Fluorouracil, doxorubicin, and cyclophosphamide (FAC) in 189 patients • Paclitaxel followed by FAC (T/FAC) in 243 patients. The RCB was calculated as an index that combines pathology measurements of the primary tumour (size and cellularity) and nodal metastases (number and size). Four RCB categories [RCB-0 (pCR) to RCB-3 (chemoresistant)] and post-treatment revised American Joint Committee on Cancer (AJCC) stage (0-III) for prediction of distant relapse-free survival (DRFS) were compared in multivariate Cox regression analyses stratified by estrogen receptor status (ER) status. The RCB was found to be a continuous predictor of DRFS and to predict relapse more strongly than AJCC stage did. In univariate Cox regression analyses, the four parameters of residual tumour (d prim , f inv , LN, and d met ) were individually associated with significantly higher risk of distant relapse (p < 0.001) after T/FAC chemotherapy. They maintained significance as independent predictors in the main-effects multivariate Cox regression model. Patients had an almost-doubled relapse risk for each unit of increase in the RCB index [hazard ratio (HR): 1.94; 95% confidence interval (CI): 1.47 to 2.55; p < 0.001]. When the RCB index was included in a multivariate Cox regression model that included clinical and treatment covariates, the overall predictive power of the model was significantly improved (p < 0.001), and the RCB index was significantly associated with the risk of disease recurrence (HR: 2.50; 95% CI: 1.70 to 3.69; p < 0.001). Selection Procedures After obtaining approval from the University of Cincinnati institutional review board, we conducted a retrospective chart review of the breast oncology database and reviewed the medical records of patients who received neoadjuvant chemotherapy at the University of Cincinnati between January 1, 2001, and December 31, 2005. We included consecutive patients diagnosed primarily with inoperable LABC staged as IIB, IIIA (T0-N2; T1/2-N2; T3-N1/2), IIIB (T4, N0-2), or IIIC disease (any T, N3). Patients with inflammatory breast cancer (IBC) were included. We excluded patients diagnosed with operable tumours staged as I, IIA, and IIB, even if they received neoadjuvant chemotherapy. Patients with stage IV disease were also excluded. Initially, we identified 50 patients; 5 were later excluded when found to have metastatic disease on staging workup. We evaluated 45 patients. Tumour and patient characteristics were reviewed (Table I) Treatment with trastuzumab was also noted. Tumour Response Clinical response was recorded before each chemotherapy cycle and before surgery. No clinical evidence of palpable tumour in the breast and axillary lymph nodes was defined as a clinical complete response (cCR), reduction in total tumour size of 50% or more was graded as a clinical partial response (cPR). An increase in total tumour size of more than 50% or the appearance of new suspicious ipsilateral axillary adenopathy was considered progressive disease. Tumours that did not meet the criteria for objective response or progression were considered stable disease. Pathologic response was determined at surgery. A pCR was defined as no invasive tumour in breast or axillary lymph nodes. Complete response in breast, but residual disease in lymph nodes was designated RDLN; residual disease in breast, but no disease in lymph nodes was designated RDB; and residual disease in both was designated RDBLN. Calculation of Residual Cancer Burden Pathology slides for 32 of the 45 patients were available. The characteristics of these 32 patients were very similar to those of the whole group (Table II). The slides were retrieved, reviewed, and analyzed by our pathologist (VS) for various parameters that are required to calculate RCB 10 , including • the largest two dimensions (in millimetres) of the residual tumour bed in the breast (largest tumour bed if multicentric disease). • histologic mapping of the entire largest cross-sectional area of the residual tumour bed, with specific identification of the relevant slides in the pathology report. • histologic assessment of the percentage of the tumour bed area that contains carcinoma (all carcinoma-that is, invasive and in situ), selected as one of 0%, 1%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90%. • histologic estimate of the percentage of the carcinoma in the tumour bed that is in situ, selected as one of 0%, 1%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90%. • number of positive (metastatic) lymph nodes. • largest diameter (in millimetres) of the largest nodal metastasis. Statistics Data were entered and analyzed using SPSS biostatistical software package (SPSS, Chicago, IL, U.S.A.), version 10.0.05. The study aimed primarily to evaluate the feasibility of RCB calculation from standard pathology specimens. It also aimed to study the association of RCB and pCR with EFS, defined as time to breast cancer recurrence, itself defined as local lymph node or breast recurrence, metastasis to other sites, second primary breast cancer, or any death. Analyses of OS were also performed; OS included all deaths whether they were breast cancer-related or not. Patient and disease characteristics between the different groups were compared using simple log-rank tests and Cox proportional hazards models. The EFS was considered in a multivariable setting with Cox proportional hazards models. Race, hormone receptor status [ER+/progesterone receptor positive (PR+), ER+/PR-, ER-/PR+, or ER-/PR-], chemotherapy (anthracycline, anthracycline and taxane, taxane, others, and trastuzumab), stage (IIB, IIIA, IIIB, IIIC, and inflammatory), and human epidermal growth factor receptor 2 (HER2/neu) were included in the multivariable model. The OS and EFS were estimated using the Kaplan-Meier product-limit method. The two-sided log-rank test was used to compare survival between pCR and RCB. DISCUSSION Our study highlights a few practical points pertaining to the management of LABC, including the overall treatment outcome in LABC, the need for a better classification of LABC, and the potential advantage of the RCB index as a better endpoint to measure response. With regard to treatment outcome, LABC continues to pose a significant clinical challenge, with standard available chemotherapy resulting in clinical and pathologic CRs in only a very few patients. In our study, clinical response following neoadjuvant chemotherapy did not well predict pathologic response. Of all patients, 22% achieved pCR, a result that is essentially consistent with other trials in LABC (mostly using standard anthracycline-taxane combination chemotherapy) [11][12][13][14] . None of the patients in the present study with IBC achieved pCR, indicating the aggressive-and probably distinct-nature of this disease entity that begs for novel treatment strategies. The optimal treatment algorithm, schedule, and mode of drug delivery in LABC needs to be determined. The best outcome yet reported in a randomized phase III trial in this patient population was obtained with metronomic chemotherapy given in protracted low doses 14 as reported by the Southwest Oncology Group. More research is needed to optimize treatment strategies so as to improve outcomes in LABC and IBC. A better classification of LABC is also of paramount significance. The general term LABC includes stage IIIA (T0 N2; T1/2 N2; T3 N1/2), stage IIIB (T4, N0-2), and stage IIIC (any T, N3) tumours. Some of the classical biologic prognostic factors such as size and lymph node invasion have implications that are similar in LABC to their implication in earlier disease stages. However, many others differ in LABC. For example, the prognostic significance for hormone receptor status and HER2/ neu is unclear. In the present study, among patients with hormone-responsive tumours, only 19% achieved pCR as compared with 25% for patients with ER-/PRtumours, indicating favourable chemosensitivity of hormone non-responsive tumours to neoadjuvant chemotherapy in LABC. In an evaluation of 124 patients with stage III breast cancer, Stewart and others 5 found that, among patients with inoperable tumours, ER status had no effect on prognosis. Other studies suggested that ER-tumours are more chemosensitive than ER+ tumours are [11][12][13][14][15] . In the present study, among patients with HER2/neu-positive tumours, 17% achieved pCR as compared with 25% whose tumours were HER2/ neu-negative. The patients included in our study were treated before

the use of trastuzumab became routine in the neoadjuvant setting. The foregoing result will therefore likely improve with that change in clinical practice, as did outcomes reported in operable HER2/neu-positive breast cancer with the addition of trastuzumab 16 . Overexpression of HER2/neu, otherwise known to be a poor prognostic factor, was found to be a predictor of a higher pCR with trastuzumab-based treatment 16 . It is therefore evident that reliable predictive and correlative prognostic markers for outcome are essential to the individualization and improvement of treatment outcomes in LABC. The search for such markers is by no means simplistic. It likely requires application of optimal molecular classification methods; different combinations of the various predictors are likely to lead to different prognostic entities with different treatment outcomes. Investigators at M.D. Anderson 6 and the University of Carolina 7 independently examined chemosensitivity in basal-like breast cancers, which are also known by the clinical proxy "triple negative" (ER-, PR-, HER2/neu-negative) 8 . Clinical response to neoadjuvant doxorubicin and cyclophosphamide was significantly higher among basal-like (86%) than among non-basal-like (HER2 68%, luminal 60%) breast cancers. Similarly, pCR occurred in 30% of basal-like, 27% of HER2/neu-positive, and 13% of luminal breast cancers 7 . However, basal-like breast cancers have a poor prognosis, which seems paradoxical given their sensitivity to chemotherapy. The difference in outcome appears to be a result of the more frequent early relapses seen among basal-like and HER2/neu-positive tumours that fail to achieve pCR. On the other hand, poor prognosis reflects the fewer treatment options available for ER-, PR-, and HER2/neu-negative tumours and the intrinsic biology of this subtype, which exhibits a high rate of relapse if complete eradication is not achieved and a poor outcome once relapse occurs 7 . That understanding suggests that the basal-like and HER2/neu subtypes that make up the preponderance of ER-tumours are the tumours most affected by improvements in chemotherapy. Other prognostic markers that may have different implications in LABC include tumour nuclear grade, with poorly differentiated tumours being more likely than well differentiated tumours to respond to neoadjuvant chemotherapy 9,17 . Also, increased expression of the human nuclear protein Ki-67, which is associated with cell proliferation and is used in routine pathology as a "proliferation marker" to measure the growth fraction of cells in human tumours, has been correlated with a better response to chemotherapy 18,19 . The main purpose of our study was to determine the ease and practical application of the RCB index in clinical practice as a more comprehensive and informative endpoint for residual disease following preoperative chemotherapy, based on data reported by Symmans and his colleagues at M.D. Anderson Cancer Center 10 . The identification of reliable treatment endpoints is of crucial importance. It seems logical that a good response to chemotherapy in LABC would predict for better long-term prognosis and ultimately for survival. The achievement of a pCR has been viewed as an acceptable primary endpoint for outcome following neoadjuvant chemotherapy for LABC; patients who achieve a pCR to neoadjuvant chemotherapy for breast cancer have an improved prognosis 4,20-23 . In our study, the achievement of a pCR was associated with better OS and EFS. The prognostic significance of a pCR has been confirmed recently in a large published experience including 1731 patients from M.D. Anderson 3 . However, a large trial involving 2411 patients with operable breast cancer did not show that improving the pCR affects OS significantly. The adoption of OS as a primary endpoint probably limited the ability of that trial to demonstrate a survival benefit, and OS may not have been an ideal endpoint because the trial was not powered to detect such small differences in OS or in disease-free survival 15 . Despite the focus on pCR as a surrogate endpoint in neoadjuvant trials, logic seems to suggest that non-pCR patients may derive clinical benefit from regression of the primary tumour, even if survival is not proved to be affected. Symmans et al. studied the concept of residual cancer burden (RCB) in a study including 432 patients with operable breast cancer who completed neoadjuvant chemotherapy at M.D. Anderson 10 . The RCB index was a continuous predictor of DRFS, and it predicted relapse more strongly than AJCC stage did. Our study found that RCB, defined from routine pathology materials, was easily quantifiable and appears to be better than pCR at predicting outcome after neoadjuvant chemotherapy in LABC. Higher RCB scores were significantly associated with lower EFS and a trend toward a lower rate of OS. The RCB index maintained significance as an independent predictor of EFS in the main-effects multivariate Cox regression model. What is interesting is that, in multivariate analysis, pCR did not maintain its significance as an independent predictor of EFS. That result would suggest that more meaningful prognostic implications could be derived from RCB scores in patients undergoing neoadjuvant therapy. Neither the RCB nor the pCR was statistically associated with OS-an expected result, because of the small number of patients and limited number of events in the short duration of study between 2001 and 2005, which precludes accurate OS assessment. Prospective trials are needed to further evaluate the role of RCB as an endpoint following primary chemotherapy for LABC. Because of the small number of patients and limited number of events in each group, it is not possible to draw definitive conclusions from the present study. Further analyses of other databases are required to confirm our finding of no difference in disease-free and overall survival between patients with 277 277 277 277 CURRENT ONCOLOGY-VOLUME 15,NUMBER 6 residual ductal carcinoma in situ and those with no invasive or in situ disease following neoadjuvant chemotherapy for breast cancer. CONCLUSIONS A better stratification of LABC based on specific markers is needed. The search for reliable predictive and correlative prognostic markers for outcome is essential to advance our understanding of this disease entity and consequently to improve treatment outcomes. However, the identification of reliable, informative, uniform endpoints is an essential first step that would also strengthen confidence in the value of neoadjuvant trials and anticipate the results of larger adjuvant trials. The classification of residual disease based on various pathologic responses may better classify the prognostic groups and would help to improve and individualize targeted treatment strategies. The RCB index, an easily quantifiable system, has the potential of providing a uniform method for reporting pathologic response with broad applicability. Simultaneous Integrated Boost using Conformal Radiation Therapy for Treatment of Cervical Cancer Purpose: Chemoradiation is the treatment of choice for locally advanced cervical cancer. At our institution, we have been using forward planned conformal radiation therapy to treat patients to the whole pelvis with asimultaneous integrated boost (SIB) to the uterus and parametria. Our aim is to report the local control, disease free survival, overall survival, and toxicities of definitive chemoradiation with SIB for treatment of cervical cancer. Methods: Medical records and treatment plans of patients with cervical cancer treated from 2009-2013 were reviewed using an IRB-approved database. The records of patients with cervical cancer treated with definitive chemoradiation and a three dimensional forward planned SIB were analyzed to determine local failure, distant failure, overall survival, and rate of toxicities. Results: Twenty one patients were treated with definitive chemoradiation with a SIB. Median follow up time was 18.1 months. The 2-year LC rate was 95.2%, the 2-year DFS was 80.9%, and there were no deaths, for an overall survival rate of 100%. One patient experienced Grade 3 or higher acute toxicity, and two patients experienced Grade 3 or higher late toxicities. Conclusions: This study demonstrates the feasibility and tolerability of SIB using forward planned conformal radiation therapy for the treatment of cervical cancer. This radiation technique can be used to deliver a higher dose to the area most at risk for recurrence in a shorter treatment time *Corresponding Author: Waqar Haque, MD, Deparment of Radiation Oncology, Greater Houston Physicians in Medicine, Association Houston, TX 77030, Tel: 832-367-1655/ Fax: 832-201-0602; E-mail: [email protected] Received Date: February 02, 2016 Accepted Date: April 21, 2016 Published Date: April 27, 2016 2 SIB using forward planned conformal radiation therapy for definitive chemoradiation for cervical cancer has not previously been explored. At our institution, in select patients, we have used a field-in-field technique to treat the whole pelvis with a SIB to the known areas of gross disease, cervix, and uterus. The purpose of this study is to describe the forward planned SIB technique and to report the clinical outcomes for patients with cervical cancer treated with forward planned SIB. Materials and Methods This analysis was approved by the institutional review board at the University of Texas Medical Branch and each of its affiliated hospitals. To determine the outcomes of patients with cervical cancer treated with definitive chemoradiation and SIB, we reviewed the records of all patients with cervical cancer treated with radiation therapy at the University of Texas Medical Branch and affiliated hospitals from January 2009 to June 2013. We limited our analysis specifically to the 21 patients who received definitive chemoradiation with the SIB technique. These patients' records and imaging were retrospectively reviewed to determine patient characteristics, tumor characteristics, clinical toxicities, and patient outcome. All patients underwent computed tomography (CT) Abdomen/Pelvis, Positron Emission Tomography (PET)/CT, and magnetic resonance imaging (MRI) for workup, and this information was used for treatment planning. Staging, however, was assigned based on clinical exam and plain films, per International Federation of Gynecology and Obstetrics (FIGO) guidelines. The primary endpoint was local control following chemoradiation. The response was determined by visual inspection, physical exam and palpation, or radiographic study including CT or MRI. The whole pelvis was treated using a 3-dimensional (3D) conformal technique to a dose of 45 Gy in 25 fractions (1.8 Gy per fraction), with the exception of one patient, who was treated to a dose of 50.4 Gy in 28 fractions (1.8 Gy per fraction). For the anterior / posterior field, the superior border was the L4-L5 interspace, the inferior border was 3 cm inferior to gross disease or the obturator foramen, whichever was more inferior, and the lateral border was 2 cm lateral to the pelvic rim. The lateral fields shared the same superior and inferior borders as the A/P field, were bounded posteriorly by the sacrum, and was bordered anteriorly by the pubic symphysis. The SIB was utilized to treat the gross disease as visualized on PET/CT and MRI, uterus, and remnant of the cervix, with a 1.5 cm margin, named the CTV boost, upon which a 5 mm margin was applied to construct the PTV boost, to achieve a total dose of 50 Gy in 25 fractions (2 Gy per fraction) when the whole pelvis was treated to 45 Gy. The one exception was in a patient treated to 50.4 Gy to the whole pelvis, where SIB achieved a dose of 54 Gy in 28 fractions (1.93 Gy per fraction) to the PTV boost. To accomplish this, initially the entire pelvic field was prescribed to be treated to 50 Gy, after which field-in-field technique was performed to block the pelvic volume not enclosed by the PTV boost to treat this region to a dose of 45 Gy (Figures 1-3). High dose rate (HDR) intracavitary brachytherapy was used in each patient to assure a normalized equivalent dose (EQD2) of 80-90 Gy. HDR brachytherapy was started in the 4 th week of treatment, and each patient received 3 to 6 HDR insertions, 5 to 7.5 Gy per fraction. All patients were treated using tandem and ring with CT confirmation of placement. Relevant points as determined by ICRU 38 and the American Brachytherapy Society (ABS) guidelines included point H, bladder, and rectal points [16] . In patients with nodal involvement, IMRT boost following completion of brachytherapy was used to take the nodal dose to 60-66 Gy. All patients were treated with concurrent weekly cisplatin 40 mg/m2, median 5 cycles given (range 2-6), with the determination of the number of cycles of chemotherapy administered determined by the treating gynecologic oncologist. Data for age at treatment, stage, outcome, radiation dose, systemic therapy, and toxicity were retrospectively tabulated for all patients. Survival was defined as time from date of diagnosis to date of death or last follow-up (in months). We performed Kaplan-Meier analyses to provide estimates of local control (LC) and disease free survival (DFS)

for all patients. A local failure was defined as any evidence of tumor recurrence or progression within the pelvis discovered either on physical exam or on surveillance imaging by CT or PET scan. Data analysis was performed using SAS software version 9.3 (SAS Institute, Inc., Cary, NC). Toxicities were categorized by the Common Terminology Criteria for Adverse Events (CT-CAE) version 4.0. All radiation therapy was delivered using a Varian linear accelerator with 6 MV and 18 MV photons (Varian Medical Systems, Palo Alto, California). Treatment planning was done using Pinnacle (Pinnacle Systems, Mountain View, California) or Eclipse Treatment Planning System (Varian Medical Systems, Palo Alto, California). Results Twenty one patients with cervical cancer were treated with radiation therapy at the University of Texas Medical Branch and affiliated hospitals between January 2009 -June 2013 and were treated with definitive chemoradiation with a SIB. Two (9.5%) of these patients were American Joint Committee on Cancer (AJCC) stage IB1, six (28.6%) were stage IIB, twelve (57.1%) were stage IIIB, and one (4.8%) was stage IVA. Four (19%) of the patients had nodal disease discovered on imaging by PET / CT. Every patient included in the present analysis had squamous cell carcinoma. For further information regarding patient characteristics, please see Table 1. Among the patients with IB1 disease who received chemotherapy, one of these patients had evidence of PET positive external iliac and common iliac nodal disease, as a result of which our group recommended concurrent chemotherapy along with definitive radiation therapy. The second patient with IB1 disease who was treated with concurrent chemotherapy was found to have disease limited to 2.5 cm on clinical exam, but was found to have extension of PET avid disease in the lower uterine segment, suggesting the size of her tumor was significantly larger than the exophytic mass visualized on clinical exam, and as such was deemed to be an appropriate candidate for chemoradiation. Twenty (95.2%) of the patients were treated to the whole pelvis using a 3D conformal technique to a dose of 45 Gy in 25 fractions (1.8 Gy per fraction), with a SIB to the clinical gross disease to 50 Gy in 25 fractions (2 Gy per fraction). One 4 After a median follow up time of 18.1 months (range 3-66 months), two patients recurred locally, at four months and at twenty-nine months after completion of treatment, respectively. Three patients developed distant metastases. One patient failed in the lung at three months, one in the para-aortic nodes at four months, and one in the left supra clavicular node at 20 months, respectively. The 2-year LC rate was 95.2%, the 2-year DFS was 80.9%, and there were no deaths, for an overall survival rate of 100%. Kaplan-Meier curves for LC and DFS are displayed in Figure 4. Overall, radiation therapy was well tolerated in our patient population. Five (23.8%) patients experienced no acute toxicities, twelve (57.1%) patients experienced Grade 1 acute toxicities, three (14.3%) patients experienced Grade 2 acute toxicities, and one (4.7%) patient experienced Grade 3 acute toxicities in two distinct organ systems ( Discussion The present study examined the outcomes of treating patients with cervical cancer with concurrent chemotherapy and EBRT using forward planned SIB using a field-in-field technique and HDR brachytherapy, followed by an IMRT boost if needed for PET avid lymph nodes. We demonstrated that EBRT using an SIB technique results in excellent local control, with 2-yr LC rate of 95.2%, in a group of patients with predominantly locally advanced cervical cancer ( > 75% had FIGO stage IIB or higher disease). These local control rates were achieved with low toxicity rates, with 4.7% of patients experiencing grade 3 or higher acute toxicity, and 9.4% of patients experiencing grade 3 or higher late toxicity. This suggests that using EBRT with forward planned SIB is a safe technique that leads to excellent local control. This is to our knowledge the first report describing SIB using forward planned conventional treatment for definitive chemoradiation to treat cervical cancer. Three groups have previously reported on the use of SIB with EBRT for definitive chemoradiation for treatment of cervical cancer, each of which utilized IMRT to achieve the SIB. The use of SIB in treatment of cervical cancer was first described by Vandecasteele et al. in 2009 [11] . This was a review that described the planning procedure and feasibility of using Intensity Modulated Arc Therapy (IMAT) in six patients with unresectable cervical cancer in which the outcomes were not reported. In a subsequent prospective trial, the same group treated 30 patients with cervical cancer with IMAT and SIB and chemotherapy followed by hysterectomy [17] . There were no grades, 3 or higher early toxicities attributed to radiation therapy, and five (16.6%) grade 3 or higher late toxicities. The 2-year local control rate was 96%, and 2-year distant control rate was 92%. A German group described treatment of 40 patients with cervical cancer using helical tomotherapy with SIB and chemotherapy and HDR brachytherapy followed by curettage [18,19] . Acute Grade 3 diarrhea occurred in 2.5% of patients, acute Grade 3 hematologic toxicity occurred in 20% of patients, and 2/40 patients had residual disease following curettage. Cihoric et al. described the outcomes of 10 patients with lymph node positive cervical cancer treated with IMRT and SIB followed by HDR brachytherapy [20] . At a median follow up time of 20 months, the local control rate was 90%, 10% of patients developed grade 3 acute toxicities, and 10% of patients developed grade 3 late toxicities. Patients treated in the present trial with a conformal forward planned SIB have similar rates of local control and acute and late toxicity, without the use of IMRT. Another group evaluated the use of a SIB with chemoradiation in the neo adjuvant setting for locally advanced cervi-cal carcinoma for 32 patients in a phase I dose escalation trial [21] . These patients received concurrent cisplatin and 5FU, following which clinical responders underwent radical hysterectomy and lymphadenectomy. The rate of G3 or higher acute toxicities was up to 60% in the arm with the highest dose. No patients in this trial reported Grade 3 or higher late toxicities. With a median follow up time of 18 months, the LC was 68.7%, DFS was 63.5%, and OS was 92%. It may be difficult to directly compare the toxicity rates of this trial and our present report due to the absence of brachytherapy use and inclusion of surgery. Nevertheless, the rate of Grade 3 or higher acute toxicities in both Group 2 (25%) and Group 3 (60%) were higher than the rate of Grade 3 or higher acute toxicities in the present trial (4.7%), with the reported LC rate of 68.7% lower than the LC in the present trial of 95.2%, suggesting the fractionation scheme used in the present trial plus HDR Brachytherapy may be superior in terms of both toxicity profile and local control. Based on 5 randomized trials, there is evidence for a significant overall survival benefit for treatment of cervical cancer with the concurrent chemotherapy and radiation therapy when compared to radiotherapy alone [22] . The two largest trials that demonstrated improved outcomes with definitive treatment with concurrent cisplatin based chemoradiation over radiation therapy alone were the Radiation Therapy Oncology Group (RTOG) 90-01 and the Gynecologic Oncology Group (GOG) 120 (23)(24)(25)(26). In RTOG 90-01, 403 patients with bulky IB to IVA cervical cancer were randomized to receive treatment with external beam radiation with or without with concurrent cisplatin and 5FU, with both arms receiving intracavitary brachytherapy (67% of these patients had IIB-IVA disease). Loco regional failure (18% vs. 34%), DFS (61% vs. 36%), and OS (67% vs. 41%) all favored the chemoradiation arm. The rate of Grade 3 or higher acute toxicity was 45% in the chemoradiation arm, with the rate of non-hematologic acute toxicity being 11%, and the rate of long term Grade 3 or higher toxicity was 13% [23,24] . In GOG 120,526 patients with IIB, III, or IVA cervical cancer were randomized to receive one of three chemotherapy regimens: cisplatin alone; cisplatin, 5FU, and hydroxyurea; or hydroxyurea alone along with radiation therapy [25,26] . At ten years, the rates of progression free survival (46%) and OS (53%) favored the arms containing cisplatin [27] . The rate of grade 3 or higher long term toxicity was 4.7% for the cisplatin alone arm, 0.9% for the arm of patients receiving three drug treatment, and 2.1% for those received hydroxyurea alone. The rates of acute and late toxicity in the present study compare favorable to those reported in these historical controls, further suggesting that the SIB can be safely used without causing significant toxicity. Addtionally, rates of LC and OS in the present study compare favorably to the historical controls, demonstrating the efficacy of the SIB. There are radiobiologic reasons to explain why treatment with SIB may potentially be more effective in treating patients with cervical cancer than conventionally fractionated radiation treatment followed by a cone down boost. Using SIB allows the treatment time to be decreased, limiting the effect of accelerated tumor repopulation [28] . Prolonging overall treatment time of patients with cervical cancer being treated with radiation therapy alone or with chemoradiation leads to worse cancer control outcomes [29][30][31][32][33][34] . Moreover, using the SIB treats the tumor to a slightly higher biologically effective dose (BED). Using the equation to calculate BED when taking into account solely the external beam radiation portion of treatment, [BED= nd{1+d/ (α/β)} -ln 2 (T-T k ) / (α*T p )] demonstrates that when using SIB to treat to 50 Gy in 25 fractions, the BED to the tumor is 56.92 Gy (assuming a α/β ratio of 10 for cervical cancer, T k of 19 days, T p of 4.5 days, and α of 0.3 / Gy), where as using conventional fractionation to treat to 45 Gy followed by a 5.4 Gy boost for a total of 28 daily fractions, the BED to the tumor is 54.95 Gy. The combination of the decrease in treatment time along with the increase in BED may contribute to greater tumor kill and consequently superior oncologic outcomes. The present study describes the use of SIB with forward planned conventional radiation therapy. Previous trials that described the use of SIB as part of definitive chemoradiation for management of cervical cancer used inverse planned IMRT. IMRT requires precise delineation of the target during and between treatments, which can be challenging in the case of the cervix. Studies have demonstrated that interfraction uterine motion can be up to 48 mm, and interfraction cervical motion can be up to 19 mm [34] . Population based PTV margins may be larger than needed for some patients, which can cause unnecessary radiation treatment to organs at risk. Though individualized margins may be a potential solution, an adaptive strategy has yet to be clinically validated [15] . Therefore, due to positional uncertainty of the cervix which necessitates large PTV margins, the conformality typically achieved with IMRT may not be applicable in the case of cervical cancer. Additionally, conventional radiation is associated with a decreased cost of delivery and decreased planning time. This result is especially important given the increased emphasis on decreasing health care related costs while maintaining high quality treatment. Using 2015 Medicare reimbursement rates, a 25 fraction treatment for cervical cancer using IMRT would cost approximately $12700, whereas a 25 fraction treatment regimen using 3D conformal treatment would cost approximately $4800, with a resultant net savings of $7900 per patient [35] . When extrapolated to the entire US population, treatment of all patients diagnosed with cervical cancer with 3D conformal treatment as opposed to IMRT would result in savings of $97,170,000 annually. As the present study suggests, SIB using forward planned 3D conformal radiation achieves similar oncologic outcomes to SIB using IMRT while maintaining low toxicity, which may make it a more efficient, cost-effective treatment. Our institutional data demonstrates that use of conventional 3D conformal radiation with forward planned SIB technique can achieve excellent local control with a reasonable toxicity profile. We did see a small number of Grade 3 acute and late toxicities. There was one patient who experienced who experienced Grade 3 acute nausea / vomiting which may have been partially due to chemotherapy, cisplatin. Two patients experienced Grade 3 or higher late

toxicities, one of whom experienced Grade 3 rectal bleeding, successfully managed with hyperbaric oxygen therapy. This patient had a history or rheumatoid arthritis and was on Methotrexate, and was taking Methotrexate while receiving radiation treatment. Methotrexate is known to cause intestinal mucositis and gastrointestinal bleeding, and this may have contributed to the rectal bleeding experienced by this patient [36] . It is possible that the SIB technique should be avoided inpatients on agents that damage the gastrointestinal mucosa, as this may predispose patients to significant risk for severe toxicity. 6 This study did have some weaknesses. Our cohort was a small, single-institution cohort, which may limit the generalizability of the findings. However, there was a good distribution of patients with different stages of cervical cancer within this cohort, which may make the results from this study generalizable to the general population. This was a retrospective study, which may lead to the possibility of underreporting of toxicities. Each patient was followed up closely by gynecologic oncologists and radiation oncologists specializing in treating gynecologic cancers, and as such we would expect any underreporting of toxicities to be minimal. Finally, this study had a limited follow up time, with a median follow up time of 18.1 months. Despite the limited follow up time, the excellent local control and disease free survival outcomes are significant, although longer follow up will be needed to determine the true effect of the SIB. Conclusion The present study described the technique and demonstrates the safety and efficacy of using SIB with forward planned 3D conformal radiation treatment for definitive chemoradiation for cervical cancer. This technique leads to excellent local control with minimal acute and long term toxicities. Oncologic outcomes are at least comparable to historical controls, and may be superior, due to decreased treatment time and a higher BED to the tumor. Longer follow up time will be needed to confirm the outcomes with treatment with SIB using 3D conformal radiation treatment. Conflict of Interest: There are no conflicts to disclose. Thalassaemia is paradoxically associated with a reduced risk of in‐hospital complications and mortality in COVID‐19: Data from an international registry Abstract Although numerous patient‐specific co‐factors have been shown to be associated with worse outcomes in COVID‐19, the prognostic value of thalassaemic syndromes in COVID‐19 patients remains poorly understood. We studied the outcomes of 137 COVID‐19 patients with a history of transfusion‐dependent thalassaemia (TDT) and transfusion independent thalassaemia (TIT) extracted from a large international cohort and compared them with the outcomes from a matched cohort of COVID‐19 patients with no history of thalassaemia. The mean age of thalassaemia patients included in our study was 41 ± 16 years (48.9% male). Almost 81% of these patients suffered from TDT requiring blood transfusions on a regular basis. 38.7% of patients were blood group O. Cardiac iron overload was documented in 6.8% of study patients, whereas liver iron overload was documented in 35% of study patients. 40% of thalassaemia patients had a history of splenectomy. 27.7% of study patients required hospitalization due to COVID‐19 infection. Amongst the hospitalized patients, one patient died (0.7%) and one patient required intubation. Continuous positive airway pressure (CPAP) was required in almost 5% of study patients. After adjustment for age‐, sex‐ and other known risk factors (cardiac disease, kidney disease and pulmonary disease), the rate of in‐hospital complications (supplemental oxygen use, admission to an intensive care unit for CPAP therapy or intubation) and all‐cause mortality was significantly lower in the thalassaemia group compared to the matched cohort with no history of thalassaemia. Amongst thalassaemia patients in general, the TIT group exhibited a higher rate of hospitalization compared to the TDT group (p = 0.001). In addition, the rate of complications such as acute kidney injury and need for supplemental oxygen was significantly higher in the TIT group compared to the TDT group. In the multivariable logistic regression analysis, age and history of heart or kidney disease were all found to be independent risk factors for increased in‐hospital, all‐cause mortality, whereas the presence of thalassaemia (either TDT or TIT) was found to be independently associated with reduced all‐cause mortality. The presence of thalassaemia in COVID‐19 patients was independently associated with lower in‐hospital, all‐cause mortality and few in‐hospital complications in our study. The pathophysiology of this is unclear and needs to be studied in vitro and in animal models. patients with no history of thalassaemia. The mean age of thalassaemia patients included in our study was 41 ± 16 years (48.9% male). Almost 81% of these patients suffered from TDT requiring blood transfusions on a regular basis. 38.7% of patients were blood group O. Cardiac iron overload was documented in 6.8% of study patients, whereas liver iron overload was documented in 35% of study patients. 40% of thalassaemia patients had a history of splenectomy. 27.7% of study patients required hospitalization due to COVID-19 infection. Amongst the hospitalized patients, one patient died (0.7%) and one patient required intubation. Continuous positive airway pressure (CPAP) was required in almost 5% of study patients. After adjustment for age-, sexand other known risk factors (cardiac disease, kidney disease and pulmonary disease), the rate of in-hospital complications (supplemental oxygen use, admission to an intensive care unit for CPAP therapy or intubation) and all-cause mortality was significantly lower in the thalassaemia group compared to the matched cohort with no history of thalassaemia. Amongst thalassaemia patients in general, the TIT group exhibited a higher rate of hospitalization compared to the TDT group (p = 0.001). In addition, the rate of complications such as acute kidney injury and need for supplemental oxygen was significantly higher in the TIT group compared to the TDT group. In the multivariable logistic regression analysis, age and history of heart or kidney disease were all found to be independent risk factors for increased in-hospital, all-cause mortality, whereas the presence of thalassaemia (either TDT or TIT) was found to be independently associated with reduced all-cause mortality. The presence of thalassaemia in COVID-19 patients was independently associated with lower in-hospital, all-cause | INTRODUC TI ON Since publication of the first reports on COVID-19, the highest incidence and mortality rates have been reported in the USA, Italy, Spain, China and France, 1 with nearly 2.2 million deaths and more than 103,000,000 infections having been reported worldwide to date. [2][3][4][5] Although patients may present with symptoms such as fever, myalgia, fatigue, dry cough and/or diarrhea, 6 asymptomatic patients are not rare. Different models have been developed to date to analyse the severity of COVID-19, with numerous co-factors having been reported as predictors of worse outcomes such as arterial hypertension, diabetes mellitus, heart disease, acute and chronic kidney disease, chronic pulmonary disease, advanced age, immunosuppression and inherited haemoglobin disorders. [7][8][9][10][11] Few data have been published to date on COVID-19 and thalassaemia syndrome and the existing data are still controversial due to the small number of patients included. 8,[12][13][14] Thalassaemia patients may be at high risk for infectious complications leading to increased mortality and morbidity. This may be due to splenectomy, adrenal insufficiency, stem cell transplantation and/or stress erythropoiesis. 15 Despite advances in the understanding of the pathophysiology of thalassaemia over the last decade, management of this disease has remained largely unchanged over the years. 16 In this present study, we described the clinical characteristics and outcomes of a cohort of COVID-19 patients suffering from concomitant thalassaemia using data from a large, international observational study and compared them to a matched cohort with no history of thalassaemia. | Thalassaemia The data set included different clinical forms of β-thalassaemia, including the co-inheritance of β-thalassaemia with haemoglobin E resulting in haemoglobin E/β-thalassaemia. Patients were managed according to the current recommendations for the treatment of thalassaemia. 14 | Outcomes The primary outcome was in-hospital, all-cause mortality. In addition, in-hospital complications including admission to an intensive care unit (ICU), as well as the type of respiratory support needed (nasal cannula, non-invasive mechanical ventilation, invasive mechanical ventilation or invasive mechanical support), were also evaluated in the outcome analysis. | Data extraction Epidemiological, clinical and outcome data were collected from the electronic medical records. The confidentiality of patient data was protected by storing them in an anonymous manner on a locked, mortality and few in-hospital complications in our study. The pathophysiology of this is unclear and needs to be studied in vitro and in animal models. | Statistical analysis Continuous variables with a normal distribution are shown as mean ± standard deviation, continuous variables with a non-normal distribution are shown as a median (interquartile range) and categorical variable are shown as a frequency (% Table 1. Table S1 presents the cohort of 134 patients separated into a TIT and a TDT group. The information regarding a history of transfusion was missing in three patients. Patients in the TIT group were in general older than in the TDT group (54.7 ± 18.6 years versus 38 ± 14, p < 0.001) and hospitalized more often (p = 0.001). Symptoms such as dyspnoea, fatigue, fever and cough were significantly more prevalent in the TIT group compared to TDT group. In-hospital complications were similar between groups except with regard to the need for supplemental oxygen and acute kidney injury, both of which were more common in the TIT group. | In-hospital complications 27.7% of the thalassaemia patients were hospitalized due to COVID-19 and 18% of patients required admission to an intensive care unit (ICU). The main presenting symptoms were as follows: | Comparison of outcomes in patients with and without a history of thalassaemia In the propensity score matching, 127 COVID-19 patients with concomitant thalassaemia were matched to 127 patients suffering from COVID-19 with no history of thalassaemia. Patients with no history of thalassaemia exhibited significantly higher rates of hospitalization than those with a history of thalassaemia as seen in Table 3 | Predictors of in-hospital, all-cause mortality in COVID-19 We matched the thalassaemia patients with COVID-19 patients in absence of thalassaemia to assess for predictors of in-hospital, all-cause mortality. In the univariate analysis, male gender, age, thalassaemia, pulmonary disease, cardiac disease, kidney disease and diabetes were all found to predict all-cause mortality. Using a multivariable logistic regression analysis ( | DISCUSS ION This present study examined patient characteristics, in-hospital complications and outcomes of COVID-19 patients with and without a history of thalassaemia. The main findings of our study were as follows: (1) Thalassaemia is an independent predictor of reduced in-hospital, all-cause mortality compared to patients with no history of thalassaemia. To date, various co-factors have been identified as being associated with increased mortality in COVID-19 patients, amongst these cancer and hemoglobinopathies. 6 With regard to thalassaemia, the data from a small number of case series and case reports have been In this Iranian cohort, the splenectomy rate was at least 80%. 11 On the other hand, preliminary data from 11 thalassaemia patients from different Italian centres hinted at milder outcomes in COVID-19 disease with all patients recovering completely. 13 A different study from centres in 10 different countries summarized data from at least 10 beta-thalassaemia patients. In this study, one patient (10%) died despite respiratory support and intensive care unit stay. 8 Only 40% of these patients had a history of splenectomy. The mortality rate of 0.7% in our study seems to be very low in comparison, but included a far greater sample size. Approximately 40% of patients in our study had a history of splenectomy. This is comparable to the recently published study, however, with a lower mortality rate. 13 We compared data from thalassaemia patients with COVID-19 to a matched cohort of COVID-19 patients with no history of thalassaemia and, seemingly paradoxically, found a significantly lower inhospital, all-cause mortality rate in the thalassaemia cohort and a lower rate of in-hospital complications. Although the rate of liver disease was higher in thalassaemia patients compared to the age-, gender-and cardiovascular disease-matched group of COVID-19 patients, the mortality rate was much lower. Of note, recently published data described a worse prognosis in patients with liver disease, and reported a mortality rate of approximately 15%. 17 Despite the higher prevalence of liver disease in our thalassaemia cohort compared to patients with no history of thalassaemia, the outcomes were still better in patients with thalassaemia. In a nationwide Iranian cohort, 36.4% had a heart and liver iron overload. Our present data showed an overload of iron in the

heart and liver in 6.8% and 35%. These data appear to confirm a lower risk of mortality in thalassaemia patients. In our present study cohort, 99% of patients recovered from COVID-19. Whereas the need for intubation was documented in only one patient, non-invasive respiratory support was required in seven patients in our study. Compared to the matched cohort, however, these rates were still significantly lower than in the non-thalassaemia group. The mild courses of COVID-19 in the thalassaemia cohort may be related to different factors. Recently published data showed that ORF8 and surface glycoproteins of COVID-19 could form a complex together with porphyrin. In addition, orflab, ORF10 and ORF3a proteins could attack the porphyrin, which is formed in the mitochondria on the 1beta chain of haemoglobin. 18,19 This may lead to dissociation of iron from porphyrin. Genetic mutations may result in the absence of beta chain synthesis while retaining normal alpha chain synthesis. This could overall lead to less haemoglobin A being produced. Since the beta chain s a potential target for COVID-19 infection is either absent or expressed at a much lower rate in the blood of thalassaemia patients, this may lead to decreased susceptibility to COVID-19 infection in thalassaemia patients. 17 Of note, a history of regular blood transfusions was documented more than 80% of our thalassaemia patients. But nevertheless, in vitro and animal studies are required to confirm this hypothesis. It might be speculated that in transfusiondependent patients when pretransfusion haemoglobin goal is 9-10 g/ dl, their bone marrow is supposed to be suppressed by transfusions therefore the majority of haemoglobin A is from normal transfused red cells and less produced by patients own abnormal red blood cells. It could be that protecting red cells from COVID-19 will be irrelevant in TDT patients but could apply to TIT patients. Compared to the matched cohort, COVID-19 symptoms were significantly lower amongst thalassaemia patients. In the Iranian thalassaemia registry, the prevalence of COVID-19 was 0.14% in the thalassaemia patients. On this point, we cannot provide any data regarding the prevalence of COVID-19 in thalassaemia patients since we included all patients with COVID-19 and did not analyse the number of thalassaemia patients in each and every centre. | CON CLUS IONS Analysis of the data of this large international cohort revealed a lower all-cause mortality rate in thalassaemia patients compared to a matched cohort with no history of thalassaemia. In-hospital complications relating to COVID-19 infection were also significantly lower in thalassaemia patients, although the hospitalization rate was 27.7%, which was significantly lower than in the matched cohort. | Limitations Data for this study were collected from several centres and bias regarding the heterogeneous diagnostic and treatment strategies between them cannot be excluded. In addition, there are differences in the capacity and other medical resources available in these centres. Although the data summarized in this study came from an international registry, the majority of patients included were from European countries (Italy and Spain) and outcomes may differ due to ethnic factors and standard of care in other parts of the world. Design and methods in a survey of living conditions in the Arctic – the SLiCA study Objectives The main objective of this study is to describe the methods and design of the survey of living conditions in the Arctic (SLiCA), relevant participation rates and the distribution of participants, as applicable to the survey data in Alaska, Greenland and Norway. This article briefly addresses possible selection bias in the data and also the ways to tackle it in future studies. Study design Population-based cross-sectional survey. Methods Indigenous individuals aged 16 years and older, living in Greenland, Alaska and in traditional settlement areas in Norway, were invited to participate. Random sampling methods were applied in Alaska and Greenland, while non-probability sampling methods were applied in Norway. Data were collected in 3 periods: in Alaska, from January 2002 to February 2003; in Greenland, from December 2003 to August 2006; and in Norway, in 2003 and from June 2006 to June 2008. The principal method in SLiCA was standardised face-to-face interviews using a questionnaire. Results A total of 663, 1,197 and 445 individuals were interviewed in Alaska, Greenland and Norway, respectively. Very high overall participation rates of 83% were obtained in Greenland and Alaska, while a more conventional rate of 57% was achieved in Norway. A predominance of female respondents was obtained in Alaska. Overall, the Sami cohort is older than the cohorts from Greenland and Alaska. Conclusions Preliminary assessments suggest that selection bias in the Sami sample is plausible but not a major threat. Few or no threats to validity are detected in the data from Alaska and Greenland. Despite different sampling and recruitment methods, and sociocultural differences, a unique database has been generated, which shall be used to explore relationships between health and other living conditions variables. T he Survey of Living Conditions in the Arctic: Inuit, Sami, and the Indigenous Peoples of Chukotka (SLiCA) is an international research project on health and other aspects of the living conditions of Indigenous peoples in Alaska, Canada, Greenland, Norway, Sweden and Russia. The motivation for launching SLiCA was the ambition to describe these aspects with regard to indigenous language, tradition and resource utilisation. The background of SLiCA is described in detail elsewhere (1Á3). The sampling methods used in the survey have been different due to country/regional variability in indigenous population density and register availability. The scope of this article does not allow in depth descriptions of the methods used in the involved countries/regions; we shall thus focus on the data from Alaska, Greenland and Norway, as these indigenous populations are integral to the future research questions of the first author. The main objective of this article is to describe the methods and design of SLiCA, relevant participation rates and the distribution of participants, as applicable to the survey data in the aforementioned countries/regions. We also briefly address possible selection bias in the data and describe how to tackle it in future studies. Material and methods The principal method in all SLiCA countries was standardised face-to-face interviews using a questionnaire. The questionnaire may be accessed on the project website (4) individuals aged 16 years and older (15 years and older in Canada and Greenland), residing in traditional settlements. Ethnic background was ascertained by self-report. The duration of each interview in Alaska, Greenland and Norway was approximately 1.5Á2 hours, and the respondents were almost exclusively interviewed in their homes. Data collection and sampling Alaska Data collection took place from January 2002 to February 2003. In Alaska, we did not have access to the U.S. Census 2000 population lists. The sample frame consisted of 4 components, that is, regions and communities, blocks, housing units and individuals. The Alaska sample is a probability multi-stage sample (5). The Iñupiat regions of Northwest Arctic (NA), North Slope (NS) and Bering Strait (BS) were selected in advance. In each of the 3 regions, one started with 2 strata, that is, regional centres and villages. The regional centres of Kotzebue (NA), Barrow (NS) and Nome (BS) were included. Villages in NA and BS were sampled and stratified as coastal or inland. All villages on the NS were included, as there are only 8 of them. In the regional centres, one applied a 2-stage area probability-sampling approach. First, a probability sample of blocks with probabilities proportionate to the number of Iñupiat households was selected. Second, a probability sample of Inuit households in each sample block was done. A local Inuit colleague identified the Inuit households in the sample blocks. Finally, Iñupiat adults, within each sampled household, were sampled according to the person with the next birthday. We observed a bias in favour of females that was addressed as a final sampling weight. According to the U.S. Census 2000, a total of 4,581, 3,082 and 3,505 persons lived in the regional centres of Barrow, Kotzebue and Nome, respectively. The total population number in the villages varied between 136 in Deering and 772 in Selawik (6). In the villages, the American Indians/Alaska Natives (AIAN) make up close to 100% of the population. In Barrow, Kotzebue and Nome, 64%, 77% and 59% of the population reported AIAN ethnicity, respectively. Here and in the villages, the AIAN category almost exclusively refers to the people of Iñupiaq ethnicity. The total Iñupiaq population in Alaska is estimated at 13,500Á15,700 (6Á9). Greenland is home to about 57,000 people, of which approximately 90% are Inuit (12). In terms of ethnic categorisation, the Greenlandic population may be di-vided according to the place of birth, that is, in or outside Greenland. For the adult population, this variable roughly refers to an ethnic categorisation of Greenlanders and Danes, which may be ascertained in interview settings (10). On basis of the official regional division by Greenland Statistics, 8 municipalities and their main towns were selected in advance. The main towns were Nanotarlik, Qaqortoq, Paamiut, Nuuk, Aasiaat, Ilulissat, Upernavik and Tasiilaq. Villages were chosen at random in the selected municipalities. In the selected towns and villages, a random sample of persons born in Greenland was drawn from the population register. As a minority of Greenlanders live in smaller settlements of less than 500 inhabitants (17% in 2005) (11,12), a greater sample weight was given to this population (3). In 2006, the total population in the main towns varied from 1,133 inhabitants in Upernavik to 14,583 in Nuuk, and in the villages, from 47 in Saarloq to 404 in Kullorsuaq and Kuummiut 1 (13). Sampling in Greenland is also described elsewhere (3). Sami respondents in Finnmark were selected through the representative database of the Population-based Study of Health and Living Conditions in Areas with both Sami and Norwegian Populations (SAMINOR). This study was run by the Centre for Sami Health Research and the Norwegian Institute of Public Health in 2003Á2004 and is described in detail by Lund et al. (14). A random sample was drawn from the sample frame of all SAMINOR participants in Kautokeino, Karasjok and Nesseby, who reported Sami ethnicity and gave consent to be contacted in future studies. This method was unavailable in Sami settlement areas, south of Finnmark, as permission to contact these participants was not obtained during SAMINOR. Instead, a nonprobability snowball sampling technique (15,16) was applied to list Sami living in Sami settlement areas in Troms, Nordland and the Trøndelag counties. From this sample frame random samples were drawn. This method was also applied in Finnmark to recruit individuals in the youngest age strata, as SAMINOR only included participants aged 30 and 36Á79 years in 2003Á2004. Sticking to a random sampling became challenging in areas, where 1 In Greenland, town status is not determined by population size but by the presence of the municipality headquarter, a hospital or health centre and a school (12). Bent-Martin Eliassen et al. the Sami population is a minority and lives scattered across great distances. The South Sami area is one such example. Because of funding issues, a scattered population structure and the few Sami living in each community, we had to interview a certain number of persons in each place to reach an adequate total number of completed interviews. Multi-stage probability sampling was not possible for the same reasons. Except for Røros (n05,683), all the municipalities and communities had, in 2008, less than 3,000 inhabitants (17). Among these, the Sami were majority only in Kautokeino, Karasjok and Nesseby. There is no updated demographic record on the Sami, but the population in Norway is usually and roughly estimated at 40,000 (18,19). Logistics The Iñupiat of Alaska inhabit the northern and western coasts as far south as Norton Sound (7), and most of the communities cannot be reached by car or ferry (20). Interviewers travelled by car within the regional centres, while all respondents lived within walking distance in the villages. Respondents were contacted by house visits, and the interviewer gave a brief description of the study to the person answering the door and asked to speak to the person who had the next birthday. If that person was not available, contact information (e.g. phone numbers) would then be obtained and attempts were made to contact the selected person. Those who failed to attend scheduled interviews were contacted to reschedule. The majority of the Inuit in Greenland are concentrated on the west coast and in

the south, with only 3,500 living on the east coast and less than 1,000 in the far north (11,12,21). In Greenland, the towns and villages are isolated from one another and can only be reached by boat or flight (21). As in Alaska, cars were used for transport in the towns, while interviewers could walk to interview appointments in the villages. Selected individuals were contacted and invited to participate by phone. If contact was not established by phone, interviewers would contact the person at home. Those who failed to attend scheduled interviews were re-contacted and new interviews were planned. The traditional Sami settlement area in Norway stretches from Finnmark in the north to Engerdal in Hedmark County in the south, with the majority of the Sami population settled in Finnmark (19). All communities are reachable by car. In Norway, invitation to participate in the study was presented in 2 ways. First, SAMINOR-sampled individuals in Finnmark received a letter of invitation, containing information on the study, a written consent form and a return envelope. The recipients were asked to return the signed consent form and provide their telephone number. Those who consented were contacted by phone to schedule the time and date for the interview. Those who did not return the consent form were tried contacted over the phone, if their number was accessible. Second, south of Finnmark, people were invited by phone only. During the phone conversation, the study was presented, and if preliminary consent was obtained, time and place of the interview were fixed. Those who failed to attend scheduled interviews were contacted by phone to reschedule. Questionnaire The core questionnaire consisted of 4 parts: the main questionnaire, 3 household charts Á intended to facilitate responses on questions concerning household members Á and a self-administered questionnaire used for sensitive questions. Finally, cue cards were used to efficiently present respondents with response choices. The core questionnaire was produced in collaboration with indigenous representatives and field tested in all countries/ regions. English was used as a common language for questionnaire development. Country/region-specific questions were produced to address issues, items and perspectives relevant to the respective country/region. All fieldworkers in SLiCA were trained in interviewing techniques and procedures. An interview guide was produced to optimise standardisation and training. In Norway and Greenland, the core questionnaire was translated into Northern Sami and Kalaallisut (Greenlandic), respectively, while only the cue cards were translated into Iñupiaq in Alaska. Results A total of 663, 1,197 and 445 individuals were interviewed in Alaska, Greenland and Norway, respectively (Table I). Herein, 135 participants in Greenland and 18 in Norway who reported non-indigenous backgrounds shall be excluded in future studies. Participation rates by age and sex are unavailable in Norway and Alaska due to the sampling methods used. Very high overall participation rates were obtained in Greenland and Alaska, while a more conventional rate was achieved in Norway. Tables II through IV present community-and regionalspecific participation rates. Tables V shows the distribution of participants by age and sex. A predominance of female respondents was obtained in Alaska. Overall, the Sami cohort is older than the samples from Greenland and Alaska. Discussion In this article, we have presented the methods and design, overall regional-and community-specific participation rates, and the distribution of participants by age, sex and country/region in the SLiCA survey in Alaska, Greenland and Norway. High overall participation rates were obtained in Greenland and Alaska, while a more conventional rate was observed in Norway. A conventional participation rate and non-probability sampling may have introduced selection bias in the Sami sample. Available literature stresses that person-to-person approaches usually give higher participation rates than initial telephone contacts (5,22). The different methods used in the recruitment phase may, thus, explain some of the observed discrepancy in participation rates between the countries/regions. In Norway, the participation rates in Finnmark were systematically lower compared with the rates from Troms, Nordland and Trøndelag; the snowball sampling may have led us to the more motivated respondents in these 3 counties. The only information on Sami non-responders available to us is their sex and place of residence. Nevertheless, it has been documented that the differences between responders and non-responders, generally, are important but seldom so great that studies are irrevocably undermined (23). Furthermore, the Sami sample is a nonprobability sample. Those invited were not chosen at random; we cannot rule out the possibility that our participants differ systematically from the population we want our sample to reflect (24,25). In terms of external validity, selection bias is generally a problem, if the priority is to describe the distribution of variables in the population (26). However, in the future, SLiCA data will be used to explore the relationships between health and related areas of living condition, and associated risk factors. Any association may well be biased, if the study participants have a different distribution of confounding factors than the non-participants (26). Given a large enough sample, however, this may be statistically corrected, by adjusting, for the known and relevant confounding variables. As all SAMINOR municipalities overlap with the municipalities visited in SLiCA, SAMI-NOR data may be used to assess possible skewed distributions of confounding variables in SLiCA in the relevant age strata. Assuming that SAMINOR is a plausible estimate of the total Sami population aged 30 and 36Á79 years, discrepancies in distributions of comparable variables may be tested statistically. Significant statistical differences in distributions may be used for evaluating inclusions of covariates in the model. Also, if relevant, valid estimates of the distribution of variables in the Sami population may be produced by applying weights to skewed variables by using the same method (27,28). However, preliminary assessments suggest that selection bias in the Sami sample is plausible, but not a major threat when comparing education attainment with SAMINOR data. Based on previous research (29) and the U.S. Census 2000 (30,31), few or no threats to validity are detected in the data from Alaska and Greenland. Probability sampling, in addition to high participation rates, explains this. Because of funding issues, data collection was delayed in Greenland and Norway. As a consequence, the respective data-sets stem from different periods, which may challenge comparisons. Data collection was done within a 6-year period; this may result in confounding, due to period effects. Some of this effect may be In Alaska non-indigenous persons were excluded prior to invitation. In Greenland and Norway, however, information on ethnic background was not known in advance. Thus, total participants include persons who did not report indigenous background. Of the 663 participants in Alaska, 67 in the Bering Strait and 2 in the Northwest Arctic reported exclusively Yupik background (data not shown). controlled, by adjusting, for interview year. A strength is that data collection took place in parts of the respective populations across the year, which reduced possible bias due to seasonal effects. Seasonal effects are common for various health variables in epidemiological studies (23). The involved peoples represent varied ways of life of unique histories, experiences, communities and languages. The goal of standardised measurements is central to survey research, and it has been considered essential to keep the wording of questions constant across respondents (32). But even the same question may mean different things to different people, which may produce differential respondent/reporting bias. Culture influences the ways in which information is processed and conceptualised (33), and by no means, meaning is determined by words alone (34). This issue has been addressed by SLiCA, and a joint effort from involved researchers and indigenous representatives have maximised consistency of meaning. Standardisation in SLiCA was also possible as the indigenous peoples involved share common concepts with regard to the role of household production, their strong ties to the environment and the continuing role of extended informal and formal social relationships (2). Despite different sampling and recruitment methods and sociocultural differences, a unique database has been generated, which shall be used to explore relationships between health and living condition variables. Ethics Detailed information on the project was given to the participants orally and in writing, and written informed consent was obtained before interviews took place. For respondents younger than 18 years, who took part in the study, prior written informed consent from parents or legal guardians was obtained. In Norway, the study was accredited by the Norwegian Social Science Data Service and the National Committee for Research Ethics in the Social Sciences and the Humanities. In Alaska, the study was approved by Genetic Association Studies Identify Unanticipated Gene Pathways Influencing Sepsis Outcome Sepsis triggers multiple parallel inflammatory signaling pathways. Of these pathways, which ones contribute most substantially to adverse outcomes and, therefore, are relevant targets for new therapies? > 100 clinical trials of mediator modulators in sepsis patients have failed (Marshall, 2014) suggesting that we need new information to direct our search. In other disease states a genome-wide association study design is an unbiased approach that has identified genes in key pathways (Altshuler et al., 2008). For example, PCSK9 was discovered using genetic association analysis of patients who had LDL levels measured (Abifadel et al., 2003). This has led to the introduction of PCSK9 inhibitors as a treatment for hypercholesterolemia where other treatments have failed. Can a similar genetic association strategy work in the complex milieu of sepsis? Sepsis triggers multiple parallel inflammatory signaling pathways. Of these pathways, which ones contribute most substantially to adverse outcomes and, therefore, are relevant targets for new therapies? N100 clinical trials of mediator modulators in sepsis patients have failed (Marshall, 2014) suggesting that we need new information to direct our search. In other disease states a genome-wide association study design is an unbiased approach that has identified genes in key pathways (Altshuler et al., 2008). For example, PCSK9 was discovered using genetic association analysis of patients who had LDL levels measured (Abifadel et al., 2003). This has led to the introduction of PCSK9 inhibitors as a treatment for hypercholesterolemia where other treatments have failed. Can a similar genetic association strategy work in the complex milieu of sepsis? In the current issue of EBioMedicine, (2016-in this issue) conducted a genome-wide association study for 28-day mortality, initially in 740 septic patients. The authors followed standard quality control practices to limit any potential methodological errors in their discovery GWAS. From 644,699 SNPs they imputed 7,993,459 SNPs for their GWAS analysis. These investigators found that a missense genetic variant located within the VPS13A gene was associated with 28-day mortality in sepsis. This was the strongest association observed within the primary GWAS analysis (p = 8.2 × 10 −8 ). The minor allele, associated with adverse outcome, occurred only 1% of the time so this association is very susceptible to a false positive result since only a small number of patients would carry this adverse genetic variantthe conclusions are based on a few "affected" patients. Therefore, replication of this finding was essential. These investigators took the relatively rare (within the critical care community), but crucial, step of reaching out to previous investigators with large genotyped sepsis cohorts. This VPS13A finding replicated to a reasonable extent (associated with SOFA severity of illness score) in a second large cohort of patients from the PROGRESS study (ClinicalTrials.gov: NCT02782013). Finally, taking all genetic variants found by sequencing across the VPS13A gene, the VPS13A gene was also found to be associated with 28-day mortality. A bioinformatics in silico analysis suggests that this protein-altering SNP (rs117983287) is predicted to be highly deleterious to VPS13A function. The original finding plus replication and further support from the sequencing study and in silico analysis leads to the potentially important and exciting finding that a signaling pathway involving VPS13A is associated with sepsis outcome. Not much is known about VPS13A function as it relates to sepsis (Munoz-Braceras et al., 2015) so much work remains. Indeed, there are many other genes in this region so it is not yet certain that VPS13A causally impacts sepsis outcome. The authors identified 13 other genetic variants that are promising candidates. None of these additional genetic variants reached the prespecified level of statistical significance and therefore do not meet the discovery threshold but remain as promising candidates requiring further work and validation. When tested for replication in the collaborators' genotyped sepsis cohorts, none replicated to the same extent as VPS13A. Among this set, the best candidates included CRISPLD2 (p = 5.99 × 10 −6

) and a region on chromosome 13q21.33 (p = 3.34 × 10 −7 ). Reversing the replication strategy, these investigators tested for replication of top association findings previously reported by Rautanen et al. (Rautanen et al., 2015). They did not observe directionally similar significant findings for any of the reported SNPs. Again, it must be appreciated that for genetic association studies, the currently reported cohort is quite small and therefore does not have much statistical power to find true associations. The use of previous data to "look-up" potential new discoveries is a very encouraging event. First, replication of the key result is impressive validation and greatly increases the probability that the primary discovery is biologically real and not a statistical fluke. Second, sharing of data is an exciting trend that will certainly improve the veracity of reported results. Another encouraging step was the use of gene-based analysis for replication. Single SNP associations may not identify causal SNPs and may simply be markers in linkage disequilibrium with the underlying causal genetic variants. Sequencing all SNPs within the identified gene is a more powerful approach (Lee et al., 2012). The increased Contents lists available at ScienceDirect EBioMedicine j o u r n a l h o m e p a g e : w w w . e b i o m e d i c i n e . c o m statistical power of this approach (Taudien et al., 2016) is tempered by the smaller number of patients within this substudy in the current report. Nevertheless, replication of gene association greatly reduces that chance that a SNP association is a false positive result. The current report highlights bad and good features of genetic association studies in sepsis, ARDS, and critical illness. A key bad feature is the relatively low power we currently have to make discoveries because we have not put together sufficiently large genotyped sepsis cohorts. Cohorts in the tens of thousands have successfully identified key genes in, for example, atherosclerosis and asthma. This has led to the development of highly successful new drugs. The very good feature of the current report is that these investigators, and indeed the critical care community, are now starting to coalesce in order to address the important observations arising from genetic association studies. Let's put together the first N 10,000 patient genetic association study in sepsis and start to make the really exciting discoveries that will transform patient care and outcomes. Support Canadian Institutes of Health Research (136986). Conflicts of Interest KW is an inventor on a patent application filed by the University of British Columbia (UBC) regarding the use of PCSK9 inhibitors in sepsis. KW is a founder and shareholder of Cyon Therapeutics which has licensed this IP from UBC. Initial experience of video capsule endoscopy at a tertiary center in Saudi Arabia Background/Aim: No prior experience with video capsule endoscopy (VCE) has been published from Saudi Arabia. In this study, we aim to report the first Saudi experience with VCE. Patients and Methods: A prospective study was conducted between March 2013 and September 2017 at King Abdulaziz Medical City, Riyadh, Saudi Arabia. Eligible patients underwent VCE and their data (age, sex, indication for VCE, type of obscure gastrointestinal bleeding [OGIB: overt vs occult], VCE findings, and complications) were recorded. Approval was obtained from the institutional ethics board before the study began and all patients provided verbal and signed consent for the procedure. The procedure was performed according to the established guidelines. Results: During the study period, a total 103 VCE procedures were performed on 96 patients. Overall, 60 participants (62.5%) were male (mean age, 58.8 years; range, 25–97 years) and 36 (37.5%) were female (mean age, 52.8 years; range, 18–78 years). The most frequent indication for VCE was OGIB (n = 91, 88.35%; overt, n = 46, 50.55%; occult, n = 45, 49.45%). Other indications were suspected Crohn's disease (n = 4, 3.88%), suspected complicated celiac disease (n = 4, 3.88%), and unexplained chronic abdominal pain (n = 4, 3.88%). The VCE results were categorized as incomplete (n = 2, 1.94%), poor-quality (n = 7; 6.8%), normal (n = 39, 37.86%), and abnormal (n = 55, 53.4%). The completion rate was 98.06% (n = 101), and the overall diagnostic yield was 53.4%. Of the 55 patients with abnormal VCE results, 43 (78.2%) had small bowel (SB) abnormalities and 12 (21.8%) had abnormalities in the proximal or distal gut. The most frequent SB abnormalities were angiodysplasia (n = 22, 40.0%) and tumors (n = 7, 12.7%). Conclusion: The diagnostic yield of VCE for Saudi patients with OGIB is comparable to that reported internationally; however, data for other VCE indications, including inflammatory bowel disease, are still lacking. gastrointestinal (GI) bleeding (OGIB), but it also aids the diagnosis of inflammatory bowel disease (IBD), celiac disease, and SB neoplasia. OGIB is the most frequently studied indication for VCE. A systematic review published in 2010 that included 227 studies (equivalent to 22,840 VCE procedures) [2] reported a pooled diagnostic yield of approximately 60% for OGIB. The most common etiology for OGIB was angiodysplasia (50%), whereas inflammatory lesions or ulcerations and SB tumors, accounted for 26.8% and 8.8%, respectively, of all cases. [2] For the indication of OGIB, VCE has higher diagnostic yield compared to SB barium study, computerized tomography (CT) enteroclysis, SB magnetic resonance imaging (MRI), and push enteroscopy; and similar diagnostic yield to mesenteric angiography and intraoperative enteroscopy. [3] The role of VCE in IBD is growing rapidly. It might be helpful for diagnosing Crohn's disease among patients with clinical suspicion of this condition but with negative ileocolonoscopy and abdominal imaging results by detecting early mucosal lesions in the SB. [4] In established cases of Crohn's disease, VCE can potentially modify the treatment and clinical outcome when it is used to assess mucosal healing. [5,6] In celiac disease, VCE has been used successfully for defining the extent, severity, and complications of the disease. [7,8] Despite the potential benefits of VCE, this procedure is not without risk. The most common complication associated with VCE is capsule retention. The rate of capsule retention depends on the clinical indication. A systematic review and meta-analysis published in 2017 [9] reported a capsule retention rate of 2% for OGIB, 3.6% for suspected Crohn's disease, and 8.2% for established Crohn's disease. The most frequent cause of capsule retention was Crohn's disease stricture (46%), followed by SB neoplasms (17%). [9] Although VCE has been used widely in most tertiary centers in Saudi Arabia, experience with VCE has never been reported from this country. In the present study, we report the initial experience with VCE at a tertiary center in Saudi Arabia. Our objective is to see its growing scope locally and to compare with international published data. PATIENTS AND METHODS The primary objectives of the study were to determine the findings and diagnostic yield of VCE in Saudi patients. The secondary objectives included procedure completion rate and complications in Saudi patients. A prospective study was conducted between March 2013 and September 2017 at King Abdulaziz Medical City (KAMC), Riyadh, SA. Eligible patients underwent VCE and their data (age, sex, indication for VCE, type of OGIB [overt vs occult], VCE findings, complications) were recorded. Approval was obtained from the institutional ethics board before the study began and all patients provided verbal and signed consent for the procedure. All adult patients (age >14 years) with OGIB, defined by the American Gastroenterological Association (AGA) as recurrent or persistent GI bleeding and negative esophagogastroduodenoscopy (EGD) and colonoscopy, [10] were included. Patients suspected of Crohn's disease based on a combination of clinical features, elevated inflammatory markers, and negative results by EGD, colonoscopy, and CT enterography were also eligible for the study. Patients with celiac disease who were unresponsive to a gluten-free diet even after counselling and dietary consultation were marked as "probably complicated celiac disease" and were included as well. Patients with abdominal pain not explained by medical history, physical examination, abdominal imaging, and endoscopy were also included. Patients with suspected or confirmed SB obstruction were excluded from the study. The presence of relative contraindication to VCE, such as an implanted cardiac device or pregnancy, did not preclude patients from inclusion if VCE was indicated. VCE system At the KAMC endoscopy unit, we use the MiroCam capsule system (IntroMedic, Republic of Korea). The capsule measures 24.5 mm × 10.8 mm and has a 170° field of view. It has 12-h operation time and captures three images per second. The capsule depends on the human body as a conductor to transmit data from its antenna to the sensors and does not depend on radiofrequency transmission, eliminating potential interference between the capsule and implanted cardiac devices. The signals emitted by the capsule also enable the approximation of its position in the abdominal cavity. Study protocol All participants were instructed to cease iron therapy 3-5 days before undergoing VCE and to avoid any medications that can affect gut motility, including narcotic and anticholinergic medications, if possible. The day before the procedure, the patients were instructed to consume only liquid diet and to take at least 1 L polyethylene glycol-based bowel preparation (Moviprep; Norgine Limited, the Netherlands) in the afternoon. The patients then fasted for 8 h before ingesting the capsule. On the day of the procedure and after the sensors had been attached to their body, the patients swallowed the capsule with water mixed with 5 ml simethicone (Salinal, Julphar Gulf Pharmaceutical Industries, Diqdaqah, Ras Al Khaimah, UAE). At the time of hospital discharge, the patients were given written instructions in both English and Arabic to follow over the next 12 h [ Table 1]. All VCE images were reviewed by at least two expert gastroenterologists and any questionable findings were discussed. An abnormal VCE result was defined as any detected abnormality that could be related to the patient's presenting problem. The presence of blood alone was also considered a positive finding, as blood usually helps to localize the lesion and supports the use of invasive procedures such as balloon-assisted enteroscopy (BAE) or surgery. Diagnostic yield was defined as the percentage of cases with abnormal VCE results. A poor-quality study was arbitrarily defined if less than 50% of mucosa could be observed. A normal study indicated that no abnormal findings could be detected in a good-quality study. Complete study was defined as the capsule passing through the ileocecal valve or into the colon on imaging, whereas incomplete study meant that the capsule had expired in the stomach or could not pass through the ileocecal valve during its working time and that further radiology was required to confirm its excretion. [2] Capsule retention was defined as nonpassage of the capsule into the cecum within 2 weeks of capsule ingestion. [9] The data were analyzed using SPSS version 20.0 (IBM, Armonk, NY, USA). RESULTS A total of 103 procedures were conducted on 96 patients during the study period. The participants comprised 60 men (62.5%) and 36 women (37.5%). The mean age for men was 58.8 years (range, 25-97 years) and for women was 52.8 years (range, 18-78 years), Overall, two patients had pacemakers and one was a pregnant woman in her second trimester. The overall completion rate for the 103 VCE procedures was 98.0% (n = 101) and the overall diagnostic yield was 53.4% (n = 55). Table1. Post procedure patient instructions During the day long procedure, you may return home, work and do regular activities. You can drink water 2 hours after capsule ingestion You can take very light meal and your usual medications 4 hours after capsule ingestion but you must avoid red -colored fluids and food. You should exercise care with the sensor array when changing clothes or in the bathroom. It is advised to avoid strenuous exercise or other activities which may dislodge the leads. You should avoid other patients undergoing capsule endoscopy as this may result in interference with your study. After 12 hours of capsule ingestion, the procedure is considered to be finished and you can eat and drink normally. You will need to come back to endoscopy unit next day to remove the sensors and the data recorder. You should avoid MRI scanners until capsule passage is confirmed. two (1.94%) with GI stromal tumors (GIST), one (0.97%) with multiple neuroendocrine tumors (NET), one

(0.97%) with Burkett's lymphoma, and one (0.97%) with SB metastasis from colonic adenocarcinoma, where the patient had previously undergone colonic tumor resection. One of the four patients with suspected Crohn's disease had SB ulcerations; repeated ileocolonoscopy and biopsy confirmed the diagnosis of Crohn's disease. Three of the four patients with suspected complicated celiac disease had ulcerative ileojejunitis. By contrast, none of the four patients with unexplained chronic abdominal pain had abnormal VCE findings. The capsule was retained in three patients (2.9%), where there was one case each of SB adenocarcinoma, SB diverticulum, and multiple SB strictures related to the use of non-steroidal anti-inflammatory drugs (NSAIDs). However, none of these patients developed SB obstruction or perforation. Capsule retention was symptomatic in one patient who presented with severe abdominal pain without signs of bowel obstruction or perforation 7 h after capsule ingestion. This patient required emergency laparoscopy to push the capsule out of a distal SB diverticulum to the cecum. Double-balloon enteroscopy (DBE) was used to retrieve the retained capsule from the patient with SB adenocarcinoma; however, this approach was unsuccessful in the patient with multiple NSAIDs-related SB strictures, who required surgery for capsule retrieval. DISCUSSION In this study of VCE in Saudi patients, both the overall diagnostic yield (53.4%) and the diagnostic yield for OGIB (56.0%) were within the ranges reported worldwide. [2] The diagnostic yield was significantly (P = 0.00052) higher for overt OGIB (73.9%) compared to occult OGIB (37.8%). These data are comparable with that of previous studies. [11][12][13] Although we included only four patients with unexplained abdominal pain, the diagnostic yield of VCE was zero for this indication. This suggests that VCE is less helpful for unexplained abdominal pain than for other indications. In fact, a systematic review of 21 studies that evaluated the diagnostic yield of VCE among patients with chronic unexplained abdominal pain found a pooled diagnostic yield of only 20.9%. [14] The completion rate in this study was high (98.0%) even though the study protocol did not include the administration of prokinetic medications to enhance procedure completion. This reduces the importance of such intervention. However, it has been previously shown that prokinetic medications increase the completion rate without improving the diagnostic yield. [15,16] Approximately 22% of the abnormalities detected by VCE in the present study were in the stomach or colon. This suggests that lesions were missed by the initial upper and lower endoscopies. All the missed lesions were vascular in nature. In fact, missing nonbleeding vascular lesions during upper and lower endoscopies is not uncommon; air insufflation and the required sedation are potential reasons for missing such lesions. [17,18] One valuable aspect of VCE is its physiological nature, as it does not require air insufflation or sedation, which increases the sensitivity for detecting vascular lesions. In the present study, three patients experienced capsule retention despite two of them having had normal abdominal CT results before undergoing VCE. Several studies have suggested that SB barium study and regular abdominal CT are inaccurate tools for ensuring SB patency, whereas CT/MR enterography and the agile patency capsule (Given Imaging, Yokneam, Israel) are considered the best options for this purpose. [19][20][21] This study has several limitations. Although we confirmed the positive VCE findings by DBE and/or surgery, not all of the negative VCE results were verified by the gold standard. Thus, we could not calculate the accuracy, sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of VCE. Instead, we used crude measures, such as positive findings and diagnostic yield. In addition, the number of patients in the study was small for non-OGIB indications. Consequently, no firm conclusions could be drawn for these indications. The strengths of this study include the prospective design and it being the first study on this topic to be conducted in our population. We anticipate that our results will encourage other national centers to share their data and help develop national guidelines on the use of VCE. CONCLUSION The diagnostic yield of VCE for Saudi patients with OGIB is comparable to that reported internationally; however, Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest. Novel Glycosylation Sites Localized in Campylobacter jejuni Flagellin FlaA by Liquid Chromatography Electron Capture Dissociation Tandem Mass Spectrometry Glycosylation of flagellin in Campylobacter jejuni is essential for motility and virulence. It is well-known that flagellin from C. jejuni 81−176 is glycosylated by pseudaminic acid and its acetamidino derivative, and that Campylobactor coli VC167 flagellin is glycosylated by legionaminic acid and its derivatives. Recently, it was shown, by use of a metabolomics approach, that C. jejuni 11168 is glycosylated by dimethyl glyceric acid derivatives of pseudaminic acid, but the sites of glycosylation were not confirmed. Here, we apply an online liquid chromatography electron capture dissociation (ECD) tandem mass spectrometry approach to localize sites of glycosylation in flagellin from C. jejuni 11168. Flagellin A is glycosylated by a dimethyl glyceric acid derivative of pseudaminic acid at Ser181, Ser207 and either Thr464 or Thr 465; and by a dimethyl glyceric acid derivative of acetamidino pseudaminic acid at Ser181 and Ser207. For comparison, on-line liquid chromatography collision-induced dissociation of the tryptic digests was performed, but it was not possible to assign sites of glycosylation by that method. ' INTRODUCTION The food-borne pathogen Campylobacter jejuni is one of the leading causes of bacterial gastroenteritis worldwide. Campylobacter are characterized by a rapid, darting motility, mediated by bipolar flagella, which is essential for virulence. The flagellum comprises a basal body linked to the flagellar filament, which consists of thousands of copies of the flagellin proteins FlaA and FlaB, with FlaA being the major component. Glycosylation of flagellins appears to be essential for flagellar biogenesis, as mutations in Cj1293, encoding a putative UDP-GlcNAc C6-dehydratase/C4-reductase involved in flagellin glycosylation, were aflagellate and nonmotile. 1 The Cj1293 orthologue in Campylobacter coli VC167 was not essential for flagellin glycosylation or flagellin biogenesis but a double mutant in this gene and ptmD involved in pseudaminic acid biosynthesis was also aflagellate and nonmotile. 1 These observations indicate that flagellin glycosylation with pseudaminic acid and derivatives is essential for targeting and/or secretion of flagellin in C. jejuni and C. coli. More recently, Howard et al. 2 have demonstrated the importance of specific structural modifications to the flagellin glycoform in the biological fitness of C. jejuni in colonization of chickens. In 2001, Thibault et al. 3 analyzed intact flagellins from three strains of C. jejuni (81-176, NCTC 11168 and OH4384) and one strain of C. coli (VC167 T2) by mass spectrometry. They showed that, in each case, the mass of the monomeric glycoprotein was ∼6 kDa greater than that predicted from the primary sequence. Liquid chromatography mass spectrometry (LC-MS) of tryptic peptides of C. jejuni 81-176 flagellin FlaA revealed the presence of seven glycopeptides. The nature of the glycans was assigned following collision-induced dissociation (CID) tandem mass spectrometry (MS/MS) of HPLC fractions containing the suspected glycopeptides. The major glycosylation component was pseudaminic acid (Pse5Ac7Ac), with 5-acetamidino pseudaminic acid (Pse5Am7Ac) and 5,7-N-(2,3-dihydroxyproprionyl)-pseudaminic acid (Pse5Pr7Pr) also present. Sites of glycosylation were determined by β-elimination reactions and subsequent MS/MS. In total, 19 serine and threonine glycosylation sites were identified on C. jejuni 81-176 FlaA. In further work, Thibault and co-workers applied a top-down approach for the identification of glycosylation in C. jejuni 81-176 flagellin. 4 Again, the major glycans were Pse5Ac7Ac and Pse5Am7Ac. In addition, novel glycans, Pse5Am7Ac8GlnAc and Pse5Ac7Ac8OAc, were identified and localized to tryptic peptide . The exact sites of the novel modification, however, were not determined. Soo and co-workers 5 developed a targeted metabolomics approach for the analysis of glycosylation in C. jejuni 81-176, confirming the presence of pseudaminic acid and its acetamidino derivative. Cell lysates were investigated for the presence of sugarnucleotide metabolites by combining hydrophilic interaction liquid chromatography (HILIC) with precursor ion scanning mass spectrometry and the structures of the metabolites were confirmed by NMR spectroscopy. The method was subsequently applied to C. coli VC167. 6 In addition to pseudaminic acid, acetamidino legionaminic acid (Leg5Am7Ac) and N-methylacetimidoyl legionaminic (Leg5-AmNMe7Ac) were identified as flagellar glycans. Most recently, the metabolomics approach was applied to the analysis of C. jejuni 11168 flagellin. 7 Although 11168 was the first C. jejuni strain to have its genome sequenced, 8 the least is known about its flagellin glycosylation. Moreover, its glycosylation locus is far more complex than either C. jejuni 81-176 or C. coli VC167. Logan et al. 7 identified two novel glycans, see Scheme 1, corresponding to dimethylglyceric acid derivatives of pseudaminic acid and 7-acetamidino pseudaminic acid, hereafter referred to as I and II, respectively. Tryptic peptide [204-223] was shown to be modified by either I or II but the site(s) of glycosylation were not confirmed. Tryptic peptide [464][465][466][467][468][469][470][471][472][473][474][475][476][477][478][479][480] was shown to be modified by I. Again, the site of modification was not localized. Hitchen et al. 9 used a`knock-in' approach to demonstrate that gene Cj1295 is responsible for the presence of the dimethylglyceric acid derivative of pseudaminic acid. Here, we have applied on-line reversed-phase liquid chromatography electron capture dissociation (ECD) tandem mass spectrometry 10 to the characterization of glycosylation of flagellin A from C. jejuni 11168. ECD 11-13 is a radically driven fragmentation technique which results in cleavage of the peptide backbone N-CR to produce c and z• (or c• and z) fragments. Unlike the more commonly used collision-induced dissociation (CID), labile post-translational modifications, such as glycosylation, are retained on the peptide backbone fragments following ECD. 14,15 It is possible, therefore, to localize sites of modification unambiguously. We show that flagellin A is glycosylated by I at Ser181, Ser207 and either Thr464 or Thr465, and by II at Ser181 and Ser207. Glycosylation of Ser181 of flagellin A from C. jejuni 11168 has not been observed previously, although LC-MS/MS analysis of flagellins from the knockout strain 11168 Cj1295:: aphA and the`knock-in' strain 111168 Cj1295::aphA cmCj1295 revealed the tryptic peptide [179][180][181][182][183][184][185][186][187][188][189][190] to be modified. 9 Interestingly, both Ser181 and Ser207 fall within the sequence motif ISTS (S is modified) suggesting the possibility of a consensus sequence for targeting of glycosylation. The equivalent site (also Ser207) has been shown to be modified by the glycans Pse5A-c7Ac and Pse5Am7Ac in flagellin A from C. jejuni strain 81-176. 3 Finally, ECD revealed that flagellin A is glycosylated by I at Thr464 or Thr465. The equivalent residues in flagellin A from C. jejuni strain 81-176 have not been observed to be glycosylated. 3 There was no evidence for modification of these residues by glycan II. Additionally, there was no evidence for the modification of flagellin A from strain 11168 by any of the glycans known for other strains of Campylobacter, for example, pseudaminic acid, legionaminic acid. For comparison, on-line liquid chromatography collision-induced dissociation tandem mass spectrometry was performed. Although the glycopeptides were observed, it was not possible to localize the sites of modification due to the preferential loss of the glycan modification. ' EXPERIMENTAL SECTION Preparation of Flagellin A The method was based on that of Power et al. 16 Cultures were grown in Mueller-Hinton broth for 24 h in an atmosphere of 5% O 2 , 10% CO 2 , and 85% N 2 , with shaking at 37°C. Cells were recovered from approximately 200 mL of culture by centrifugation at 6000g for 20 min at 4°C and resuspended to a calculated OD of 20. Suspensions were homogenized at 20 500 rpm for 3 min using an Ultra-Turrax T-25 high speed homogenizer (IKA Labortechnik) before removal of cells and debris by centrifugation at 6000g for 20 min at 4°C. The supernatant was then subjected to ultracentrifugation at 100 000g for 60 min at 5°C. The pellet was resuspended in 1% (w/v) sodium dodecyl sulfate in distilled water to solubilize membrane fragments, incubated at 37°C for 2 h, and again subjected to ultracentrifugation as above. The cycle of SDS washing and ultracentrifugation was repeated three times and purity checked by SDS-PAGE and electron microscopy before suspending in water and ultracentrifugation as above for three more cycles to remove SDS. Flagellar fragments were stored in water at -20°C until required for analysis. Flagellar proteins isolated from C. jejuni strain 11168 were separated by polyacrylamide gel (12%) electrophoresis and visualized with Coomassie stain R250 (BioRad). Bands with a molecular mass of approximately 60 kDa were cut out of the gel and stored at -20°C until